Auxiliary iron-sulfur cofactors in radical SAM enzymes.

Biochimica et Biophysica Acta 1853 (2015) 1316–1334

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta journal homepage: www.elsevier.com/locate/bbamcr

Review

Auxiliary iron–sulfur cofactors in radical SAM enzymes☆ Nicholas D. Lanz a, Squire J. Booker a,b,⁎ a b

Department of Biochemistry and Molecular Biology, The Pennsylvania State University, University Park, PA 16802, United States Department of Chemistry, The Pennsylvania State University, University Park, PA 16802, United States

a r t i c l e

i n f o

Article history: Received 19 September 2014 Received in revised form 15 December 2014 Accepted 6 January 2015 Available online 15 January 2015 Keywords: Radical SAM Iron–sulfur cluster S-adenosylmethionine Cofactor maturation Radical-dependent post-translational modification SPASM/twitch domain

a b s t r a c t A vast number of enzymes are now known to belong to a superfamily known as radical SAM, which all contain a [4Fe–4S] cluster ligated by three cysteine residues. The remaining, unligated, iron ion of the cluster binds in contact with the α-amino and α-carboxylate groups of S-adenosyl-L-methionine (SAM). This binding mode facilitates inner-sphere electron transfer from the reduced form of the cluster into the sulfur atom of SAM, resulting in a reductive cleavage of SAM to methionine and a 5′-deoxyadenosyl radical. The 5′-deoxyadenosyl radical then abstracts a target substrate hydrogen atom, initiating a wide variety of radical-based transformations. A subset of radical SAM enzymes contains one or more additional iron–sulfur clusters that are required for the reactions they catalyze. However, outside of a subset of sulfur insertion reactions, very little is known about the roles of these additional clusters. This review will highlight the most recent advances in the identification and characterization of radical SAM enzymes that harbor auxiliary iron–sulfur clusters. This article is part of a Special Issue entitled: Fe/S proteins: Analysis, structure, function, biogenesis and diseases. © 2015 Elsevier B.V. All rights reserved.

1. Introduction The radical-SAM (RS) superfamily of enzymes is defined by the ability of its members to cleave S-adenosyl-L-methionine (SAM) to methionine and a 5′-deoxyadenosyl radical (5′-dA•), a potent oxidant, using an electron provided by a reduced, [4Fe–4S]1+, cluster (Fig. 1) [1–7]. Cysteines residing in a CX3CX2C motif ligate three of the four iron ions of the cluster, leaving an available coordination site to which SAM binds in a bidentate fashion via its α-carboxylate and α-amino groups [8–12]. All structurally characterized RS enzymes that reside in pfam04055 contain either a full or partial Triose-phosphate Isomerase Mutase (TIM) barrel fold in which the SAM/[4Fe–4S] cluster cofactor is bound. Additional domains that interact with the core fold are also frequently found in various subclasses of RS enzymes [13–25]. While RS enzymes share a common fold, they diverge greatly in the reactions they catalyze and the mechanisms by which they operate. A subset of RS enzymes contains one or more iron–sulfur (Fe/S) clusters in addition to the cluster required for SAM cleavage [26]. This group of enzymes highlights both the diversity of reactions catalyzed by RS enzymes as well as the range of functions of Fe/S clusters. The authors have already penned a review describing the identification and functions of

☆ This article is part of a Special Issue entitled: Fe/S proteins: Analysis, structure, function, biogenesis and diseases. ⁎ Corresponding author at: 104 Chemistry Building University Park, PA 16802, United States. Tel.: +1 814 865 8793; fax: +814 865 2927. E-mail address: [email protected] (S.J. Booker).

http://dx.doi.org/10.1016/j.bbamcr.2015.01.002 0167-4889/© 2015 Elsevier B.V. All rights reserved.

RS enzymes containing auxiliary Fe/S clusters [26]. Therefore, this work will not strive to reproduce what has already been written, but rather focus on the most recent advances made in the field. While the function of many of the auxiliary Fe/S clusters remains unknown, recent work does provide insight into their roles in several enzymes. Notably, diverging roles for the auxiliary clusters in sulfur-donating enzymes have been presented, multiple crystal structures of enzymes containing the SPASM/twitch domain have been reported, mechanistic insight has been provided for the enzymes responsible for the maturation of the molybdenum cofactor and various hydrogenase cofactors, and several new RS enzymes have been added to the class (vide infra).

2. Sulfur donating enzymes Two distinct types of sulfur-insertion reactions have been described to date (Fig. 2). The first type involves the insertion of one or two sulfur atoms into an organic substrate. Biotin synthase (BioB) generates a thioether bond between carbons 6 and 9 of dethiobiotin in the final step of the biosynthesis of the biotin cofactor [27–29]. The other member of the group, lipoyl synthase (LipA), inserts sulfur atoms at carbons 6 and 8 of a protein-bound n-octanoyl chain [30–33]. The second type of reaction involves the generation of a methylthio moiety and its subsequent attachment to a substrate carbon atom [34]. Three classes of methylthiotransferases (MTTases), represented by RimO, MiaB, and MtaB, catalyze this type of reaction, of which only RimO and MiaB have been studied mechanistically [34–36]. RimO catalyzes the methylthiolation of the β-carbon of a specific aspartyl residue of the

N.D. Lanz, S.J. Booker / Biochimica et Biophysica Acta 1853 (2015) 1316–1334

1317

attachment to an unactivated carbon atom, evidence now points to differing roles of the auxiliary clusters in these enzymes.

2.1. Lipoyl synthase LipA and BioB catalyze highly similar reactions. The reaction of BioB (Fig. 2A), the insertion of a single sulfur atom between C6 and C9 of dethiobiotin, requires the presence of a [2Fe–2S] cluster in addition to the RS [4Fe–4S] cluster. The [2Fe–2S] cluster has a unique ligation environment comprised of three cysteines and one arginine [14]. A number of studies were conducted to determine the cluster content of BioB [38–42], the sacrificial role of its [2Fe–2S] cluster [43–50], as well as intermediates in its reaction [42,51]. As there have been no recent developments, the focus of this section will be on the related enzyme, LipA. LipA catalyzes the stepwise insertion of two sulfur atoms into an octanoyl chain attached to a lipoyl carrier protein (LCP) (Fig. 2B) [30–33]. As is the case for BioB, LipA cleaves two molecules of SAM for each molecule of product produced [52]. However, in place of the [2Fe–2S] cluster found in BioB, LipA contains an auxiliary [4Fe–4S] cluster [52]. In the proposed mechanism, the first 5′-dA• generated abstracts a hydrogen atom (H•) from C6 of the octanoyl chain, resulting in a transient carbon-centered substrate radical (Fig. 3) [31,52]. The substrate radical attacks a bridging μ-sulfido ion of the auxiliary [4Fe–4S] cluster of LipA, generating a covalently cross-linked intermediate between the octanoyl chain of the LCP and the [4Fe–4S] cluster of LipA. The second 5′dA• abstracts a H• from C8 of the octanoyl substrate, and the resulting C8 radical attacks a second bridging μ-sulfido ion of the auxiliary cluster, complete the synthesis of the lipoyl cofactor upon the addition of two protons. Significant mechanistic detail of the LipA reaction was obtained from several studies on the enzyme. Cicchillo et al. provided evidence that both sulfur atoms are derived from the same polypeptide in an Fig. 1. Cleavage of SAM by RS enzymes.

S12 subunit of the ribosome [37]. MiaB performs a similar reaction, catalyzing the methylthiolation of an adenosine nucleotide at position 37 (A37) on several tRNAs [35,36]. While the reactions all involve sulfur

Fig. 2. Sulfur insertion and methylthiolation reactions catalyzed by RS enzymes.

Fig. 3. Proposed mechanism for lipoyl synthase.

1318


experiment where two forms of LipA, overproduced separately in media containing either natural abundance sulfide or 34S-labeled sulfide, were mixed together and allowed to react under turnover conditions [32]. The sulfur atoms in the resulting lipoyl products were predominantly both 32S-labeled or 34S-labeled. If each polypeptide contributed only one sulfur atom, the lipoyl products would have been expected to have mixtures of the two isotopes in a ratio of 1:2:1 for the 32S/32S-, 32 34 S/ S-, and 34S/34S-labeled species, which were not observed. Importantly, a significant isotope effect was observed on the LipA reaction when a substrate containing a perdeuterated octanoyl moiety was used [52]. Approximately one-half of an equivalent of singlydeuterated 5’-deoxyadenosine (5′-dA) was observed after 1 h of incubation time; however, the lipoyl product was not detected. It was determined that only one of the two deuterium atoms was abstracted, effectively arresting the reaction at the halfway point. Work by Bryant et al. demonstrated that a thermostable LipA from Sulfolobus solfataricus could act on a short tripeptide substrate containing the requisite octanoyl-lysine [53]. This substrate provided a more manageable means of measuring LipA activity, because it could be readily synthesized and reliably quantified by chromatographic methods. Douglas et al. first identified the 6-mercapto-octanoyl-peptide intermediate using a combination of mass spectrometry and two-dimensional nuclear magnetic resonance (NMR) spectroscopy after quenching the reaction in acid and purifying the intermediate by high-performance liquid chromatography HPLC [31]. The authors measured the mass shift of the products of a reaction in the presence of 8,8,8-2H3octanoyl-peptide. All three deuteriums were retained in the monothiolated product, indicating that a H• was not abstracted from C8. Sulfur insertion at C6 of the intermediate was confirmed by characterization of the intermediate isolated from a reaction containing unlabeled substrate by COSY NMR spectroscopy. More recently, the auxiliary cluster of LipA was characterized at the intermediate stage of the reaction by a combination of mass spectrometry and EPR and Mössbauer spectroscopies [54]. The intermediate was generated using an 8,8,8-2H3-octanoyl-LCP, and the protein complex, which was observed as a cross-link between LipA and the octanoylLCP, was purified by anion-exchange chromatography. Formation of the cross-link required incubation of LipA with the octanoyl-LCP substrate under turnover conditions. Moreover, the cross-link was labile in the presence of acid or detergent, but stable under chromatographic conditions that separate LipA and the LCP, which bind tightly, before initiating the reaction. This data suggests that the cross-link is mediated by an iron–sulfur cluster of LipA rather than a direct covalent interaction between the two polypeptides. The cross-linked intermediate was also generated with an eight-amino acid peptide substrate under three different conditions [54]. In the first instance, an 8,8,8-2H3-octanoyl-peptide additionally tagged with biotin was used. LipA from Thermus thermophilus was incubated with this peptide under turnover conditions, and the resulting cross-linked species was separated from unreacted enzyme using a streptavidin-mutein column. In the second instance, the same peptide was used to arrest turnover after the first half of the reaction using enzyme from Escherichia coli (EcLipA). The yield of the intermediate species was sufficiently high that further purification was unnecessary. In the last instance, taking advantage of the fact that the first half of the LipA reaction proceeds much more rapidly than the second half, EcLipA was reacted with unlabeled octanoyl-peptide in the presence of only one equivalent of SAM, resulting in the formation of the crosslinked intermediate as the predominant species. Using the unlabeled peptide substrate, the authors were also able to demonstrate the chemical and kinetic competence of the isolated intermediate. Upon formation of the cross-link using one equivalent of SAM, excess peptide and other small-molecule reaction components were removed by gelfiltration chromatography. After reintroduction of SAM and dithionite, a low-potential reductant used to initiate turnover, the lipoyl product was observed to form in the absence of additional octanoyl-peptide,

demonstrating the catalytic competence of the cross-linked intermediate. Furthermore, the rate of formation of the lipoyl product from the crosslinked species was consistent with the overall rate of the reaction, indicating the kinetic competence of the reaction from the cross-linked species. For each of the methods of isolating the cross-linked species, samples were prepared with 57Fe-labeled LipA for analysis by EPR and Mössbauer spectroscopies [54]. Although very small quantities of paramagnetic species were detected, the Mössbauer spectra provided interesting insight. As expected, the spectrum of the as-isolated enzyme was composed of a quadrupole doublet with parameters indicative of the presence of [4Fe–4S] clusters. Unexpectedly, the Mössbauer spectra revealed the presence of a [3Fe–4S] cluster present in the cross-linked species regardless of the method of isolation. Though the fraction of [3Fe–4S] cluster present in the anion-exchange purified sample suggested that greater than one equivalent of 3Fe species was present, the amount in excess of the first equivalent was presumed to be caused by the purification method. In the other three cases, where gentler purification methods were used or no purification was necessary, the amount of 3Fe species present correlated with the amount of cross-linked species. The remaining iron in the spectrum was split between ferrous iron that corresponded to the iron lost from the auxiliary cluster, and a mixture of [4Fe–4S] and site-differentiated [4Fe–4S] clusters (SAM or methionine-bound) that correspond to the unaffected RS cluster. Upon completion of the reaction, the Mössbauer spectra indicated the presence of a single [4Fe–4S] cluster remaining in LipA. Again, this observation is consistent with the RS cluster remaining stable throughout the reaction. The remainder of the iron is a mixture of thiolate-ligated ferrous iron, N/O-ligated ferrous iron, and 2Fe-clusters that are likely the byproduct of the destruction of the auxiliary cluster. Like the [2Fe–2S] cluster of BioB, the auxiliary cluster of LipA appears to be sacrificed for the sake of a single turnover. The recent crystal structure of LipA from Thermosynechococcus elongates showed that the protein is composed of a partial TIM barrel that is typical of RS enzymes, but also revealed a novel serine ligand to the auxiliary cluster (Fig. 4A) [21]. As expected, the RS cluster is coordinated by the CX3CX2C motif, and the auxiliary cluster is ligated by the N-terminal CX4CX5C motif as well as by a conserved serine residue at the C-terminus of the enzyme (Fig. 4B). Two structures of LipA were reported. In one structure, the enzyme was crystallized in the presence of SAH instead of SAM. S-adenosylhomocysteine (SAH), which lacks the sulfonium methyl group of SAM, was bound to the cluster similarly to how SAM binds in other RS enzymes. In the second case, LipA was crystallized in the presence of SAM, but a breakdown product of SAM, 5′-methylthioadenosine (5′-MTA), was bound in the active site, and dithiothreitol was coordinated to the unique iron site of the RS cluster. Using the information in these structures, SAM could be modeled into the structure as well as octanoyl-lysine methylamide as a substrate surrogate. In this model, the sulfonium ion of SAM is 4.4 Å from the nearest iron ion of the RS cluster, and the 5′ carbon of SAM is 4.4 Å and 5.2 Å from C6 and C8 of the octanoyl chain, respectively, positioning the substrates suitably for the reaction to occur. Substitution of the serine ligand to the auxiliary cluster with either alanine or cysteine abolished the ability of the enzyme to generate a lipoyl product [21]. However, abortive cleavage of SAM was still observed. The importance of the serine ligation to the auxiliary cluster is consistent with the spectroscopic data previously observed in which an iron ion is lost from the auxiliary cluster during the course of the formation of the monothiolated intermediate (vide supra) [54]. Serine has a higher pKa value than cysteine, and its protonation, while coordinated to the unique Fe ion, might trigger its dissociation from the cluster, allowing for release of the Fe ion to which it was ligated. Perhaps the most intriguing question left to be answered with regard to the BioB and LipA reactions is that of the mechanism of cluster reassembly. One possibility is that the enzymes responsible for building and inserting Fe/S clusters are also capable of reassembling the auxiliary clusters after each turnover. These proteins are encoded in the isc or suf


