Recent Discoveries in Mono- and Dinuclear Nonheme Iron Enzymes: Emerging Mechanisms and Catalytic Strategies

Abstract

Nonheme iron enzymes encompass one of the most chemically versatile families of metalloenzymes, catalyzing an extraordinary range of oxidative transformations essential to metabolism, natural product biosynthesis, and environmental biodegradation. Despite their mechanistic diversity, these enzymes share a unifying principle: the controlled activation of molecular oxygen to drive selective substrate oxidation. Many mononuclear nonheme iron enzymes employ a conserved 2-His-1-carboxylate facial triad and frequently utilize 2-oxoglutarate (2OG) as a co-substrate to generate Fe(IV)=O intermediates that mediate diverse C–H and C=C bond oxidations. Organic cofactor-independent nonheme iron enzymes have been recognized for decades, with early examples including catechol dioxygenases and intradiol-cleaving enzymes. Structural and mechanistic advances over the past two decades have revealed that many of these cofactor-independent enzymes feature alternative 3-His or 4-His metal-binding motifs, which confer distinct electronic properties and enable oxygen activation pathways beyond those accessible to 2-His-1-carboxylate systems. Parallel advances in the study of dinuclear nonheme iron enzymes have illuminated equally rich mechanistic diversity, particularly within the ferritin-like (FDO) and heme-oxygenase-like (HDO) families. These diiron systems employ cooperative metal-metal interactions to form diiron-oxygen intermediates that enable multielectron oxidation chemistry beyond the reach of mononuclear centers. Structural and mechanistic studies of HDO-type enzymes involved in natural product biosynthetic pathways have revealed noncanonical coordination environments and substrate selectivity, underscoring the evolutionary and functional plasticity of nonheme iron catalysis. This review summarizes emerging insights into His-rich mononuclear enzymes and HDO-type diiron systems, highlighting their structural innovations, mechanistic principles, and collective contributions to expanding the frontiers of oxygen activation chemistry.

Introduction

Nonheme iron enzymes represent one of the most functionally diverse and mechanistically versatile classes of metalloenzymes, catalyzing a wide spectrum of oxidative transformations that include hydroxylation, halogenation, desaturation, epoxidation, ring-opening, and ring-expansion reactions [1–2]. These enzymes play central roles in primary metabolism, natural product biosynthesis, and the biodegradation of complex organic substrates [3–4]. Despite the variety of transformations they mediate, their unifying feature lies in the ability to couple molecular oxygen activation to substrate oxidation, frequently achieving regio- and stereo-chemical control.

In canonical mononuclear systems, the iron center is typically ligated by a highly conserved 2-His-1-carboxylate facial triad, which defines the archetypal coordination environment. This arrangement enables efficient binding and activation of oxygen, often in concert with an organic cofactor such as 2-oxoglutarate (2OG) that serves as a co-substrate and electron donor [5–6]. The 2OG-dependent nonheme iron enzymes constitute one of the most extensively studied families, with well-characterized mechanisms involving a high-valent Fe(IV)-oxo intermediate to catalyze substrate oxidation [6–9]. However, organic cofactor-independent nonheme iron enzymes that operate without 2OG or other redox-active cofactors have been recognized for decades, with early examples including catechol dioxygenases and intradiol-cleaving enzymes. More recently, structural characterization has revealed that many of these cofactor-independent enzymes utilize alternative metal-binding motifs, such as 3-His or 4-His, which alter the electronic environment and redox potential of the iron center [10–11]. These variations enable unique pathways for oxygen activation and substrate oxidation, broadening the catalytic landscape of nonheme iron chemistry beyond the constraints of Fe/2OG enzymes.

Beyond mononuclear systems, dinuclear nonheme iron enzymes represent another major branch of oxygen-dependent enzymes, employing cooperative mechanisms between two metal centers to achieve multi-electron oxidation of substrates. Diiron enzymes encompass multiple families including ferritin-like diiron oxygenases/oxidases (FDOs), heme-oxygenase-like diiron oxygenases/oxidases (HDOs), bacterial multicomponent monooxygenases, ribonucleotide reductases, fatty acid desaturases, and additional classes reviewed elsewhere [12–15]. Both FDO and HDO families share the capacity to bind and activate O₂, but differ in their protein folding, ligand environment, and catalytic strategies. Through electron transfer between the two irons located in the active site, these enzymes form reactive diiron-oxo, μ-peroxo, or bis-μ-oxo intermediates that serve as key oxidants for C–H, C=C, or C–C bond conversions [12]. Such dimetal arrangement may allow these enzymes to perform transformations that are often inaccessible to mononuclear enzymes. While extensive studies on FDOs such as methane monooxygenase and ribonucleotide reductase have established foundational principles, structural and mechanistic characterization of HDO-type diiron enzymes involved in the biosynthesis of natural products and the modification of secondary metabolites has revealed novel iron coordination, redox flexibility, and substrate selectivity [12]. Together, these features point to an evolutionary continuum linking mononuclear enzymes with alternative His-rich coordination motifs and diiron systems with nontraditional ligand environments—both representing parallel solutions to the challenge of controlled oxygen activation.

Mononuclear enzymes employing 3- or 4-His metal-binding motifs and HDO-type diiron enzymes exemplify the expanding frontiers of nonheme iron catalysis (Figure 1). Despite their distinct architecture, both families share common mechanistic features such as tuning iron redox potential, controlling reactive intermediate lifetime, and orchestrating precise substrate activation. In this review, we summarize advances in identifying, characterizing, and mechanistically dissecting these enzymes, with particular emphasis on structural and mechanistic studies reported within the past decade. Section A focuses on mononuclear nonheme iron enzymes featuring 3- or 4-His coordination motifs and their emerging catalytic profiles, while Section B highlights the structural and mechanistic landscape of HDO-type diiron enzymes.

Figure 1.

Distinct metal-binding motifs are found in organic-cofactor independent mono- and di-nuclear iron enzymes. These enzymes catalyze a broad range of oxidative transformations, including heteroatom (N or S) oxidation, C–X (X = C, N or O) bond formation, rearrangement, and oxidative decarboxylation among other reactions.

Section A: Mononuclear Nonheme Iron and Organic Cofactor-Independent Enzymes with a 3- or 4-His Fe Binding Motif

Coordination Sphere Effects on Oxygen Reactivity: Electronic and Mechanistic Principles

Mononuclear nonheme iron and oxygen-dependent enzymes typically share a common 2-His-1-carboxylate facial triad, which provides three open coordination sites for exogeneous ligands such as molecular oxygen and an organic cofactor (e.g., 2OG). This metal-binding motif establishes an optimal geometric and electronic environment for O₂ binding and activation, enabling the formation of high-valent Fe(IV)-oxo species central to oxidative catalysis [6–9]. Several mononuclear nonheme iron enzymes have been identified that diverge from this canonical architecture, featuring unconventional 3- and 4-His coordination motifs [10–11]. These alternative ligand environments modify the electronic structure of the metal center, thereby influencing its reactivity profile and oxygen activation mechanism. The discoveries of these His-rich motifs reveal that multiple mechanisms are observed in nature to achieve oxidative transformations, broadening the chemical and structural diversity of mononuclear nonheme iron enzymes.

Replacement of a carboxylate ligand with histidine fundamentally alters the electronic structure and reactivity of nonheme iron centers. In the canonical 2-His-1-carboxylate facial triad, the anionic, π-donating carboxylate (Glu or Asp) stabilizes higher oxidation states and facilitates formation of high-valent Fe(IV)-oxo species through strong electron donation to the metal d-orbitals [16]. By contrast, the neutral, σ-donating histidine residues characteristic of 3-His and 4-His motifs may reduce the overall anionic character of the coordination sphere, typically increasing the reduction potential of Fe(III) and destabilizing Fe(IV)-oxo intermediates. This shift in redox tuning can favor alternative oxygen activation pathways, including stabilization of Fe(III)-superoxo intermediates and altered O–O bond cleavage mechanisms [17].

