Advances in the DMSO Reductase Family: From Discovery to Mechanism

Abstract

DMSO reductase (DMSOR) family found in prokaryotes is the most diverse group of molybdenum cofactor (Moco) containing enzymes from - the structural as well as the catalytic viewpoints. While DMSOR, the prototype member of its family, is well characterized thanks to decades of rigorous research, a large majority of this family are yet to be characterized. These enzymes are important in biogeochemical cycles, biotechnology and health. With the discovery of new DMSOR family members, the family continues to grow further underscoring their importance. This review will focus on the latest developments in the DMSOR reductase family.

Graphical Abstract

1. Introduction

The DMSOR superfamily comprises the most structurally and functionally diverse class of molybdenum cofactor (Moco) containing enzymes catalyzing redox transformation of substrates spanning ~1.5 V. [1] Although oxygen atom transfer (OAT) is the predominant reaction, members also catalyze hydride transfer, hydroxylation, dehydroxylation, and sulfur transfer (Figure 1). Expansion of genomic databases has revealed that >99% of predicted DMSOR homologs remain experimentally uncharacterized, making it difficult to access the full extent of catalytic diversity of this superfamily. [2] In several organisms, paralogs have been identified, but the function of their gene products remains poorly understood. [3,4] For some homologs, including E. coli YdhV, the enzyme has been characterized, yet its biological function has not been established. [5]

Figure 1.

DMSOR family catalyzes a diverse array of reactions. Only the formate dehydrogenase catalyzed reaction is shown as reversible, as the physiological direction is organism dependent.

The catalytic subunits of this family have two pyranopterins coordinated to the Mo-center via dithiolate moieties. The structural diversity in the catalytic subunit results from the Mo-coordinating amino acid residue and the nature of the sixth ligand (e.g., terminal oxido, sulfido, or selenido). In the catalytic subunit, in addition to the Moco, a 4Fe4S cluster is often found; however, the simplest form has only the Moco as in Rhodobacter capsulatus DMSOR.

2. Reactions catalyzed by the DMSOR family

2.1. Sulfoxide reduction

Anaerobic bacterial growth on DMSO relies on DMSOR (DmsABC), although roles beyond anaerobic respiration, such as involvement of E. coli DmsABC in host-pathogen interaction and associated stress response, have been suggested.[6] The genome of nontyphoidal Salmonella enterica has three putative dmsABC homologs in operons STM0964-STM0966, STM2530-STM2527 and STM4305-STM4308.[7] STM0964-STM0966 is the closest homolog of the E. coli DmsABC and exhibits the highest DMSOR activity (Figure 2).[7] A deletion mutant of all three dmsA genes (STM2530, STM4305 and STM0964) in S. typhimurium results in loss of colonization in mammalian intestines.[7] The ∆STM0964 mutant complemented with STM0964 restored the ability of the bacteria to respond to DMSO at WT levels. The growth enhancement was not restored when the ∆STM0964 mutant was complemented with STM4305, suggesting STM4305 does not restore functionality of STM0964. In contrast, STM2530 has the lowest sequence identity with E. coli DmsABC (32–55%) and its role in DMSO reduction is unclear.

Figure 2.

Sequence alignment of three putative DMSOR in S. typhimurium and E. coli dmsA showing the 4Fe4S cluster coordinating residues on top and Mo coordinating residues at the bottom.

Recently, DmsEFABGH has been implicated in iodate reduction in Shewanella Oneidensis,[8] where iodate is reduced to HOI and H₂O₂. Although DMSOR is known to work via OAT, it is unclear whether the formation of HOI involves OAT. The mechanism of iodate reduction remains unresolved because current models do not fully explain H₂O₂ formation.

