To access the private area of this site, please log in.

Username: 

Password:  

 

Chapter 42

Biosynthesis of the Molybdenum Cofactor

K. V. RAJAGOPALAN

 

INTRODUCTION

Studies on the molybdenum cofactor of liver sulfite oxidase led to the structural characterization of a novel pterin, termed molybdopterin (MPT), bearing a unique dithiolene group on its 6-alkyl side chain (22). MPT is the invariant component of a family of cofactors and is found in all enzymes containing Mo or W, with the sole exception of nitrogenase (22). The active cofactor has never been isolated in a pure state because of its extreme instability. However, structural studies on a dicarboxamidomethyl derivative of MPT from sulfite oxidase provided strong evidence for the validity of the proposed elemental constituents of the pterin (22). The proposed structure (9) for the molybdenum cofactor of liver sulfite oxidase is shown in Fig. 1 and features coordination of the dithiolene group to Mo. Analyses of a number of molybdoenzymes have shown that there is one MPT per Mo (22). Prokaryotic variants of the cofactor contain GMP, CMP, AMP, or IMP linked to MPT by a PP_i bond (23). Evidence has been presented for a dihydro form of the pterin in sulfite oxidase and xanthine oxidase (6). Verification of the proposed structure by direct characterization of the active cofactor itself has not been possible.

Chan et al. (4) have recently elucidated the X-ray crystallographic structure of aldehyde ferredoxin oxidoreductase, a tungsten-containing enzyme (13, 14) from the hyperthermophilic archaean Pyrococcus furiosus, and have refined the details of the W site. Their analysis supports every aspect of the proposed cofactor structure, including coordination of the metal to dithiolene. The only surprise is that in the aldehyde ferredoxin oxidoreductase there are two MPTs per W, and both dithiolenes are coordinated to the metal.

A number of inducible molybdenum-containing enzymes have been identified in Escherichia coli. The physiology of formate dehydrogenases N and H and nitrate reductase has been thoroughly and critically reviewed by Stewart (28). Dimethyl sulfoxide reductase (31) and trimethylamine N-oxide reductase (27) are also induced in anaerobically grown E. coli in the presence of the respective electron-accepting substrates. A constitutive enzyme, biotin sulfoxide reductase (5), serves to scavenge biotin sulfoxide by reductively converting it to biotin. Analysis of nitrate reductase, formate dehydrogenase, and dimethyl sulfoxide reductase has shown that all of them contain MPT guanine dinucleotide (MGD) as the pterin component of their molybdenum cofactor (23, 25).

Because of the existence of molybdoenzyme-independent anaerobic respiratory pathways, none of the above enzymes is essential for the growth of E. coli. The absence of any or all of these enzymes is therefore nonlethal for the organism, permitting the isolation of a number of pleiotropic mutants. Several genetic loci, previously designated as chl and now termed moa, mob, mod, moe, or mog, have been implicated in the pleiotropy of the molybdenum enzymes, most of them being involved in the biosynthesis of the molybdenum cofactor. The genetic aspects of these loci have also been discussed in detail by Stewart (28).

In 1987, Johnson and Rajagopalan (10) described two new pleiotropic mutants of E. coli, chlM and chlN, that had lesions in the moa and moe loci, respectively. They made the crucial observation that chlM cells contained a low-molecular-weight compound that could be converted to MPT by proteins present in extracts of a previously known mutant with the chlA1 allele. Subsequent isolation and structural characterization of the putative precursor, termed precursor Z (Fig. 2), by Wuebbens and Rajagopalan (32) provided a crucial breakthrough for delineating many of the aspects of the MPT biosynthetic pathway in E. coli, summarized in Fig. 2.

In parallel developments, two of the major loci involved in molybdenum cofactor biosynthesis were cloned and their nucleotide and deduced amino acid sequences were determined. The sequence of the moa operon was determined by Rivers et al. (24), revealing five open reading frames. Nohno and coworkers (16) cloned the moe operon and reported two open reading frames in the clone. The confluence of information from these cloning studies with the biochemical information on the components of the terminal step in MPT synthesis has provided considerable insight into the roles of the moa and moe gene products in the reactions of the pathway, as discussed below. The deduced masses of the proteins encoded by the five moa genes and the two moe genes are listed in Table 1. The assignment of a series of moa Mu insertion mutants (29) as reported by Rivers et al. is also shown. The two mutants isolated by Johnson and Rajagopalan (10), chlM and chlN, correspond to moaD and moeB, respectively.