In addition, the nature of substrate coordination to the iron center further modulates dioxygen reactivity. In enzymes such as thiol dioxygenases (CDO, MDO, MSDO), gentisate 1,2-dioxygenase (GDO), and diketone-cleaving enzyme (Dke1), the substrate itself directly coordinates to the iron through anionic functional groups (thiolate, phenolate, enolate). This substrate-derived ligation introduces significant electronic effects: the anionic substrate ligand increases electron density at the metal center, facilitating O₂ binding and activation while simultaneously positioning the substrate for direct attack by oxygen-derived intermediates [18–19]. The ordered sequential binding of substrate prior to dioxygen in these systems reflects the requirement for substrate coordination to prepare the iron center for O₂ activation. Conversely, in enzymes where substrates bind outside the first coordination sphere, such as anthraquinone ring-cleavage enzymes and carotenoid cleavage dioxygenases, substrate coordination effects are absent, and oxygen activation relies solely on protein-derived ligands. This fundamental distinction between substrate-as-ligand versus substrate-in-second-sphere mechanisms could thus account for significant differences in reaction kinetics, spin-state preferences, and intermediate stability across His-rich nonheme iron enzyme families.

Mononuclear nonheme iron and oxygen-dependent enzymes that diverge from the canonical 2-His-1-carboxylate architecture have been shown to catalyze diverse oxidative transformations including thiol oxidation, oxidative C–C or C–N bond cleavage, oxidative decarboxylation, and molecular rearrangements [10]. These reactions proceed independently of organic cofactors, with the protein scaffold providing the necessary redox tuning and active-site architecture for dioxygen activation. The mechanistic characterization of these systems offers valuable insights into how subtle changes in active-site geometry, metal coordination, and substrate binding mode expand the catalytic repertoire of nonheme iron enzymes. In this section, we discuss several examples of mononuclear nonheme iron and oxygen-dependent enzymes bearing 3- or 4-His motifs, organized by structural class and catalytic function as summarized in Figure 2.

Figure 2.

Summary of different reaction types catalyzed by mononuclear nonheme iron and organic cofactor-independent enzymes bearing a 3- or 4-His motif.

Thiol Dioxygenases

Thiol dioxygenases constitute a functionally important class of nonheme iron enzymes that catalyze the oxidation of thiols to sulfinates. Multiple families of thiol dioxygenases have been characterized, each with distinct substrate specificities, physiological roles, and evolutionary origins. Small-molecule thiol dioxygenases belonging to the CDO_I family (Pfam PF05995) include cysteine dioxygenase (CDO), which catalyzes the first committed step in cysteine catabolism by converting cysteine (A1) to cysteine sulfinic acid (A2), a key precursor in taurine biosynthesis [20]; 3-mercaptopropionate dioxygenase (MDO), which oxidizes 3-mercaptopropionate (A3) into 3-sulfinopropionate (A4) [21]; and mercaptosuccinate dioxygenase (MSDO), which oxidizes mercaptosuccinate (A5) into sulfinosuccinate (A6) [22]. In addition to these small-molecule thiol dioxygenases, distinct thiol dioxygenase subclasses include persulfide dioxygenase and 3-phosphoadenosine 5′-phosphosulfate (PAPS) reductase-associated thiol dioxygenase (also referred to as 3-mercaptopyruvate sulfurtransferase cysteine persulfide dioxygenase), which acts on protein-derived cysteine persulfides and plays roles in sulfide catabolism and mitochondrial function [23]. A mechanistically and structurally distinct family is represented by cysteamine dioxygenase (ADO), which belongs to the N-cysteinyl oxidase family (Pfam PF07847) and converts cysteamine (A7) into hypotaurine (A8) [24–25].

The functional distinctions between CDO_I and N-cysteinyl oxidase families are significant. In addition to hypotaurine biosynthesis, ADO catalyzes post-translational oxidation of N-terminal cysteine residues to trigger the Cys-Arg/N-degron pathway and functions as an O₂ sensor, exhibiting a dioxygen affinity (K_M, O₂ ~ 200-500 μM) orders of magnitude lower than CDO_I enzymes (K_M, O₂ ~ 10-30 μM) despite sharing a 3-His metal-binding motif [26–27]. This difference in O₂ affinity reflects the distinct physiological roles: CDO functions constitutively in cysteine catabolism, whereas ADO acts as an oxygen-responsive switch controlling protein degradation. The overall oxidation reactions executed by representative thiol dioxygenases are depicted in Figure 3A. As essential regulators of cellular thiol homeostasis in both prokaryotic and eukaryotic systems, thiol dioxygenases play important roles in sulfur metabolism.

Figure 3.

(A) Reactions catalyzed by four TDOs: CDO, MDO, MSDO, and ADO respectively. (B) Active site of CDO (PDB ID: 2ATF), depicting the chelation of metal center, nickel (green sphere), by three histidine residues (cyan sticks) and three water molecules (red spheres). The crosslink between a cysteine and a proximal tyrosine residue (pink sticks) is depicted. (C) Proposed mechanisms for cysteine oxidation catalyzed by CDO, illustrating key intermediates and possible routes which also serve as general models for other TDO-mediated reactions.

The catalytic requirements and oxygen incorporation stoichiometry of thiol dioxygenases have been established through biochemical and isotopic labeling studies. In vitro reconstitution demonstrated that thiol dioxygenase activity required only ferrous iron and molecular oxygen [28]. Isotopic labeling experiments with ¹⁸O₂ confirmed that both oxygen atoms in the sulfinate product originate from molecular oxygen, firmly establishing the dioxygenase function [29]. Electron paramagnetic resonance (EPR) studies established that CDO maintained a relatively stable low-spin Fe(II) resting state under aerobic conditions, with substrate binding to the enzyme preceding O₂ coordination, further confirmed by spectroscopic and kinetic analyses [30]. The obligate-ordered binding was believed to tune the electronic structure of the iron center, enabling controlled O₂ activation. To probe the reactive species involved in the mechanistic pathway, transient stopped-flow spectroscopy was utilized, and an intermediate absorbing near 500 and 640 nm during the reaction of CDO with cysteine under aerobic conditions was detected, though attempts to trap and characterize this intermediate using freeze-quench Mössbauer methods were unsuccessful [31]. To further gain some insights into potential reactive iron species, metal substitution studies revealed that ADO retained substantial catalytic activity when Fe(II) was replaced by Co(II), providing evidence that catalysis proceeded via a non-high-valent pathway rather than through an Fe(IV)-oxo intermediate typical of 2-His-1-carboxylate enzymes [32].

Crystal structures of thiol dioxygenases revealed that the metal center adopts a 3-His facial coordination geometry, with three histidine residues ligating the iron in an approximately octahedral arrangement, as exemplified by CDO (PDB ID: 2ATF) in Figure 3B [33]. Structural studies further identified a conserved Ser-His-Tyr (SHY) motif in small-molecule thiol dioxygenases (CDO_I family), with the serine and tyrosine positioned to facilitate proton transfer and hydrogen bonding interactions that modulate substrate and ligand coordination [34]. A unique feature of mammalian CDO is a covalent Cys-Tyr crosslink between Cys93 and Tyr157 adjacent to the metal-binding site (Figure 3B) [33]. This crosslink forms autocatalytically during enzyme maturation and enhances catalytic efficiency by approximately 10-fold [35–36]. Mutagenesis and EPR spectroscopy supported a role for the crosslinked residues in orienting cysteine toward the iron center for selective oxidation [37]. Interestingly, crosslink formation has been demonstrated in mammalian ADO through incorporation of 3,5-difluoro-tyrosine (F2-Tyr) and detection of fluoride release [38], though the proposed Cys–Tyr crosslink reported in one ADO crystal structure has not been observed in any of the ten other available ADO structures, leaving its functional relevance uncertain [38–39].