RS Cluster

A

1319

B Substrate binding channel

Auxiliary Cluster

RS Cluster Methionine

Ser283

Auxiliary Cluster Fig. 4. Structure of LipA (A) and coordination of methionine and a serine residue to the RS and auxiliary clusters of LipA, respectively (B) (PDB accession code 4U0P). The RS domain of LipA is shaded in green while the N-terminal and C-terminal extensions that contain the ligands for the auxiliary cluster are shaded in blue and pink, respectively.

operons in E. coli and other organisms. ISA1 and ISA2, two of the enzymes from the mitochondrial Fe/S cluster assembly system in Saccharomyces cerevisiae, were found to support catalysis of BioB, but were not required for the de novo biosynthesis of the Fe/S clusters of the protein. It is possible that these proteins aid in the regeneration of the sacrificial [2Fe–2S] cluster of BioB [55]. It has also been proposed that ISA1 and ISA2 may be required for the biosynthesis of [4Fe–4S] clusters while a second set of proteins, ISU1 and ISU2, are responsible for generating [2Fe–2S] clusters as well as the precursors for [4Fe–4S] clusters [56]. In this model, ISU1 and ISU2 may be necessary to regenerate the auxiliary cluster of BioB. The well-studied iron–sulfur cluster assembly systems of both E. coli and S. cerevisiae include an ATP-dependent chaperone, HscA (Ssq1 in S. cerevisiae) [57]. Unlike homologous chaperones, HscA is constitutively expressed and likely plays an important housekeeping role not associated with stress [58,59]. In addition to the previously studied interaction of HscA with the iron–sulfur cluster scaffold IscU, it was found that HscA also interacts with BioB [60]. HscA can form a complex with both IscU and BioB simultaneously and displays a higher affinity for BioB lacking both iron–sulfur clusters. These results suggest that HscA may recruit IscU to BioB and aid in the regeneration of the auxiliary cluster. A gene encoding an interacting partner of ISA1 and ISA2 in S. cerevisiae, IBA57, was identified by Gelling et al. [61]. Deletion of the IBA57 gene prevented insertion of the iron–sulfur clusters of aconitase and homoaconitase and also inhibited the functions of BioB and LipA in vivo. Defective versions of the human homologs, ISCA1, ISCA2, and IBA57, are associated with severe disease and premature death as well [62,63]. Misformed mitochondria and decreased activity of a subset of mitochondrial Fe/S proteins, including LipA, were observed when these three genes were silenced using RNAi [62]. Given that these genes are also found in E. coli, this pathway is likely equally important in prokaryotes [61]. Two additional human genes, NFU1 and BOLA3, have similar phenotypes [64,65]. Defects in these genes result in decreased protein lipoylation and low activity of lipoyl-dependent multienzyme complexes, again leading to disease and premature death [64,65]. NFU1 is proposed to be an alternative Fe/S scaffold, and BOLA3 is thought to be a reductase that aids in the transfer of clusters from the scaffold to the target protein [64]. RNAi knockdown of NFU1 did not affect the activities of most mitochondrial Fe/S proteins, but impacted LipA and succinate dehydrogenase, suggesting that the role of NFU1 may be specific for LipA. It is yet unknown whether NFU1 and BOLA3 act in the de novo biosynthesis of the iron–sulfur clusters in LipA or if they may be involved in cluster regeneration during the catalytic cycle.

2.2. Methylthiotransferases The MTTases are often compared with LipA and BioB, because the reactions they catalyze involve sulfur attachment to unactivated carbon atoms. Two MTTases, RimO and MiaB, have been studied mechanistically. RimO catalyzes the methylthiolation of the β-carbon of Asp89 (E. coli numbering) of the S12 subunit of the ribosome (Fig. 2D) [37]. Like BioB and LipA, this reaction requires the abstraction of a H• from an sp3-hybridized carbon and the subsequent formation of a carbon\sulfur bond. Similarly, MiaB catalyzes the methylthiolation of C2 of an isopentenylated adenosine (i6A) immediately 3′ to the anticodon on several tRNAs, forming 2-methylthio-i6A (ms2i6A) (Fig. 2C) [66–68]. In this case, the methylthio group replaces an H• on an sp2-hybridized carbon atom. A significant difference between the MTTases and sulfur insertion enzymes lies in their usage of SAM. Both LipA and BioB cleave two molecules of SAM in a stepwise fashion, generating two equivalents of 5′-dA. MTTases also utilize two equivalents of SAM; however, only one equivalent is used to generate a 5′-dA•, while the second equivalent is used as the methyl donor to cap the sulfur atom. The co-product of this typical SN2-based methyl transferase reaction is SAH. Like LipA, RimO and MiaB harbor two [4Fe–4S] clusters. Given the similarities in the reaction to LipA and BioB, it is tempting to assign a sacrificial role to the auxiliary clusters of the MTTases as well. Because early studies reported no more than one turnover per polypeptide for MiaB and RimO, it was postulated that MTTases may insert the sulfur atom into the substrate in a mechanism that parallels that of LipA, and that this inserted sulfur atom would be subsequently capped with a methyl group. Early evidence supported this model [69]. Strains of E. coli auxotrophic for methionine and cultured under methionine starvation conditions produced a derivative of the isopentenylated adenosine that could be methylated in a separate reaction. This observation led to the conclusion that the unknown adenosine derivative was a sulfurated intermediate. An additional mechanism for methylthiolation invokes the generation of a protein-bound methylthio moiety followed by its transfer to the substrate in a radical reaction. The observation of SAH production in the absence of a low-potential reductant or substrate favored this possibility [25,70]. Landgraf et al. reported the transfer of a methyl group from SAM to RimO or MiaB in the absence of a low-potential reductant, and identified a protein-bound methylthio intermediate, lending credence to the latter mechanism [71]. The chemical and kinetic competence of the protein-bound methylthio intermediate was tested in differential labeling experiments. MiaB and RimO were incubated

1320


with unlabeled SAM in the absence of substrate and reductant to allow for methyl transfer. Excess SAM and SAH were removed by gel-filtration chromatography and the enzymes were reacted with substrate and S-adenosyl-L-[methyl-2H3]methionine (2H3-SAM) under turnover conditions. For both MiaB and RimO, a burst of unlabeled product was observed followed by formation of 2H3-labeled product, indicating that the unlabeled methyl group is transferred initially to the product followed by the labeled methyl group in subsequent turnovers (Fig. 5A). Transfer of the methyl group from SAM to the enzyme was also studied using S-adenosyl-L-[methyl-14C]methionine (methyl-14CSAM) or S-[8-14C]adenosyl-L-methionine ([adenosyl-14C]SAM). The reactions were then subjected to gel-filtration chromatography, and the radioactivity in each fraction was counted. When MiaB or RimO were reacted with methyl-14C-SAM, the radioactivity eluted in the protein fractions. In the control experiments, in which adenosyl-14C-SAM was used, the radioactivity was found in the small-molecule fractions, indicating that the first result was not due to tight binding of SAM to the enzymes. If reactions with MiaB or RimO and methyl-14C-SAM were treated with urea prior to separation on the gel-filtration column, the radioactivity appeared in the small-molecule fraction. Thus, the methyl acceptor is labile to denaturation and is not likely to be an amino acid of the protein. It was also determined that exogenously added methanethiol could compete with the enzymatically-derived methylthio moiety [71]. Reactions containing MiaB or RimO, 2H3-SAM, reductant, and substrate were supplemented with methanethiol, and the RNA or peptide products containing either an unlabeled- or a 2H3-methyl group were quantified by LC–MS. A mixture of the labeled and unlabeled products was observed, indicating that both enzymes are capable of taking up and utilizing free methanethiol. Finally, direct detection of methanethiol produced by RimO and MiaB was observed. In this experiment, reactions of RimO and MiaB with SAM were conducted in sealed vials and quenched in acid, and then samples of the headspace were analyzed by GC–MS for the presence of methanethiol. Parallel samples were analyzed by LC–MS in order to quantify SAH production. The amount

of SAH and methanethiol produced was tightly correlated and approximately half an equivalent with respect to enzyme concentration. Taken together, these results demonstrate the presence of a protein-bound methanethiol intermediate, which implies that the mechanism in which the methylthio group, formed by transfer of a methyl group from SAM to a protein sulfur atom, is transferred intact to the substrate in a radical-dependent reaction. Furthermore, the source of the sulfur atom must either be sulfide ligated to one of the Fe/S clusters or a sulfur atom from the Fe/S cluster itself, given that the intermediate becomes labile when the enzyme is denatured. Forouhar et al. also studied the mechanism of the RimO and MiaB reactions, and reported that MiaB catalyzed formation of up to four equivalents of ms2i6A per mol of enzyme [24]. The authors noted that the enzyme was reconstituted with iron and sulfide and contained sulfide in excess of that required for the two [4Fe–4S] clusters. Upon further supplementation of the reaction with sodium sulfide, MiaB was observed to catalyze 12 or more turnovers. Similarly, addition of methanethiol or methaneselenol to reactions also increased the amount of product generated, though not to the same extent as the addition of sulfide. In similar experiments, multiple turnovers were also observed for RimO. These observations suggest that, unlike BioB and LipA, the MTTases do not use their auxiliary Fe/S clusters in a sacrificial manner. Consistent with this premise, HYSCORE experiments performed on MiaB and RimO variants containing Cys → Ser substitutions at all cysteines in the CX3CX2C motif that ligate the RS cluster, revealed a coupling between the 77Se isotope in methaneselenol and the electron spin in the reduced form of the auxiliary cluster. The observed spectrum was consistent with methaneselenol binding to the open coordination site of the auxiliary cluster and DFT calculations predict that the selenide is not intercalating into the cluster by replacing one of the bridging μ-sulfido ions. Forouhar et al. also reported a crystal structure of holo-RimO. Although a crystal structure had been previously determined, it lacked both of the Fe/S clusters and portions of the polypeptide containing the cysteines that ligate the clusters [25]. Perhaps the most intriguing

Fig. 5. Proposed mechanisms of methylthiolation of Asp88 of the S12 subunit of the E. coli ribosome by RimO. In both proposed mechanisms, a 5′-dA• abstracts a hydrogen atom from Cβ of the substrate aspartyl residue. The resulting carbon-centered radical attacks either a bridging μ-sulfido ion of the auxiliary cluster capped with a methyl group (A) or a methyl-capped persulfide species coordinated to the unique iron ion of the auxiliary cluster (B).


aspect of the updated RimO structure is a pentasulfide bridge that connects the two unique iron ions of each cluster. Although the bridge is most likely an artifact of the reconstitution procedure used to generate the holo-form of the enzyme, it does provide possible insight into the mechanism of catalysis utilized by MTTases. First, it explains the source of the extra sulfide found in reconstituted RimO and MiaB and accounts for the multiple turnovers performed by these enzymes. While the pentasulfide bridge would likely preclude SAM from binding to the RS cluster, a sulfide ion or a shorter polysulfide likely binds to the open coordination site of the auxiliary cluster and is then capped by a methyl group donated by SAM (Fig. 5B). The enzyme then uses radical chemistry to transfer the methylthio moiety to the substrate. A new sulfide ion can then be appended to the cluster and catalysis can resume without the need to repair or reassemble an iron–sulfur cluster. The difference between the mechanisms of LipA and BioB versus the MTTases may derive from the inherent difference in the abundance of their products. Relatively small quantities of cofactors such as lipoic acid and biotin are required for cellular survival. Therefore, the inefficient process of rebuilding an Fe/S cluster after its destruction between each turnover may be permissible. A sacrificial role for the auxiliary cluster may not be tolerated for enzymes such as MiaB and RimO, which must operate on the cellularly-abundant tRNA and ribosome substrates. Therefore, nature may have evolved a distinct mechanism for the MTTases that allows for multiple turnovers without the need to rebuild an iron–sulfur cluster after each round. While RimO and MiaB may be catalytic, some evidence suggests one or both of the clusters of MiaB may be unstable during turnover. Recent studies have found additional factors that are required for optimal activity of MiaB in vivo. YgfZ, the E. coli homolog of IBA57 (vide supra), is a folate-dependent enzyme that can maintain activity in Fe/S enzymes, including MiaB [72,73]. While it has been determined that it is tetrahydrofolate-dependent, its mechanism of action is yet unknown [73–75]. Two additional proteins, a monothiol glutaredoxin (GrxD) and an iron–sulfur carrier protein (NfuA), have been shown to interact with MiaB [76]. It was discovered that untagged MiaB could be co-purified with his-tagged GrxD or NfuA or vice versa. Additionally, surface plasmon resonance was used to measure a KD of 12 μM for apo-MiaB binding to apo-GrxD and a KD of 44 μM for apo-MiaB binding to apo-NfuA. Holo-GrxD and NfuA were also tested for the ability to transfer an iron–sulfur cluster to apo-MiaB. GrxD was unable to do so; however, NfuA was able to transfer a [4Fe–4S] cluster to apo-MiaB and the triple variant containing Cys → Ala substitutions in the CX3CX2C motif, indicating that the auxiliary cluster could be restored by NfuA. Finally, the ratio of ms2i6A to i6A in tRNAs in E. coli was measured for WT and grxD and nfuA deletion strains. Both the grxD and nfuA deletion strains showed a modest decrease in the ratio of the modified bases; however, the double deletion showed a five-fold difference in the ratio, suggesting that GrxD and NfuA are important for MiaB activity in vivo. If the auxiliary Fe/S clusters of MTTases are not utilized in a sacrificial manner as Forouhar et al. suggests [24], one can speculate that these additional factors are not required for reassembly of an Fe/S cluster, but instead may supply the sulfide for the reaction. 3. SPASM domain-containing enzymes A subfamily of RS enzymes contains a motif of seven cysteines that ligate two additional Fe/S clusters. The SPASM domain, named for the founding members that function in the maturation of subtilosin A, pyrroloquinoline quinone, anaerobic sulfatase, and mycofactocin, consists of a CX9–15GXC–gap–CX2CX5CX3C–gap–C motif [20,77,78]. The SPASM subfamily also includes enzymes with a truncated version of the SPASM domain termed the twitch domain. The shorter twitch domain lacks the CX2CX5CX3C motif of the SPASM domain and coordinates a single auxiliary Fe/S cluster; however, it contains many of the structural features of the larger SPASM domain (vide supra). Bioinformatic approaches have identified more than 18,000 unique sequences of