 

BIOSYNTHESIS OF THE MOLYBDENUM COFACTOR

The reactions of the MPT biosynthetic pathway comprise four stages. The first of these leads to the formation of the sulfur-free pterin precursor Z. The second stage involves the conversion of the precursor to MPT. The uptake and processing of molybdenum are required for the formation of the Mo-MPT complex. These three processes are presumably common to all organisms utilizing molybdoenzymes. The final stage in the biosynthetic pathway involves the formation of the dinucleotide variants—MGD in the case of E. coli. This chapter will summarize the recent developments in the cloning of the mo genes and identification of their biosynthetic function.

 

Mechanism of Conversion of Precursor Z to MPT (moaDE)

Precursor Z isolated from moaD and moeB mutants is a dihydropterin. Unlike MPT in which the side chain contains a terminal monophosphate and the unique dithiolene moiety, precursor Z has a side chain with a six-member cyclic phosphate and no sulfur. Pitterle and Rajagopalan (20) purified a protein from the moaA mutant of E. coli by its apparent ability to convert precursor Z to MPT. The heteromeric protein is composed of two different subunits of 16.8 and 8.5 kDa, both of which are required for activity (19, 20). N-terminal amino acid sequence analysis of the two subunits showed that they are encoded by moaD (8.5 kDa) and moaE (16.8 kDa). Pitterle et al. (18) showed that admixture of purified MPT synthase and precursor Z resulted in the formation of MPT, identified by derivatization to dicarboxamidomethyl MPT. The sensitivity of MoaD but not MoaE to the sulfhydryl reagents N-ethylmaleimide, N-bromobimane, and iodoacetamide suggested that the smaller subunit of MPT synthase is the donor of the dithiolene sulfurs (20).

Pitterle and Rajagopalan (20) also purified inactive MPT synthase from the moeB mutant and compared its properties with those of active enzyme by electrospray mass spectrometry. The large subunits of the active and inactive synthase had identical molecular masses; however, the small subunit isolated from the moeB mutant was lower in mass by 16 Da. Since the key reaction carried out by MPT synthase is the addition of the dithiolene sulfurs to the pterin side chain, the difference in mass could be accounted for by the substitution of a sulfur for an oxygen in the active form of the synthase. A reactive, transferable form of sulfur, such as a thiocarboxylate, incorporated into the small subunit could serve as the direct sulfur source for dithiolene formation. This conclusion is supported by the sensitivity of active MoaD, but not the inactive protein from the moeB mutant, to sulfhydryl reagents (20) and the observation that the MPT synthase from the moeB mutant can be activated by incubation with another protein, MPT synthase sulfurylase (see below).

Since MPT synthase serves as the sole source of sulfur for dithiolene formation, as studied in the purified system, the conversion of precursor Z to MPT by the synthase can only be stoichiometric. For the synthase to act catalytically, it is necessary to regenerate its transferable sulfur. Since the moeB mutant contains the nonsulfurated form of the synthase, it appeared that the MoeB protein is involved in the activation of the synthase by sulfur transfer.

 

MPT Synthase Sulfurylase (moeB)

Plasmid pTN525, isolated by Nohno et al., contains both moeA (44 kDa) and moeB (26.7 kDa) genes (16). This plasmid was able to correct the moeB mutant. To identify the complementing gene, the moaD mutant RK5200, which lacks the small subunit of the converting factor, was transformed with the plasmid. Since this plasmid does not correct moa mutants, no MPT is produced in the transformed cell. When RK5200/pTN525 extract was mixed with MPT synthase from the moeB mutant, in vitro formation of MPT was observed (D. M. Pitterle, Ph.D. thesis, Duke University, Durham, N.C., 1991). The levels of MoeB protein in untransformed RK5200 are very low so that, in the absence of the plasmid, in vitro complementation with moeB extracts was not detectable. The activating factor purified from PK5200/pTN552 extracts displayed a single band with a molecular weight of 26,000 upon sodium dodecyl sulfate-gel electrophoresis, identifying it as the moeB gene product. The purified protein has been shown to contain zinc (34). Since activation of MPT synthase requires transfer of the mobile sulfur atom to the small subunit of the synthase, the moeB gene product has been named MPT synthase sulfurylase. Details of the mechanism of action of the sulfurylase, including the identity of the sulfur donor for the protein, are as yet unknown.