Mechanistic understanding of thiol dioxygenase catalysis has evolved significantly over the past 15 years. A cysteine persulfenate intermediate was observed in a crystal structure of CDO soaked with cysteine (PDB ID: 3ELN, 1.4 Å resolution), initially suggesting a mechanistic pathway involving sulfur attack on the Fe-proximal oxygen to form a thiadioxirane intermediate [40]. Independent replication of persulfenate-bound CDO structures under variable pH conditions further supported this structural observation [41]. However, subsequent studies challenged the catalytic relevance of this species. Thawing of crystals containing the putative persulfenate did not yield cysteine sulfinic acid, indicating that this species was likely off-pathway or not catalytically competent [42]. Computational analysis demonstrated that the energy barrier for the persulfenate pathway was prohibitively high (~30 kcal/mol) relative to alternative mechanisms involving superoxide chemistry [43].

To further investigate the mechanism, chemical rescue experiments were performed to generate a putative Fe(III)-superoxide intermediate in CDO, characterized by UV-Vis absorption at 565 nm and parallel-mode EPR spectroscopy revealing an integer-spin (S = 3/2) signal inconsistent with an Fe(II)-persulfenate [44]. The decay kinetics of this intermediate matched product formation rates, and both the ligand-to-metal charge transfer (LMCT) intensity and EPR characteristics supported assignment as an Fe(III)-superoxo species. Moreover, stopped-flow UV-Vis studies identified transient LMCT features (500 and 640 nm) upon rapid mixing of substrate-bound CDO with dioxygen [31]. Subsequent DFT calculations identified a cyclic [Fe–O–O–S]-cysteine intermediate with a quintet spin state that explained these spectroscopic features [34].

More recent computational [45], spectroscopic [34], and synthetic model complex studies [46] have further refined the consensus mechanism. The current mechanistic picture centers on an Fe(III)-superoxo intermediate that forms a four-membered cyclic [Fe–O–O–S] species (Figure 3C, Path B). Subsequent O–O bond cleavage generates a transient Fe(IV)-oxo intermediate that completes sulfinate formation through oxygen atom transfer. The persulfenate pathway (Path A, shown with reduced emphasis in Figure 3C) is now considered unlikely based on the combined weight of experimental and computational evidence accumulated over the past decade and a half. The consensus view among researchers in the field has shifted toward Path B, which better accounts for spectroscopic observations, kinetic data, and computational energetics.

Sulfoxide Synthases

Ergothioneine and ovothiol A are histidine-derived thiol-containing natural products known for their potent antioxidant and anti-inflammatory activities [47]. Ergothioneine was originally isolated from ergot [48], and its biosynthetic gene cluster was recently identified [49]. On the other hand, ovothiol A was isolated from the eggs and ovaries of marine invertebrates and trypanosomatids [50], with its complete biosynthetic gene cluster still under investigation [51]. The biosynthesis of both compounds involves an unusual oxidative carbon-sulfur (C–S) bond formation catalyzed by sulfoxide synthases. Specifically, Egt1 and OvoA can catalyze the oxidative coupling of l-cysteine and hercynine (A7) to afford A8. Additionally, OvoA is capable of catalyzing the oxidative coupling between l-cysteine and histidine (A9) to generate A10. EgtB catalyzes a related oxidative coupling between A7 and γ-glutamate-cysteine to yield (A11) (Figure 4A) [49, 51–52]. Given the physiological significance of ergothioneine and ovothiol A, considerable efforts have been devoted to elucidating the mechanistic basis of sulfoxide synthase-catalyzed oxidative C–S bond formation.

Figure 4.

(A) Reactions catalyzed by sulfoxide synthases, Egt1, EgtB, and OvoA, depicting oxidative C–S bond formation between thiol and imidazole substrates. (B) Structure of EgtB (PDB ID: 6O6M) showing the chelation of metal center, iron (orange sphere), by three histidine residues (cyan sticks) and three water molecules (red spheres). (C) Proposed mechanisms of OvoA-catalyzed C–S bond formation, representing the catalysis mediated by sulfoxide synthases.

The crystal structures of EgtB (PDB ID: 4X8D) and OvoA (PDB ID: 8KHQ) (Figure 4B) revealed that sulfoxide synthases share a three-histidine iron-binding motif, akin to that observed in thiol dioxygenases [53–54]. Additionally, crystal structures identified a catalytically important tyrosine residue positioned near the metal center, implicated in molecular oxygen binding and activation [53–54]. Intriguingly, a single point mutation of this tyrosine residue to O-methyltyrosine (OvoA_Y417MtTyr) or phenylalanine (EgtB_Y337F) enabled sulfoxide synthases to function as thiol dioxygenases, suggesting that both enzyme classes could share a common reaction intermediate [55–56]. Investigations employing isotope-labeled substrates further refined mechanistic understanding [54–55, 57–59]. Briefly, an inverse deuterium kinetic isotope effect was observed when OvoA_Y417MtTyr utilized deuterium-labeled histidine as the substrate, indicating the catalytic tyrosine acted as a redox mediator and that C–S bond formation preceded sulfur oxidation [55]. However, because both the pK_a and redox potentials of tyrosine and O-methyltyrosine are different, it is difficult to differentiate the contributions of the two factors. In addition, alternative explanations rooted in steric effects between the methyl group and hydrogen atom have left this interpretation under debate.

Recently, an intermediate spin (S = 1) Fe(IV) species was trapped and characterized through transient kinetics and spectroscopic techniques including Mössbauer, EPR, and X-ray absorption near edge structure (XANES) [60]. This species supported a mechanistic model in which O₂ activation coincides with oxidation of the cysteine sulfur to generate an Fe(IV)–sulfoxide intermediate (A12 or A13) (Figure 4C). This complex may be regarded as in resonance with a ferric sulfur-centered radical complex (A14) that performs a radical attack on imidazole of histidine to form the C–S bond. Subsequently, a catalytic tyrosine residue assists proton transfer from the imidazole nitrogen to Fe(III)=O, followed by re-aromatization of imidazole to yield the final sulfoxide product.

Diketone-cleaving Enzyme

Diketone-cleaving enzyme (Dke1) from Acinetobacter johnsonii was first isolated and characterized about two decades ago [61]. Dke1 mediates the cleavage of β-dicarbonyl substrates, such as acetylacetone (A15), a compound with neurotoxic and immunomodulatory properties (Figure 5A) [62], thus playing an essential role in the microbial degradation and detoxification of acetylacetone.

Figure 5.

(A) Reaction catalyzed by Dke1, illustrating the oxidative cleavage of β-diketone substrates (e.g., acetylacetone into methylglyoxal and acetate). (B) Proposed mechanisms for Dke1-mediated C–C bond cleavage, including the formation of key intermediates such as enzyme-bound enolate (A16), peroxidate (A17), dioxetane (A19) (Path A), or epoxide (A20) and Fe(IV)-oxo species (Path B).

