The Molybdenum Cofactor*

The transition element molybdenum needs to be complexed by a special cofactor to gain catalytic activity. Molybdenum is bound to a unique pterin, thus forming the molybdenum cofactor (Moco), which, in different variants, is the active compound at the catalytic site of all molybdenum-containing enzymes in nature, except bacterial molybdenum nitrogenase. The biosynthesis of Moco involves the complex interaction of six proteins and is a process of four steps, which also require iron, ATP, and copper. After its synthesis, Moco is distributed, involving Moco-binding proteins. A deficiency in the biosynthesis of Moco has lethal consequences for the respective organisms.

The transition element molybdenum is an essential micronutrient for microorganisms, plants, and animals (1). Surprisingly, molybdenum itself is catalytically inactive in biological systems until it is complexed by a special scaffold (2). One type of scaffold is the ubiquitous pterin-based molybdenum cofactor (Moco), 2 which, in different variants, forms part of the active centers of all molybdenum enzymes in living organisms, except one. This exception is bacterial nitrogenase, which harbors the other type of cofactor, namely the Fe-S cluster-based FeMo cofactor, which is found only once in nature (for details, compare the minireview of Hu and Ribbe (62) in this thematic series). Molybdenum belongs to the group of trace elements, i.e. the organism needs it only in minute amounts; however, unavailability of molybdenum is lethal for the organism. Molybdenum is very abundant in the oceans in the form of the molybdate anion (3). In soils, the molybdate anion is the only form of molybdenum that is available for bacteria, plants, and fungi. More than 50 enzymes are known to be molybdenumdependent. The vast majority of them are found in bacteria, whereas only seven have been identified in eukaryotes (4,5). It is somewhat surprising that not all organisms need molybdenum. The commonly used eukaryotic model organism yeast plays no role in molybdenum research, as Saccharomyces cerevisiae does not contain either molybdenum enzymes or the Moco biosynthesis pathway. Also Schizosaccharomyces pombe does not use molybdenum, whereas Pichia pastoris needs molybdenum. Genome-wide database analyses revealed a significant number of bacteria and unicellular eukaryotes that do not need molybdenum, whereas all multicellular eukaryotes are dependent on molybdenum (6). In addition, mainly anaerobic archaea and some bacteria are molybdenum-independent, but they require tungsten for their growth (7). In the periodic table of elements, tungsten lies directly below molybdenum and features chemical properties similar to molybdenum.
Molybdenum metabolism is tightly connected to Fe-S cluster synthesis in that some of the molybdenum enzymes and Moco biosynthesis itself depend on Fe-S enzymes and on a mitochondrial transporter that is known to be crucial for the maturation of cytosolic Fe-S proteins. Moreover, Moco biosynthesis recruits mechanisms previously known from Fe-S cluster synthesis, which involves the mobilization of sulfide for the formation of a Mo-S center for specific molybdenum enzymes, which will be touched upon below.

