Structural and functional studies of ferredoxin and oxygenase components of 3-nitrotoluene dioxygenase from Diaphorobacter sp. strain DS2

Functional and spectroscopic approaches to determining thermal limitations of Rieske oxygenases

The biotechnological potential of Rieske Oxygenases (ROs) and their cognate reductases remains unmet, in part because these systems can be functionally short-lived. Here, we describe a set of experiments aimed at identifying both the functional and structural stability limitations of ROs, using terephthalate (TPA) dioxygenase (from Comamonas strain E6) as a model system. Successful expression and purification of a cofactor-complete, histidine-tagged TPA dioxygenase and reductase protein system requires induction with the Escherichia coli host at stationary phase as well as a chaperone inducing cold-shock and supplementation with additional iron, sulfur, and flavin. The relative stability of the Rieske cluster and mononuclear iron center can then be assessed using spectroscopic and functional measurements following dialysis in an iron chelating buffer. These experiments involve measurements of the overall lifetime of the system via total turnover number using both UV-Visible absorbance and HPLC analyses, as well specific activity as a function of temperature. Important methods for assessing the stability of these multi-cofactor, multi-protein dependent systems at multiple levels of structure (secondary to quaternary) include differential scanning calorimetry, circular dichroism, and metallospectroscopy. Results can be rationalized in terms of three-dimensional structures and bioinformatics. The experiments described here provide a roadmap to a detailed characterization of the limitations of ROs. With a few notable exceptions, these issues are not widely addressed in current literature.

Semi-rational design of nitroarene dioxygenase for catalytic ability toward 2,4-dichloronitrobenzene

Rieske non-heme dioxygenase family enzymes play an important role in the aerobic biodegradation of nitroaromatic pollutants, but no active dioxygenases are available in nature for initial reactions in the degradation of many refractory pollutants like 2,4-dichloronitrobenzene (24DCNB). Here, we report the engineering of hotspots in 2,3-dichloronitrobenzene dioxygenase from Diaphorobacter sp. strain JS3051, achieved through molecular dynamic simulation analysis and site-directed mutagenesis, with the aim of enhancing its catalytic activity toward 24DCNB. The computationally predicted activity scores were largely consistent with the detected activities in wet experiments. Among them, the two most beneficial mutations (E204M and M248I) were obtained, and the combined mutant reached up to a 62-fold increase in activity toward 24DCNB, generating a single product, 3,5-dichlorocatechol, which is a naturally occurring compound. In silico analysis confirmed that residue 204 affected the substrate preference for meta-substituted nitroarenes, while residue 248 may influence substrate preference by interaction with residue 295. Overall, this study provides a framework for manipulating nitroarene dioxygenases using computational methods to address various nitroarene contamination problems.IMPORTANCEAs a result of human activities, various nitroaromatic pollutants continue to enter the biosphere with poor degradability, and dioxygenation is an important kickoff step to remove toxic nitro-groups and convert them into degradable products. The biodegradation of many nitroarenes has been reported over the decades; however, many others still lack corresponding enzymes to initiate their degradation. Although rieske non-heme dioxygenase family enzymes play extraordinarily important roles in the aerobic biodegradation of various nitroaromatic pollutants, prediction of their substrate specificity is difficult. This work greatly improved the catalytic activity of dioxygenase against 2,4-dichloronitrobenzene by computer-aided semi-rational design, paving a new way for the evolution strategy of nitroarene dioxygenase. This study highlights the potential for using enzyme structure-function information with computational pre-screening methods to rapidly tailor the catalytic functions of enzymes toward poorly biodegradable contaminants.

Biotransformation of 2,4,6-trinitrotoluene by Diaphorobacter sp. strain DS2

Diaphorobacter strain DS2 degrades 3-nitrotoluene and 2-nitrotoluene via ring oxidation with 3-nitrotoluene dioxygenase (3NTDO). In the current study, we hypothesized that 3NTDO might also be involved in the degradation of 2,4,6-trinitrotoluene (TNT), a major nitroaromatic explosive contaminant in soil and groundwater. Strain DS2 transforms TNT as a sole carbon and nitrogen source when grown on it. Ammonium chloride and succinate in the medium accelerated the TNT degradation rate. A resting cell experiment suggested that TNT does not compete with 3NT degradation (no negative impact of TNT on the reaction velocity for 3NT). Enzyme assay with 3NTDO did not exhibit TNT transformation activity. The above results confirmed that 3NTDO of DS2 is not responsible for TNT degradation. In the resting cell experiment, within 10 h, 4ADNT completely degraded. The degradation of 2ADNT was 97% at the same time. We hypothesized that 3NTDO involve in this reaction. Based on the DS2 genome, we proposed that the N-ethylmaleimide reductases (nemA) were involved in the initial reduction of the nitro group and aromatic ring of TNT. Our findings suggest that strain DS2 could be helpful for the removal of TNT from contaminated sites with or without any additional carbon and nitrogen source and with minimal accumulation of undesirable intermediates.

