Multinuclear non-heme iron dependent oxidative enzymes (MNIOs) involved in unusual peptide modifications

Abstract

Multinuclear non-heme iron dependent oxidative enzymes (MNIOs), formerly known as domain of unknown function 692 (DUF692), are involved in the post-translational modification of peptides during the biosynthesis of peptide-based natural products. These enzymes catalyze highly unusual and diverse chemical modifications. Several class-defining features of this large family (>14,000 members) are beginning to emerge. Structurally, the enzymes are characterized by a TIM-barrel fold and a set of conserved residues for a di- or tri-iron binding site. They use molecular oxygen to modify peptide substrates in a four-electron oxidation, often taking place at a cysteine residue. This review summarizes the current understanding of MNIOs. Four modifications are discussed in detail: oxazolone-thioamide formation, β-carbon excision, hydantoin-macrocycle formation, and 5-thiooxazole formation. Briefly discussed are two reactions that do not take place on Cys residues.

Graphical Abstract

Introduction

Several families of non-heme diiron enzymes are involved in the biosynthesis of natural products, such as the ferritin-like oxidases and oxygenases (FDOs), the heme oxygenase-like oxidases and oxygenases (HDOs), and the mixed-valent diiron oxygenases (MVDOs) [1–4]. Recently, members from a different of iron-dependent enzymes, initially annotated as domain of unknown function 692 (DUF692), have been identified in the biosynthesis of ribosomally synthesized and post-translationally modified peptides (RiPPs). These natural products display a diverse array of bioactivities [5–7]. RiPPs are made following a common biosynthetic logic in which a precursor peptide is modified by one or more post-translational modifying enzymes, and then proteolytically processed to yield the mature natural product. Several classes of post-translational modifying enzymes have been characterized in RiPP biosynthesis [8], with the DUF692 family enzymes implicated relatively recently. The first member to be characterized was MbnB, involved in the biosynthesis of the copper-chelating peptide methanobactin [9]. Following recent characterization of more members, DUF692 enzymes were renamed multinuclear non-heme iron dependent oxidative enzymes (MNIOs) [10]. While most characterized MNIOs have been shown to be oxidases, some MNIOs may function as monooxygenases or dioxygenases (Figure 1A).

Figure 1.

Diverse post-translational modifications performed by members of the MNIO family. (A) Reactions catalyzed by MbnBC, TglHI, ChrHI, BufBC, MovX, and ApyHI. (B) Sequence similarity network (SSN) of MNIOs (DUF692), colored by the top ten phyla in which MNIOs are most frequently found. Percentages represent the distribution of MNIO-containing species across phyla. The SSN was generated by using the Enzyme Function Initiative-Enzyme Similarity Tool with over 14 000 members [15,16].

MNIOs are around 30–45 kDa in size and typically form a heterodimer with a partner protein, although the identity of the partner protein is not conserved. The structures of several MNIOs have been elucidated, revealing a triose-phosphate isomerase (TIM)-barrel fold with a di- or tri-iron binding site located within the barrel [11–13]. MNIOs use molecular oxygen, appear to function without the need of external reductant, and in most cases catalyze a four-electron oxidation of substrate [9,10,14].

The growing family of MNIOs contains over 14 000 members across both Gram-negative and Gram-positive bacteria. A sequence similarity network (SSN) [15,16] shows that the majority of MNIOs are found in Pseudomonadota (Proteobacteria), followed by Actinomycetota (Actinobacteria), Myxococcota, Bacteroidota, and other phyla (Figure 1B). Several subfamilies remain to be characterized, the majority of which are predicted to be involved in RiPP biosynthesis (for a co-occurrence SSN of MNIO-RiPPs, see refs [17,18]).

In this review, we summarize the unusual rearrangements catalyzed by currently characterized MNIOs (Figure 1A) and discuss patterns across MNIO structures, substrates, and mechanistic proposals.

Cysteine-modifying MNIOs

MbnB

The first characterized MNIO was MbnB, an enzyme involved in the maturation of the RiPP methanobactin, a copper-chelator (chalkophore) that is produced by methanotrophic bacteria under copper-limiting conditions [19,20]. The peptide is secreted into the environment in its apo-form, then taken back up by the cell in its copper-bound form [21]. Structurally, most methanobactins contain a pair of oxazolone-thioamide moieties (Figure 1A) that bind copper [19,22].