Dke1 requires only O₂ and iron as cofactors for its catalytic function [61]. Stoichiometric analysis confirmed consumption of a single equivalent of O₂ per substrate converted, and isotope-labeling experiment using ¹⁸O₂ established that both oxygen atoms incorporated into the products were derived from O₂, confirming its classification as a dioxygenase [61, 63]. The crystal structure of Dke1 (PDB ID: 3BAL) showed that three histidine residues coordinate the metal center [64–65]. Furthermore, substrate binding studies demonstrated that acetylacetone displaced a water ligand, generating a five-coordinate site for O₂ binding at the Fe(II) center [64, 66].

Biophysical and spectroscopic investigation using UV-Vis absorption revealed a weak band, which was indicative of an Fe(II)-β-keto-enolate complex (A16). Kinetic analyses suggested that the conversion of this enzyme-bound enolate to an alkyl peroxidate intermediate (A17) constituted the rate-limiting step of the reaction [67]. Further mechanistic insight was gained by substituting electronic groups on the substrate. Specifically, replacing the acetyl group in A18 with a trifluoroacetyl group altered the bond cleavage site, implying that the bond cleavage was favored adjacent to the most electron-deficient carbonyl group [63]. Mechanistically, two reaction pathways have been proposed (Figure 5B). Path A features nucleophilic attack by superoxide on the carbonyl carbon to generate a dioxetane intermediate (A19), which then decomposes to the final products (methylglyoxal and acetate). Path B undergoes the formation of an epoxide intermediate (A20) and an Fe(IV)-oxo species, followed by epoxide ring-opening and further oxidation to yield an ester intermediate with subsequent oxo attack by Fe(IV)-oxo at a newly formed C=C bond [66–67].

Gentisate 1,2-Dioxygenases

Gentisic acid serves as a central intermediate in the aerobic biodegradation pathways of a diverse range of aromatic compounds including dibenzofuran, naphthalene, salicylate, and anthranilate. The microbial degradation of gentisate is enabled by gentisate 1,2-dioxygenase (GDO), which catalyzes an oxidative aromatic ring cleavage of gentisic acid (A21) between the carboxyl and vicinal hydroxyl groups to produce maleylpyruvate (A22) (Figure 6A) [68]. Maleylpyruvate is then isomerized to fumarylpyruvate and hydrolyzed to fumarate and pyruvate, allowing the integration of aromatic compounds into the citric acid cycle [69]. Although GDOs from various species have been isolated and characterized [68, 70–75], only a few mechanistic studies have been conducted to elucidate the oxidative cleavage transformations enabled by GDOs.

Figure 6.

(A) Reaction catalyzed by GDO, showing oxidative cleavage of the aromatic ring of A21 between the carboxyl and adjacent hydroxyl group to form A22. (B) Proposed mechanism for reactions catalyzed by GDO.

GDOs were established as dioxygenases by ¹⁸O₂ isotope-labeling experiments in which both oxygen atoms were derived from O₂ [68, 75]. EPR spectroscopy revealed an Fe(II) center at the active site, and the enzyme formed an EPR-active complex (S = 3/2) upon nitric oxide binding. In addition, hyperfine broadening from ¹⁷O-labeled substrate was observed when ¹⁷O was incorporated into either the carboxylate or hydroxyl group, implying that substrate bound to the iron in a bidentate fashion through its carboxylate and adjacent hydroxyl groups [68]. Structural and mutational analyses across species confirmed that GDOs universally possessed a three-histidine iron-binding motif. Site-directed mutagenesis of any of these histidines abolished catalytic activity [71, 73–74]. The crystal structure of GDO (PDB ID: 1EY2) depicted the 3-His ligand arrangement [72]. Taken together, the ring cleavage is proposed to proceed through simultaneous binding of A21 and O₂ to Fe(II) at the enzyme active site. Subsequently, nucleophilic attack at C2 carbon of A21, dioxygen bond cleavage, and oxygen insertion assisted by ketonization of the C5 hydroxyl group. Lastly, the reaction undergoes ring opening to afford the final product A22 (Figure 6B) [68].

Anthraquinone Ring Cleavage Enzymes

Anthraquinones are widespread natural products recognized for their potent biological activities including anticancer, antiviral, antioxidative, and anti-inflammatory effects [76]. The enzymatic cleavage of the anthraquinone ring is the key transformation in the biosynthesis of structurally diverse seco-anthraquinone products [77]. Until now, only four anthraquinone ring cleavage enzymes have been characterized. Of these, AgnL3, GedK, and BTG13 process structurally similar anthraquinones, converting A23 or A24 into A25 or A26, respectively [78–81], while PaD acts on a monocyclic hydroquinone A27 to generate A28 (Figure 7A) [82].

Figure 7.

(A) Reactions catalyzed by anthraquinone ring cleavage enzymes, AgnL3, GedK, BTG13, and PaD, illustrating oxidative ring opening of core anthraquinones (AgnL3/GedK/BTG13) and a monocyclic hydroquinone (PaD). (B) Crystal structure of BTG13 (PDB ID: 7Y3W) showing the iron center (orange sphere) coordinated by four histidine residues (cyan sticks), a carboxylated lysine residue (pink sticks), and a water molecule (red sphere) within the active site. (C) Proposed mechanisms for anthraquinone ring cleavage enzymes.

Early mechanistic proposals suggested a Baeyer-Villiger oxidation, in which an oxygen atom insertion yielded a seven-membered lactone that was subsequently hydrolyzed [78–79]. However, through gene disruption and enzymatic assays, it was later demonstrated that a reductase first catalyzed reduction of the C10 carbonyl of anthraquinone to a hydroxyl group before C–C bond cleavage takes place. Anthraquinone ring cleavage enzymes are iron- and O₂-dependent [80]. Crystal structures of BTG13 (PDB ID: 7Y3W) and PaD (PDB ID: 8Z4Q) revealed an unusual active site coordination environment, where the metal center was ligated by four histidine residues, a carboxylated lysine, and a water molecule (Figure 7B) [81–82]. In addition, ¹⁸O₂ isotope-labeling experiment confirmed that both oxygen atoms in the product were derived from O₂ [80].

Along with quantum mechanical/molecular mechanical (QM/MM) calculations and molecular dynamics (MD) simulations, two reaction mechanisms for anthraquinone ring cleavage were proposed (Figure 7C) [80–84]. While O₂ activation of Fe(II) was involved in both proposed mechanisms, Path A is initiated by deprotonation at C10 position, likely by a basic residue at the active site, to generate a conjugated anion intermediate. Subsequently, a single electron transfer to the Fe(III)-superoxo occurs to generate a radical intermediate. This radical reacts with the Fe(II)-superoxo to form A29 followed by a Michael addition to generate a dioxetane (A30). Subsequent rearrangement and protonation lead to the formation of the final product. Differently, Path B is initiated by a hydrogen atom abstraction by Fe(III)-superoxo, followed by O–O bond cleavage to generate Fe(IV)-oxo. This Fe(IV)-oxo abstracts a hydrogen atom from the hydroxyl group, generating an oxygen radical that attacks the adjacent double bond to form an epoxide (A31). Next, C–C bond cleavage and proton-coupled electron transfer afford a lactone intermediate (A32), yielding the final product with the release of water.

Carotenoid Cleavage Oxygenases

Carotenoids function as accessory pigments in photosynthesis and act as antioxidants in plants and animals. Carotenoid cleavage oxygenases (CCOs) catalyze regioselective oxidative cleavage of a double bond within a polyene backbone, introducing carbonyl functionality and generating apocarotenoids (Figure 8A). Apocarotenoids function as regulatory molecules in cellular signaling and growth regulation [85]. Cleavage of β,β-carotene to form retinal was the first CCO reaction characterized, but numerous CCOs spanning bacteria, plants, and animals have since been described, each exhibiting varied regioselectivity and substrate scope [86–87].