The NADH recycling enzymes TsaC and TsaD regenerate reducing equivalents for Rieske oxygenase chemistry

Many microorganisms use both biological and nonbiological molecules as sources of carbon and energy. This resourcefulness means that some microorganisms have mechanisms to assimilate pollutants found in the environment. One such organism is Comamonas testosteroni, which metabolizes 4-methylbenzenesulfonate and 4-methylbenzoate using the TsaMBCD pathway. TsaM is a Rieske oxygenase, which in concert with the reductase TsaB consumes a molar equivalent of NADH. Following this step, the annotated short-chain dehydrogenase/reductase and aldehyde dehydrogenase enzymes TsaC and TsaD each regenerate a molar equivalent of NADH. This co-occurrence ameliorates the need for stoichiometric addition of reducing equivalents and thus represents an attractive strategy for integration of Rieske oxygenase chemistry into biocatalytic applications. Therefore, in this work, to overcome the lack of information regarding NADH recycling enzymes that function in partnership with Rieske non-heme iron oxygenases (Rieske oxygenases), we solved the X-ray crystal structure of TsaC to a resolution of 2.18 Å. Using this structure, a series of substrate analog and protein variant combination reactions, and differential scanning fluorimetry experiments, we identified active site features involved in binding NAD+ and controlling substrate specificity. Further in vitro enzyme cascade experiments demonstrated the efficient TsaC- and TsaD-mediated regeneration of NADH to support Rieske oxygenase chemistry. Finally, through in-depth bioinformatic analyses, we illustrate the widespread co-occurrence of Rieske oxygenases with TsaC-like enzymes. This work thus demonstrates the utility of these NADH recycling enzymes and identifies a library of short-chain dehydrogenase/reductase enzyme prospects that can be used in Rieske oxygenase pathways for in situ regeneration of NADH.

Custom tuning of Rieske oxygenase reactivity

Rieske oxygenases use a Rieske-type [2Fe-2S] cluster and a mononuclear iron center to initiate a range of chemical transformations. However, few details exist regarding how this catalytic scaffold can be predictively tuned to catalyze divergent reactions. Therefore, in this work, using a combination of structural analyses, as well as substrate and rational protein-based engineering campaigns, we elucidate the architectural trends that govern catalytic outcome in the Rieske monooxygenase TsaM. We identify structural features that permit a substrate to be functionalized by TsaM and pinpoint active-site residues that can be targeted to manipulate reactivity. Exploiting these findings allowed for custom tuning of TsaM reactivity: substrates are identified that support divergent TsaM-catalyzed reactions and variants are created that exclusively catalyze dioxygenation or sequential monooxygenation chemistry. Importantly, we further leverage these trends to tune the reactivity of additional monooxygenase and dioxygenase enzymes, and thereby provide strategies to custom tune Rieske oxygenase reaction outcomes.

Rieske Oxygenases and Other Ferredoxin-Dependent Enzymes: Electron Transfer Principles and Catalytic Capabilities

Enzymes that depend on sophisticated electron transfer via ferredoxins (Fds) exhibit outstanding catalytic capabilities, but despite decades of research, many of them are still not well understood or exploited for synthetic applications. This review aims to provide a general overview of the most important Fd-dependent enzymes and the electron transfer processes involved. While several examples are discussed, we focus in particular on the family of Rieske non-heme iron-dependent oxygenases (ROs). In addition to illustrating their electron transfer principles and catalytic potential, the current state of knowledge on structure-function relationships and the mode of interaction between the redox partner proteins is reviewed. Moreover, we highlight several key catalyzed transformations, but also take a deeper dive into their engineerability for biocatalytic applications. The overall findings from these case studies highlight the catalytic capabilities of these biocatalysts and could stimulate future interest in developing additional Fd-dependent enzyme classes for synthetic applications.