What Is the Molybdenum Cofactor?
Molybdenum has a versatile redox chemistry that is used by the enzymes to catalyze diverse redox reactions. Molybdenum enzymes generally catalyze the transfer of an oxygen atom (ultimately derived from or incorporated into water) to or from a substrate (5). Each reaction, either reduction or oxidation, involves the transfer of two electrons, thereby causing a change in the oxidation state of the molybdenum atom in the substratebinding site from IV to VI or vice versa. Most remarkably, it turns out that the metal is not directly attached to the catalytic site; rather, the molybdenum atom is complexed within a specific low molecular scaffold to fulfill its catalytic function. This compound is a unique tricyclic pterin called molybdopterin or metal-containing pterin (MPT) (Fig. 1), the latter reflecting the fact that, in bacteria, not only molybdenum but also tungsten can be coordinated by this pterin scaffold, which seems to be the rule in archaea that prefer tungsten instead of molybdenum. As the result of molybdenum coordination by MPT, Moco is formed (2). The chemical nature of Moco was elucidated by the work of J. L. Johnson and K. V. Rajagopalan ( Fig. 1) (8). As Moco turned out to be very labile and sensitive to air oxidation, its stable oxidation products were used to uncover its pterin nature. However, Moco proved to be a unique pterin derivative because it possesses a four-carbon side chain as a C6 substituent that coordinates molybdenum via an enedithiolate group located within the four-carbon side chain of the cofactor (9). Finally, crystal structures of molybdenum enzymes confirmed this core structure and showed the existence of a third pyrano ring between the OH group at C3Ј of the side chain and the pterin C7 atom (10). Once the pyrano ring is closed, a fully reduced hydrogenated pterin is formed. In prokaryotes, variants of Moco occur: (i) Moco can bind a nucleotide (GMP or CMP) to its phosphate, thus forming a dinucleotide cofactor; and (ii) molybdenum can be coordinated by two pterin or two dinucleotide equivalents, thus forming bis forms of Moco. In this minireview, I will focus on higher organisms (eukaryotes). Bacterial Moco variants were recently reviewed in detail (11). What could be the task of the pterin moiety of Moco? The fusion of a pterin with a pyrano ring as identified for Moco and its direct precursor, the metal-free MPT, is unique in nature and may have evolved to position the catalytic metal correctly within the active center of a given molybdenum enzyme. Another possible role of the pterin moiety could be control of the redox behavior of the molybdenum atom. In addition, the pterin might also participate in the electron transfer to or from molybdenum via the delocalized electrons within the pterin (5). X-ray crystallographic analyses of molybdenum enzymes revealed that the cofactor is not located on the surface of the protein but is buried deeply within the interior of the enzyme, and a tunnel-like structure makes it accessible to the cognate substrates (12,13). During its lifetime, the molybdenum enzyme does not liberate Moco. In vitro, however, Moco can be removed from its protein environment, where after it easily loses the molybdenum atom and becomes rapidly oxidized, resulting in an irreversible loss of function. The demolybdo forms of molybdenum enzymes are catalytically inactive. To this end, there are no indications for a Moco-recycling mechanism in the cell.

Molybdenum Uptake into Cells
Molybdenum is taken up in the form of its oxyanion molybdate. In the presence of competing anions, it requires specific uptake systems, which have been studied in detail in bacteria, where high affinity ATP-binding cassette (ABC)-type transporters have been described that require ATP hydrolysis for operation (7). In higher organisms such as in algae and plants, only recently have the first molybdate-transporting proteins been identified (14 -16). In the large sulfate carrier superfamily, two proteins, Mot1 and Mot2, were shown to transport molybdate with ultrahigh affinity (nanomolar k m value) across cell membranes. Surprisingly, neither of them was found to reside in the plasma membrane. Contradictory reports localized Mot1 to the endomembrane system (14) and to the mitochondrial envelope (16). The latter is questionable, however, as the insertion of molybdenum into the Moco backbone takes place in the cytosol. For Mot2, GFP fusion proteins have been used to show that the protein localizes to the vacuolar membrane (17). Molybdate quantification in isolated vacuoles demonstrated that this organelle serves as an important molybdate storage compartment in Arabidopsis thaliana cells, where Mot2 was shown to be required for vacuolar molybdate export into the cytosol. The major question that remains is how molybdate enters the cell. The answer might come from recent results obtained with the alga Chlamydomonas reinhardtii, where another molybdate transporter has been identified that, unlike Mot1 and Mot2, is not found exclusively in algae and higher plants but also occurs in humans (18). Although still not localized, it is likely that this transporter serves as the general molybdate importer for the cell. Furthermore, it appears that, in addition to the specific high affinity uptake system, molybdate may also enter the cell nonspecifically through the sulfate and phosphate uptake system. Molybdate uptake through a sulfate transporter has recently been described, thus supporting this assumption (19).