Significant sequence similarities between MoeB and a number of other proteins have been identified and are illustrated in Fig. 3. ThiF is a protein encoded by the thi operon which functions in the biosynthesis of thiamine. Since the thiF mutant is defective in the synthesis of the thiazole ring, the product of that gene must carry out a function similar to that of MoeB (30). HesA was identified by Borthakur et al. (2) as an open reading frame in a cDNA clone isolated from heterocysts of Anabaena sp. strain PCC 7120 and may be involved in the assembly of an iron-sulfur protein and is likely to resemble MoeB in its function. All of these proteins contain two Cys-X-X-Cys sequences, possibly representing the zinc-binding site. A highly interesting and suggestive homology is that between MoeB and UBA1, the gene coding for ubiquitin-activating enzyme E1 (12). The amino acid sequences of MoeB and the yeast activating enzyme show 23% identity. Similar levels of identity are noted with each of the other activating enzymes from wheat and humans, but as a group the three proteins show 50% identity to MoeB. The homology to ubiquitin-activating enzymes has direct implications as to the sulfur transfer mechanism of the sulfurylase, since ATP-dependent thiocarboxylate ester formation is an essential feature of the ubiquitin-dependent protein degradation system (7). To initiate the process by which ubiquitin is attached to a protein targeted for degradation, the carboxyl group of ubiquitin’s C-terminal glycine is linked to a cysteine of the activating enzyme E1, forming a thioester. The activating enzyme then transfers the ubiquitin to a carrier protein, E2, forming a new thioester linkage. Ubiquitin itself is a highly conserved protein of 76 amino acids. Its sequence bears little homology to those of any of the known molybdenum cofactor-forming enzymes, with the notable exception that its carboxy-terminal Gly-Gly sequence is identical to that of the small subunit of MPT synthase. The MoeB, ThiF, and HesA proteins may carry out reactions similar to that of UBA1, with the exception that the second step of the reaction would generate a free thiocarboxylate rather than the thioester created by UBA1. This reaction would require transfer of the carboxyl to a sulfide ion rather than a cysteine residue. Presumably, the metal center of MoeB, and by inference ThiF and HesA, serves to create the sulfite ion. The absence of the two Cys-X-X-Cys sequences in UBA1 is in accord with this hypothesis.

 

Dinucleotide Formation (mobAB)

The pleiotropy of the mob locus of E. coli has long been a puzzle since mutants with lesions therein contain high levels of active molybdenum cofactor (1). Identification of MGD as the pterin component of the E. coli molybdoenzymes nitrate reductase and formate dehydrogenase (23) suggested the possibility that the function of the mob locus is to synthesize MGD from MPT. Johnson et al. (8) analyzed the MPT pools of wild-type and mob mutant cells and found that wild-type E. coli contained both MPT and MGD, whereas the mob mutant contained only MPT. The pleiotropy of the mob locus shows that all molybdoenzymes of E. coli require MGD for their activity. Santini et al. (26) have reported that purified nitrate reductase from a mob mutant contains Mo and MPT. They also reported that the inactive enzyme in crude extracts of a mob mutant could be activated by a factor designated FA, present in wild-type E. coli and in any mo mutants other than mob (11). More recently they have cloned a fragment of E. coli DNA capable of complementing mob mutants and have expressed and purified factor FA (17). The purified protein was capable of activating nitrate reductase in crude extracts of a mob mutant but could not activate purified mob nitrate reductase. The activation in crude extracts required GTP (17). The sequence of the region of E. coli DNA covering the mob locus has recently been determined and found to contain two contiguous open reading frames, presumably constituting the mob operon (21). The partial mob sequence determined by Palmer et al. (17) for factor FA corresponds to the first gene of the mob operon. Accordingly, the gene encoding FA has been designated mobA. Palmer et al. (17) observed that reconstitution of mob nitrate reductase in mob extracts required stoichiometric amounts of purified FA. They suggested that for FA to act catalytically, an additional factor(s) would be required. Whether the protein encoded by mobB serves such a function remains to be determined.

 

Formation of Precursor Z (moaAto moaC)

The presence of the 6-alkyl side chain in MPT raised the possibility that the early steps in the biosynthetic pathway of the molybdenum cofactor are similar or identical to those of other pteridine biosynthetic pathways. Currently, three major routes for the synthesis of pteridines have been identified which result in the formation of folates in plants and microorganisms (3) and the synthesis of tetrahydrobiopterin and other nonconjugated pterins in animals (15). In all cases, the initial step is effected by the enzyme GTP cyclohydrolase I, which converts GTP to the reduced pterin H₂-neopterin triphosphate (H₂NTP). In this reaction, as shown in Fig. 2, C-8 of guanine is eliminated as formate and C-1' and C-2' of ribose are converted to the C-6 and C-7 atoms of the pterin ring, while C-3', C-4', and C-5' of ribose become C-1', C-2', and C-3', respectively, of the 6-alkyl side chain of H₂NTP. During tetrahydrobiopterin synthesis in animals, H₂NTP is dephosphorylated and rearranged by internal oxidoreduction to yield 6-pyruvoyl tetrahydropterin. Further reduction of the side chain then results in the production of l-erythro-tetrahydrobiopterin, the cofactor utilized by the aromatic amino acid hydroxylases in animals.