Leveraging a Structural Blueprint to Rationally Engineer the Rieske Oxygenase TsaM

Rieske nonheme iron oxygenases use two metallocenters, a Rieske-type [2Fe-2S] cluster and a mononuclear iron center, to catalyze oxidation reactions on a broad range of substrates. These enzymes are widely used by microorganisms to degrade environmental pollutants and to build complexity in a myriad of biosynthetic pathways that are industrially interesting. However, despite the value of this chemistry, there is a dearth of understanding regarding the structure-function relationships in this enzyme class, which limits our ability to rationally redesign, optimize, and ultimately exploit the chemistry of these enzymes. Therefore, in this work, by leveraging a combination of available structural information and state-of-the-art protein modeling tools, we show that three "hotspot" regions can be targeted to alter the site selectivity, substrate preference, and substrate scope of the Rieske oxygenase p-toluenesulfonate methyl monooxygenase (TsaM). Through mutation of six to 10 residues distributed between three protein regions, TsaM was engineered to behave as either vanillate monooxygenase (VanA) or dicamba monooxygenase (DdmC). This engineering feat means that TsaM was rationally engineered to catalyze an oxidation reaction at the meta and ortho positions of an aromatic substrate, rather than its favored native para position, and that TsaM was redesigned to perform chemistry on dicamba, a substrate that is not natively accepted by the enzyme. This work thus contributes to unlocking our understanding of structure-function relationships in the Rieske oxygenase enzyme class and expands foundational principles for future engineering of these metalloenzymes.

Engineering Rieske oxygenase activity one piece at a time

Enzyme engineering plays a central role in the development of biocatalysts for biotechnology, chemical and pharmaceutical manufacturing, and environmental remediation. Rational design of proteins has historically relied on targeting active site residues to confer a protein with desirable catalytic properties. However, additional "hotspots" are also known to exist beyond the active site. Structural elements such as subunit-subunit interactions, entrance tunnels, and flexible loops influence enzyme catalysis and serve as potential "hotspots" for engineering. For the Rieske oxygenases, which use a Rieske cluster and mononuclear iron center to catalyze a challenging set of reactions, these outside of the active site regions are increasingly being shown to drive catalytic outcomes. Therefore, here, we highlight recent work on structurally characterized Rieske oxygenases that implicates architectural pieces inside and outside of the active site as key dictators of catalysis, and we suggest that these features may warrant attention in efforts aimed at Rieske oxygenase engineering.

Molecular Basis and Evolutionary Origin of 1-Nitronaphthalene Catabolism in Sphingobium sp. Strain JS3065

Nitrated polycyclic aromatic hydrocarbons (nitro-PAHs) enter the environment from natural sources and anthropogenic activities. To date, microorganisms able to mineralize nitro-PAHs have not been reported. Here, Sphingobium sp. strain JS3065 was isolated by selective enrichment for its ability to grow on 1-nitronaphthalene as the sole carbon, nitrogen, and energy source. Analysis of the complete genome of strain JS3065 indicated that the gene cluster encoding 1-nitronaphthalene catabolism (nin) is located on a plasmid. Based on the genetic and biochemical evidence, the nin genes share an origin with the nag-like genes encoding naphthalene degradation in Ralstonia sp. strain U2. The initial step in degradation of 1-nitronaphthalene is catalyzed by a three-component dioxygenase, NinAaAbAcAd, resulting in formation of 1,2-dihydroxynaphthalene which is also an early intermediate in the naphthalene degradation pathway. Introduction of the ninAaAbAcAd genes into strain U2 enabled its growth on 1-nitronaphthalene. Phylogenic analysis of NinAc suggested that an ancestral 1-nitronaphthalene dioxygenase was an early step in the evolution of nitroarene dioxygenases. Based on bioinformatic analysis and enzyme assays, the subsequent assimilation of 1,2-dihydroxynaphthalene seems to follow the well-established pathway for naphthalene degradation by Ralstonia sp. strain U2. This is the first report of catabolic pathway for 1-nitronaphthalene and is another example of how expanding the substrate range of Rieske type dioxygenase enables bacteria to grow on recalcitrant nitroaromatic compounds. IMPORTANCE Nitrated polycyclic aromatic hydrocarbons (nitro-PAHs) have been widely detected in the environment and they are more toxic than their corresponding parent PAHs. Although biodegradation of many PAHs has been extensively described at genetic and biochemical levels, little is known about the microbial degradation of nitro-PAHs. This work reports the isolation of a Sphingobium strain growing on 1-nitronaphthalene and the genetic basis for the catabolic pathway. The pathway evolved from an ancestral naphthalene catabolic pathway by a remarkably small modification in the specificity of the initial dioxygenase. Data presented here not only shed light on the biochemical processes involved in the microbial degradation of globally important nitrated polycyclic aromatic hydrocarbons, but also provide an evolutionary paradigm for how bacteria evolve a novel catabolic pathway with minimal alteration of preexisting pathways for natural organic compounds.