Methionine sulfoxide (MetSO) reductase (MtsZ) supports the virulence of Haemophilus influenzae.[9] In addition to MetSO, it reduces S-biotin sulfoxide and trimethylamine N-oxide (TMAO). Activity measurement with varying concentration of S_C-(R+S)_S-MetSO led to a k_cat of 45.5 s⁻¹ and an apparent K_M of 0.114 mM (S_C indicates S configuration at the α-carbon of MetSO and (R+S)_S indicates the presence of both R and S configuration at sulfur center). i S_C-R_S-MetSO is unlikely to be converted to the product and thus many have inhibitory effects due to the formation of a nonproductive enzyme-substrate complex. This hypothesis is supported by the observation that racMetSO, where three out of four stereoisomers of MetSO are not substrates, has a fourfold higher apparent K_M (0.4 mM). Thus, the apparent K_M for the likely natural substrate (S_C-S_S-MetSO) is lower than the value determined for S_C-(R+S)_S-MetSO. The Mo center in MtsZ is coordinated by two pyranopterin units, one Ser (Ser187) and a water-derived ligand (2.34 Å) (Figure 3). [10] The distance (2.34 Å) of the oxygen from Mo suggests a reduced form of the enzyme. [10] Ser187 adopts both bound and unbound conformations and its dissociation at lower pH might explain the reduced catalytic activity of the enzyme. [10] Docking studies indicate that the S-sulfoxide specificity of this enzyme is due to the binding pocket asymmetry that enables favorable H-bonding interaction between the lone pair on S and conserved Tyr. These findings indicate that subtle change in the substrate binding pocket plays a key role in the evolution of sulfoxide specificity.

Figure 3.

Left: The active structure of E. coli MtsZ with bound (O atom bonded to Mo) and unbound (O atom of Set is 4.2 Å away from Mo) Ser at pH 7; right: The active site structure with unbound Ser (O atom of Ser is 4.1 Å away from Mo) at pH 5.5.

2.2. N-oxide/nitrate reduction

E. coli trimethylamine N-oxide reductase (TorA) reduces TMAO to trimethylamine during anaerobic respiration. TorA receives electrons from a pentaheme c-type cytochrome protein, TorC, which receives electrons from the menaquinol pool. Although TorA has been structurally characterized already (PDB: 1TOR, 9H4T), the details of electron transfer from TorC to TorA are not clear. However, the oxidation of reduced TorC upon the addition of TorA and TMAO demonstrates direct electron transfer from TorC to TorA. [11] Steady-state catalytic parameters have been determined using menadiol as an alternate electron donor for menaquinol. The K_M value obtained with menadiol (10 μM) is comparable to those measured with reduced methyl viologen (MV, 32 μM) and benzyl viologen (BV, 22 μM). However, the k_cat value in the presence of menadiol (0.15 s⁻¹) is substantially lower than those observed with MV (464 s⁻¹) or BV (228 s⁻¹). The lower k_cat in the case of menadiol has been attributed to its lower reduction potential (−189 mV compared to −450 mV for MV or −374 mV for BV). Binding studies revealed a relatively tight interaction between TorAD₂ and TorC, with a binding constant (K_D) of 0.3 μM. The observation of TorAD₂ is interesting because generally the signal peptide interacts with one chaperone; here two copies of the chaperone are involved, raising questions about their recognition sites. This is different from the previous binding studies (between TorC and TorA), which led to the determination of two binding constants 17 nM and 3 μM..[12] It has been suggested that these two binding constants might correspond to the interaction of TorC_N (N-terminal domain of TorC) and TorC_C (C-terminal domain of TorC) to TorA. However, both studies propose a model where TorC_N binds TorA first, positioning TorC appropriately for electron transfer to TorA.

Periplasmic nitrate reductase (NapA) catalyzes nitrate reduction to nitrite, the first step in denitrification as well as assimilatory and dissimilatory nitrate reduction to ammonia.[13] Thus, NapA plays an important role in the global nitrogen cycle. Like other members of the DMSOR family, NapA accepts a variety of alternate substrates including TMAO, chlorate and perchlorate,[14,15] albeit with a lower catalytic efficiency, primarily due to their higher K_M values except for TMAO, for which k_cat is also significantly lower (Table 1). [16,17]

Table 1.

catalytic parameters of NapA for various substrates

substrate	kcat(S−1)	Km(μM)	kcat/KM(M−1s−1)
nitrate	5.9	3.4	1.7×106
TMAO	0.092	161	5.7×102
chlorate	2.13	968	2.2×103
perchlorate	2.52	22907	1×102