Molybdenum Cofactor Biosynthesis
Early genetic work demonstrated that mutations in the genes for Moco biosynthesis result in the pleiotropic loss of all molybdenum-dependent cellular processes. Analysis of Moco-deficient mutants in a given organism ranging from bacteria to plants and humans resulted in the identification of several gene loci involved in Moco biosynthesis (2,11). Along with the conserved structure of Moco, these findings provided a basis to propose an evolutionary old multistep biosynthesis pathway (20). The first model for Moco biosynthesis was presented by Rajagopalan and Johnson (21) for the bacterium Escherichia coli. Later studies of Moco biosynthesis uncovered a more complex picture of this pathway in higher organisms, where molecular, biochemical, and genetic analyses of Moco mutants were advanced most in plants (22). These results formed the basis to decipher Moco biosynthesis also in humans, and it turned out that the pathways of Moco biosynthesis have a high degree of similarity in both organisms (2).
In all higher organisms studied so far, Moco is synthesized by a conserved biosynthesis pathway that can be divided into four steps according to the biosynthetic intermediates cyclic pyranopterin monophosphate (cPMP; also known previously as precursor Z), MPT, adenylated MPT (MPT-AMP), and Moco, respectively ( Fig. 2). In eukaryotes, six proteins catalyzing Moco biosynthesis have been identified in plants (23), fungi (24), and humans (25)(26)(27). These genes are homologous to their counterparts in bacteria, and some (but not all) of the eukaryotic Moco biosynthetic genes are able to functionally complement the matching bacterial mutants. Different nomenclatures were introduced for genes and gene products involved in Moco formation. Genes and the encoded proteins were named in plants according to the cnx nomenclature (cofactor for nitrate reductase and xanthine dehydrogenase). For human Moco synthetic genes, a different MOCS (molybdenum cofactor synthesis) nomenclature was introduced, and the names for both the plant and human proteins are given in Fig. 2. A third nomenclature is used in bacteria. For comparison, Fig. 2 also shows the names of the corresponding bacterial proteins. The individual steps of Moco biosynthesis are briefly characterized below.

Biosynthesis Step 1: Conversion of GTP to cPMP
Like the biosynthesis of other pteridines, Moco synthesis starts with 5Ј-GTP, which is converted by a complex reaction sequence into cPMP (Fig. 2). In contrast to the other pteridine pathways (producing three-carbon side chains), MPT is unique There are two different molybdenum enzyme families known in eukaryotes. In enzymes of the sulfite oxidase family, X is represented by a single-bonded sulfur provided by a cysteine residue of the respective protein, whereas Y corresponds to a double-bonded oxygen. In enzymes of the xanthine oxidase family, X is represented by a double-bonded inorganic sulfur, whereas Y corresponds to a hydroxyl group.
in having a four-carbon side chain. cPMP is the most stable intermediate of Moco biosynthesis, with an estimated half-life of several hours at low pH (28). Therefore, it was possible to solve its structure (E. coli) using 1 H NMR, whereas structural elucidation of MPT, MPT-AMP, and Moco required crystalli-zation of protein-ligand complexes (29). cPMP already possesses a fully reduced tricyclic tetrahydropyranopterin structure and is predominantly hydrated at C1Ј, resulting in a geminal diol (30). GTP labeling studies and 1 H NMR demonstrated that each carbon atom of the ribose and of the guanine FIGURE 2. Biosynthesis of eukaryotic Moco. The biosynthesis pathway is divided into four steps, as shown in italics on the right. On the left, the names for the proteins from humans (red), plants (green), and E. coli (black) catalyzing the respective steps are given. For MPT and MPT-AMP, the ligands of the dithiolate sulfurs are indicated by R, as it is currently unknown at which step copper is bound to the dithiolate. In GTP, the C8 atom of the purine is labeled with an asterisk. This carbon is inserted between the 2Ј-and 3Ј-ribose carbon atoms, thus forming the new C1Ј position in the four-carbon side chain of the pterin (labeled with an asterisk in cPMP). In step 2, the heterotetrameric MPT synthase complex converts cPMP into MPT. In this process, two sulfur atoms need to be transferred from the thiocarboxylated C termini of the small subunits, which later form the dithiolene group of MPT. Having transferred their sulfur atoms, the small subunits need to be reloaded with sulfur, which is facilitated by the MPT synthase sulfurase. This enzyme consists of two domains, with the N-terminal adenylation domain (AD) catalyzing the Mg-ATP-dependent adenylation at the C-terminal carboxyl group of the small subunit of MPT synthase and the C-terminal RLD being responsible for subsequent sulfur transfer. In step 3, the two-domain protein molybdenum insertase catalyzes the Mg-ATP-dependent adenylation of MPT at its G-domain, with the subsequent transfer of MPT-AMP to its E-domain. In step 4 and occurring at the E-domain, MPT-AMP is deadenylated, and the molybdate anion is incorporated to form mature Moco. Geph-G and Geph-E, gephyrin G-and E-domains.
ring are incorporated into cPMP. The underlying mechanism involves a complex radical-based rearrangement reaction in which the C8 atom of the purine is inserted between the 2Ј-and 3Ј-ribose carbon atoms, thus forming the new C1Ј position in the four-carbon side chain of the pterin (28,31).
cPMP is first intermediate of Moco biosynthesis. It is still sulfur-free, but it already has the tricyclic pyranopterin structure similar to the mature cofactor. In all organisms, the conversion of GTP to cPMP is catalyzed by two proteins; one of them (Cnx2 in plants and MOCS1A in humans) belongs to the superfamily of S-adenosylmethionine (SAM)-dependent radical enzymes (32). Members of this protein family catalyze the formation of protein and/or substrate radicals by reductive cleavage of SAM involving a [4Fe-4S] cluster. The MOCS1A protein contains two oxygen-sensitive Fe-S clusters, each coordinated by only three cysteine residues (32). For its bacterial homolog (MoaA protein in E. coli and Staphylococcus aureus), the complex reaction mechanism has been deciphered in detail (33,34). As the plant gene cnx2 (35) and the human gene MOCS1A (26) are able to functionally complement their bacterial counterparts, one can assume that the reaction mechanism is likely to occur similarly in eukaryotes. The N-terminal [4Fe-4S] cluster binds SAM and carries out the reductive cleavage of SAM to generate the 5Ј-deoxyadenosyl radical, which subsequently initiates the transformation of 5Ј-GTP bound through the C-terminal [4Fe-4S] cluster. The function of the second protein involved in catalysis step 1 (i.e. Cnx3 in plants and MOCS1B in humans) is as yet unknown, but it is believed to participate in pyrophosphate release upon the rearrangement reaction.