Biosynthetic gene clusters (BGCs) responsible for methanobactin biosynthesis encode MbnB and MbnC as well as the gene for the precursor peptide MbnA [23]. In vitro assays demonstrated that MbnB, together with its partner protein MbnC, installs the thioamide and oxazolone moieties onto MbnA [9]. An MbnC deletion mutant in Methylosinus trichosporium OB3b showed that while both MbnBC are required for the formation of the C-terminal oxazolone-thioamide, the N-terminal oxazolone-thioamide could be formed without MbnC [24]. Furthermore, in contrast to MsMbnB, which is only stable in a complex with MbnC, homologs from Vibrio fluvalis (MovB and MovC) could be purified independently [9,18].

The active form of MbnB is believed to be a mixed-valent Fe(II)/Fe(III) diiron species instead of a triiron species, supported by spectroscopic data detailed below [12,25]. The mechanism is proposed to feature coordination of the cysteine thiolate of MbnA to Fe(III) (Figure 2A), which was observed crystallographically in an MbnABC complex [11] and supported by electron nuclear double resonance spectroscopy [25]. Then, in a mechanism reminiscent of isopenicillin N synthase (IPNS) [26,27], binding of O₂ to the ferrous iron has been proposed to generate an Fe(III)-superoxo intermediate, which abstracts a hydrogen atom from the β-carbon of cysteine. The radical is then oxidized to a thioaldehyde that is attacked by the adjacent amide-nitrogen, forming a β-lactam and an Fe(IV)-oxo (ferryl) intermediate. This potent oxidizing species performs a second H-atom abstraction from the β-lactam to form a thio analog of a cyclic imide, which was proposed to be opened by an enzyme nucleophile [11,12]. Substitution and tautomerization form the final oxazolone-thioamide moiety [12].

Figure 2.

Proposed reaction mechanisms of MNIOs. (A) Proposed mechanisms of the cysteine modifying MNIOs MbnB, TglH, ChrH, and BufB. The mechanisms depicted here are consistent with the current evidence and are drawn to show possible similarities across the family. For example, all four mechanisms involving cysteine are drawn with a thiolate-coordinating ferric iron being different from the ferrous iron that activates O₂. Other mechanisms can be drawn in which the same Fe coordinates Cys and activates O₂ [14]. The fate of the carbonyl carbons during the ChrH reaction is indicated with an asterisk; X = enzyme residue. (B) Proposed mechanisms of the asparagine/aspartate modifying MNIOs MovX and ApyH. B = enzyme base. We note that for all reactions shown, other mechanisms are possible, with several steps having both homolytic or heterolytic possibilities.

TglH

The second MNIO to be characterized, TglH, was discovered as part of a RiPP pathway that produces 3-thiaglutamate (or a derivative) in the plant pathogen Pseudomonas syringae [14]. TglH, together with its partner protein TglI, performs an unusual excision of the β-carbon of a C-terminal cysteine, releasing the carbon as formate and reattaching the thiol group to the ⍺-carbon (Figure 1A) [14]. These findings suggest that TglH could be an oxygenase if molecular oxygen is incorporated into formate. Use of a substrate analog perdeuterated at the ⍺- and β-carbon of cysteine resulted in only one deuterium in the peptide product, suggesting that the ⍺ hydrogen is likely not removed in the reaction. These findings support a mechanism with similarity to myo-inositol oxygenase (MIOX), 2-hydroxyethylphosphonate dioxygenase (HEPD), and IPNS catalyzed reactions [26,28–30]. The proposed mechanism again involves H-atom abstraction from the β-carbon of cysteine by an Fe(III)-superoxo (Figure 2A). Hydration of the thioaldehyde generates a thioacetal and Fe(IV)-oxo. Proton-coupled electron transfer (PCET) of the hydroxyl group to the ferryl generates an oxygen radical that undergoes β-scission resulting in a resonance stabilized radical at the ⍺-carbon. Recombination of the sulfur of the thioformyl group with the ⍺-carbon radical generates the S-C⍺ bond. Finally, hydrolysis of the thioester by an iron-bound hydroxide results in the production of formate and the peptide product [14].

A follow up study showed that an analog of the precursor peptide TglA in which cysteine was replaced with selenocysteine inhibited TglHI [31]. How selenocysteine acts as an inhibitor, and whether this strategy could be applied to generate other MNIO inhibitors remains to be explored. In another study, a homolog of TglHI, TmoHI from Tistrella mobilis, was found to perform the same reaction on a C-terminal cysteine of a scaffold peptide [32]. In this case the later biosynthetic steps diverged, resulting in the production of 3-thiahomoleucine (instead of 3-thiaglutamate) attached to the scaffold peptide TmoA [32].