Structural studies on NapA revealed a Mo center coordinated by two pyranopterin cofactors and a Cys residue. While the identity of the sixth ligand has been debated, a detailed EXAFS analysis supports the presence of a terminal oxo group as the sixth ligand. [18] The structure of NapA from C. jejuni contains several lysine-rich polypeptide insertions in the substrate channel that may contribute to its high affinity for nitrate.[14] Although all NapA structures show the presence of tricyclic pyranopterin rings, the planarity of the two pyranopterin rings is different. These pyranopterin rings are designated as proximal or distal based on their proximity to the 4Fe4S cluster. The proximal pyranopterin is more distorted than the distal one. [19] This degree of distortion has also been linked to differing oxidation states (e.g., dihydro vs tetrahydro) of the pyranopterin. The more distorted nature of the tetrahydro/reduced proximal pyranopterin might favor cyclization leading to a tricyclic pyranopterin ring.[20] In contrast, the more planar distal pyranopterin is thought to be electronically coupled to the Mo-center, and modulate its electronic environment. In NarG (PDB ID 1Q16), EBDH (PDB ID 2IVF), PcrA (PDB ID 4YDD) and Nxr (PDB ID 7B04), the distal pyran ring is found to be open (in Nxr, the proximal pyran ring is also open), and in these cases, the electronic coupling will largely be lost. Thus, the enzymes may use ring opening and closing to adjust the redox properties of Moco, and in turn, control catalytic activity.[21] However, direct experimental evidence supporting the hypothesis on catalytic control is yet to be reported.

Over the years, three structure-based mechanisms have been proposed for NapA-catalyzed nitrate reduction, and all of which involve the OAT process. A key difference arises from the different assignment of the sixth ligand. In particular, assigning the ligand as a terminal sulfur gave rise to the ‘sulfur shift’ mechanism. Recently, isotope labeling and mass spectrometry provided direct evidence for the OAT and argue against the ‘sulfur shift’ mechanism.[22] Consistent with the absence of a terminal sulfido group, cyanide inhibition of NapA is reversible and does not produce ^—SCN, unlike sulfido-containing molybdoenzyme xanthine oxidase. The Mo(VI)=O formed under single turnover conditions is partially reduced by the reduced 4Fe4S cluster producing Mo(V)=O, which has been characterized by EPR.[23] Nitrite oxidation by NapA is less efficient compared to its nitrate reduction. In contrast, nitrite oxidoreductase (Nxr), which is also a member of the DMSOR family, can carry out both nitrate reduction and nitrite oxidation with similar ease. However, the reasons for this difference remain unclear.

2.3. Antimonite oxidation/antimonate reduction

The role of DMSOR family members in the arsenic biogeochemical cycle is well established,[24] whereas their role in antimony biogeochemical cycling is less well understood, although more than 90 Sb(III) oxidizing bacteria have been reported.[25] Most of these detoxify Sb(III) by oxidizing it to less toxic Sb(V). Heterotrophic bacteria use periplasmic As(III) oxidase AioAB and cytoplasmic Sb(III) oxidase (antimonite oxidase, AnoA) to oxidize Sb(III).[26] Recently, a nitrate-dependent antimonite oxidase (NaoABC) has been shown to couple Sb(III) oxidation with nitrate reduction.[27] NaoABC works in conjunction with NarGHI, enabling electron transfer from Sb(III) to nitrate and generating the energy required for autotrophic growth. NaoABC has the periplasmic catalytic subunit NaoA. Despite differences in operon structure between NaoABC and AioAB, both NaoA and AioA are closely related and cluster within the NapA lineage. Likewise, the electron transfer subunits NaoBC and AioB are related to DmsB. This phylogenetic relationship may rationalize why AioAB has a similar role in Sb(III) oxidation.

In contrast to Sb(III) oxidizing bacteria, relatively little is known about Sb(V) reducing bacteria. Since the discovery of Desulfuribacillus stibiiarsenatis, only a handful of Sb(V) reducing bacteria have been reported.[28] In D. stibiiarsenatis, a putative antimonate reductase (AnrA), which contains a twin arginine translocase (TAT) motif indicating its periplasmic location, has been identified.[29] A multi-omics study revealed four operons encoding antimonate reductase, where the catalytic subunit has been designated as SbrA.[30] The gene organization in one operon (Ope.55) found in the Sporomusacea family is distinct from the other three operons and the gene directly downstream of sbrA encodes a TorD family chaperone protein. This operon is similar to that of the anr gene cluster in D. stibiiarsenatis, suggesting its potential as an antimonate reductase. A putative DMSOR in Pseudomonas, which is closely related to antimonate reductase SbrA (66.7% sequence identity), has been proposed to reduce Sb(V).[31] A recently identified DMSOR family protein from Geobacter sp, (WP_173201954.1), which contains an N-terminal TAT motif, a 4Fe4S binding motif (DX₂CX₃CX_nC) and a Moco binding motif, has been shown to be involved in antimonate reduction.[28] WP_173201954.1, which is closely related to AnrA, is encoded within the anrXSRBAD operon. Transcription of anrA is upregulated under Sb(V) respiring conditions, indicating that AnrA likely functions as an Sb(V) specific respiratory reductase.[28]