Figure 8.

(A) Reactions catalyzed by CCOs, illustrating regioselective oxidative cleavage of double bonds in carotenoid substrates to generate apocarotenoid products. (B) Crystal structure (PDB ID: 2BIW) shows chelation of iron (orange sphere) by four histidine residues (cyan sticks) and a water molecule (red sphere) within the active site. (C) Proposed monooxygenase mechanism for CCOs-mediated catalysis. (D) Proposed dioxygenase mechanism for CCOs-mediated catalysis.

Structural studies revealed four histidine residues coordinating the iron center within the CCO active site (Figure 8B) [88–95]. Additionally, crystal structures of two CCOs, VP14 (PDB ID: 3NPE) and NOV1 (PDB ID: 5J53), revealed that iron bound to dioxygen in the four-His motif despite debate over the nature of O₂ binding in the resting state of the enzyme [90, 92]. EPR spectroscopic investigations of Fe(II)-NOV1 exposed to NO identified a stable S = 3/2 {Fe-NO}⁷ complex, supporting the model that O₂ bound iron in the absence of the substrate [92]. However, Mössbauer and X-ray absorption spectroscopic studies, along with the protein structure of NOV2, suggested that the resting state may feature a five-coordinate iron where O₂ did not bind in the absence of substrate [93]. Isotope-labeling experiments with ¹⁸O₂ have been extensively used to distinguish monooxygenase from dioxygenase mechanisms. However, rapid isotopic exchange between labeled carbonyl products and water has made mechanistic resolution difficult, leading to two distinct reaction possibilities, namely mono- vs. di-oxygenase pathway (Figure 8C and 8D) [89, 96–97]. The monooxygenase mechanism features formation of an epoxide intermediate at the scissile alkene bond, followed by water-mediated ring-opening and oxidative cleavage to yield the final products (Figure 8C). In contrast, the dioxygenase mechanism (Figure 8D) is initiated by O₂ activation of Fe(II), creating a Fe(III)-superoxo species, followed by stepwise electron transfer to the Fe(III) species, generating a carbocation intermediate (A33), which marks a branching point for two different reaction paths. In Path A, the carbocation undergoes nucleophilic attack by water followed by Criegee rearrangement, forming a secondary carbocation (A34), which is subsequently quenched by the Fe(II)-OH folowed by further breakdown of the intermediate to afford final products. Alternatively, Path B involves the dioxetane (A35) formation followed by [2+2] cycloelimination to yield the final products.

Others

Beyond thiol oxidation, C–S bond formation, and oxidative cleavage discussed above, enzymes featuring three- or four-histidine ligand motifs catalyze additional mechanistically diverse transformations. For example, TqaM was proposed to utilize a three-His iron-binding motif to catalyze an oxidative decarboxylation for conversion of A36, A37, and A38 into non-canonical amino acids, A39, A40, and A41, respectively, as shown in Figure 9A [98–99]. A crystal structure of TqaM in complex with Mn and substrate supported this proposal, demonstrating that the metal center was chelated by three histidine residues. In addition, the carboxylate and hydroxyl groups of the substrate also coordinated the metal (Figure 9C). Isotope-labeling experiments using ¹⁸O₂ revealed that one oxygen atom from O₂ was incorporated into the final product. Building on these results, biochemical assays using substrate analogs and kinetic studies using a C2-deuterated isotopologue demonstrated that TqaM was stereo- and regioselective for H atom abstraction from the C2 position of the substrate, and that C–H bond cleavage was at least partially rate-limiting [99]. Taken together, a mechanism was proposed for TqaM catalysis (Figure 9D). The reaction is initiated by the coordination of substrate and molecular oxygen to Fe(II) to generate A42, which abstracts a hydrogen atom from the C2 position of the substrate to generate a C2 radical (A43). Then, A43 further forms a keto acid intermediate (A44). Subsequently, C–O bond formation followed by decarboxylation results in the generation of product.

Figure 9.

(A) Reaction catalyzed by TqaM, highlighting the oxidative decarboxylation involved in the biosynthesis of 2-aminoisobutyric acid and its derivatives. (B) Reaction scheme of OzmD, demonstrating the oxidative rearrangement in oxazinomycin biosynthesis. (C) Crystal structure of TqaM (PDB ID: 9M09) showing chelation of manganese (purple sphere) by three histidine residues (cyan sticks), A36 (magenta sticks), and a water molecule (orange sphere) within the active site. (D) Proposed mechanism of oxidative decarboxylation catalyzed by TqaM. (E) Proposed mechanism of oxidative rearrangement catalyzed by OzmD.

Another notable enzyme, OzmD, catalyzes an oxidative rearrangement that converts A45 to A46 in the biosynthesis of oxazinomycin (Figure 9B) [100]. Structural modeling suggested that OzmD features a four-His ligand environment, a hypothesis supported by mutagenesis studies that individual replacement of any four predicted iron-binding histidine residues with alanine abolished catalytic activity. To probe OzmD’s mechanism, ¹⁸O₂ isotope-labeling experiment revealed that only one oxygen atom from molecular oxygen was incorporated into the final product, followed by determination of the incorporation site by derivatization methods. Mechanistic analysis also uncovered the generation of cyanide as a byproduct, and further isotope-labeling experiment confirmed that this cyanide was derived from the C3-amino group of the substrate [100]. A current mechanistic proposal suggests that OzmD catalysis is initiated by substrate coordination to iron center, followed by dioxygen activation. Subsequently, a semiquinone radical (A47) forms an endoperoxide bridge with the superoxo radical (A48). Heterolysis of the O–O bond then results in nitrile formation (A49). Subsequently, an intramolecular Michael addition occurs on the carbamate group, forming the 1,3-oxazine ring (A50), and aromatization yields final product and cyanide (Figure 9E).

Section B: Heme-oxygenase-like Diiron Oxygenases/Oxidases (HDOs)

Structural Architecture and Iron-Iron Distance Effects

Both ferritin-like diiron oxygenases/oxidases (FDOs) and heme-oxygenase-like diiron oxygenases/oxidases (HDOs) utilize dinuclear irons [12–14], yet they have evolved distinct structural frameworks. FDOs feature a highly conserved four-helix bundle as their structural scaffold, which coordinates the dinuclear irons using a set of carboxylate (usually glutamate) and histidine residues, often described as a 4E/3H ligand pattern (Figure 10A and 10B) [101]. These residues are positioned within four core α-helices that form the active site scaffold. In contrast, HDO enzymes feature labile metal centers and a dimeric architecture characterized by a seven-helix bundle fold. HDOs coordinate dinuclear metals via three histidines and three carboxylates (Figure 10C and 10D) [102].

Figure 10.

(A–B) Structure of the FDO diiron scaffold with AurF as an example (PDB ID: 3CHH). (C–D) Structure of the HDO diiron scaffold using SznF as an example (PDB ID: 6VZY). (E) Sequence alignment of the core metal-binding α-helices from HDOs discussed in this review.

A critical structural parameter influencing dioxygen activation and reactivity in HDO-type enzymes is the iron-iron distance within the diiron cluster. Variations in metal-metal separation directly impact the stability and electronic structure of bridging oxygen intermediates, including μ-peroxo and bis-μ-oxo species. In general, shorter iron-iron distances (2.5–3.0 Å) favor formation of diamond-core bis-μ-oxo intermediates capable of multi-electron oxidation chemistry, while longer distances (3.5–4.5 Å) tend to stabilize μ-peroxo intermediates with distinct reactivity profiles. The protein scaffold controls iron-iron separation through the geometry and identity of bridging and terminal ligands, including carboxylates, histidines, and water molecules.