Export of cPMP from Mitochondria and a Link between Moco and Fe-S Cluster Synthesis
In eukaryotes, the two proteins involved in step 1 of Moco biosynthesis carry N-terminal extensions, predicting a mitochondrial localization of these proteins. Subfractionation of mitochondria demonstrated that Cnx2 and Cnx3 reside in the matrix, where 5Ј-GTP is available as a substrate for cPMP synthesis and where Fe-S clusters are synthesized as the essential prosthetic group for Cnx2 (36). In contrast to the first step, all subsequent steps of Moco biosynthesis were demonstrated to be localized in the cytosol (37)(38)(39), and thus, after its synthesis in mitochondria, cPMP has to pass the mitochondrial membranes to enable its further processing to Moco (Fig. 3). Although cPMP is hydrophobic enough to pass through biological membranes simply by diffusion, recent work on plants demonstrated that a specific transporter in the inner membrane of mitochondria is involved in the transport of cPMP into the cytosol (36). The respective transporter belongs to the ABC transporter family and is referred to as ATM3 in plants. Inter-

FIGURE 3. Organization of biosynthesis and distribution of Moco in higher organisms (plants). Moco biosynthesis starts in the mitochondria. The enzymes
Cnx2 and Cnx3 catalyze the SAM-dependent conversion of GTP to cPMP. Cnx2 requires Fe-S clusters provided by the mitochondrial Fe-S synthesis machinery, which also generates an as yet unknown compound that is exported by the ABC-type transporter ATM3 to allow synthesis of cytosolic Fe-S clusters. However, ATM3 is also involved in the transport of cPMP from mitochondria into the cytosol, where MPT synthase adds two sulfur atoms and converts cPMP to MPT. The small subunit Cnx7 of MPT synthase, which holds the transfer-ready sulfur atom, has received this sulfur from the enzyme MPT synthase sulfurase Cnx5. AD denotes the adenylation domain of Cnx5, which is required for adenylation and activation of Cnx7, whereas the sulfur is mobilized by the RLD. It is assumed that copper is inserted directly after dithiolene formation. The individual reactions of the molybdenum insertase Cnx1 and its products (Moco, pyrophosphate (PP i ), AMP, and copper) are indicated. Mature Moco can be bound to a MoBP, directly to the molybdenum enzymes, or to the Moco-binding domain (MocoBD) of the Moco sulfurase ABA3. The Moco sulfurase generates a protein-bound persulfide, which is the source of the terminal sulfur ligand of Moco in xanthine oxidoreductase and aldehyde oxidase. Unlike Cnx2, the latter two enzymes depend on cytosolic Fe-S clusters. mARC, mitochondrial amidoxime-reducing component.
estingly, together with yeast Atm1p and mammalian ABCB7, this transporter was originally identified to be essential for the maturation of extramitochondrial Fe-S proteins by transporting an as yet unknown compound generated during mitochondrial Fe-S cluster synthesis into the cytosol, where it serves as a substrate for the cytosolic Fe-S assembly machinery (63). Loss of ATM3 function is associated not only with a deficiency in extramitochondrial Fe-S proteins but also with an accumulation of cPMP in mitochondria, which has the consequence that the cytosol becomes short of cPMP, thus leading to decreased Moco levels and molybdenum enzyme activities in the cell (36). The precise role of ATM3 is still unknown, however.