Contrasting Mechanisms of Aromatic and Aryl-Methyl Substituent Hydroxylation by the Rieske Monooxygenase Salicylate 5-Hydroxylase

The hydroxylase component (S5HH) of salicylate-5-hydroxylase catalyzes C5 ring hydroxylation of salicylate but switches to methyl hydroxylation when a C5 methyl substituent is present. The use of 18O2 reveals that both aromatic and aryl-methyl hydroxylations result from monooxygenase chemistry. The functional unit of S5HH comprises a nonheme Fe(II) site located 12 Å across a subunit boundary from a one-electron reduced Rieske-type iron-sulfur cluster. Past studies determined that substrates bind near the Fe(II), followed by O2 binding to the iron to initiate catalysis. Stopped-flow-single-turnover reactions (STOs) demonstrated that the Rieske cluster transfers an electron to the iron site during catalysis. It is shown here that fluorine ring substituents decrease the rate constant for Rieske electron transfer, implying a prior reaction of an Fe(III)-superoxo intermediate with a substrate. We propose that the iron becomes fully oxidized in the resulting Fe(III)-peroxo-substrate-radical intermediate, allowing Rieske electron transfer to occur. STO using 5-CD3-salicylate-d8 occurs with an inverse kinetic isotope effect (KIE). In contrast, STO of a 1:1 mixture of unlabeled and 5-CD3-salicylate-d8 yields a normal product isotope effect. It is proposed that aromatic and aryl-methyl hydroxylation reactions both begin with the Fe(III)-superoxo reaction with a ring carbon, yielding the inverse KIE due to sp2 → sp3 carbon hybridization. After Rieske electron transfer, the resulting Fe(III)-peroxo-salicylate intermediate can continue to aromatic hydroxylation, whereas the equivalent aryl-methyl intermediate formation must be reversible to allow the substrate exchange necessary to yield a normal product isotope effect. The resulting Fe(III)-(hydro)peroxo intermediate may be reactive or evolve through a high-valent iron intermediate to complete the aryl-methyl hydroxylation.

The α- and β-Subunit Boundary at the Stem of the Mushroom-Like α 3 β 3 -Type Oxygenase Component of Rieske Non-Heme Iron Oxygenases Is the Rieske-Type Ferredoxin-Binding Site

Cumene dioxygenase (CumDO) is an initial enzyme in the cumene degradation pathway of Pseudomonas fluorescens IP01 and is a Rieske non-heme iron oxygenase (RO) that comprises two electron transfer components (reductase [CumDO-R] and Rieske-type ferredoxin [CumDO-F]) and one catalytic component (α₃β₃-type oxygenase [CumDO-O]). Catalysis is triggered by electrons that are transferred from NAD(P)H to CumDO-O by CumDO-R and CumDO-F. To investigate the binding mode between CumDO-F and CumDO-O and to identify the key CumDO-O amino acid residues for binding, we simulated docking between the CumDO-O crystal structure and predicted model of CumDO-F and identified two potential binding sites: one is at the side-wise site and the other is at the top-wise site in mushroom-like CumDO-O. Then, we performed alanine mutagenesis of 16 surface amino acid residues at two potential binding sites. The results of reduction efficiency analyses using the purified components indicated that CumDO-F bound at the side-wise site of CumDO-O, and K117 of the α-subunit and R65 of the β-subunit were critical for the interaction. Moreover, these two positively charged residues are well conserved in α₃β₃-type oxygenase components of ROs whose electron donors are Rieske-type ferredoxins. Given that these residues were not conserved if the electron donors were different types of ferredoxins or reductases, the side-wise site of the mushroom-like structure is thought to be the common binding site between Rieske-type ferredoxin and α₃β₃-type oxygenase components in ROs.