2.4. Selenite/selenate reduction

A respiratory selenite reductase (Srr) has been identified from Bacillus selenitireducens with canonical catalytic, electron-transfer, and membrane-anchor subunits.[32] Two rhodanese domain-containing proteins, SrrE and SrrF, are also part of the enzyme complex and may assist in substrate delivery or product removal.[33] SrrA is specific for Se(IV) with a k_cat of 23 s⁻¹ and K_M of 0.145 mM. It does not accept selenate, arsenate, or thiosulfate as a substrate. This high specificity is notable given the significantly higher concentration of As (200 μM vs 38 nM for Se) in the organism’s natural habitat. While Moco-containing enzymes typically catalyze two-electron OAT reactions, the reduction of Se(IV) to elemental selenium Se(0) requires a four-electron transfer. Consequently, the mechanism underlying selenite reduction by SrrA remains unclear. B. selenitireducens can also respire on Se(0), presumably by reducing Se(0) to Se(-II). However, a putative Se(0) reductase in B. selenitireducens has not yet been identified.[34]

Selenate reductase consists of a catalytic subunit (SerA) containing Moco and a 4Fe4S cluster, an electron transfer subunit (SerB), and SerC containing a heme b. The periplasmic location of SerC suggests its role in electron transfer. In this pathway, a diheme cytochrome, Cyt c₄, accepts electrons from the quinol pool and donates electrons to SerABC.[35] A second respiratory selenate reductase (Srd) has been identified from B. selenatarsenatis SF-1.[36] Srd is periplasmic and it consists of a catalytic subunit (SrdA), an electron transfer subunit (SrdB), and a membrane anchor (SrdC), which provides the electron to the SrdA from the quinol pool. Recently, a periplasmic selenate reductase (SrnABCD) has been identified in Enterobacter cloacae.[37] It consists of a catalytic subunit (SrnA), an electron transfer subunit (SrnB), a membrane anchor (SrnC) and a chaperone protein (SrnD). SerA and SrnA are only distantly related despite catalyzing the same reaction, suggesting convergent evolution. Finally, the E. coli ynfEFGH gene cluster encodes two Moco-containing proteins, which are involved in selenate/tellurate reduction.[38] Reductase assays complemented with gene disruption experiments showed that YnfE reduces selenate to selenite, while both YnfE and YnfF reduce tellurate to tellurite.

2.5. Sulfate reduction

Three previously unknown periplasmic sulfate reductases (psr1A, psr2A and psr3A) from nontyphoidal Salmonella have been reported.[39] These enzymes facilitate the formation of H₂S and iron-sulfur clusters. Although all three enzymes contain the Moco and a conserved Mo-coordinating Ser residue, deletion of the Moco biosynthesis gene moaA or mutation of the conserved Ser does not abolish the activity. These results led to the proposal that sulfate reduction does not depend on Moco, although the 4Fe4S cluster is essential for the activity, as disruption of this cluster led to loss of sulfate reductase activity. However, it is not clear how this cluster catalyzes the reduction of sulfate to H₂S. At present, the role of Moco in an unknown activity cannot be ruled out.

Sulfate is also an important part of microbial sulfur disproportionation (MSD). Sulfate is produced during the oxidation of thiosulfate, sulfite or sulfur, whereas sulfide is produced during the reduction of the same species. Recent work suggests a gene cluster encoding molybdopterin reductase A and B subunits (designated as MolyA and MolyB) and a TorD/DmsD chaperone protein (designated as MolyC), as the one of the best candidates for MSD in Desulfobacterota and Nitrospirota.[40] Based on the distribution and phylogeny of the genes in the moly gene cluster, it was suggested that moly gene cluster in Desulfobacterota and Nitrospirota likely existed in the last universal common ancestor of these phyla. If this gene cluster is indeed involved in MSD, it will indicate a very ancient emergence of MSD.