ChrH

ChrH from Chryseobacterium sp. JM1 is part of another MNIO subfamily that performs vastly different chemistry compared to MbnB and TglH (Figure 1A). ChrH performs a remarkably complex rearrangement of a peptide ChrA containing two conserved Cys residues to install a peptide macrocycle, a hydantoin (imidazolidine-2,4-dione) heterocycle, and a thiomethyl group (Figure 1A) [10]. ¹³C-labelling showed that the carbonyl carbon of the 2^nd Cys in the sequence migrates to the hydantoin moiety in the product (Figure 2A). Furthermore, S-adenosylmethionine (SAM) is required for the reaction in vitro [10]. ChrH copurifies with ChrI, annotated as a member of the anthrone oxygenase-like protein family (formerly DUF1772) [33]. Whereas the original report suggested that ChrI was required for ChrH activity, our recent unpublished work shows that ChrI is not needed. Furthermore, while the structure of the ChrH-modified peptide has been determined, the final product of the chr BGC has yet to be elucidated.

Several possible mechanisms have been proposed for the complex rearrangement by ChrH [6]. Here, we propose a modified mechanism that is more similar to those proposed for MbnB and TglH. The first step again invokes hydrogen atom abstraction from the β-carbon of Cys₆₆ by an Fe(III)-superoxo intermediate (Figure 2A). Formation of a thioaldehyde, subsequent β-lactam formation, and formation of a ferryl could occur as in the MbnB mechanism [12]. Then, N-H bond cleavage at a Gly amide nitrogen by the Fe(IV)-oxo could result in a rearrangement to generate a bicyclic system and oxygen radical. Next, β-scission akin to the C-C bond cleavage step in TglHI would form the hydantoin heterocycle and a radical on the former ⍺-carbon of the initial cysteine residue. Radical recombination with an iron-bound thiolate could then generate an epi-sulfide. Finally, the thiol of the second cysteine (Cys₆₃) attacks the epi-sulfide, which undergoes ring opening and S-methylation by SAM to generate the final product.

BufB

BufB is part of the largest subfamily of MNIOs and is involved in the biosynthesis of bufferins (Figure 1), which are copper-chelating metallophores [34]. Unlike methanobactins, which are involved in copper import, bufferins are thought to be involved in copper export during copper-induced stress. BufB1 from Caulobacter vibrioides, together with its partner protein BufC1 (a DUF2063 family protein), converts cysteine residues to 5-thiooxazoles on the precursor peptide BufA1 (a DUF2282 family protein) (Figure 1A) [34]. BufA1 contains four cysteine residues, the middle two of which are modified to 5-thiooxazoles that enable copper binding.

While little is known about the mechanism of BufB, the reaction is proposed to similarly start from a mixed-valent state that involves oxygen activation and H-atom abstraction from the β-carbon of Cys (Figure 2A). Cyclization followed by ring oxidation and tautomerization is proposed to form the 5-thiooxazole [34]. Interestingly, this motif is precedented in RiPP biosynthesis, but is formed by a radical SAM enzyme [35].

Asparagine/Aspartate-modifying MNIOs

MovX

MovX (previously termed MbnX) is part of a methanobactin-like gene cluster in Vibrio fluvalis [23]. MovX was shown to act on an asparagine residue, performing a N-C⍺ bond cleavage to generate a C-terminal amide (Figure 1A) [18]. This discovery showed that MNIOs can modify residues other than cysteine. Furthermore, MovX can catalyze this transformation without the need for a partner protein, likely because it encodes a C-terminal extension that is predicted to recruit substrate [18].

One proposed mechanism for this transformation involves oxygen activation to generate a Fe(III)-superoxo that abstracts the ⍺-hydrogen atom from asparagine (Figure 2B) [18]. Peroxide rebound is proposed to be followed by H₂O₂ elimination and generation of an imine. Hydrolysis of the imine generates the C-terminally amidated peptide and a diketone dipeptide. Although the dipeptide coproduct eluded detection, several lines of evidence are consistent with this mechanism. The reaction required O₂, hydrogen peroxide was detected, and use of ¹⁵N- labelled asparagine (at the side chain amide) resulted in loss of the label in the C-terminally amidated peptide product [18]. This mechanism deviates from the other characterized MNIOs in that the enzyme would perform a two-electron oxidation of substrate.