IMPORTANCE We clarified the critical amino acid residues of the oxygenase component (Oxy) of Rieske non-heme iron oxygenase (RO) for binding with Rieske-type ferredoxin (Fd). Our results showed that Rieske-type Fd-binding site is commonly located at the stem (side-wise site) of the mushroom-like α₃β₃ quaternary structure in many ROs. The resultant binding site was totally different from those reported at the top-wise site of the doughnut-like α₃-type Oxy, although α₃-type Oxys correspond to the cap (α₃ subunit part) of the mushroom-like α₃β₃-type Oxys. Critical amino acid residues detected in this study were not conserved if the electron donors of Oxys were different types of Fds or reductases. Altogether, we can suggest that unique binding modes between Oxys and electron donors have evolved, depending on the nature of the electron donors, despite Oxy molecules having shared α₃β₃ quaternary structures.

Design principles for site-selective hydroxylation by a Rieske oxygenase

Rieske oxygenases exploit the reactivity of iron to perform chemically challenging C–H bond functionalization reactions. Thus far, only a handful of Rieske oxygenases have been structurally characterized and remarkably little information exists regarding how these enzymes use a common architecture and set of metallocenters to facilitate a diverse range of reactions. Herein, we detail how two Rieske oxygenases SxtT and GxtA use different protein regions to influence the site-selectivity of their catalyzed monohydroxylation reactions. We present high resolution crystal structures of SxtT and GxtA with the native β-saxitoxinol and saxitoxin substrates bound in addition to a Xenon-pressurized structure of GxtA that reveals the location of a substrate access tunnel to the active site. Ultimately, this structural information allowed for the identification of six residues distributed between three regions of SxtT that together control the selectivity of the C–H hydroxylation event. Substitution of these residues produces a SxtT variant that is fully adapted to exhibit the non-native site-selectivity and substrate scope of GxtA. Importantly, we also found that these selectivity regions are conserved in other structurally characterized Rieske oxygenases, providing a framework for predictively repurposing and manipulating Rieske oxygenases as biocatalysts.

SxtT and GxtA are Rieske oxygenases that are involved in paralytic shellfish toxin biosynthesis and catalyze monohydroxylation reactions at different positions on the toxin scaffold. Here, the authors present crystal structures of SxtT and GxtA with the native substrates β-saxitoxinol and saxitoxin as well as a Xenon-pressurized structure of GxtA, which reveal a substrate access tunnel to the active site. Through structure-based mutagenesis studies the authors identify six residues in three different protein regions that determine the substrate specificity and site selectivity of SxtT and GxtA. These findings will aid the rational engineering of other Rieske oxygenases.

A Recently Assembled Degradation Pathway for 2,3-Dichloronitrobenzene in Diaphorobacter sp. Strain JS3051

Diaphorobacter sp. strain JS3051 utilizes 2,3-dichloronitrobenzene (23DCNB), a toxic anthropogenic compound, as the sole carbon, nitrogen, and energy source for growth, but the metabolic pathway and its origins are unknown. Here, we establish that a gene cluster (dcb), encoding a Nag-like dioxygenase, is responsible for the initial oxidation of the 23DCNB molecule. The 2,3-dichloronitrobenzene dioxygenase system (DcbAaAbAcAd) catalyzes conversion of 23DCNB to 3,4-dichlorocatechol (34DCC). Site-directed mutagenesis studies indicated that residue 204 of DcbAc is crucial for the substrate specificity of 23DCNB dioxygenase. The presence of glutamic acid at position 204 of 23DCNB dioxygenase is unique among Nag-like dioxygenases. Genetic, biochemical, and structural evidence indicate that the 23DCNB dioxygenase is more closely related to 2-nitrotoluene dioxygenase from Acidovorax sp. strain JS42 than to the 34DCNB dioxygenase from Diaphorobacter sp. strain JS3050, which was isolated from the same site as strain JS3051. A gene cluster (dcc) encoding the enzymes for 34DCC catabolism, homologous to a clc operon in Pseudomonas knackmussii strain B13, is also on the chromosome at a distance of 2.5 Mb from the dcb genes. Heterologously expressed DccA catalyzed ring cleavage of 34DCC with high affinity and catalytic efficiency. This work not only establishes the molecular mechanism for 23DCNB mineralization, but also enhances the understanding of the recent evolution of the catabolic pathways for nitroarenes.