2.6. Iodate reduction

Iodate reductase (idrABP1P2) plays an important role in the global iodine cycle.[41] An idrA knockout mutant of Denitromonas does not grow in iodate, demonstrating its essential role in iodate reduction.[42] IdrA and IdrB are related to the catalytic and electron transfer subunits of AioAB, although IdrB, instead of IdrA, contains an N-terminal TAT motif.[43] IdrAB reduces IO₃^- to HOI and H₂O₂. HOI subsequently disproportionates to I^– and IO₃^–. The associated idrP1 and idrP2 encode diheme proteins related to cytochrome c peroxidase and are proposed to detoxify H₂O₂. However, the mechanism of iodate reduction remains unclear, particularly the mechanism of formation of H₂O₂ and the roles of IdrP1P2. An alternate pathway involving a chlorite dismutase (Cld) like protein (PSCT_04366), which was one of the most abundant proteins expressed in iodate respiring cells, has been proposed. However, comparative genomics does not favor the involvement of a Cld-like protein in iodate reduction.[42]

2.7. CO₂ reduction/Formate oxidation

Formate dehydrogenase catalyzes reversible interconversion of CO₂ and formate. The enzyme is thought to contain a sulfido ligand, which accepts hydride during formate oxidation. In vitro sulfuration of R. capsulatus Fdh has been shown to reactivate the enzyme, which has enzymatic activity comparable to Fdh obtained from heterologous expression in E. coli.[44] Insertion of sulfido ligand was confirmed using EPR spectroscopy, which shows a Mo(V) center ligated by six sulfur ligands. Recent EXAFS analysis on oxidized C. necator FdsDABG does not support the presence of a sulfido group in the Moco. Instead, DFT calculations indicate that the oxidized enzyme has a cysteine persulfdo structure. Following hydride transfer, the persulfido bond cleaves to form Mo-SH, suggesting that the redox chemistry does not involve the metal center.[45] This conclusion is further supported by EXAFS analysis, showing that both the oxidized and reduced enzyme contain Mo(IV).

Previous studies suggested the presence of two substrate access tunnels in formate dehydrogenase,[46] although experimental evidence for the second tunnel is lacking. Recent studies on FdhAB from N. vulgaris show the presence of a substrate retention site at the junction of two substrate tunnels, which leads to the active site.[47] This retention site plays a role in increasing CO₂ reduction by increasing substrate concentration near the active site. Variants affecting this retention site selectivity affect CO₂ reduction but have little impact on formate oxidation. Thus, these studies provide experimental support for a CO₂-specific tunnel and identify structural features responsible for efficient CO₂ reduction.

2.8. C-S bond formation

Moco-containing ergothioneine synthase (MES) in Caldithrix abyssi catalyzes sulfur atom transfer during the conversion of hercynine to ergothioneine.[48] This enzyme has two domains: an N-terminal Moco binding domain and a C-terminal pyridoxal-5ʹ-phosphate dependent cysteine desulfurase domain. The proposed mechanism of ergothioneine biosynthesis involves transfer of the sulfur atom from the C-terminal cysteine desulfurase domain to the N-terminal Moco via an intramolecular persulfide relay. In this process, the Mo(IV) reacts with persulfide to produce Mo(VI)=S. Then, the imidazole ring of hercynine attacks the sulfido ligand of Mo(VI)=S to form ergothioneine. Another DMSOR family member, thiosulfate reductase, which reduces thiosulfate to H₂S and bisulfite, might also function via sulfur atom transfer (Figure 4). [49] However, this hypothesis needs to be tested.

Figure 4.

Proposed mechanism of sulfur atom transfer during thiosulfate reduction.

2.9. Dehydroxylation

Recent discovery of catechol dehydroxylases expanded the functional diversity of the DMSOR family.[50] Catechol dehydroxylases, found in intestinal bacteria such as Eggerthella lenta and Gordonibacter pamelaeae, metabolize dietary compounds, altering their biological activity and bioavailability. For example, Hydrocaffeic acid dehydroxylase (Hcdh) from Gordonibacter dehydroxylates hydrocaffeic acid.[51] The oxidized Hcdh contains a Mo center coordinated by four thiolate ligands from two pyranopterin units, one cysteine and one oxo group (1.69 Å).[52] Upon the reduction of the enzyme, Mo=O is reduced to produce Mo(IV) containing a water derived ligand (2.1 Å). Mutagenesis studies identified the Mo-coordinating Cys and a conserved Asp210 as essential for catalytic activity. In the proposed mechanism of dehydroxylation, Asp210 plays an important role by lowering the activation barrier for catechol ring dearomatization. Interestingly, the UV-vis spectrum of the aerobically purified enzyme did not show features of 4Fe4S cluster, indicating the lack of 4Fe4S in Hcdh.