ApyH

ApyH from Burkholderia thailandensis is another example of an MNIO that modifies a non-Cys residue. ApyH, together with its partner protein ApyI, converts a C-terminal aspartate on the precursor peptide ApyA to an ⍺-keto acid moiety during aminopyruvatide biosynthesis (Figure 1A) [36]. In addition to the conserved C-terminal aspartate, ApyA homologs contain a highly conserved cysteine and glycine in the middle of the sequence. While no modification was detected at these residues, various mutations of the cysteine abolished activity, suggesting that the cysteine may be involved in substrate binding or substrate-assisted catalysis [36].

The ApyHI mechanism is thought to occur by oxygen activation and H-atom abstraction from the β-carbon of aspartate (Figure 2B) [36]. Like the Cys-modifying enzymes, a Fe(III)-hydroperoxo is proposed to initiate hydroxylation of the β-carbon with generation of an Fe(IV)-oxo. The ferryl may then abstract the remaining hydrogen atom from the β-carbon. Decarboxylation involving an electron transfer to Fe(III) then produces the enol form of the ketoacid. Electron transfer coupled to decarboxylation to form an alkene has been previously proposed in other iron-dependent enzymes such as iron(II)/2-oxoglutarate-dependent and radical SAM enzymes [37–39]. Finally, tautomerization of the enol would form the C-terminal ⍺-ketoacid [36].

Structural basis of catalysis and substrate binding

Several MNIO structures have been solved in the iron-free, diiron, or triiron forms, or in a complex with their partner protein and substrate (Table 1; Figure 3A) [11–13]. The active site structure generally shows that iron 1 (Fe1) and Fe2 are six-coordinate, with Fe3 (when it is present) having five or six-coordinate geometry (Figure 3B). Fe1 is coordinated by the side chains of two His and one Glu, which is bidentate and bridges Fe1 and Fe2. Fe2 is coordinated by two Glu, one His, one Asp residue, and Fe3 is coordinated by His, Asp, and Asn. Combined structural and mutagenesis data have also shown that the active site residues Asp240 in MbnB and Asn73 in TglH are essential, likely playing a role as a base or for substrate positioning, respectively [11,13].

Table 1.

Summary of the characterized MNIOs to date.

MNIO	Organism	Modification	Modified residue	Representative structure(s) (PDB ID)	Pathway
U.C.a	Haemophilus somnus (Histophilus somni) strain 129pt	Unknown	Unknown	3BWW (diiron)	Unknown
MbnB	Methylosinus trichosporium OB3b, Rugamonas rubra ATCC-43154, Vibrio caribbeanicus ATCC BAA-2122	Oxazolone-thioamide installation	Cysteine	7DZ9, 7FC0 (triiron MbnABC complex)	Methanobactin biosynthesis
TglH	Pseudomonas syringae pv. maculicola ES4326	β-carbon excision	Cysteine	8HCI (apo TglHI) 8HI7 (diiron TglHI) 8HI8 (triiron TglHI)	3-Thiaglutamate biosynthesis
TmoH	Tistrella mobilis KA081020-065	β-carbon excision	Cysteine	N.A.a	3-Thiahomoleucine biosynthesis
ChrH	Chryseobacterium sp. JM1	Hydantoin-macrocycle formation	Cysteine	N.A.	Unknown
ApyH	Burkholderia thailandensis E264	Aminopyruvic acid formation	Aspartate	N.A.	Aminopyruvatide biosynthesis
MovX (MbnX)	Vibrio fluvalis 33809	C-terminal amidation via N-C⍺ bond cleavage	Asparagine	N.A.	Unknown
BufB	Caulobacter vibrioides NA1000	5-thiooxazole installation	Cysteine	N.A.	Bufferin biosynthesis

^aU.C., uncharacterized; N.A., not available

Figure 3.

Structural insights into MNIO catalysis and substrate binding. (A) Structure of the MbnABC complex from Rugamonas rubra (PDB:7FC0), showing the common TIM-barrel fold shared across MNIOs and the three bound irons (orange spheres). The dotted box highlights the contact between MbnA and the MbnC partner protein that is required for substrate binding [11]. (B) Active site structure of RrMbnB showing coordinating side chains (teal sticks) and water molecules (red spheres). The cysteine residue of the substrate MbnA coordinating with Fe1 is shown in pink. A water molecule occupies this coordination site in the absence of substrate.