Biosynthesis Step 2: Synthesis of Molybdopterin
In the second step, sulfur is transferred to cPMP to generate MPT. This reaction is catalyzed by the enzyme MPT synthase, a heterotetrameric complex (Fig. 2) of two small (Cnx7 and MOCS2B) and two large (Cnx6 and MOCS2A) subunits that stoichiometrically converts cPMP into MPT. The sulfur is bound to the C terminus of the small subunits as thiocarboxylate. Due to the fact that each small subunit of MPT synthase carries a single sulfur atom, a two-step mechanism for the formation of the MPT dithiolate has been proposed that was deciphered in detail in bacteria (11). Among all small subunits so far analyzed from diverse species, the C-terminal region is highly conserved and includes a terminal double glycine that is of functional importance for thiocarboxylation (40,41). E. coli MPT synthase was found to be an elongated protein complex in which the thiocarboxylated C termini of the small subunits are deeply inserted into the large subunits to form two clearly separated active sites (42). Obviously, the two sulfur atoms are not simultaneously transferred to cPMP; rather, the sulfurs become sequentially inserted, starting with C2Ј of cPMP, with the consequence that a monosulfurated reaction intermediate will occur (43). Whether the intermediate will be transferred within the MPT synthase to the other active site or whether the enzyme dissociates to host another sulfurated small subunit remains to be seen. Again, as in step 1, also the reaction mechanism of MPT synthase is conserved between bacteria and higher organisms, as, at least for the large subunits, proteins can be exchanged between organisms (41).
After MPT synthase has transferred the two sulfurs to cPMP, it has to be resulfurated by the enzyme MPT synthase sulfurase (Cnx5 and MOCS3, respectively) (Fig. 2) to reactivate the enzyme for the next reaction cycle of cPMP conversion. This resulfuration involves adenylation of the MPT synthase small subunit followed by the sulfur transfer. At this stage, the sulfur transfer reaction in higher organisms appears to involve different protein components, as the eukaryotic genes cannot complement their bacterial counterparts. MPT synthase sulfurase is a two-domain protein consisting of an N-terminal adenylation domain (homologous to E. coli MoeB) and a C-terminal rhodanese-like domain (RLD), where the sulfur is bound to a conserved cysteine in the form of a persulfide (38,44). In analogy to the bacterial mechanism, this enzyme is supposed to activate the small subunit of MPT synthase by adenylation followed by sulfur transfer (coming from the RLD), thus forming the thiocarboxylate at the C terminus of the small subunit (2). There-fore, MPT synthase sulfurase can be viewed as a multifunctional protein combining two subsequent reactions carried out by two domains fused to each other (Fig. 2), thus representing a good example of product-substrate channeling. In humans, cysteine desulfurase Nfs1 is a likely candidate to function as sulfur donor for the MOCS3-catalyzed resulfuration step (45).