3. Role of DMSOR family on biogeochemical cycles

Recent analysis of 47,000 DMSOR homologs suggests that these enzymes are defined by a conserved structural fold rather than by the Mo or W cofactor itself.[1] This analysis also indicates that formate dehydrogenase and NasC/NasA are the most abundant families, representing 21.56% and 18.6% of homologs, respectively. Phylogenetic analysis identifies formylmethanofuran dehydrogenase subunit B (FwdB) as the most ancient lineage and reveals four major evolutionary clades distinguished by substrate preference and active site ligation..[53] Evolution within the DMSOR superfamily occurred through the loss of the metal cofactor or the gain or loss of protein domains. Finally, phylogenetic analysis also provides information about poorly characterized enzymes. For example, IdrA clusters with AioA, suggesting a similar active site architecture containing a 3Fe4S cluster and lacking a Mo-coordinating residue.

4. Health relevance of DMSO family members

DMSOR family also contributes to bacterial pathogenesis.[54] For example, DmsABC supports virulence in Non-typeable H. influenzae.[55] NapA deletion reduces the ability of Salmonella typhimurium to infect host cells, supporting the role of Nap in pathogenicity.[56] Recent studies showed that molybdate transport protein ModA regulates nitrate reductase activity to increase intestinal colonization and extraintestinal dissemination of Klebsiella pneumoniae.[57] Further, host-derived nitrate boosts E. coli growth and drives kidney disease progression by enhancing indole production.[58,59] Previously, it has been proposed that the inhibition of indole production might enhance the activity of quinolone antibiotics against E. coli persisters.[60] Thus, selective inhibition of nitrate reduction might be a potential drug target. This assertion is further supported by the work showing that the oral administration of sodium tungstate inhibited molybdoenzymes and selectively decreased gut colonization of Gammaproteobacteria, including E. coli.[61] The work also showed that reduced colonization of Gammaproteobacteria decreased intestinal inflammation and reduced the incidence of colonic tumors in mouse models.

5. Biotechnological application of DMSO family members

DMSOR family members have attracted interest in biotechnological applications.[62,63] DMSOR has been used for preparing chiral sulfoxide through kinetic resolution, while Mannheimia succiniciproducens has been metabolically engineered to produce malic acid by introducing Actinobacillus succinogens DMSOR.[64,65] Electrochemical reversible interconversion of CO₂ and formate has been demonstrated for formate dehydrogenase.[66–68] Although O₂-induced deactivation of enzymes remains a challenge, an O₂-tolerant formate dehydrogenase from S. oneidensis has been used to construct a direct bioelectrocatalytic CO₂ reduction system.[69] Also, steroid C25 dehydrogenase has been used for electrobiosynthesis of 25-hydroxy vitamin D3.[70] Thus, in principle, other Moco-containing enzymes (O₂ tolerance is desirable) can also be optimized for bioelectrocatalysis.

Conclusions and future outlook

DMSO family members catalyze a diverse array of reactions important to biogeochemical cycles, biotechnology, and health. Despite decades of work, details of how some enzymes of this family work are emerging. For example, recently, it has been reported that formate dehydrogenase works via ligand based redox chemistry rather than metal center redox. At the same time, new modes of reactivity, such as sulfur atom transfer during ergothioneine biosynthesis, are discovered. Likewise, catechol dehydroxylase is of general interest due to the scarcity of general and mild synthesis methods for dehydroxylation of aromatic rings.

For several DMSOR family enzymes, both the catalytic mechanism and the enzyme itself remain poorly understood. For example, sulfite reductase, sulfate reductase, and iodate reductase catalyze the reduction of sulfite, sulfate, and iodate, respectively. For sulfate reductase, the role of Moco remains unclear, whereas for selenite and iodate reduction, the functions of both the Moco-containing subunit and accessory proteins are still unresolved. Similarly, for TMAO reductase, determination of the TorA complex structure would provide critical insight into subunit interactions and intersubunit electron transfer. Continued biochemical, structural, and mechanistic studies are therefore essential to elucidate how these enzymes function. Such knowledge will not only deepen our understanding of metalloenzyme catalysis but also expand opportunities for biotechnological and therapeutic applications.