RiPP BGCs frequently encode a RiPP recognition element (RRE) to recruit substrates [40]. RREs contain three ⍺-helices and three β-sheets that bind an N-terminal leader sequence of the precursor peptide. The partner proteins TglI, ApyI, and BufC1 contain RREs that recruit their respective substrates. Furthermore, MovX contains a C-terminal extension with a winged helix-turn-helix fold that is predicted to act as an RRE. However, both MbnB and ChrH lack an obvious RRE encoded in their clusters [9,10,14]. The structure of the MbnABC complex shows that while a canonical RRE is not present, the leader peptide of MbnA binds to the partner protein MbnC in an analogous fashion to RRE recognition (Figure 3A) [11]. Notably, contacts between MbnH and TglH and their substrates also seem to be important. This observation may help to explain why ChrH can function without a partner protein to recruit substrate. The mechanism of substrate binding of ChrA by ChrH is still unclear due to lack of a co-crystal structure, but is likely to occur via a different mechanism than TglAHI and MbnABC.

MNIO substrates

In general, MNIOs appear to have relaxed substrate specificity with respect to peptide length and sequence. TglH can modify a truncated substrate consisting of the last 19 amino acids of the precursor peptide TglA, and mutations at several positions within this 19-mer were well tolerated [31]. BufBC could modify the core peptide lacking the N-terminal signal sequence, albeit with lower efficiency [34]. Similarly, N-terminal truncation of MbnA to a 22-mer still resulted in oxazolone-thioamide formation by MbnBC [9]. MbnBC from Ms. trichosporium was also shown to modify MbnA precursor peptides from four out of the five phylogenetic classes of methanobactin-like operons, despite substantially different sequences [9].

While initial studies of MNIOs described modifications of cysteine residues [41], characterization of MovX and ApyH showed that MNIOs can also act on asparagine and aspartate residues. These observations suggest that the breadth of MNIO reactivity is larger than currently realized, and that other amino acid residues could potentially be substrates. At present, no examples have been reported of MNIOs that use a non-peptide substrate.

Mechanistic commonalities and differences

Despite the seemingly highly diverse reactions in Figure 1A, some parallels can be drawn from the transformations that involve Cys residues. MbnB, TglH, ChrH, and BufB all carry out a four-electron oxidation of the substrate and initial H-atom abstraction from the β-carbon of cysteine [9,10,14]. In addition, all three enzymes are proposed to use an Fe(III) ion as a Lewis acid to activate the initially formed thioaldehyde. Other similarities between TglH and ChrH are that they both catalyze transfer of the sulfur of Cys from the β-carbon to the α-carbon. As shown in Figure 2A, Fe(III)-superoxo and Fe(IV)-oxo species have been proposed as the oxidizing species in the catalytic cycle, but both have yet to be detected spectroscopically. Prior studies on oxygen activation in a variety of non-heme iron enzymes have established several spectroscopic features that may inform future studies on O₂ activation in MNIOs [1,4]. Currently also unresolved is whether oxygen binding is substrate triggered as observed for other non-heme diiron enzymes [42–44]. At present, any similarities of MovX and ApyH with other MNIOs are less obvious [18,36].

The mechanisms of the cysteine modifying MNIOs all invoke one ferrous ion in the resting state of the enzyme that activates oxygen to form both Fe(III)-superoxo and Fe(IV)-oxo intermediates (Figure 2A). Such use of a single iron center to generate consecutive oxidizing species has precedent in a subset of non-heme iron-dependent enzymes, including IPNS and HEPD [1,45]. The other irons in the MNIO active site are proposed to not be used for oxygen activation but for electron transfer, Lewis acid activation of sulfur, and potentially substrate binding and orientation. EPR and Mössbauer studies on MbnB coupled to activity assays have revealed that a mixed valent Fe(II)/Fe(III) is the active species [12]. The Fe(II)/Fe(III) irons in MbnB are antiferromagnetically coupled, giving a S=1/2 rhombic signal with a g_ave-value of 1.9. The sample also showed the presence of a rhombic S=5/2 Fe(III) signal at g_eff=4.3 [9,12]. Although no external reductant is needed in the proposed MNIO mechanisms, the addition of ascorbate to MbnB has been shown to increase enzyme activity in vitro [12,18]. This observation may be attributed to the reduction of a ferric state to the active mixed-valent species, and is consistent with the Fe(III) EPR signal decreasing as ascorbate was titrated into MbnBC.