1321

enzymes that contain either the SPASM or twitch domains, which accounts for a significant percentage of the greater RS superfamily [79]. While a large number of enzymes contain this motif, only a few have been well characterized. Among those, two dehydrogenases, the anaerobic sulfatase maturating enzymes (anSMEs) and 2-deoxy-scylloinosamine dehydrogenase (BtrN), as well as the heme b synthase for Desulfovibrio vulgaris will be highlighted here. 3.1. Anaerobic sulfatase maturating enzymes Arylsulfatases require a distinctive formylglycyl (FGly) residue in their active sites to carry out their function of hydrolyzing organosulfate monoesters. The FGly residue is generated by a two-electron oxidation of either a cysteine or serine residue of the arylsulfatase (Fig. 6A). Two distinct types of maturases catalyze this reaction, the first of which, known as FGly generating enzymes or FGEs, is oxygen-dependent and found both in eukaryotes and prokaryotes [80,81]. The oxygenindependent maturases, designated anaerobic sulfatase maturating enzymes (anSMEs), are SPASM domain-containing RS enzymes [82–84]. Initial characterization of the anSMEs from Clostridium perfringens(anSMEcp) and Bacteroides thetaiotaomicron (anSMEbt) suggested that these enzymes contained only one [4Fe–4S] cluster per protein [85,86]. A more rigorous characterization of the enzyme from Klebsiella pneumoniae (anSMEkp) revealed that it harbored three [4Fe–4S] clusters per polypeptide, which was later confirmed for the enzyme from B. thetaiotaomicron [77,78]. However, the function of the auxiliary clusters was not immediately apparent. The mechanisms proposed for anSMEs involve abstraction of H• from the β-carbon of the cysteine or serine residue followed by loss of a proton and an electron to generate the aldehyde or a thioaldehyde that undergoes hydrolysis to the aldehyde and H2S (Fig. 6). Benjdia et al. demonstrated direct abstraction of the Cβ H•, providing evidence for this mechanism [87]. The role of the auxiliary clusters in this mechanism is more cryptic. Two proposed possibilities invoke a role in electron transfer for the auxiliary clusters, while a third suggests that they play only a structural role in the enzyme. In one instance, it has been suggested that the auxiliary clusters provide an electron-transport pathway to provide the requisite electron to reduce the RS cluster and subsequently cleave SAM (Fig. 6C). Benjdia et al. found that variants of the cysteines that coordinate the auxiliary clusters of anSMEbt exhibited a 50- to 100-fold decrease in SAM cleavage, hinting that reduction of the RS cluster is affected in the absence of the auxiliary clusters [78]. However, given that the flavodoxin/flavodoxin reductase reducing system is capable of providing the electron to reduce the RS cluster of other RS enzymes that do not have auxiliary clusters, it is unlikely that anSMEs would require additional clusters for this purpose [77]. A second possibility is that the second electron during the two-electron oxidation is shuttled to an electron acceptor via the two auxiliary clusters (Fig. 6B). Grove et al. also characterized an anSME from C. perfringens and found that it harbors three [4Fe–4S] clusters [88]. The as-isolated enzyme contained 9.6 iron ions and 10.0 sulfide ions per polypeptide. Approximately 95% of the iron in the sample was assigned to [4Fe–4S] clusters by Mössbauer spectroscopy, yielding 2.3 clusters per protein in the as-isolated sample. After reconstitution with iron and sulfide, anSMEcp contained 14.1 iron ions and 12.8 sulfide ions per monomer. Approximately 75% of the iron in the reconstituted protein was in the form of [4Fe–4S] clusters, resulting in 2.7 clusters per protein. These results suggested that anSMEcp also has three [4Fe–4S] clusters. Triple variants containing Cys → Ala substitutions of the cysteines in the CX3CX2C motif were generated and the as-isolated and reconstituted enzymes contained 3.2 and 8.8 iron ions and 7.5 and 15.1 sulfide ions per polypeptide, respectively. This stoichiometry together with the Mössbauer spectrum of the as-isolated sample indicated the presence of 0.6 [4Fe–4S] clusters and 0.3 [2Fe–2S] clusters per monomer, suggesting that the [4Fe–4S] clusters are less stable in the variant enzyme. After

1322


Fig. 6. Reaction (A) and potential mechanisms of anaerobic sulfatase maturating enzymes (B and C).

reconstitution, the variant contained 1.5 [4Fe–4S] clusters per monomer, confirming the presence of multiple auxiliary clusters, though the less than full complement of two clusters still suggests a greater instability of the clusters in the variant. The activity of anSMEcp was measured using peptide substrates containing the physiological cysteine, a serine in place of the cysteine, or a selenocysteine in place of the cysteine [88]. The enzyme was capable of oxidizing all three substrates, though the use of the selenocysteine-containing substrate led to significant uncoupling of 5′-dA and product formation after 10 min. Although anSMEcp is capable of oxidizing the serine-containing substrate, the rate of the reaction is approximately three-fold slower than for the natural substrate. The enzyme was also able to utilize both dithionite and the flavodoxin/ flavodoxin reductase reducing system as the source of the electron required to cleave SAM, though, as was the case for anSMEkp, the rate of the reaction with the flavodoxin reducing system was significantly slower. Reductive cleavage of SAM requires a single electron, and the abstraction of a H• by the 5′-dA• is a one-electron oxidation of the substrate. Because the substrate ultimately undergoes a two-electron oxidation, it was hypothesized that the second electron could be returned to the RS cluster to initiate the next round of catalysis, or alternatively, transferred to an unknown electron acceptor. In order to address the fate of the second electron, the authors generated the flavodoxin semiquinone (Fld SQ) by incubating flavodoxin with a stoichiometric amount of dithionite [88]. A two-fold excess of the Fld SQ was incubated with SAM, peptide and anSMEcp, and the amount of Fld SQ consumed was measured by EPR in parallel to quantification of the peptide and 5′-dA products by LC–MS. During the course of the 30 min reaction, approximately three equivalents of 5′-dA and product were formed; however, only a slight decrease in the EPR signal for the Fld SQ was observed. Furthermore, a signal for a reduced [4Fe–4S] cluster was not detected, indicating that after each round of turnover, the electron was returned to flavodoxin. Thus, the electron generated in the reaction can be used to initiate subsequent rounds of catalysis in a flavodoxin-dependent manner. The authors also attempted to address the coordination of cysteines to the auxiliary clusters of anSMEcp and anSMEkp [88]. The enzyme from C. perfringes contains 18 cysteine residues compared to 13 cysteines present in anSMEkp. In order to limit the number of variants required, Cys to Ala variants were made of each of the ten cysteines not found in the RS motif of the enzyme from K. pneumoniae. While the authors intended to purify and measure the activity of each variant, they found that only two of the variants, C127A and C245A (K. pneumoniae numbering), were soluble. Both of these variants supported activity equivalent to that of the WT enzyme, so it was assumed that these two cysteines are unimportant for catalysis. A third variant, C291A, was sparingly soluble; however, it did not exhibit significant activity

above the limit of detection of the assay. The corresponding variant of anSMEcp, C276A, was also generated with similar results. The enzyme was soluble in this case; however, the activity was greatly diminished. The remaining seven variants were completely insoluble and could not be analyzed for activity. It was surmised that the seven cysteines that yielded insoluble enzyme when substituted with alanine act as ligands to the auxiliary clusters. The role of C291 in anSMEkp could not be assigned given the data available. The crystal structures of anSMEcp in the presence and absence of its peptide substrate resolved the issue concerning the cysteines that coordinate the auxiliary clusters (Fig. 7) [20]. The N-terminal auxiliary cluster (Aux I) is ligated by C255, C261, C276, and C317 (C. perfringens numbering), while the C-terminal auxiliary cluster (Aux II) is ligated by C320, C326, C330, and C348. One of the cysteines ligating Aux I, C255, lies outside of the SPASM domain, while the other three are part of the seven-cysteine motif. The first three cysteines of the CX2CX5CX3C motif ligate Aux II, with the final ligand coming from the final cysteine in the greater CX9 -15GXC–gap–CX2CX5CX3C–gap–C motif. The crystal structure also allows for an examination of the hypothesized role of the auxiliary clusters in electron transfer. Both the RS cluster and Aux I are within reasonable electron transfer distance (8.9 and 8.6 Å, respectively) of Cβ of the substrate cysteine, respectively (Fig. 7B). Assuming that the electron is passed to Aux I, the only reasonable candidate to receive this electron is Aux II, because the RS cluster is 16.9 Å away and the path to solvent is likely obscured by the substrate. Aux II is 8.3 Å from bulk solvent, so it could transfer the electron to flavodoxin, which in turn could pass the electron to the RS cluster. At 8.6 Å and 20.8 Å, both Aux I and Aux II are too far from the substrate cysteine to ligate the substrate thiolate, suggesting that the auxiliary clusters are either involved in electron transfer or play a purely structural role. 3.2. Butirosin biosynthesis The 2-deoxy-scyllo-inosamine (DOIA) dehydrogenase from Bacillus circulans (BtrN) catalyzes the two-electron oxidation of DOIA to form amino-dideoxy-scyllo-inosose (amino-DOI) as a step in the biosynthesis of the antibacterial butirosin B [89]. BtrN contains a segment of the SPASM domain containing four of the seven cysteines, which has been amusingly coined the twitch domain. Although it was originally thought to contain only one [4Fe–4S] cluster [90], it was later demonstrated that the enzyme actually harbors two clusters [91]. Despite the differences in the substrates for anSME and BtrN—a protein substrate for the former and a small-molecule substrate for the latter—their mechanisms of catalysis are believed to be quite similar [91]. In one mechanism, abstraction of a hydrogen atom from the carbon center undergoing oxidation is followed by loss of a proton and an electron, while in a second mechanism, deprotonation of the alcohol could precede hydrogen atom abstraction [15,89]. Given that these reactions are two-electron


A

1323

B

Substrate Peptide Aux 2 8.9 Å Aux 2

SAM 20.8 Å

Aux 1

8.3 Å

RS Cluster RS Cluster

16.9 Å

12.9 Å

8.6 Å Aux 1

Solvent

Fig. 7. Structure of anSMEcpe (A) and the active site of the enzyme demonstrating coordination of SAM to the RS cluster and the distances between the iron–sulfur clusters and the peptide substrate (B) (PDB accession code 4K39). The RS domain of anSMEcpe is shaded in green while the SPASM domain is shaded in pink. The carbon backbones of SAM and the peptide substrate are depicted in cyan and blue, respectively.

oxidations, an electron must be passed to an unknown acceptor. It has been proposed that the auxiliary cluster is the immediate electron acceptor and that it might deliver the electron back to the RS cluster for subsequent rounds of turnover via the intermediacy of flavodoxin [91]. The crystal structure of BtrN shows the structural similarity of the C-terminal domains of BtrN and anSMEcpe (Fig. 8) [15]. The twitch domain of BtrN and the SPASM domain of anSMEcpe are each composed of a β-hairpin and an α-helix. In BtrN, the lone auxiliary cluster is completely ligated by the four cysteines of the twitch domain. The first cysteine ligand is just upstream of the β-hairpin, the second ligand resides between the β-hairpin and the α-helix, and the last two ligands fall on a loop after the α-helix (Fig. 9). This arrangement is similar to that in anSMEcpe, in which the first ligand to Aux I is just upstream of the start of the SPASM domain, the second resides upstream of the β-hairpin, and the third resides between the β-hairpin and the α-helix. The fourth ligand to Aux I, as well as all four ligands to Aux II, reside downstream of the α-helix. The structure of BtrN also reveals that the role of the cluster is likely the same as that in the two auxiliary clusters in anSMEs. Both the RS and auxiliary clusters are sufficiently close to C3 of DOIA to receive an electron upon H• abstraction. However, at 9.6 Å, the auxiliary cluster is not properly positioned to ligate the

A

substrate as previously hypothesized (Fig. 8B) [91]. Because the distance between the two clusters is 15.9 Å, direct transfer of the electron from the auxiliary cluster to the RS cluster is unlikely. The rationale for the lack of a third Fe/S cluster in BtrN likely lies in the nature of its substrate. The macromolecular protein substrate of anSMEs obstructs access of Aux I to solvent. In the case of BtrN, the small molecule substrate does not have this affect, rendering the extra electron transfer step unnecessary.

3.3. Alternative heme biosynthesis An alternative pathway for the biosynthesis of heme in sulfatereducing Desulfovibrio bacteria and methanogenic bacteria converts siroheme to heme b in a four-enzyme process [92–94]. The last step, catalyzed by AhbD, is the conversion of two propionyl side groups of iron-coproporphyrin III (Fe-CpIII) to vinyl groups, yielding heme b (Fig. 10) [92]. In addition to the canonical CX3CX2C RS motif, AhbD also contains a SPASM domain [95]. Initial characterization of the enzyme by UV/visible spectroscopy revealed features consistent with the presence of one or more iron–sulfur clusters [95]. After confirming

B 2-DOIA SAM

8.6 Å 9.6 Å

15.9 Å

Auxiliary Cluster RS Cluster

RS Cluster

Auxiliary Cluster Fig. 8. Structure of BtrN (A) and the active site of the enzyme demonstrating coordination of SAM to the RS cluster and the distances between the iron–sulfur clusters and the 2-DOIA substrate (B) (PDB accession code 4M7T). The RS domain of BtrN is shaded in purple while the twitch domain is shaded in cyan. The carbon backbones of SAM and the 2-DOIA substrates are depicted in green and blue, respectively.