Structural characterization of HDO enzymes has revealed significant variations in iron-iron distances across different family members. For example, the diferric state of SznF exhibits an iron-iron distance of approximately 3.3 Å [102], while BesC shows distances ranging from 3.4–3.8 Å depending on substrate occupancy [103]. UndA structures reveal iron-iron distances of ~3.5 Å in the resting state [104]. These variations correlate with differences in intermediate stability and substrate selectivity. In SznF, the relatively short iron-iron distance facilitates formation of a stable μ-peroxodiiron(III/III) intermediate (P) that mediates sequential N-hydroxylations [105]. In contrast, the longer iron-iron distances observed in UndA and BesC may contribute to the formation of more reactive peroxo intermediates that undergo rapid O–O bond cleavage to generate high-valent iron-oxo species capable of C–H abstraction and C–C bond cleavage.

The structural plasticity of the HDO scaffold, particularly the conformational flexibility of core helix α3, allows for dynamic changes in iron-iron distance upon substrate binding and cofactor assembly [102–104, 106–107]. This dynamic regulation of metal-metal spacing represents a key mechanism by which HDO enzymes tune their reactivity toward diverse substrates and oxidative transformations. Understanding how small changes in coordination geometry propagate to altered iron-iron distances and reactivity patterns remains an active area of investigation with implications for both mechanistic enzymology and biocatalyst design. At the structural core, helices α1–α3 constitute the primary metal-binding scaffold. Core helix α1 typically contributes one histidine and one carboxylate ligand; α2 provides an additional carboxylate residue; and α3 donates three more metal-binding residues (Figure 10E). Surrounding these core elements are four auxiliary helices (aux α1–α4), which lend additional stability and define the substrate-access channel. The conformation of core helix α3 is sensitive to metal occupancy, adopting an ordered α-helical structure in the presence of the second metal but becoming disordered or conformationally variable in metal- or substrate-free forms [102–104, 106–107].

As a newly emerging family of diiron-cluster oxygenases/oxidases, HDO enzymes exhibit catalytic diversity, performing a wide range of oxidative transformations on structurally varied substrates (Figure 11). For instance, SznF HDO domain functions as an N^ω-methyl-L-arginine N-hydroxylase, catalyzing B1 to form B2 [102, 105, 108–109]. RohS acts as an N-oxygenase, converting 2-aminoimidazole (B3) to azomycin (B4) [110–112]. BelK and HrmI are l-lysine N-oxygenases, catalyzing regioselective N^ε-oxygenation of B5 to produce the nitro group in B6 [107, 113–115]. UndA catalyzes oxidative decarboxylation of medium-chain fatty acids (B7) to generate terminal alkenes (B8) [104, 116–117]. VarO mediates peptide dehydrogenation (B9→B10) in thioamidated ribosomal peptide biosynthesis [118]. In the biosynthetic pathway of l-β-ethynylserine, BesC mediates C–C bond cleavage of chlorinated lysine (B11), yielding the terminal alkyne intermediate (B12) [103, 119–120]. FlcE catalyzes both an oxidative decarboxylation and N-hydroxylation of substrate B13 to yield product B14 [121]. Additionally, AetD and SktA convert tryptophan derivatives into nitrile (B15→16) and the methylated product (B17→B18) [106, 122–124]. PolF performs sequential desaturation and azetidination to yield compound (B19→B20) [125–127], and CADD catalyzes the formation of p-aminobenzoate (B22) from a protein-derived tyrosine residue B21 [128–132]. These examples highlight the extraordinary functional versatility of HDO enzymes. This section summarizes the mechanistic and structural characteristics of representative HDO enzymes, emphasizing cofactor coordination, substrate specificity, and reaction diversity.

Figure 11.

Summary of reactions catalyzed by HDO enzymes discussed in this review.

SznF

SznF, also known as StzF, is a multidomain metalloenzyme essential for the biosynthesis of streptozotocin [105, 108–109]. SznF catalyzes both hydroxylation and rearrangement of the guanidine moiety of N^ω-methyl-l-arginine (l-NMA, B1), leading to the formation of an N-nitrosourea derivative (B24) (Figure 12A). LC-MS analysis using synthetic standards confirmed the reaction intermediates, indicating that SznF catalyzes two sequential N-hydroxylation events at the N^δ and N^ω′ positions, followed by an oxidative rearrangement to generate the N-nitrosourea derivative (B1→B23→B2→B24, Figure 12A). Isotopic labeling experiments with ¹⁸O₂ and ¹⁵N-labeled B1 revealed that each N-hydroxylation event consumes one equivalent of molecular oxygen, and the third equivalent provides the ureido oxygen incorporated into the product B24 [108–109].

Figure 12.

(A) Sequential N-oxygenations and oxidative rearrangement catalyzed by HDO and cupin domain in SznF, respectively. (B) Two possible mechanisms proposed for SznF HDO domain-catalyzed N-hydroxylation.

X-ray crystallography revealed that SznF is a homodimeric, tridomain protein comprising two N-terminal HDO domains and a C-terminal cupin domain. Site-directed mutagenesis further demonstrated that HDO domains catalyze N-hydroxylations, whereas the C-terminal cupin domain mediates the subsequent rearrangement and N–N bond formation [102, 109]. Freeze-quench coupled Mössbauer spectroscopy and stopped-flow absorption identified a catalytically relevant μ-peroxodiferric intermediate (P), shared with FDO enzymes [105]. This peroxo intermediate is used for both hydroxylation steps observed in SznF. Structures of SznF with both iron cofactors in the active revealed the presence of a seventh coordinating ligand in the HDO domain, providing insight into domain-specific features and cofactor dynamics [105].

Despite significant progress, key aspects of the SznF mechanism remain to be fully defined. The current proposed models focus on initial N-hydroxylation step, implicating guanidine protonation and substrate-mediated proton transfer to the peroxide intermediate, possibly via sequential or concerted mechanisms (Figure 12B). Following guanidine protonation, in Path A, the proton at N^δ position of B1 protonates intermediate P whereas the proton at N^ω′ position serves as the proton donor in Path B. In either case, subsequent nucleophilic attack from N^δ to oxygen, tautomerization and protonation steps complete the reaction, yielding N-hydroxylated product B23.

RohS, HrmI, and BelK

In addition to SznF, three HDO-catalyzed N-oxygenases have been identified, including RohS, HrmI and BelK. RohS was first reported in 2018 during studies on the biosynthesis of the nitroimidazole antibiotic azomycin, where it was shown to catalyze the oxidation of 2-aminoimidazole (B3) to azomycin (B4) (Figure 13A) [110–112]. While the detailed mechanism of RohS remains to be fully elucidated, mechanistic proposals based on related N-oxygenases suggest that the conversion B3 to B4 occurs via a stepwise process involving hydroxylamine (B25) and nitroso (B26) intermediates [110].

Figure 13.

(A–B) N-oxygenation of 2-aminoimidazole (B3) to azomycin (B4) and l-lysine (B5) to generate 6-nitronorleucine (B6) catalyzed by RohS, HrmI and BelK, respectively. (C) Possible peroxodiferric intermediate (P’) structures in the HrmI catalyzed reaction.

HrmI and BelK were identified as N-oxygenases in the biosynthetic pathways of hormaomycin and belactosin A where they catalyze the N^ε-oxygenation of l-lysine (B5) to generate 6-nitronorleucine (B6) (Figure 13B) [107, 113–115]. BelK and HrmI represent the first identified diiron nitro-forming enzymes that act on free amino acids. The observed conversion of N-6-hydroxy-l-lysine (B27) indicates that B27 likely serves as an on-pathway intermediate enroute to nitro product B6.