Diiron or triiron enzyme?

An unresolved question is whether MNIOs are di- or triiron enzymes, or whether examples exist of both within the family. For BufB1, enzyme that modified the precursor peptide in vitro co-purified with two irons [34]. Similarly, in MbnB, higher activity was correlated with higher abundance of the diiron form, and has led to its functional classification as an MVDO [12,25]. However, MbnB and TglH can bind up to three irons based on reported structures, and Mössbauer and mass spectrometry data [11–13]. A co-crystal structure of the MbnABC complex indicated ligation of the substrate cysteine thiolate to Fe1, identifying it as crucial to the reaction (Figure 3B) [11]. Both Fe1 and Fe2 sites showed full occupancy whereas the Fe3 site often shows partial occupancy [11,12]. Thus, the importance and role of Fe3 remain unresolved.

In both MbnB and TglH, mutating some of the Fe3 ligands to alanine or serine decreases activity but does not completely abolish it, suggesting Fe3 may not be critical [12,13]. However, mutation of His ligands to Fe1 or Fe2 in TglH to Ala also decreased but did not completely abolish activity [13]. Thus, Fe3 could still play a role in facilitating substrate positioning, or orientating side chains of active site residues in TglH. Detailed spectroscopic studies on TglH, ChrH, MovX, ApyH, and BufB will be needed to investigate whether the active species differs from MbnB. If MNIOs are diiron enzymes, it begs the question of why the ligand set for binding three irons is conserved across most MNIOs.

Conclusions and Outlook

Recent reports have demonstrated the remarkable breadth of MNIO-catalyzed reactions, all housed within a common TIM-barrel fold. MNIOs have been shown to modify cysteine, asparagine, and aspartate residues in RiPP biosynthesis. Several questions remain in our understanding of this enzyme family. Firstly, what other chemical transformations are performed by MNIOs? Despite their prevalence across different phyla, several subfamilies of MNIOs remain functionally uncharacterized (Figure 1B) [17]. The diverse and unprecedented reactivity of the enzymes characterized thus far suggests that exciting new post-translational modifications are to be discovered. Secondly, what is the mechanism behind these transformations, and what determines the type of modification performed? Probing the mechanisms may not only reveal fundamental insights into an unusual di-/tri-iron catalytic center, but will be key for the potential development of MNIOs as biocatalysts or the design of mechanism-based inhibitors. Finally, which biological processes are mediated by the products of MNIO-containing BGCs? Methanobactins play a key role in copper acquisition in methanotrophic bacteria, while bufferins have been shown to provide protection against copper toxicity [9,34]. Recent studies also suggest that an MNIO-modified RiPP is a virulence factor in the human respiratory pathogen, Haeomophilus influenzae [46,47]. On the other hand, the physiological roles for 3-thia-⍺-amino acids, aminopyruvatides, and the chr product remain unknown.

Reports on the few MNIOs characterized thus far have demonstrated their emerging role in performing unusual and challenging peptide modifications. Future genome mining efforts, functional characterizations, and protein engineering studies will elucidate the full catalytic breadth of this enzyme family.

Highlights.

• Remarkable chemistry catalyzed by multinuclear non-heme iron dependent oxidative enzymes (MNIOs)

• MNIOs convert Cys to an oxazolone-thioamide in methanobactin biosynthesis

• MNIOs excise the β-carbon of Cys in 3-thiaglutamate biosynthesis

• MNIOs convert Cys residues into macrocyclic thioaminals and hydantoins

• MNIOs convert Cys residues to 5-thiooxazoles in bufferin biosynthesis

• MNIOs cleave the peptide backbone at Asn to generate a C-terminal amide

• MNIOs convert C-terminal Asp into aminopyruvic acid

Abbreviations

BGC

biosynthetic gene cluster

DUF692

domain of unknown function 692

EPR

electron paramagnetic resonance

HEPD

2-hydroxyethylphosphonate dioxygenase

IPNS

isopenicillin N synthase

MIOX

myo-inositol oxygenase

MNIO

multinuclear non-heme iron dependent oxidative enzyme

PCET

proton-coupled electron transfer

RiPP

ribosomally synthesized and post-translationally modified peptides

RRE

RiPP recognition element

SAM

S-adenosylmethionine

SSN

sequence similarity network

TIM

triose-phosphate isomerase

Data availability

No data was used for the research described in the article.