1324


A

B

BtrN

anSMEcpe

Fig. 9. SPASM and twitch domains of anSMEcpe and BtrN. The topologies of the SPASM and twitch domains of BtrN and anSMEcpe are shown in cartood form with the cysteine ligands to Aux I depicted as green dots and the cysteine ligands to Aux II depicted as pink dots (A). Structures of the twitch and SPASM domains of BtrN and anSMEcpe, respectively, are shaded in cyan (B). The cysteine ligands to Aux I are shaded in green and the cysteine ligands to Aux II are shaded in pink.

that the enzyme binds Fe-CpIII, it was demonstrated that AhbD catalyzes the conversion of CpIII to heme b and generates 5′-dA. EPR spectroscopy was used to probe for the presence of multiple clusters [95]. The observed EPR signal and its temperature dependence were consistent with the presence of one or more [4Fe–4S] clusters. When the microwave power was increased, an additional resonance was observed. A slower relaxing signal was observed with g values of 2.057 and 1.925, while a faster relaxing signal was observed at g = 2.083, 1.921, and 1.925 at low temperatures. When both substrates, SAM and Fe-CpIII, were incubated with the samples, one of the signals, with g values of 2.064 and 1.92, remained unperturbed. The other signal, with g values of 2.13 and 1.92, shifted somewhat. The changes were suggested to be caused by SAM binding to the RS cluster. A model of the structure of AhbD was created based on the crystal structures of anSMEcp and MoaA [95]. In this model, the ligands for the RS cluster and one of the two auxiliary clusters are oriented to bind an Fe/S cluster. The binding pocket for Aux I lacks two of the four cysteines required to ligate the third cluster. Because the substrate is a small molecule, it is possible that only one auxiliary cluster is required as is the case for BtrN. It should be noted that the sequence alignment

Fig. 10. The final step in the alternative pathway for heme b biosynthesis.

of AhbD indicates the presence of the full SPASM motif and all eight cysteines that ligate the auxiliary clusters in anSMEcp. Therefore, the model may not accurately predict the locations of all of the cysteines, and AhbD may actually contain three [4Fe–4S] clusters. A quantitative analysis of iron and sulfide content paired with Mössbauer spectroscopy may be required to definitively assign the cluster content of AhbD. 4. Complex cofactor maturases 4.1. Molybdopterin cofactor biosynthesis The molybdopterin cofactor (MoCo) is a widely conserved redox cofactor composed of a tricyclic pyranopterin with a dithiolene moiety that coordinates a mononuclear molybdenum center [96]. In humans, three complexes require MoCo for their activity: sulfite oxidase, aldehyde oxidase, and xanthine dehydrogenase [97]. The absence of MoCo in these complexes results in neurological symptoms and usually death [98]. The first step in the conversion of GTP to MoCo is catalyzed by the RS enzyme MoaA (Fig. 11A). In conjunction with MoaC, MoaA converts GTP to cyclic pyranopterin monophosphate (cPMP) [16, 97–100]. The reactions of MoaA and MoaC have proven difficult to separate due to the lability of the intermediates and products, resulting in multiple proposed structures of both the chemical intermediate and the enzyme products [22,101–104]. MoaA harbors two [4Fe–4S] clusters, both of which are required for activity of the enzyme [105–107]. MoaA is another member of the twitch family of RS enzymes, though in contrast to BtrN, it lacks a fourth cysteine to ligate the auxiliary Fe/S cluster [15]. The open coordination site created by the absent cysteine is used to bind the GTP substrate (Fig. 12) [22,108]. The crystal structure of MoaA indicates that both N1 and the exocyclic amine are within bonding distance of the open coordination site of the auxiliary cluster; however, ENDOR studies suggest that N1 is in fact bound to the cluster [22,108]. Furthermore, spectroscopic evidence suggests that GTP is bound in the enol form rather than the keto form, which may modulate


1325

Fig. 11. Pathway for molybdenum cofactor biosynthesis (A) and proposed mechanism for the biosynthesis of cPMP by MoaA and MoaC. In the mechanism proposed by Mehta, et al., MoaA catalyzes all but the last step (the second shaded intermediate) in the biosynthesis of cPMP. The work of Hover et al. suggests that 3,8-cH2GTP is the product of MoaA, and MoaC is responsible for the remaining steps.

the reactivity of the substrate [108]. Coordination of GTP to the Fe/S cluster may also be important for substrate specificity, because ATP, which is present in much higher concentrations in the cell than GTP, must be discriminated against [22]. A series of recent studies probed the mechanistic details of the MoaA reaction, the first of which revealed the identity of the H• abstracted by the 5′-dA• [103]. A series of deuterium-labeled GTP substrates were incubated with SAM and MoaA, and the incorporation of deuterium into 5′-dA was measured by LC–MS. A single deuterium atom was transferred to 5′-dA from universally labeled GTP, demonstrating that the 5′-dA• directly abstracts an H• from the GTP substrate. No deuterium transfer was detected from [8-2H]-GTP or [2′-2H]-GTP, but a single deuterium was incorporated from [3′,4′,5′,5′-2H4]-GTP. The position was finally narrowed down to the 3′ H• with observation of deuterium transfer from [3′-2H]-GTP. Additionally, analysis of the cPMP product of MoaA and MoaC with [3′,4′,5′,5′-2H4]-GTP by LC–MS showed an increase of 3 Da in the pterin product, indicating that the 4′ and 5′ hydrogens are retained. With the knowledge that an H• is abstracted from C3′ by the 5′-dA•, the authors proposed a mechanism in which the resulting C3′ radical adds to C8 of GTP (Fig. 11B). Shortly after, Hover et al. identified an intermediate product of the MoaA reaction that could be transferred to MoaC and then converted to cPMP [104]. The reaction was performed stepwise by incubating MoaA with SAM and GTP, and then removing MoaA by ultrafiltration.

MoaC was then added to the filtrate, and cPMP formation was detected by HPLC. The authors then set out to isolate this soluble intermediate. Large-scale reactions were conducted, and the intermediate was purified by anion-exchange chromatography before being analyzed by ESI–TOF–MS, 1H NMR, 13C NMR, and multiple 2D NMR techniques to identify its molecular structure. It was ultimately determined to be (8S)-3′,8-cyclo-7,8-dihydroguanosine 5′-triphosphate (3′,8-cH2GTP, Fig. 11B). To verify that 3′,8-cH2GTP is the product of the MoaA reaction, the purified intermediate was incubated with MoaC, which was able to convert 3′,8-cH2GTP to cPMP. The authors proposed a mechanism in which an intermediate proposed by Mehta et al. [103] is actually the final product of MoaA, which suggests that MoaC has a much larger role in the conversion of GTP to cPMP than previously thought. Mehta and coworkers were also able to trap a similar reaction intermediate using a substrate analogue [101]. In the proposed mechanism, the 3′ hydroxyl of the 3′,8-cH2GTP intermediate must be deprotonated to allow for the rearrangement of the five-membered ribose ring to form a six-membered ring. The analogue, 2′,3′-dideoxy-GTP, lacks the 3′ hydroxyl group and should therefore arrest the reaction at the stage of the 3′,8-cH2GTP intermediate. MoaA was incubated with 2′,3′dideoxyGTP and SAM and analyzed by LCMS. The product of this reaction was consistent with the predicted intermediate. The potential redox role of the auxiliary Fe/S cluster was noted both by the authors and the previous report by Hover et al. The radical intermediate

1326


A

B

5’-dA GTP

RS Cluster

Methionine

Auxiliary Cluster

Fig. 12. Structure of MoaA (A) and coordination of methionine and GTP to the RS and auxiliary clusters of MoaA, respectively (B) (PDB accession code 2FB3). The RS domain of MoaA is shaded in orange while the twitch domain is shaded in blue. The carbon backbones of methionine and 5′-dA are depicted in green and the carbon backbone of GTP is depicted in salmon.

generated by addition of the C3′ radical to C8 would need to be reduced by one electron to form the stable intermediate observed by both groups. The auxiliary [4Fe–4S] cluster to which the substrate is ligated was suggested to serve this purpose. A more recent study by Mehta et al. identified a different product of the MoaA reaction and suggested a lesser role for MoaC [102]. In this study, MoaA was overproduced in a moaC deletion strain to prevent contamination from MoaC in MoaA preparations. When MoaA prepared in this manner was incubated with SAM, GTP and dithionite, the product detected after derivatization was inconsistent with 3′,8-cH2GTP, but instead appeared to be the tricyclic pterin triphosphate intermediate that directly precedes the cPMP product. The authors also used two substrate analogs, 2′-chloroGTP and 2′-deoxyGTP, to trap intermediates in the MoaA reaction. The 2′-chloroGTP substrate was used to provide evidence for the 3′ radical that is proposed to initiate the reaction. Upon formation of the 3′ radical, the chlorine at the 2′ position leaves as a radical species, generating a labile compound. The expected ribosyl derivative was detected as the product of this reaction. The 2′deoxyGTP substrate was predicted to prevent the dehydration reaction in which the 2′ hydroxyl is the leaving group. In the absence of the 2′ hydroxyl, an alternative product is formed, which lacks the third ring of cPMP. The authors proposed a mechanism for MoaA that accounts for the product observed with the substrate analogue (Fig. 11B). The discrepancy between the current data lies in the nature of the product of the MoaA reaction and the role of MoaC in generating cPMP. Hover et al. suggests that MoaA generates 3′,8-cH2GTP, and MoaC catalyzes the remaining steps of the conversion of 3′,8-cH2GTP to cPMP [104]. The work of Mehta et al. indicates that MoaA can catalyze all but the last step of the reaction, which is the hydrolysis of the triphosphate to form pyrophosphate and the fourth ring of cPMP [102]. Further study will be required to reconcile the existing data and elucidate the true product of the MoaA reaction. 4.2. Fe–Fe hydrogenase cofactor maturation The interconversion of protons and molecular hydrogen is catalyzed by hydrogenases using a metal center-containing active site. One of the three types of hydrogenases, the Fe–Fe hydrogenase (HydA), contains a complex cofactor known as the H-cluster. The H-cluster is composed of a cubane [4Fe–4S] cluster bridged by a cysteine thiol to a diiron cluster that contains three carbon monoxide (CO) ligands, two cyanide (CN)

ligands, and a dithiomethylamine (DTN) moiety that bridges the two iron ions (Fig. 13A) [109–112]. The [4Fe–4S] component is built on HydA first, presumably by the standard Fe/S cluster assembly machinery (i.e. the gene products of the isc operon) [113]. Assembly of the diiron cluster is more complicated and requires an additional three maturases. The first of the maturases, HydF, is a GTPase that is likely the scaffold and insertase for the diiron center [114–119]. The other two maturases, HydE and HydG, are RS enzymes that each potentially contain an auxiliary Fe/S cluster [120]. While the reaction catalyzed by HydE is still uncertain, it has been proposed that it may function in the synthesis of the bridging DTN moiety. This is in large part inferred from the fact that HydG catalyzes the formation of the CO and CN ligands bound to the diiron center (vide infra), which leaves only the synthesis of the DTN moiety unaccounted for. In addition to the canonical RS motif, HydE contains a CX7CX2C motif that ligates an auxiliary cluster [121]. In the crystal structure, a [2Fe–2S] cluster is coordinated by this motif; however, it was suggested that this may be a breakdown product of a [4Fe–4S] cluster [121]. The role the auxiliary cluster might play in the mechanism of HydE is entirely unknown, and it may not be required at all. The CX7CX2C motif is not conserved among HydEs, and variants with single substitutions of the cysteines in the motif still retained near WT activity [121]. It was proposed that the second iron–sulfur cluster might have a regulatory role or may simply be vestigial. HydG also contains a second cysteine rich motif that harbors an Fe/S cluster, and the cysteines residing in the CX2CX22C motif are essential for HydG activity [122]. The reaction catalyzed by HydG has been much better characterized. It was first noted that HydG could cleave tyrosine to form p-cresol [123]. Further studies, using isotopically-labeled tyrosine substrates, indicated that HydG generated both CO and CN from tyrosine [124,125]. Additionally, a variant in which the first two cysteines of the CX2CX22C motif were substituted with serine was able to generate CN but not CO [126]. It was therefore proposed that dehydroglycine derived from tyrosine was cleaved in a radicaldependent fashion to generate formaldehyde imine and •CO− 2 . The •CO− 2 would need to coordinate to the auxiliary cluster to be reduced by one electron to CO. To further investigate the mechanism by which HydG operates, Driesner et al. constructed a triple Cys → Ala variant of cysteines in the CX3CX2C motif and a truncated variant of the C-terminal domain of HydG (ΔCTD) from Clostridium acetobutylicum that lacks the


1327

Fig. 13. Reaction pathway of H-cluster biosynthesis (A) and proposed mechanism of HydG (B).

CX2CX22C motif [127]. The truncated HydG and triple variant each contained 3.1 and 3.4 iron ions, respectively, and an EPR spectrum of each sample confirmed the presence of a [4Fe–4S] cluster. The spectrum of the ΔCTD variant was similar to that of the WT enzyme. In the absence of SAM, the WT spectrum displays g values 2.03, 1.92, and 1.90, while the addition of SAM results in a more complex spectrum with two distinct sets of g values (g1 = 2.02, 1.93, and 1.91, g2 = 2.00, 1.87, and 1.83). The spectrum of the ΔCTD variant has g values identical to the WT enzyme in the absence of SAM, and nearly identical for the SAM-containing sample (g1 = 2.04, 1.92, and 1.90, g2 = 2.00, 1.88, and 1.84). The spectrum of the triple variant enzyme was also similar to that of the WT enzyme (g = 2.03, 1.92, and 1.88), except that addition of SAM no longer perturbed it. The binding and kinetic constants of the WT and variants of HydG with respect both to SAM and tyrosine were measured in order to determine if the auxiliary cluster has a role in tyrosine cleavage, SAM cleavage, or both [127]. Both the ΔCTD variant as well as a C386S variant, in which one of the ligands to the auxiliary cluster was substituted with a serine, were compared with the reaction performed by the WT enzyme. The effects of the variants on binding and rates of cleavage of SAM were modest; however, the apparent KM for tyrosine was significantly higher for the C386S variant, and especially for the ΔCTD truncation. The rate of tyrosine cleavage as measured by formation of p-cresol was not greatly affected by disruption of the auxiliary cluster; however, the rate of CN production by the ΔCTD variant was negligible, and neither variant produced any CO, furthering the notion that cleavage of dehydroglycine to CN and CO requires the assistance of the auxiliary cluster. Another study of HydG from Shewanella oneidensis revealed the presence of a radical intermediate in the pathway of tyrosine cleavage [128]. Rapid-mix freeze-quench coupled with EPR spectroscopy was

used to trap and detect the radical intermediate. Formation and decay of the transient radical was observed with its maximum intensity at approximately 2 s after the reaction was initiated. The g = 2 radical signal was further investigated using HYSCORE and ENDOR techniques. Coupling observed with 13C- and 15N-labeled tyrosine confirmed that the radical signal was derived from tyrosine. The exact identity of the radical species was identified using specifically-labeled tyrosine substrates. Labeling of Cα of tyrosine with 13C does not affect the shape of the spectrum; however, the signal collapses when [2H7,15N]-labeled tyrosine is used as a substrate. This observation is consistent with previously identified tyrosyl radicals; however, further isotopic labeling experiments indicate that the radical in HydG is different than the protein tyrosyl radicals that have been characterized. The spectrum of HydG in the presence of 3,5-2H2-tyrosine is nearly identical to that of the enzyme with unlabeled tyrosine. Conversely, a 1:2:1 pattern is observed in the spectrum of the 2,3,5,6-2H4-tyrosine-containing sample, indicating the presence of two equivalently coupled protons. A sample containing β,β-2H2-tyrosine further collapses the signal, indicating that the greatest amount of spin is centered on Cβ. Taken together, the spectra are inconsistent with a tyrosyl radical, but instead indicate the presence of a 4-oxidobenzyl radical. Using this knowledge, the authors propose a mechanism that accounts for this radical intermediate and the previously accumulated data (Fig. 13B). A subsequent study adds to the mechanistic understanding of the steps that follow the formation of the 4-oxidobenzyl radical [129]. Stopped-flow Fourier transform infrared (SF–FTIR) and ENDOR spectroscopies were used to track the formation of a new intermediate in the formation of the diiron center of HydA. SF–FTIR is ideal for detecting CO and CN formation from HydG. In the WT enzyme, two discrete intermediates were detected. The first was assigned to an Fe–CO and an Fe–CN complex. This complex formed on the same timescale as the