The crystal structure of HrmI revealed a diiron cluster coordinated by three histidine and three carboxylate ligands. Biochemical and spectroscopic analyses showed that diferrous HrmI reacts with O₂ in the absence of substrate to form intermediate (P) (Figure 13C) [107]. This intermediate subsequently coverts into a second species exhibiting optical and Mössbauer features consistent with an activated peroxodiferric adduct (P’), though its precise structure remains to be carefully evaluated.

UndA

UndA was discovered during screens for enzymes capable of generating terminal alkenes from fatty acid precursors. It catalyzes the oxidative decarboxylation of dodecanoic acid (B7) to produce undecane (B8) and carbon dioxide (Figure 14A) [104, 116–117]. UndA also catalyzes conversion of 2-hydroxydodecanoic acid (B28) to 1-undecanal (B29), while exhibiting minimal activity for structurally similar 3-hydroxydodecanoic acid (B30) or 2,3-dodecenoic acid (B31), indicating essential roles of C3 carbon substitution [117]. Mechanistic analysis using deuterated substrates revealed retention of both deuterium atoms at C2, whereas deuterium labeling at C3 resulted in the loss of one deuterium atom. Single-turnover experiments demonstrated that the formation of terminal alkene and CO₂ required UndA, reductant, and fatty acid substrate, but no accessory proteins [117]. Moreover, upon exposure to O₂, the diferrous enzyme forms an intermediate (P), which converts to an Fe₂(III/IV) species [104, 116]. Computational and spectroscopic studies suggest that a peroxo or hydroperoxo species initiates either C3-hydrogen abstraction (Path A) or carbonyloxy radical (Path B) formation. In Path A, a C3 centered radical enables alkene formation via one of four possible intermediates. Alternatively, a Kolbe-like decarboxylation pathway (Path B) initiated by carbonyloxy radical can be envisioned (Figure 14B) [104, 116].

Figure 14.

(A) UndA catalyzes oxidative decarboxylation of selected fatty acid substrates. (B) Possible mechanisms for UndA-mediated oxidative decarboxylation.

VarO

VarO identified in Actinobacteria catalyzes the dehydrogenation of thiovarsolin B and D to form thiovarsolin A and C via introducing a trans double bond between the Cβ and Cγ positions of the arginine side chain (Figure 15) [118]. Gene knockout experiments demonstrated that inactivation of VarO led to the accumulation of thiovarsolin B and D, confirming the role of VarO [118]. However, it remains unclear whether VarO acts on the precursor peptide before or after proteolytic maturation.

Figure 15.

Dehydrogenation of thiovarsolins catalyzed by VarO.

BesC

BesC identified from Streptomyces cattleya was found to catalyze oxidative C–C bond cleavage of 4-chloro-l-lysine (B11) to yield 4-chloroallylglycine (B12), along with ammonia and formaldehyde (Figure 16A) [103, 119–120]. Spectroscopic studies including transient absorption and Mössbauer spectroscopies supported the formation of the intermediate P upon reaction of the diferrous BesC with O₂ in the presence of substrate [103, 119]. Kinetic isotope effect experiments with [4,4,5,5-²H₄]-l-lysine (d₄-l-lysine) further demonstrated that this intermediate (or its reversibly connected successor) abstracted a hydrogen atom from C4. These results helped formulate possible reaction mechanism for BesC catalysis. Following hydrogen atom abstraction from C4, several plausible pathways can be envisioned. Briefly, these mechanisms include either a carbocation-mediated rearrangement (Path A) or a radical fragmentation process (Path B) (Figure 16B) [103]. To elucidate reaction mechanism, studies using substrate analogs were carried out. In the presence of l-lysine, instead of C–C bond cleavage, the reaction yielded 4-hydroxy-l-lysine (B35) (Figure 16A) [119]. The isotope-labeling experiments using ¹⁸O₂ and H₂¹⁸O demonstrated that the oxygen atom in B35 originated exclusively from the solvent rather than molecular oxygen, supporting a mechanism involving a carbocation intermediate.

Figure 16.

(A) BesC-catalyzed oxidative C–C cleavage reaction using B11 as the substrate, while B33 and B35 are produced when B32 and B34 are utilized. (B) Proposed mechanisms for the BesC-mediated C–C cleavage of B11.

FlcE

FlcE was discovered in the biosynthetic pathway of fluopsin C, a copper-containing antibiotic from Pseudomonas aeruginosa [121]. Sequence alignment indicated that FlcE contains an HDO-like domain, and biochemical characterization revealed that FlcE functions as a bifunctional enzyme, mediating both oxidative decarboxylation and N-hydroxylation. In the FlcE-catalyzed reaction, (R)-S-succinyl-l-cysteine (B13) is converted into the oxime product B14, accompanied by decarboxylation of the cysteine moiety (Figure 17). Stereochemical analysis demonstrated that FlcE acts selectively on the (R)-isomer but not on the (S)-isomer of B13.

Figure 17.

Reactions catalyzed by FlcE.

AetD and SktA

AetD, identified in the biosynthetic pathway of the cyanobacterial neurotoxin aetokthonotoxin, catalyzes the conversion of the 2-aminopropionate moiety of 5,7-dibromo-l-tryptophan (B15) into a nitrile product (B16) (Figure 18A) [106, 122–123, 133]. The reaction proceeds independently of bromination, indicating that indole substitution does not affect catalysis. SktA, an AetD homologue, functions as skatole synthase that converts 5-bromo-l-tryptophan (B17) into 5-bromoskatole (B18) (Figure 18B) [124]. Similar to other HDOs, a substrate-triggered intermediate (P) was observed in the AetD catalyzed reaction.

Figure 18.

(A) AetD catalyzes conversion of the alanyl side chain of l-tryptophan derivatives into corresponding nitrile and aldehyde products. (B) SktA catalyzes the transformation of 5-bromo-l-tryptophan (B17) into 5-bromoskatole (B18). (C) Proposed mechanisms for AetD-catalyzed reaction, illustrated using B15 as an example. (D) Proposed mechanism for SktA-catalyzed transformation.

In AetD, ¹⁵N isotopic labeling experiments demonstrated that nitrogen bound to C2 migrates to C3 to from the nitrile group. Mechanistically, the intermediate (P) was proposed to abstract a β-hydrogen to generate a radical (B37), followed by an aziridine intermediate (B38) and ring opening to yield a nitrone (B39). A second oxidative step likely proceeds through N-hydroxylation which dehydrates to form the nitrile product (B16) with glycolic acid as the byproduct. An off-pathway C-hydroxylation affords 5,7-dibromo-indole-3-carbaldehyde (B36) with glyoxylic acid as a byproduct (Figure 18C) [122]. For SktA, isotopic labeling experiments with ¹³C and ²H-labeled substrates confirmed that the skatole methyl carbon and proton originated from the β-carbon and α-proton of tryptophan with cyanide and bicarbonate as byproducts. The reaction was proposed to begin with hydrogen atom abstraction from the amino group to generate a nitrogen-centered radical (B40). This intermediate then undergoes β-scission to yield a skatole-3-methylene radical (B41) and iminoglycine (B42). Subsequent hydrogen atom transfer produces B18 and imidoyl radical which is degraded to cyanide and carbon dioxide (Figure 18D) [124].