A pteridine is also formed as an intermediate during the synthesis of the isoalloxazine ring system of riboflavin in plants and microorganisms (3). Again, the initial step in synthesis is the loss of C-8 of GTP as formate. Although all of the original guanosine atoms except C-8 are again incorporated into the final structure of riboflavin, the ribose carbons of GTP are not involved in the closure of the pteridine ring system in this case but are retained in toto as the ribityl group of riboflavin.

In light of this information, it was not improbable that a guanine derivative also serves as the in vivo precursor of MPT, with compounds such as H₂NTP serving as common intermediates in the MPT and folate pathways. The discovery and structural characterization of the MPT precursor precursor Z and its oxidation product compound Z greatly increased the feasibility of in vivo labeling experiments in whole cells for a number of reasons. In particular, precursor Z accumulates in the E. coli mutants moeB and moaE in amounts much higher than the MPT content of wild-type cells, and compound Z is stable and easily purified. In addition, since both compound Z and precursor Z contain all 10 of the carbon atoms present in MPT, any information derived from in vivo labeling of carbons in compound Z would be immediately relevant to the elucidation of the early steps in MPT biosynthesis.

Wuebbens and Rajagopalan (33) devised procedures for the release and purification of compound Z from total E. coli culture followed by the analysis for the presence of label in both the side chain and pterin ring carbons. Growth of moeB in medium containing U-[¹⁴C]guanosine led to the formation of precursor Z labeled in both the pterin ring and the 6-alkyl side chain, suggesting a GTP cyclohydrolase I type of reaction in the conversion of the guanosine ring to the pterin ring. Such a mechanism should lead to the loss of C-8 of guanosine as formate. Unexpectedly, when [8-¹⁴C]guanosine was used as the tracer, precursor Z contained high levels of radioactivity, all of which was found to reside on C-1' of the side chain. Pterin-6-carboxylase isolated from the bulk pterins of the moeB culture contained only trace levels of radioactivity. The results obtained from these experiments indicate that although a guanosine derivative is indeed the initial in vivo precursor of MPT biosynthesis in E. coli, the pathway by which this guanosine derivative is converted to MPT is unlike all previously known pterin biosynthetic pathways.

Analysis of the Mu insertion mutants listed in Table 1 showed that moaA and moaC mutants do not produce precursor Z which accumulates in the moaD and moaE mutants (Pitterle, Ph.D. thesis). It would thus appear that the MoaA, MoaB, and MoaC proteins are involved in the synthesis of precursor Z from a guanosine derivative. It is not known whether these proteins carry out distinct reactions or whether some of them form heteromeric complexes. Rivers et al. have reported limited sequence similarities of each of these sequences to various other known sequences (24), but no insights can be gleaned from these data. The MoeB sequence is reported to contain a folate-binding motif. The fact that folate coenzymes function as single-carbon carriers is of significance in relation to the observed retention of C-8 of the guanine ring as C-1' of the side chain of precursor Z. It is possible that MoeB uses a folate compound to bind and transfer the guanine C-8 during the synthesis of the pterin ring.

 

Molybdenum Transport and Processing (mod, mog)

It is generally believed that the mod locus codes for a molybdate transporter and that the mog gene product is involved in processing intracellular molybdate. The phenotypes resulting from mutations in both operons are corrected by growth on high molybdate (28). There has been no new information on the biochemistry of these loci.

 

moeA

The original moeA mutant, chlE5, was known to be pleiotropic. Johnson and Rajagopalan (10) found low, but detectable, levels of MPT in moeA mutant cells. No further studies have been carried out on the role of this locus in the synthesis and processing of the molybdenum cofactor.

 

CONCLUSION

The foregoing discussion has summarized the current state of knowledge regarding the biosynthesis of the molybdenum cofactor. There are still many aspects of the pathways to be understood, including the identity of the guanosine derivative serving as the initial substrate, the fascinating mechanism of conversion of the guanosine compound to precursor Z, the individual steps in the sulfur mobilization pathway, and the reactions involved in the uptake and processing of molybdate. The availability of DNA clones containing most of the loci involved in these processes should facilitate progress in these areas.

ACKNOWLEDGMENTS

The work from this laboratory has been supported by grant GM00091 from the National Institutes of Health.