1328


decay of the 4-oxidobenzyl radical previously characterized. A second complex was detected that forms concomitant with the decay of the first. This complex exhibited two Fe–CO vibrational bands and an Fe-CN band in the FTIR spectrum. Thus, it was predicted to be a cuboidal [3Fe–4S] cluster with an Fe(CO)2(CN) at the unique iron site of the auxiliary cluster. ENDOR spectroscopy revealed that the iron ions of the diiron center of HydA are derived from the [4Fe–4S] cluster of HydG. In this experiment, in vitro assembly of the diiron center was performed using HydG that had been overproduced in the presence of 57Fe and HydE and HydF that had been overproduced in the presence of natural abundance iron. The presence of 57Fe in HydA indicates that the iron was derived from HydG, and the signal was consistent with the diiron portion of the H cluster. The authors describe the product of HydG as a Fe(CO)2(CN) synthon, two of which must be transferred to HydA with loss of one CO molecule and the addition of the DTN moiety, presumably from HydE, to form the diiron center of HydA (Fig. 13A). This reaction can be thought of as parallel to those of LipA and BioB. Whereas LipA and BioB donate one or two sulfur atoms from their auxiliary Fe/S clusters, HydG donates an iron ion. Some of the same questions that have been asked of the sulfur-donating enzymes also apply to HydG. For instance, are both iron ions of the diiron center donated by the same HydG polypeptide, how is the auxiliary cluster reassembled for subsequent rounds of turnover, and what factors are required to repair the cluster? 5. Glycyl radical enzyme activating enzymes Glycyl radical enzymes (GREs) contain a specific glycyl residue that is converted to a glycyl radical via H• abstraction from Cα by an activating enzyme [130–133]. The radical is stabilized by the captodative effect, and can have a half life of several hours under anaerobic conditions [134]. The enzymes that generate the glycyl radical cofactor belong to the RS superfamily. Two of the best studied GRE-activating enzymes (GRE-AEs), pyruvate formate-lyase activating enzyme (Pfl-AE) and anaerobic ribonucleotide reductase activating enzyme (Nrd-AE), require only the RS [4Fe-4S] cluster to carry out this transformation [131,135]. The activases for benzylsuccinate synthase (BssD), 4-hydroxyphenylacetate decarboxylase (HPD-AE), B12-independent glycerol dehydratase (GDH-AE), and the more recently discovered choline trimethylamine-lyase (CutD) each contain two additional cysteinerich motifs that may harbor one or more additional Fe/S clusters [136–139]. Here, we will focus on the reactions catalyzed by HPD-AE and CutD for which new information has become available.

cleavage products of SAM were determined. Unlike GDH-AE, HPLC and TLC analysis identified 5′-dA and methionine products of SAM generated by HPD-AE. The amount of 5′-dA was quantified by HPLC and correlated to the amount of the glycyl radical produced on HPD as determined by integration of the EPR signal for the radical. The amount of 5′-dA produced was similar to the amount of radical present until the maximum activation of HPD was achieved. No other cleavage products of SAM were detected. Thus, the unusual mode of cleavage of SAM by GDH-AE is not conserved among GRE-AEs of this class. 5.2. Choline trimethylamine-lyase activating enzyme Choline trimethylamine-lyase (CutC) is an enzyme found in bacteria in the human gastrointestinal tract that cleaves choline to trimethlamine [142]. It is a GRE that has an associated activating enzyme (CutD) that belongs to the RS superfamily. Like HPD-AE discussed above (as well as BssD and GDH AE), CutD from Desulfovibrio alaskensis contains two CX2–5CX2–4CX3C sequences that have been termed “Clostridial” type motifs [139,141]. As hypothesized for the other GRE-AEs of this class, the two additional cysteine-rich motifs of CutD are also thought to coordinate auxiliary Fe/S clusters. When incubated with SAM and dithionite, purified CutD cleaved SAM, forming 5′-dA and methionine, confirming that CutD is in fact a member of the RS superfamily. Although the CutC substrate was not present in this experiment, abortive cleavage of SAM in the absence of substrate is relatively common among RS enzymes. When CutC was added to the reaction, formation of the glycyl radical on CutC could be measured using EPR spectroscopy. Finally, CutD was found to contain 8.4 mol of iron and 7.6 mol of sulfide per mol of protein after reconstitution with iron and sulfide, suggesting that multiple clusters may be present [139]. The presence of three distinct cysteine motifs has led many to believe that the GRE-AEs with these extra motifs contain three [4Fe–4S] clusters. Indeed, the iron content of the enzymes that have been characterized hints that one or more auxiliary Fe/S clusters may be coordinated by the cysteines in these motifs. However, definitive evidence that the GRE-AEs harbor multiple clusters, or that the extra cysteine motifs are required for activity is still lacking. Furthermore, the function of the putative Fe/S clusters is also a mystery, especially given that the better studied GRE-AEs, Pfl-AE and Nrd-AE do not need an auxiliary cluster to generate the glycyl radical cofactor on their respective substrates. 6. Complex heterocycle biosynthesis 6.1. Wybutosine biosynthesis

5.1. 4-Hydroxyphenylacetate decarboxylase activating enzyme A study of GDH-AE suggested that instead of cleaving SAM to generate 5′-dA• and methionine, the enzyme produced a 3-amino-3carboxypropyl radical and 5′-deoxy-5′-methylthioadenosine [140]. Given that none of the GRE-AEs that contain extra cysteine motifs had been characterized, it was postulated that the others may share this unusual mechanism. In order to investigate this hypothesis, HPD-AE was characterized further [141]. The as-isolated enzyme contained 4.0 iron ions per polypeptide, but could be reconstituted with up to 12.0 mol of iron per mol of protein when treated with cell lysate in which the Fe/S cluster assembly machinery had been over-expressed. The EPR spectra revealed the presence of small quantities of [3Fe–4S] clusters, but suggested that the majority of the iron was in the form of [4Fe–4S] clusters. When reduced with dithionite, the reconstituted sample contained 1.6 [4Fe–4S]+ clusters per protein, strongly indicating the presence of at least two [4Fe–4S] clusters in the enzyme. The Mössbauer spectrum was also consistent with the majority of the iron residing in the form of [4Fe–4S] clusters, with the caveat that the sample on which Mössbauer spectroscopy was performed only contained 4 to 6 mol of iron per mol of protein. The enzyme was tested for the ability to activate its cognate protein substrate, HPD, and the

Wybutosine (yW) is a tricyclic base found 3′ to the anticodon of certain tRNAs in archaea and eukaryotes [143]. The heavily modified yW base promotes translational reading frame fidelity due to its high hydrophobicity, which reinforces the interaction of the codon and anticodon. The modifications also prevent the base from forming Watson–Crick base pairing interactions with the first base of the codon [144–148]. Wybutosine is synthesized from guanosine via methylation of N1 by the SAM-dependent enzyme TRM5 (Fig. 14A) [149–152]. Formation of the tricyclic ring structure (4-demethylwyosine or imG-14) is catalyzed by TYW1, which is an RS enzyme [23,153]. TYW2, TYW3, and TYW4 catalyze formation of the remaining substituents on C7 of yW [154]. TYW1 contains the canonical CX3CX2C RS motif and an additional CX12CX12C motif that may harbor an auxiliary iron–sulfur cluster. Eukaryotic, but not archaeal TYW1 enzymes have an additional flavodoxin-like domain that may be required for reduction of the RS cluster [23]. Prior to any biochemical characterization of TYW1, multiple crystal structures provided a hint of how this enzyme might work [13,23]. The first structure of the enzyme, from Methanocaldococcus jannaschii, showed that TYW1 is composed of the partial TIM barrel typical of other RS enzymes [13]. The Fe/S clusters were not strongly present in the structure, given that the crystals were grown under aerobic


1329

Fig. 14. Biosynthetic pathway of wybutosine (A) and proposed mechanisms of TYW1 (B and C).

conditions. Even after soaking with iron and sulfide under anaerobic conditions cluster incorporation was poor. Very weak density for the RS cluster was observed, and no electron density was observed for the predicted auxiliary cluster. However, the cysteines of this motif were aligned such that they could coordinate a [4Fe–4S] cluster, and therefore, the authors modeled two [4Fe–4S] clusters in the structure. In this model, the unligated iron site faces towards the active site and the RS cluster, potentially as a site for substrate binding as observed for MoaA. Moreover, activity of TYW1 was absolutely dependent on the presence of all six conserved cysteines as well as a conserved lysine in the active site. A second structure provided evidence that both motifs ligate an Fe/S cluster [23]. In this study, anomalous data was used to map the Fe/S clusters. Two anomalous peaks corresponding to the cavities created by the cysteine motifs were observed. The anomalous peak corresponding to the RS motif was much larger than the other peak, and was assigned to a [4Fe–4S] cluster. Because the peak for the auxiliary cluster was much weaker, it was modeled as a [2Fe–2S] cluster; however, iron and sulfide content was not measured, so the lack of density for the auxiliary cluster may be due to poor incorporation. Young and Bandarian elucidated the source of the final two carbons and proposed a mechanism for the formation of the third ring of imG-14 [155]. Activity of TYW1 from M. jannaschii was measured with a number of potential two-carbon donors in order to determine the source of the two carbons that complete the third ring of imG-14. Acetyl-CoA, acetyl phosphate, phosphoenolpyruvate, and pyruvate were each added to

reactions containing TYW1, tRNAPhe, dithionite, as well as TRM5 and SAM to catalyze the initial methylation of N1. Only the addition of pyruvate allowed for formation of imG-14. Furthermore, pyruvate labeled with 13C at C2 or C3 resulted in imG-14 with a mass increase of 1 Da, while C1-13C labeled pyruvate did not result in a mass shift. Addition of C1,2,3-13C3 labeled pyruvate to the reaction resulted in a mass shift of 2 Da, indicating that two carbons from pyruvate are incorporated into imG-14. In the mechanism proposed, the conserved lysine forms a Schiff base with pyruvate (Fig. 14B). Abstraction of a H• from the N1 methyl of N-methylguanosine (m1G) is followed by radical recombination of the N1 methyl with C2 of pyruvate and homolytic cleavage of the C1\C2 bond. Finally, transamination and deprotonation releases imG14. Spectroscopic characterization revealed more about the potential role of the auxiliary cluster of TYW1 [156]. Aerobic purification of TYW1 from Pyrococcus abyssi resulted in less than 0.7 iron and sulfide ions per monomer; however, anaerobic reconstitution of the protein yielded 8.2 iron and sulfide ions per protein. Analysis by Mössbauer spectroscopy indicated that approximately 95% of the iron in the sample was in the form of a [4Fe–4S] cluster; thus, TYW1 harbors two [4Fe–4S] clusters per protein as the sequence analysis would suggest. A complex EPR spectrum of dithionite-reduced TYW1 results from two distinct species that can be simulated well. Addition of SAM alters the spectrum of the RS cluster while addition of both SAM and pyruvate results in only one signal corresponding to the SAM-bound RS cluster. These results

1330


imply that SAM and pyruvate interact with the RS and auxiliary clusters, respectively. The HYSCORE spectrum of SAM bound to TYW1 also showed strong interactions of a nitrogen atom interacting with the reduced [4Fe–4S] cluster, indicating that SAM likely binds in a similar manner as shown for other RS enzymes with the α-amino and α-carboxylate groups ligated to the unique iron site (Figs. 7B and 8B, structures of SAM bound to anSMEcpe and BtrN). Further characterization of the SAM and SAM plus pyruvate bound species was carried out by Mössbauer spectroscopy. The subtraction of the two spectra revealed that the interaction of pyruvate to the auxiliary cluster induces reoxidation of the cluster to the S = 0 [4Fe-4S]2+ state in addition to altering the environment of the cluster. The results of this study strongly imply that pyruvate binds to the auxiliary [4Fe–4S] cluster of TYW1 and led to an alternative mechanism to that proposed by Young and Bandarian, in which the Fe/S cluster is used in place of the lysine to facilitate catalysis (Fig. 14C). 6.2. F0 synthase The F420 cofactor is a widely found deazaflavin cofactor found in numerous organisms that participates in redox reactions [157–161]. It is composed of a tricyclic core that is structurally similar to that of flavins, and a variable number of glutamic acid residues linked together by γ-glutamyl bonds appended to the core via the phospholactate group (Fig. 15A). The core of the cofactor is generated by F0 synthase, which fuses tyrosine with 5-amino-6-ribitylamino-2,4(1H,3H)pyrimidinedione (diaminouracil), a precursor to riboflavin [162–164]. F0 synthase is a single polypeptide in actinobacteria (FbiC), but is also found as two separate subunits in archaea and cyanobacteria (CofG and CofH) [165]. F0 synthase contains two distinct RS motifs, either within the same protein for FbiC, or split between the CofG and CofH subunits.

Characterization of both types of F0 synthases revealed that two Fe/S clusters are present in each system, and provided insight into the mechanism of F0 biosynthesis [163]. FbiC from Thermobifida fusca contained only 1.2 iron ions per polypeptide, but could be reconstituted with up to 9.5 iron ions per monomer. Quantification of the products of a reaction containing FbiC, SAM, tyrosine, diaminouracil, and the flavodoxin/ flavodoxin reductase reducing system yielded approximately 2.5 equivalents of 5′-dA produced for every equivalent of F0. This result is consistent with the proposal that one molecule of SAM is cleaved at each of the two RS cluster sites. When dithionite is used as the reducing reagent in place of the flavodoxin/flavodoxin reductase reducing system, 3.6 equivalents of 5′-dA was produced for each equivalent of F0 produced; however, this stoichiometry is likely due to abortive cleavage of SAM. Abortive cleavage of SAM was also observed for each enzyme of the CofG and CofH pair. In this study, CofG from M. jannaschii and CofH from Nostoc punctiforme were purified and characterized and shown to abortively cleave SAM in vitro. CofG and CofH contained 1.7 and 1.8 iron ions per polypeptide, respectively, indicating incomplete incorporation of the [4Fe–4S] clusters in these enzymes but still demonstrating that both enzymes are likely to contain an Fe/S cluster. Reactions containing both enzymes produced F0, though the ratio of 5′-dA to F0 was 4:1 or greater. It was also demonstrated that CofH generates a diffusible intermediate that can be converted to F0 by CofG. Each enzyme was independently reacted with SAM, dithionite, tyrosine and diaminouracil and then subjected to ultrafiltration. The filtrate was then reacted with the opposite enzyme, SAM, dithionite, tyrosine, and diaminouracil. F0 was produced only when the product of CofH was provided to CofG. Based on the available information, the authors proposed a mechanism for F0 synthase (Fig. 15B) [163]. Abstraction of the phenoylic H• of tyrosine by 5′-dA• leads to radical fragmentation of tyrosine to a glycyl radical and a quinine methide, which can then

Fig. 15. Structure of the F420 cofactor (A) and the proposed mechanism for F0 biosynthesis (B). Shaded in blue and yellow are the reactions catalyzed by CofH and CofG, respectively, in the multi-subunit system.


add to diaminouracil. The second 5′-dA• abstracts a H• from the bridging methylene carbon, which allows for release of ammonia, deprotonation of the phenol, and subsequent cyclization. Loss of an additional H• (or a proton and an electron) results in F0. The authors propose that the electron is transferred back to the [4Fe–4S] cluster, but do not account for the fate of the glycyl radical that is generated.