A Nag-like dioxygenase initiates 3,4-dichloronitrobenzene degradation via 4,5-dichlorocatechol in Diaphorobacter sp. strain JS3050

The chemical synthesis intermediate 3,4-dichloronitrobenzene (3,4-DCNB) is an environmental pollutant. Diaphorobacter sp. strain JS3050 utilizes 3,4-DCNB as a sole source of carbon, nitrogen and energy. However, the molecular determinants of its catabolism are poorly understood. Here, the complete genome of strain JS3050 was sequenced and key genes were expressed heterologously to establish the details of its degradation pathway. A chromosome-encoded three-component nitroarene dioxygenase (DcnAaAbAcAd) converted 3,4-DCNB stoichiometrically to 4,5-dichlorocatechol, which was transformed to 3,4-dichloromuconate by a plasmid-borne ring-cleavage chlorocatechol 1,2-dioxygenase (DcnC). On the chromosome, there are also genes encoding enzymes (DcnDEF) responsible for the subsequent transformation of 3,4-dichloromuconate to β-ketoadipic acid. The fact that the genes responsible for the catabolic pathway are separately located on plasmid and chromosome indicates that recent assembly and ongoing evolution of the genes encoding the pathway is likely. The regiospecificity of 4,5-dichlorocatechol formation from 3,4-DCNB by DcnAaAbAcAd represents a sophisticated evolution of the nitroarene dioxygenase that avoids misrouting of toxic intermediates. The findings enhance the understanding of microbial catabolic diversity during adaptive evolution in response to xenobiotics released into the environment.

Structural basis for divergent C-H hydroxylation selectivity in two Rieske oxygenases

Biocatalysts that perform C–H hydroxylation exhibit exceptional substrate specificity and site-selectivity, often through the use of high valent oxidants to activate these inert bonds. Rieske oxygenases are examples of enzymes with the ability to perform precise mono- or dioxygenation reactions on a variety of substrates. Understanding the structural features of Rieske oxygenases responsible for control over selectivity is essential to enable the development of this class of enzymes for biocatalytic applications. Decades of research has illuminated the critical features common to Rieske oxygenases, however, structural information for enzymes that functionalize diverse scaffolds is limited. Here, we report the structures of two Rieske monooxygenases involved in the biosynthesis of paralytic shellfish toxins (PSTs), SxtT and GxtA, adding to the short list of structurally characterized Rieske oxygenases. Based on these structures, substrate-bound structures, and mutagenesis experiments, we implicate specific residues in substrate positioning and the divergent reaction selectivity observed in these two enzymes.

Rieske oxygenases are iron-dependent enzymes that catalyse C–H mono- and dihydroxylation reactions. Here, the authors characterise two cyanobacterial Rieske oxygenases, SxtT and GxtA that are involved in the biosynthesis of paralytic shellfish toxins and determine their substrate free and saxitoxin analog-bound crystal structures and by using mutagenesis experiments identify residues, which are important for substrate positioning and reaction selectivity.

Structural and Mechanistic Insights into Caffeine Degradation by the Bacterial N-Demethylase Complex

Caffeine, found in many foods, beverages, and pharmaceuticals, is the most used chemical compound for mental alertness. It is originally a natural product of plants and exists widely in environmental soil. Some bacteria, such as Pseudomonas putida CBB5, utilize caffeine as a sole carbon and nitrogen source by degrading it through sequential N-demethylation catalyzed by five enzymes (NdmA, NdmB, NdmC, NdmD, and NdmE). The environmentally friendly enzymatic reaction products, methylxanthines, are high-value biochemicals that are used in the pharmaceutical and cosmetic industries. However, the structures and biochemical properties of bacterial N-demethylases remain largely unknown. Here, we report the structures of NdmA and NdmB, the initial N₁- and N₃-specific demethylases, respectively. Reverse-oriented substrate bindings were observed in the substrate-complexed structures, offering methyl position specificity for proper N-demethylation. For efficient sequential degradation of caffeine, these enzymes form a unique heterocomplex with 3:3 stoichiometry, which was confirmed by enzymatic assays, fluorescent labeling, and small-angle x-ray scattering. The binary structure of NdmA with the ferredoxin domain of NdmD, which is the first structural information for the plant-type ferredoxin domain in a complex state, was also determined to better understand electron transport during N-demethylation. These findings broaden our understanding of the caffeine degradation mechanism by bacterial enzymes and will enable their use for industrial applications.