PolF

PolF was recently identified in the biosynthetic pathway of polyoxins, a family of peptidyl nucleoside antibiotics produced by Streptomyces cacaoi and S. aureochromogenes [125–127, 134]. PolF catalyzes sequential desaturation and azetidination of l-isoleucine (B19) to install the ethylidene-azetidine (B20) moiety, a key element of polyoxins (Figure 19A). Interestingly, the olefinic intermediate (B43) can be generated either by PolF or by an Fe and pterin-dependent oxidase, PolE, which assists PolF by enhancing the flux of l-isoleucine desaturation [125, 127]. PolF also employs the intermediate (P) in a substrate-triggered manner [125–126]. Mechanistic insights into PolF were obtained through deuterium kinetic isotope effect (D-KIE) experiments using deuterium-labeled substrates. In the desaturation, following hydrogen atom abstraction, double bond installation likely occurs via a proton-coupled electron transfer (PCET) mechanism. For the azetindation reaction, C–N bond formation step might involve dissociation of the substrate α-NH₂ from the Fe1 center to interact with the allylic radical and could be assisted by the surrounding tryptophan residue(s), as suggested by mutagenesis studies [125, 127]. However, more experiments are needed to verify the proposed reaction mechanisms.

Figure 19.

(A) Desaturation and azetidination catalyzed by PolF. (B) Proposed mechanisms accounting for PolF reaction.

CADD

CADD (Chlamydia protein associating with death domains) functions as a p-aminobenzoate (pABA, B22) synthase in tetrahydrofolate biosynthetic pathway in Chlamydia trachomatis, generating B22 from a protein-derived tyrosine residue (Figure 20A) [128–132]. Although CADD was initially proposed to utilize a diiron cofactor, recent studies showed that maximal activity requires both iron and manganese, implicating a heterobimetallic Fe/Mn cluster as the catalytically active form [128, 130, 132]. Crystal structures of Fe(II)/Mn(II) bound CtCADD (PDB ID: 8VA9) along with kinetic and EPR analyses revealed that the heterobimetallic cluster reacts with O₂, forming a tyrosyl radical and the Mn(III)/Fe(III) cluster state [128, 132].

Figure 20.

(A) p-aminobenzoate (B22) formation catalyzed by CADD. (B) Proposed mechanisms for the CADD catalyzed reactions.

Isotopic labeling and proteomic studies established that the aromatic scaffold of B22 originated from a protein-encoded tyrosine with a neighboring lysine serving as the amine donor [129, 132]. The ¹⁸O₂ tracer experiments confirmed oxygen incorporation into the carboxylate of B22. Furthermore, detection of intermediates B44 and B45 supported a mechanism in which amine installation precedes oxidation of hydroxyl group (Path A, Figure 20B) [132]. Nonetheless, an alternative reaction sequence involving hydroxyl group oxidation followed by amine transfer (Path B) was also proposed [129].

Conclusion

Taken together with the case studies discussed above, these themes illustrate a continuum of mechanisms by which nonheme iron enzymes solve the problem of controlled oxygen activation. The structural and mechanistic diversity of mono- and dinuclear nonheme iron enzymes reflects the remarkable evolutionary plasticity of metalloenzyme active sites in tuning oxygen activation chemistry. His-rich mononuclear enzymes and HDO-type diiron systems illustrate how variations in first-sphere coordination—replacement of anionic carboxylate ligands with neutral histidines, incorporation of substrate-derived ligands, and modulation of metal-metal distances—directly influence redox potentials, oxygen binding affinities, spin-state preferences, and intermediate stability. These coordination sphere effects enable access to iron-oxygen intermediates beyond the canonical Fe(IV)-ferryl species characteristic of 2-His-1-carboxylate enzymes, including Fe(III)-superoxo, Fe(IV)-sulfoxide, μ-peroxo-diiron(III/III), and bis-μ-oxo-diiron(IV/IV) intermediates with distinct spin states and reactivity profiles.

A key mechanistic principle emerging from studies of His-rich enzymes is the role of substrate coordination in modulating oxygen activation. In thiol dioxygenases, gentisate dioxygenase, and diketone-cleaving enzymes, direct coordination of anionic substrate functional groups (thiolate, phenolate, enolate) to the iron center provides electronic activation of both the metal and the substrate, enabling obligate-ordered substrate binding prior to O₂ and facilitating direct oxygen insertion or radical attack mechanisms. By contrast, enzymes where substrates bind in the second coordination sphere—such as anthraquinone and carotenoid cleavage dioxygenases—rely entirely on protein-derived ligands for oxygen activation, resulting in fundamentally different kinetic and mechanistic signatures. This substrate-as-ligand versus substrate-as-target distinction represents an important principle underlying the activity of cofactor-independent nonheme iron enzymes.

For HDO-type diiron enzymes, cooperative interactions between the two metal centers enable formation of bridging oxygen intermediates that perform multi-electron oxidations that may be inaccessible to mononuclear systems. The iron-iron distance, controlled by the geometry and identity of bridging ligands, serves as a key structural determinant of intermediate stability and reactivity. Variations in metal-metal separation (ranging from ~2.5 Å in compact diamond cores to >4.0 Å in more open structures) correlate with distinct oxygen-activation strategies and enable different oxidative transformations within natural product biosynthetic pathways.

Although the catalytic consequences of replacing carboxylate with histidine or adding an “extra” histidine remain incompletely understood, current results indicate that 3- and 4-His motifs alter the redox potential of the intermediary iron complexes and active-site geometry to fine-tune reactivity. The lower overall anionic charge of His-rich coordination environments typically destabilizes high-valent Fe(IV)-oxo intermediates relative to 2-His-1-carboxylate systems, favoring alternative pathways involving Fe(III)-superoxo species and altered O–O bond cleavage mechanisms.

Outstanding Questions and Future Directions:

Despite significant progress, several fundamental questions remain open for investigation:

1. Second-sphere effects: How do residues beyond the metal-binding motif contribute to fine-tuning redox potentials and intermediate lifetimes? What roles do outer-sphere hydrogen bonding networks, electrostatic interactions, and hydrophobic packing play in modulating reactivity?

2. Pathway partitioning: What factors govern the partitioning between competing oxygen activation pathways (e.g., heterolytic vs. homolytic O–O bond cleavage) in His-rich enzymes? Can predictive frameworks be developed based on coordination environment, substrate electronics, and active-site architecture?

3. Biocatalyst design: Can the mechanistic principles learned from these enzymes be translated into designed biocatalysts with tailored substrate selectivity and reaction outcomes? What minimal structural features are required to recapitulate specific oxygen-activation chemistries in simplified scaffolds?

4. Dynamic regulation: How do dynamic conformational changes in protein scaffolds gate substrate and oxygen access to tune reactivity? What is the role of protein dynamics in controlling the lifetime and reactivity of iron-oxygen intermediates?

5. Substrate selectivity in HDOs: What structural features determine substrate recognition and regioselectivity in the diverse HDO family? How do variations in substrate-access channels, active-site architecture, and iron-iron distances contribute to the remarkable functional diversity observed across HDO enzymes?

6. Evolutionary trajectory: How did His-rich coordination motifs and HDO-type diiron clusters evolve from ancestral scaffolds? What selective pressures drove the emergence of these alternative metal-binding strategies, and how do they complement the catalytic repertoire of canonical 2-His-1-carboxylate enzymes?

Addressing these questions will require integration of crystallographic snapshots with time-resolved spectroscopy, advanced computational modeling, and protein engineering approaches. The continued exploration of nonheme iron enzyme diversity promises to reveal new catalytic mechanisms, inform biocatalyst design for sustainable chemistry applications, and deepen our understanding of how the precise control over reactive oxygen chemistry is achieved in nature. Future studies employing biochemical, biophysical, structural, and genomic approaches will expand the mechanistic understanding of these enzymes and likely lead to the discovery of new nonheme iron enzymes with previously unknown chemistries, further broadening the opportunity for catalytic and biotechnological innovation.