6.3. Methanopterin biosynthesis Methanopterin (MPT) and its analogs are one-carbon carrier coenzymes found in methanotrophic bacteria and methanogenic archaea (Fig. 16A) [166–169]. It is structurally similar to folate; however, the enzymes that synthesize and use MPT are not homologous to those of the folate cofactor [170,171]. MPT is distinguished by methyl groups on C7 and C9 of its heterocyclic ring structure. The enzyme responsible for methylation of the pterin precursor was only recently identified in M. jannaschii as MJ0619 [172]. MJ0619 was recombinantly expressed and purified from E. coli, and the modifications to the natural folate cofactors in E. coli were identified [172]. Folate derivatives were extracted from E. coli expressing MJ0619 and analyzed by HPLC after being oxidatively or reductively cleaved. Oxidation of folates results in cleavage of the C6\C9 bond while reductive cleavage results in breakage of the C9\N10 bond (Fig. 16). 7-methylpterin (Fig. 16B) and 6-ethyl-7-methylpterin (Fig. 16C) were found in the respective reactions, indicating that MJ0619 is capable of methylating folates in E. coli. The lack of singly-methylated

1331

intermediates in the reductive cleavage reaction suggests that both methylations are catalyzed by the same enzyme. Like F0 synthase, MJ0619 contains two canonical RS motifs [172]. Single Cys → Ala variants in each of the cysteine motifs were created in order to determine if both motifs are essential. The variant in which one of the cysteines in the N-terminal motif was substituted with an alanine was able to generate 7-methylpterin; however, the dimethylated product was not observed. No methylated products were observed in extracts of E. coli expressing the variant of the C-terminal motif. It was hypothesized that the N-terminal cluster was required for methylation of C9 and that methylation of C7 precedes methylation at C9. The nature of the methyl donor was also probed [172]. A methylcobalamin intermediate was ruled out because MJ0619 does not contain a cobalamin-binding motif. Moreover, when the mj0619 gene is expressed in E. coli cultured in minimal media, 7-methylpterin was generated even though E. coli cannot synthesize cobalamin. SAM was predicted to be the methyl donor as is the case in the well-characterized RlmN and Cfr systems [173,174]. However, when E. coli expressing the mj0619 gene was supplemented with [methyl-2H3]-methionine, no deuterium incorporation was observed in 7-methylpterin. However, when the cells were supplemented with deuterated acetate, a pattern of deuterium incorporation was observed that was vastly different than that of the methyl group of methionine extracted from cells grown in media containing deuterated acetate as the sole carbon source. Given that the methyl group of methionine is derived from N5-methyl-tetrahydrofolate, N5-methyl-tetrahydrofolate is not likely to be the source of the methyl group for MJ0619. Given these results, the authors proposed a mechanism for MJ0619 that

Fig. 16. Structures of methanopterin (A), 7-methylpterin (B) and 6-ethyl-7-methylpterin (C). The proposed mechanism for MJ0619 in E. coli (D). In the native organism, M. jannaschii, the N5,N10-methylene-tetrahydrofolate is likely replaced with N5,N10-methylene-methanopterin.

1332


accounts for the isotopic labeling data, in which methylenetetrahydrofolate donates the appended methyl carbon (Fig. 16D). 7. Conclusions The number of RS enzymes for which extra cysteine-containing motifs have been identified in their primary structures continues to grow. Many of these enzymes have been purified and shown to harbor auxiliary clusters. In almost every case, the auxiliary cluster(s) has proven to be essential for the activity of enzymes that contain them. The greater challenge lies in assigning a function to the auxiliary clusters in each of the enzymes that require one. For most of the known systems the function of the auxiliary clusters is still undefined. For LipA and BioB, it has become clear that the auxiliary cluster is required as a source of sulfur for the reaction. In the cases of MoaA and TYW1, the cluster likely binds and/or activates the substrate for catalysis. In a number of systems such as the SPASM domain-containing enzymes, it is hypothesized that the auxiliary clusters are required for electron transfer, but experimental evidence is limited. Beyond the functions already listed, Fe/S clusters may function in substrate positioning or Lewis acid catalysis. We also have to consider that some Fe/S clusters may be structurally important, or in rare cases, simply vestigial. The example of HydG in which the auxiliary Fe/S cluster both coordinates the substrate and sacrifices an iron ion to the H-cluster of HydA demonstrates the variety of ways in which nature uses Fe/S clusters. The functional diversity of Fe/S clusters may not yet be fully understood, which makes study of RS enzymes containing auxiliary clusters even more appealing. Acknowledgments This work was supported by grants from the National Institutes of Health (GM-63847 and GM-103268) and the American Chemical Society (ACS-PRF 46065-AC4) to S.J.B. References [1] H.J. Sofia, et al., Radical SAM, a novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods, Nucleic Acids Res. 29 (5) (2001) 1097–1106. [2] P.A. Frey, S.J. Booker, Radical mechanisms of S-adenosylmethionine-dependent enzymes, Adv. Protein Chem. 58 (2001) 1–45. [3] P.A. Frey, A.D. Hegeman, G.H. Reed, Free radical mechanisms in enzymology, Chem. Rev. 106 (8) (2006) 3302–3316. [4] S.J. Booker, Anaerobic functionalization of unactivated C\H bonds, Curr. Opin. Chem. Biol. 13 (1) (2009) 58–73. [5] S.J. Booker, T.L. Grove, Mechanistic and functional versatility of radical SAM enzymes. F1000, Biol. Reprod. 2 (2010) 52. [6] M.R. Challand, R.C. Driesener, P.L. Roach, Radical S-adenosylmethionine enzymes: mechanism, control and function, Nat. Prod. Rep. 28 (2011) 1696–1721. [7] J.B. Broderick, et al., Radical S-adenosylmethionine enzymes, Chem. Rev. 114 (8) (2014) 4229–4317. [8] N.J. Cosper, et al., Direct FeS cluster involvement in generation of a radical in lysine 2,3-aminomutase, Biochemistry 39 (51) (2000) 15668–15673. [9] D. Chen, et al., Coordination and mechanism of reversible cleavage of S-adenosylmethionine by the (4Fe–4S) center in lysine 2,3-aminomutase, J. Am. Chem. Soc. 125 (39) (2003) 11788–11789. [10] C. Krebs, et al., Coordination of adenosylmethionine to a unique iron site of the (4Fe–4S) of pyruvate formate-lyase activating enzyme: a Mössbauer spectroscopic study, J. Am. Chem. Soc. 124 (6) (2002) 912–913. [11] C.J. Walsby, et al., An anchoring role for FeS clusters: chelation of the amino acid moiety of S-adenosylmethionine to the unique iron site of the (4Fe–4S) cluster of pyruvate formate-lyase activating enzyme, J. Am. Chem. Soc. 124 (38) (2002) 11270–11271. [12] C.J. Walsby, et al., Electron-nuclear double resonance spectroscopic evidence that S-adenosylmethionine binds in contact with the catalytically active (4Fe–4S)+ cluster of pyruvate formate-lyase activating enzyme, J. Am. Chem. Soc. 124 (12) (2002) 3143–3151. [13] Y. Suzuki, et al., Crystal structure of the radical SAM enzyme catalyzing tricyclic modified base formation in tRNA, J. Mol. Biol. 372 (5) (2007) 1204–1214. [14] F. Berkovitch, et al., Crystal structure of biotin synthase, an S-adenosylmethioninedependent radical enzyme, Science 303 (5654) (2004) 76–79. [15] P.J. Goldman, et al., X-ray analysis of butirosin biosynthetic enzyme BtrN redefines structural motifs for AdoMet radical chemistry, Proc. Natl. Acad. Sci. U. S. A. 110 (40) (2013) 15949–15954.

[16] P. Hänzelmann, H. Schindelin, Crystal structure of the S-adenosylmethioninedependent enzyme MoaA and its implications for molybdenum cofactor deficiency in humans, Proc. Natl. Acad. Sci. U. S. A. 101 (35) (2004) 12870–12875. [17] A.K. Boal, et al., Structural basis for methyl transfer by a radical SAM enzyme, Science 332 (2011) 1089–1092. [18] J.L. Vey, C.L. Drennan, Structural insights into radical generation by the radical SAM superfamily, Chem. Rev. 111 (2011) 2487–2506. [19] J.L. Vey, et al., Structural basis for glycyl radical formation by pruvate formate–lyase activating enzyme, Proc. Natl. Acad. Sci. U. S. A. 105 (2008) 16137–16141. [20] P.J. Goldman, et al., X-ray structure of an AdoMet radical activase reveals an anaerobic solution for formylglycine posttranslational modification, Proc. Natl. Acad. Sci. U. S. A. 110 (21) (2013) 8519–8524. [21] J.E. Harmer, et al., Structures of lipoyl synthase reveal a compact active site for controlling sequential sulfur insertion reactions, Biochem. J. 464 (2014) 123–133. [22] P. Hänzelmann, H. Schindelin, Binding of 5′-GTP to the C-terminal FeS cluster of the radical S-adenosylmethionine enzyme MoaA provides insights into its mechanism, Proc. Natl. Acad. Sci. U. S. A. 103 (18) (2006) 6829–6834. [23] S. Goto-Ito, et al., Structure of an archaeal TYW1, the enzyme catalyzing the second step of wye-base biosynthesis, Acta Crystallogr. D Biol. Crystallogr. 63 (Pt 10) (2007) 1059–1068. [24] F. Forouhar, et al., Two Fe–S clusters catalyze sulfur insertion by radical-SAM methylthiotransferases, Nat. Chem. Biol. 9 (5) (2013) 333–338. [25] S. Arragain, et al., Post-translational modification of ribosomal proteins: structural and functional characterization of RimO from Thermotoga maritima, a radical Sadenosylmethionine methylthiotransferase, J. Biol. Chem. 285 (2010) 5792–5801. [26] N.D. Lanz, S.J. Booker, Identification and function of auxiliary iron-sulfur clusters in radical SAM enzymes, Biochim. Biophys. Acta 1824 (11, Sp. Iss. SI) (2012) 1196–1212. [27] A. Marquet, B.T. Bui, D. Florentin, Biosynthesis of biotin and lipoic acid, Vitam. Horm. 61 (2001) 51–101. [28] A. Marquet, et al., In vivo formation of C\S bonds in biotin. An example of radical chemistry under reducing conditions, J. Phys. Org. Chem. 11 (8–9) (1998) 529–535. [29] C.J. Fugate, J.T. Jarrett, Biotin synthase: insights into radical-mediated carbonâ€sulfur bond formation, Biochim. Biophys. Acta (BBA), Proteins Proteomics 1824 (11) (2012) 1213–1222. [30] J.R. Miller, et al., Escherichia coli LipA is a lipoyl synthase: in vitro biosynthesis of lipoylated pyruvate dehydrogenase complex from octanoyl-acyl carrier protein, Biochemistry 39 (49) (2000) 15166–15178. [31] P. Douglas, et al., Lipoyl synthase inserts sulfur atoms into an octanoyl substrate in a stepwise manner, Angew. Chem. Int. Ed. Engl. 45 (31) (2006) 5197–5199. [32] R.M. Cicchillo, S.J. Booker, Mechanistic investigations of lipoic acid biosynthesis in Escherichia coli: both sulfur atoms in lipoic acid are contributed by the same lipoyl synthase polypeptide, J. Am. Chem. Soc. 127 (9) (2005) 2860–2861. [33] X. Zhao, et al., Assembly of the covalent linkage between lipoic acid and its cognate enzymes, Chem. Biol. 10 (12) (2003) 1293–1302. [34] M. Atta, et al., S-adenosylmethionine-dependent radical-based modification of biological macromolecules, Curr. Opin. Struct. Biol. 20 (2010) 1–9. [35] B.P. Anton, et al., Functional characterization of the YmcB and YqeV tRNA methylthiotransferases of Bacillus subtilis, Nucleic Acids Res. 38 (2010) 6195–6205. [36] S. Arragain, et al., Identification of eukaryotic and prokaryotic methylthiotransferase for biosynthesis of 2-methylthio-N6-threonylcarbamoyladenosine in tRNA, J. Biol. Chem. 285 (37) (2010) 28425–28433. [37] B.P. Anton, et al., RimO, a MiaB-like enzyme, methylthiolates the universally conserved Asp88 residue of ribosomal protein S12 in Escherichia coli, Proc. Natl. Acad. Sci. U. S. A. 105 (6) (2008) 1826–1831. [38] I. Sanyal, G. Cohen, D.H. Flint, Biotin synthase: purification, characterization as a (2Fe–2S) cluster protein, and in vitro activity of the Escherichia coli bioB gene product, Biochemistry 33 (12) (1994) 3625–3631. [39] E.C. Duin, et al., (2Fe–2S) to (4Fe–4S) cluster conversion in Escherichia coli biotin synthase, Biochemistry 36 (39) (1997) 11811–11820. [40] B. Tse Sum Bui, et al., Mössbauer studies of Escherichia coli biotin synthase: evidence for reversible interconversion between [2Fe–2S]2+ and [4Fe–4S]2+ clusters, FEBS Lett. 459 (3) (1999) 411–414. [41] S. Ollagnier-de Choudens, et al., Iron-sulfur center of biotin synthase and lipoate synthase, Biochemistry 39 (14) (2000) 4165–4173. [42] C.J. Fugate, et al., 9-Mercaptodethiobiotin is generated as a ligand to the [2Fe–2S]+ cluster during the reaction catalyzed by biotin synthase from Escherichia coli, J. Am. Chem. Soc. 134 (2012) 9042–9045. [43] M.M. Cosper, et al., Recombinant Escherichia coli biotin synthase is a [2Fe–2S]2+ protein in whole cells, FEBS Lett. 529 (2–3) (2002) 332–336. [44] M.M. Cosper, et al., Characterization of the cofactor composition of Escherichia coli biotin synthase, Biochemistry 43 (7) (2004) 2007–2021. [45] G.N.L. Jameson, et al., Role of the (2Fe–2S) cluster in recombinant Escherichia coli biotin synthase, Biochemistry 43 (7) (2004) 2022–2031. [46] N.B. Ugulava, C.J. Sacanell, J.T. Jarrett, Spectroscopic changes during a single turnover of biotin synthase: destruction of a (2Fe–2S) cluster accompanies sulfur insertion, Biochemistry 40 (28) (2001) 8352–8358. [47] N.B. Ugulava, K.K. Surerus, J.T. Jarrett, Evidence from Mössbauer spectroscopy for distinct [2Fe–2S]2+ and [4Fe–4S]2+ cluster binding sites in biotin synthase from Escherichia coli, J. Am. Chem. Soc. 124 (31) (2002) 9050–9051. [48] R. Benda, et al., Iron–sulfur clusters of biotin synthase in vivo: a Mössbauer study, Biochemistry 41 (50) (2002) 15000–15006. [49] B. Tse Sum Bui, et al., Biotin synthase mechanism: on the origin of sulphur, FEBS Lett. 440 (1–2) (1998) 226–230. [50] K.J. Gibson, D.A. Pelletier, I.M. Turner Sr., Transfer of sulfur to biotin from biotin synthase (BioB protein), Biochem. Biophys. Res. Commun. 254 (3) (1999) 632–635.

N.D. Lanz, S.J. Booker / Biochimica et Biophysica Acta 1853 (2015) 1316–1334 [51] A.M. Taylor, et al., Reduction of the [2Fe–2S] cluster accompanies formation of the intermediate 9-mercaptodethiobiotin in Escherichia coli biotin synthase, Biochemistry 50 (37) (2011) 7953–7963. [52] R.M. Cicchillo, et al., Lipoyl synthase requires two equivalents of S-adenosyl-Lmethionine to synthesize one equivalent of lipoic acid, Biochemistry 43 (21) (2004) 6378–6386. [53] P. Bryant, et al., The activity of a thermostable lipoyl synthase from Sulfolobus solfataricus with a synthetic octanoyl substrate, Anal. Biochem. 351 (1) (2006) 44–49. [54] N.D. Lanz, et al., Evidence for a catalytically and kinetically competent enzymesubstrate cross-linked intermediate in catalysis by lipoyl synthase, Biochemistry 53 (28) (2014) 4557–4572. [55] U. Mühlenhoff, et al., The iron–sulfur proteins Isa1 and Isa2 are required for the function but not for the de novo synthesis of the Fe/S clusters of biotin synthase in Saccharomyces cerevisiae, Eukaryot. Cell 6 (2007) 495–504. [56] U. Mühlenhoff, et al., Specialized function of yeast Isa1 and Isa2 proteins in the maturation of mitochondrial [4Fe–4S] proteins, J. Biol. Chem. 286 (48) (2011) 41205–41216. [57] L.E. Vickery, J.R. Cupp-Vickery, Molecular chaperones HscA/Ssq1 and HscB/Jac1 and their roles in iron–sulfur protein maturation, Crit. Rev. Biochem. Mol. Biol. 42 (2) (2007) 95–111. [58] J.J. Silberg, K.G. Hoff, L.E. Vickery, The Hsc66–Hsc20 chaperone system in Escherichia coli: chaperone activity and interactions with the DnaK–DnaJ–grpE system, J. Bacteriol. 180 (24) (1998) 6617–6624. [59] L.E. Vickery, J.J. Silberg, D.T. Ta, Hsc66 and Hsc20, a new heat shock cognate molecular chaperone system from Escherichia coli, Protein Sci. 6 (5) (1997) 1047–1056. [60] M.R. Reyda, C.J. Fugate, J.T. Jarrett, A complex between biotin synthase and the iron–sulfur cluster assembly chaperone HscA that enhances in vivo cluster assembly, Biochemistry 48 (45) (2009) 10782–10792. [61] C. Gelling, et al., Mitochondrial Iba57p is required for Fe/S cluster formation on aconitase and activation of radical SAM enzymes, Mol. Cell. Biol. 28 (5) (2008) 1851–1861. [62] A.D. Sheftel, et al., The human mitochondrial ISCA1, ISCA2, and IBA57 proteins are required for [4Fe–4S] protein maturation, Mol. Biol. Cell 23 (7) (2013) 1157–1166. [63] N. Ajit Bolar, et al., Mutation of the iron–sulfur cluster assembly gene IBA57 causes severe myopathy and encephalopathy, Hum. Mol. Genet. 22 (13) (2013) 2590–2602. [64] J.M. Cameron, et al., Mutations in iron–sulfur cluster scaffold genes NFU1 and BOLA3 cause a fatal deficiency of multiple respiratory chain and 2-oxoacid dehydrogenase enzymes, Am. J. Hum. Genet. 89 (4) (2011) 486–495. [65] A. Navarro-Sastre, et al., A fatal mitochondrial disease is associated with defective NFU1 function in the maturation of a subset of mitochondrial Fe–S proteins, Am. J. Hum. Genet. 89 (5) (2011) 656–667. [66] F. Pierrel, et al., Enzymatic modification of tRNAs, J. Biol. Chem. 277 (16) (2002) 13367–13370. [67] F. Pierrel, et al., MiaB protein from Thermotoga maritima, J. Biol. Chem. 278 (32) (2003) 29515–29524. [68] B. Esberg, et al., Identification of the miaB gene, involved in methylthiolation of isopentenylated A37 derivatives in the tRNA of Salmonella typhimurium and Escherichia coli, J. Bacteriol. 181 (23) (1999) 7256–7265. [69] P.F. Agris, et al., Maturation of a hypermodified nucleoside in transfer RNA, Nucleic Acids Res. 2 (5) (1975) 691–698. [70] K.H. Lee, et al., Characterization of RimO, a new member of the methylthiotransferase subclass of the radical SAM superfamily, Biochemistry 48 (42) (2009) 10162–10174. [71] B.J. Landgraf, et al., Identification of an intermediate methyl carrier in the radical S-adenosylmethionine methylthiotransferases RimO and MiaB, J. Am. Chem. Soc. 135 (41) (2013) 15404–15416. [72] T. Ote, et al., Involvement of the Escherichia coli folate-binding protein YgfZ in RNA modification and regulation of chromosomal replication initiation, Mol. Microbiol. 59 (1) (2006) 265–275. [73] J.C. Waller, et al., A role for tetrahydrofolates in the metabolism of iron–sulfur clusters in all domains of life, Proc. Natl. Acad. Sci. U. S. A. 107 (23) (2010) 10412–10417. [74] J.C. Waller, et al., Evidence that the folate-dependent proteins YgfZ and MnmEG have opposing effects on growth and on activity of the iron–sulfur enzyme MiaB, J. Bacteriol. 194 (2) (2011) 362–367. [75] C.E. Gelling, Tetrahydrofolate and Iron–Sulfur Metabolism in Saccharomyces cerevisiae, 2008. (Ph. D. thesis). [76] S. Boutigny, et al., Physical and functional interactions of a monothiol glutaredoxin and an iron sulfur cluster carrier protein with the sulfur-donating radical S-adenosyl-Lmethionine enzyme MiaB, J. Biol. Chem. 288 (20) (2013) 14200–14211. [77] T.L. Grove, et al., In vitro characterization of AtsB, a radical SAM formylglycinegenerating enzyme that contains three [4Fe–4S] clusters, Biochemistry 47 (28) (2008) 7523–7538. [78] A. Benjdia, et al., Anaerobic sulfatase-maturating enzyme—a mechanistic link with glycyl radical-activating enzymes? FEBS J. 277 (8) (2010) 1906–1920. [79] S.C. Pegg, et al., Leveraging enzyme structure–function relationships for functional inference and experimental design: the structure–function linkage database, Biochemistry 45 (8) (2006) 2545–2555. [80] M.P. Cosma, et al., The multiple sulfatase deficiency gene encodes an essential and limiting factor for the activity of sulfatases, Cell 113 (4) (2003) 445–456. [81] T. Dierks, et al., Multiple sulfatase deficiency is caused by mutations in the gene encoding the human Cα-formylglycine generating enzyme, Cell 113 (4) (2003) 435–444. [82] Q. Fang, J. Peng, T. Dierks, Post-translational formylglycine modification of bacterial sulfatases by the radical S-adenosylmethionine protein AtsB, J. Biol. Chem. 279 (15) (2004) 14570–14578.

1333

[83] C. Marquordt, et al., Posttranslational modification of serine to formylglycine in bacterial sulfatases. Recognition of the modification motif by the iron–sulfur protein AtsB, J. Biol. Chem. 278 (4) (2003) 2212–2218. [84] C. Szameit, et al., The iron sulfur protein AtsB is required for posttranslational formation of formylglycine in the Klebsiella sulfatase, J. Biol. Chem. 274 (22) (1999) 15375–15381. [85] A. Benjdia, et al., Anaerobic sulfatase-maturating enzymes: radical SAM enzymes able to catalyze in vitro sulfatase post-translational modification, J. Am. Chem. Soc. 129 (12) (2007) 3462–3463. [86] A. Benjdia, et al., Anaerobic sulfatase-maturating enzymes, first dual substrate radical S-adenosylmethionine enzymes, J. Biol. Chem. 283 (26) (2008) 17815–17826. [87] A. Benjdia, et al., Mechanistic investigations of anaerobic sulfatase-maturating enzyme: direct Cβ H-atom abstraction catalyzed by a radical AdoMet enzyme, J. Am. Chem. Soc. 131 (24) (2009) 8348–8349. [88] T.L. Grove, et al., Further characterization of Cys-type and Ser-type anaerobic sulfatase maturating enzymes suggests a commonality in the mechanism of catalysis, Biochemistry 52 (17) (2013) 2874–2887. [89] K. Yokoyama, et al., Characterization and mechanistic study of a radical SAM dehydrogenase in the biosynthesis of butirosin, J. Am. Chem. Soc. 129 (49) (2007) 15147–15155. [90] K. Yokoyama, et al., Mechanistic study on the reaction of a radical SAM dehydrogenase BtrN by electron paramagnetic resonance spectroscopy, Biochemistry 47 (34) (2008) 8950–8960. [91] T.L. Grove, et al., A consensus mechanism for radical SAM-dependent dehydrogenation? BtrN contains two [4Fe–4S] clusters, Biochemistry 49 (18) (2010) 3783–3785. [92] S. Bali, et al., Molecular hijacking of siroheme for the synthesis of heme and d1 heme, Proc. Natl. Acad. Sci. U. S. A. 108 (45) (2011) 18260–18265. [93] S. Storbeck, et al., A novel pathway for the biosynthesis of heme in Archaea: genome-based bioinformatic predictions and experimental evidence, Archaea 2010 (2010) 175050. [94] B. Buchenau, et al., Heme biosynthesis in Methanosarcina barkeri via a pathway involving two methylation reactions, J. Bacteriol. 188 (24) (2006) 8666–8668. [95] S.A. Lobo, et al., Characterisation of Desulfovibrio vulgaris haem b synthase, a radical SAM family member, Biochim. Biophys. Acta 1844 (7) (2014) 1238–1247. [96] K.V. Rajagopalan, J.L. Johnson, The pterin molybdenum cofactors, J. Biol. Chem. 267 (15) (1992) 10199–10202. [97] J. Reiss, Genetics of molybdenum cofactor deficiency, Hum. Genet. 106 (2) (2000) 157–163. [98] J. Reiss, J.L. Johnson, Mutations in the molybdenum cofactor biosynthetic genes MOCS1, MOCS2, and GEPH, Hum. Mutat. 21 (6) (2003) 569–576. [99] M. Fischer, et al., 7.17 — the biosynthesis of folate and pterins and their enzymology, Comprehensive Natural Products II, Elsevier, Oxford, 2010. 599–648. [100] M.M. Wuebbens, K.V. Rajagopalan, Investigation of the early steps of molybdopterin biosynthesis in Escherichia coli through the use of in vivo labeling studies, J. Biol. Chem. 270 (3) (1995) 1082–1087. [101] A.P. Mehta, S.H. Abdelwahed, T.P. Begley, Molybdopterin biosynthesis: trapping an unusual purine ribose adduct in the MoaA-catalyzed reaction, J. Am. Chem. Soc. 135 (30) (2013) 10883–10885. [102] A.P. Mehta, et al., Molybdopterin biosynthesis: trapping of Intermediates for the MoaA-catalyzed reaction using 2′-DeoxyGTP and 2′-ChloroGTP as substrate analogues, J. Am. Chem. Soc. 136 (30) (2014) 10609–10614. [103] A.P. Mehta, et al., Catalysis of a new ribose carbon-insertion reaction by the molybdenum cofactor biosynthetic enzyme MoaA, Biochemistry 52 (7) (2013) 1134–1136. [104] B.M. Hover, et al., Identification of a cyclic nucleotide as a cryptic intermediate in molybdenum cofactor biosynthesis, J. Am. Chem. Soc. 135 (18) (2013) 7019–7032. [105] P. Hänzelmann, et al., Characterization of MOCS1A, an oxygen-sensitive iron–sulfur protein involved in human molybdenum cofactor biosynthesis, J. Biol. Chem. 279 (33) (2004) 34721–34732. [106] C. Menendez, et al., A pAO1-encoded molybdopterin cofactor gene (moaA) of Arthrobacter nicotinovorans: characterization and site-directed mutagenesis of the encoded protein, Arch. Microbiol. 164 (2) (1995) 142–151. [107] C. Menendez, D. Siebert, R. Brandsch, MoaA of Arthrobacter nicotinovorans pAO1 involved in Mo-pterin cofactor synthesis is an Fe–S protein, FEBS Lett. 391 (1–2) (1996) 101–103. [108] N.S. Lees, et al., ENDOR spectroscopy shows that guanine N1 binds to [4Fe–4S] cluster II of the S-adenosylmethionine-dependent enzyme MoaA: mechanistic implications, J. Am. Chem. Soc. 131 (26) (2009) 9184–9185. [109] A. Silakov, et al., (14)N HYSCORE investigation of the H-cluster of [FeFe] hydrogenase: evidence for a nitrogen in the dithiol bridge, Phys. Chem. Chem. Phys. 11 (31) (2009) 6592–6599. [110] J.W. Peters, et al., X-ray crystal structure of the Fe-only hydrogenase (CpI) from Clostridium pasteurianum to 1.8 angstrom resolution, Science 282 (5395) (1998) 1853–1858. [111] Y. Nicolet, et al., Desulfovibrio desulfuricans iron hydrogenase: the structure shows unusual coordination to an active site Fe binuclear center, Structure 7 (1) (1999) 13–23. [112] A.J. Pierik, et al., Carbon monoxide and cyanide as intrinsic ligands to iron in the active site of [NiFe]-hydrogenases. NiFe(CN)2CO, biology's way to activate H2, J. Biol. Chem. 274 (6) (1999) 3331–3337. [113] D.W. Mulder, et al., Stepwise [FeFe]-hydrogenase H-cluster assembly revealed in the structure of HydAΔEFG, Nature 465 (7295) (2010) 248–251. [114] X. Brazzolotto, et al., The [Fe–Fe]-hydrogenase maturation protein HydF from Thermotoga maritima is a GTPase with an iron–sulfur cluster, J. Biol. Chem. 281 (2) (2006) 769–774.

1334


[115] S.E. McGlynn, et al., In vitro activation of [FeFe] hydrogenase: new insights into hydrogenase maturation, J. Biol. Inorg. Chem. 12 (4) (2007) 443–447. [116] S.E. McGlynn, et al., HydF as a scaffold protein in [FeFe] hydrogenase H-cluster biosynthesis, FEBS Lett. 582 (15) (2008) 2183–2187. [117] M.C. Posewitz, et al., Discovery of two novel radical S-adenosylmethionine proteins required for the assembly of an active [Fe] hydrogenase, J. Biol. Chem. 279 (24) (2004) 25711–25720. [118] I. Czech, et al., The [FeFe]-hydrogenase maturase HydF from Clostridium acetobutylicum contains a CO and CN− ligated iron cofactor, FEBS Lett. 584 (3) (2010) 638–642. [119] E.M. Shepard, et al., Synthesis of the 2Fe subcluster of the [FeFe]-hydrogenase H cluster on the HydF scaffold, Proc. Natl. Acad. Sci. U. S. A. 107 (23) (2010) 10448–10453. [120] J.K. Rubach, et al., Biochemical characterization of the HydE and HydG iron-only hydrogenase maturation enzymes from Thermatoga maritima, FEBS Lett. 579 (22) (2005) 5055–5060. [121] Y. Nicolet, et al., X-ray structure of the [FeFe]-hydrogenase maturase HydE from Thermotoga maritima, J. Biol. Chem. 283 (27) (2008) 18861–18872. [122] P.W. King, et al., Functional studies of [FeFe] hydrogenase maturation in an Escherichia coli biosynthetic system, J. Bacteriol. 188 (6) (2006) 2163–2172. [123] E. Pilet, et al., The role of the maturase HydG in [FeFe]-hydrogenase active site synthesis and assembly, FEBS Lett. 583 (3) (2009) 506–511. [124] R.C. Driesener, et al., [FeFe]-hydrogenase cyanide ligands derived from Sadenosylmethionine-dependent cleavage of tyrosine, Angew. Chem. Int. Ed. Engl. 49 (9) (2010) 1687–1690. [125] E.M. Shepard, et al., [FeFe]-hydrogenase maturation: HydG-catalyzed synthesis of carbon monoxide, J. Am. Chem. Soc. 132 (27) (2010) 9247–9249. [126] Y. Nicolet, et al., A glycyl free radical as the precursor in the synthesis of carbon monoxide and cyanide by the [FeFe]-hydrogenase maturase HydG, FEBS Lett. 584 (19) (2010) 4197–4202. [127] R.C. Driesener, et al., Biochemical and kinetic characterization of radical Sadenosyl-L-methionine enzyme HydG, Biochemistry 52 (48) (2013) 8696–8707. [128] J.M. Kuchenreuther, et al., A radical intermediate in tyrosine scission to the CO and CN– ligands of FeFe hydrogenase, Science 342 (6157) (2013) 472–475. [129] J.M. Kuchenreuther, et al., The HydG enzyme generates an Fe(CO)2(CN) synthon in assembly of the FeFe hydrogenase H-cluster, Science 343 (6169) (2014) 424–427. [130] H. Eklund, M. Fontecave, Glycyl radical enzymes: a conservative structural basis for radicals, Structure 7 (11) (1999) R257–R262. [131] M. Fontecave, Ribonucleotide reductases and radical reactions, Cell. Mol. Life Sci. 54 (7) (1998) 684–695. [132] G. Sawers, G. Watson, A glycyl radical solution: oxygen-dependent interconversion of pyruvate formate-lyase, Mol. Microbiol. 29 (4) (1998) 945–954. [133] T. Selmer, A.J. Pierik, J. Heider, New glycyl radical enzymes catalysing key metabolic steps in anaerobic bacteria, Biol. Chem. 386 (10) (2005) 981–988. [134] H.G. Viehe, et al., The captodative effect, Acc. Chem. Res. 18 (5) (1985) 148–154. [135] V. Unkrig, F.A. Neugebauer, J. Knappe, The free radical of pyruvate formate-lyase. Characterization by EPR spectroscopy and involvement in catalysis as studied with the substrate-analogue hypophosphite, Eur. J. Biochem. 184 (3) (1989) 723–728. [136] P.I. Andrei, et al., Subunit composition of the glycyl radical enzyme p-hydroxyphenylacetate decarboxylase. A small subunit, HpdC, is essential for catalytic activity, Eur. J. Biochem. 271 (11) (2004) 2225–2230. [137] C. Raynaud, et al., Molecular characterization of the 1,3-propanediol (1,3-PD) operon of Clostridium butyricum, Proc. Natl. Acad. Sci. U. S. A. 100 (9) (2003) 5010–5015. [138] B. Leuthner, et al., Biochemical and genetic characterization of benzylsuccinate synthase from Thauera aromatica: a new glycyl radical enzyme catalysing the first step in anaerobic toluene metabolism, Mol. Microbiol. 28 (3) (1998) 615–628. [139] S. Craciun, J.A. Marks, E.P. Balskus, Characterization of choline trimethylaminelyase expands the chemistry of glycyl radical enzymes, ACS Chem. Biol. 9 (7) (2014) 1408–1413. [140] J.M. Demick, W.N. Lanzilotta, Radical SAM activation of the B12-independent glycerol dehydratase results in formation of 5′-deoxy-5′-(methylthio)adenosine and not 5′-deoxyadenosine, Biochemistry 50 (4) (2011) 440–442. [141] B. Selvaraj, et al., 4-Hydroxyphenylacetate decarboxylase activating enzyme catalyses a classical S-adenosylmethionine reductive cleavage reaction, J. Biol. Inorg. Chem. 18 (6) (2013) 633–643. [142] S. Craciun, E.P. Balskus, Microbial conversion of choline to trimethylamine requires a glycyl radical enzyme, Proc. Natl. Acad. Sci. U. S. A. 109 (52) (2012) 21307–21312. [143] U.L. RajBhandary, S.H. Chang, Studies on polynucleotides. LXXXII. Yeast phenylalanine transfer ribonucleic acid: partial digestion with ribonuclease T-1 and derivation of the total primary structure, J. Biol. Chem. 243 (3) (1968) 598–608. [144] F.V. Murphy IV, et al., The role of modifications in codon discrimination by tRNALys UUU, Nat. Struct. Mol. Biol. 11 (12) (2004) 1186–1191. [145] J. Li, et al., Three modified nucleosides present in the anticodon stem and loop influence the in vivo aa-tRNA selection in a tRNA-dependent manner, J. Mol. Biol. 271 (2) (1997) 209–221.

[146] A.L. Konevega, et al., Purine bases at position 37 of tRNA stabilize codon–anticodon interaction in the ribosomal A site by stacking and Mg2+-dependent interactions, RNA 10 (1) (2004) 90–101. [147] C. Yarian, et al., Accurate translation of the genetic code depends on tRNA modified nucleosides, J. Biol. Chem. 277 (19) (2002) 16391–16395. [148] J. Urbonavicius, et al., Improvement of reading frame maintenance is a common function for several tRNA modifications, EMBO J. 20 (2001) 4863–4873. [149] G.R. Björk, et al., A primordial tRNA modification required for the evolution of life? EMBO J. 20 (2001) 231–239. [150] A.S. Byström, G.R. Björk, Chromosomal location and cloning of the gene (trmD) responsible for the synthesis of tRNA (m1G) methyltransferase in Escherichia coli K-12, Mol. Gen. Genet. MGG 188 (3) (1982) 440–446. [151] L. Droogmans, H. Grosjean, Enzymatic conversion of guanosine 3′ adjacent to the anticodon of yeast tRNAPhe to N1-methylguanosine and the wye nucleoside: dependence on the anticodon sequence, EMBO J. 6 (2) (1987) 477–483. [152] W.F. Waas, V. de Crecy-Lagard, P. Schimmel, Discovery of a gene family critical to wyosine base formation in a subset of phenylalanine-specific transfer RNAs, J. Biol. Chem. 280 (45) (2005) 37616–37622. [153] A. Noma, et al., Biosynthesis of wybutosine, a hyper-modified nucleoside in eukaryotic phenylalanine tRNA, EMBO J. 25 (10) (2006) 2142–2154. [154] L.J. Alderwick, et al., Structure, function and biosynthesis of the Mycobacterium tuberculosis cell wall: arabinogalactan and lipoarabinomannan assembly with a view to discovering new drug targets, Biochem. Soc. Trans. 35 (Pt 5) (2007) 1325–1328. [155] A.P. Young, V. Bandarian, Pyruvate is the source of the two carbons that are required for formation of the imidazoline ring of 4-demethylwyosine, Biochemistry 50 (49) (2011) 10573–10575. [156] P. Perche-Letuvée, et al., 4-Demethylwyosine synthase from Pyrococcus abyssi is a radical-S-adenosyl-L-methionine enzyme with an additional [4Fe–4S]+2 cluster that interacts with the pyruvate co-substrate, J. Biol. Chem. 287 (49) (2012) 41174–41185. [157] P. Cheeseman, A. Toms-Wood, R.S. Wolfe, Isolation and properties of a fluorescent compound, factor 420, from Methanobacterium strain M.o.H, J. Bacteriol. 112 (1) (1972) 527–531. [158] M.S. Kuo, et al., Isolation and identification of 7,8-didemethyl-8-hydroxy-5deazariboflavin, an unusual cosynthetic factor in streptomycetes, from Streptomyces lincolnensis, J. Antibiot. 42 (3) (1989) 475–478 (Tokyo). [159] T.B. Bair, D.W. Isabelle, L. Daniels, Structures of coenzyme F420 in Mycobacterium species, Arch. Microbiol. 176 (1–2) (2001) 37–43. [160] D.E. Graham, H. Xu, R.H. White, Identification of the 7,8-didemethyl-8-hydroxy-5deazariboflavin synthase required for coenzyme F420 biosynthesis, Arch. Microbiol. 180 (6) (2003) 455–464. [161] A.F. Glas, et al., The archaeal cofactor F0 is a light-harvesting antenna chromophore in eukaryotes, Proc. Natl. Acad. Sci. U. S. A. 106 (28) (2009) 11540–11545. [162] B. Reuke, et al., Biosynthetic precursors of deazaflavins, J. Bacteriol. 174 (12) (1992) 4042–4049. [163] L. Decamps, et al., Biosynthesis of F0, precursor of the F420 cofactor, requires a unique two radical-SAM domain enzyme and tyrosine as substrate, J. Am. Chem. Soc. 134 (44) (2012) 18173–18176. [164] R. Jaenchen, P. Schonheit, R.K. Thauer, Studies on the biosynthesis of coenzyme F420 in methanogenic bacteria, Arch. Microbiol. 137 (4) (1984) 362–365. [165] K.P. Choi, N. Kendrick, L. Daniels, Demonstration that fbiC is required by Mycobacterium bovis BCG for coenzyme F420 and F0 biosynthesis, J. Bacteriol. 184 (9) (2002) 2420–2428. [166] J.C. Escalante-Semerena, K.L. Rinehart Jr., R.S. Wolfe, Tetrahydromethanopterin, a carbon carrier in methanogenesis, J. Biol. Chem. 259 (15) (1984) 9447–9455. [167] R. Fischer, R.K. Thauer, Methyltetrahydromethanopterin as an intermediate in methanogenesis from acetate in Methanosarcina barkeri, Arch. Microbiol. 151 (5) (1989) 459–465. [168] L. Chistoserdova, et al., C1 transfer enzymes and coenzymes linking methylotrophic bacteria and methanogenic Archaea, Science 281 (5373) (1998) 99–102. [169] J.A. Vorholt, et al., Distribution of tetrahydromethanopterin-dependent enzymes in methylotrophic bacteria and phylogeny of methenyl tetrahydromethanopterin cyclohydrolases, J. Bacteriol. 181 (18) (1999) 5750–5757. [170] L.L. Grochowski, R.H. White, Biosynthesis of methanogenic coenzymes, Comprehensive Natural Products II, Elsevier, Oxford, 2010. 711–748. [171] B.E. Maden, Tetrahydrofolate and tetrahydromethanopterin compared: functionally distinct carriers in C1 metabolism, Biochem. J. 350 (Pt 3) (2000) 609–629. [172] K.D. Allen, H. Xu, R.H. White, Identification of a unique radical S-adenosylmethionine methylase likely involved in methanopterin biosynthesis in Methanocaldococcus jannaschii, J. Bacteriol. 196 (18) (2014) 3315–3323. [173] F. Yan, et al., RlmN and Cfr are radical SAM enzymes involved in methylation of ribosomal RNA, J. Am. Chem. Soc. 132 (2010) 3953–3964. [174] T.L. Grove, et al., A radically different mechanism for S-adenosylmethioninedependent methyltansferases, Science 332 (2011) 604–607.

Carbon extension in peptidylnucleoside biosynthesis by radical SAM enzymes.

SPASM and twitch domains in S-adenosylmethionine (SAM) radical enzymes.

Radical S-adenosylmethionine (SAM) enzymes in cofactor biosynthesis: a treasure trove of complex organic radical rearrangement reactions.

Introduction to the thematic minireview series on radical S-adenosylmethionine (SAM) enzymes.

Radical SAM-Mediated Methylation of Ribosomal RNA.

On the Role of Additional [4Fe-4S] Clusters with a Free Coordination Site in Radical-SAM Enzymes.

Recent advances in radical SAM enzymology: new structures and mechanisms.

The biosynthesis of thiol- and thioether-containing cofactors and secondary metabolites catalyzed by radical S-adenosylmethionine enzymes.

Mechanistic diversity of radical S-adenosylmethionine (SAM)-dependent methylation.

Radical S-adenosylmethionine enzymes.

Mechanistic Enzymology of the Radical SAM Enzyme DesII.

A B12-dependent radical SAM enzyme involved in oxetanocin A biosynthesis.

Radical S-adenosyl-L-methionine chemistry in the synthesis of hydrogenase and nitrogenase metal cofactors.

Watasemycin biosynthesis in Streptomyces venezuelae: thiazoline C-methylation by a type B radical-SAM methylase homologue.

Radical SAM enzyme QueE defines a new minimal core fold and metal-dependent mechanism.

Common themes and differences in SAM recognition among SAM riboswitches.

GAL11 protein, an auxiliary transcription activator for genes encoding galactose-metabolizing enzymes in Saccharomyces cerevisiae.

Radical SAM catalysis via an organometallic intermediate with an Fe-[5'-C]-deoxyadenosyl bond.

C-H methylation of heteroarenes inspired by radical SAM methyl transferase.

Biosynthesis of the [FeFe] Hydrogenase H Cluster: A Central Role for the Radical SAM Enzyme HydG.

Structures of the peptide-modifying radical SAM enzyme SuiB elucidate the basis of substrate recognition.

Characterisation of Desulfovibrio vulgaris haem b synthase, a radical SAM family member.

A sensitive luminescent assay for the histone methyltransferase NSD1 and other SAM-dependent enzymes.

Non-canonical active site architecture of the radical SAM thiamin pyrimidine synthase.