Biosynthesis Step 3: Molybdenum Insertion Starts with Adenylation of Molybdopterin
After synthesis of the MPT moiety, the chemical backbone is built to bind and coordinate the molybdenum atom (Fig. 2). Therefore, in the third step, molybdate is transferred to MPT to form Moco, thus linking the molybdate uptake system to the MPT pathway, which is not a spontaneous process, however, but is catalyzed by a molybdenum insertase. Mutants defective in this step accumulate MPT but can be partially rescued by growth on high molybdate (1-10 mM) medium, which is used as an assay tool to identify molybdenum insertase mutants (46). However, physiological molybdate concentrations are not sufficient to achieve any non-catalyzed molybdenum ligation by MPT. In bacteria, this step is catalyzed by two separately expressed proteins (MogA and MoeA), whereas during evolution to higher organisms, these two proteins were fused to a single two-domain molybdenum insertase (Cnx1 in plants and gephyrin in mammals). The two domains of molybdenum insertase are named the G-domain (homologous to MogA) and the E-domain (homologous to MoeA) (Fig. 2), and work carried out with the molybdenum insertase Cnx1 from the model plant A. thaliana assigned different mechanistic functions to each of these domains (47,48). The metal insertion reaction can be subdivided into two separate steps. Structural studies revealed that to coordinate molybdenum, MPT first has to be activated by adenylation. This is carried out by the Cnx1 G-domain in a Mg 2ϩ -and ATP-dependent manner, thus generating MPT-AMP. The finding that MPT-AMP represents a general reaction intermediate in Moco biosynthesis was further extended by recent studies in E. coli (11). MPT-AMP serves as a substrate for the subsequent molybdenum insertion reaction, which is carried out by the E-domain of Cnx1.

Biosynthesis Step 4: Molybdenum Insertion into Molybdopterin
In the final step, MPT-AMP is transferred from the G-domain of Cnx1 to the E-domain, which cleaves the adenylate from MPT and catalyzes the insertion of molybdate into the dithiolene group of MPT, thus yielding physiologically active Moco (Fig. 2). The MPT adenylate is hydrolyzed in a Mg 2ϩ -and molybdate-dependent way, and adenylated molybdate might occur as a hypothetical reaction intermediate (47,48). Moco formed by the Cnx1 E-domain most probably carries two oxo ligands and one OH group in a deprotonated form (48). There is no experimental evidence for a reduction of molybdenum at this stage.
The crystal structure of the Cnx1 G-domain reveals an unexpected finding, namely a copper bound to the MPT dithiolate sulfurs, whose nature was confirmed by anomalous scattering of the metal (1, 29). These structures show tetragonal coordi-nation of copper, suggesting a type I copper binding site for Cu ϩ . Given the presence of copper in MPT, the insertion of molybdenum into the MPT dithiolene group can be characterized as a metal exchange reaction, with copper presumably serving as a suitable leaving group. It is also possible that copper protects the MPT dithiolate from oxidation. The origin of this copper is still unclear, but it is reasonable to assume that it binds to the enedithiolate group just after the latter has been formed, i.e. at the end of step 2 of Moco biosynthesis. Because copper occurs in vivo exclusively protein-bound, it is likely that both copper binding to MPT and its exchange for molybdenum depend on yet unidentified cytoplasmic chaperones involved in cellular copper metabolism.

Product-Substrate Channeling in Moco Biosynthesis
Bacteria catalyze steps 3 and 4 of Moco biosynthesis by separate proteins (MogA and MoeA, respectively), whereas higher organisms combined these two consecutive steps into a single protein (plant Cnx1 and human gephyrin) with two domains. Both domains were fused at least two times during evolution, resulting in two-domain proteins with different orientations of the G-and E-domains: plants have the E-domain at the N terminus of the protein, and mammals and fungi have the G-domain at the N terminus (2,27). These evolutionarily distinct events point to a high pressure and a functional benefit of having the adenylation function and the metal insertion function coupled into one protein, where the fragile intermediate MPT-AMP is channeled from the G-domain to the E-domain (49).
Clearly, facilitated product-substrate flow seems to be a general characteristic of the Moco biosynthesis cascade. Indeed, it was recently found that Cnx5, Cnx6/7, and Cnx1 (catalyzing steps 2-4) undergo tight protein-protein interaction in the cytosol of living plant cells (39), thus supporting the idea of channeling the fragile Moco intermediates in a protected way within a multiprotein complex.

Storage and Transfer of Moco
Moco is extremely sensitive to oxidation (21) and is therefore assumed to occur permanently protein-bound in the cell. Also the fast flow of Moco to its target enzymes is an essential prerequisite to reduce the threat of Moco degradation. Both preconditions may be met by Moco-binding proteins (MoBPs), ensuring Moco binding and its directed transfer to cognate target enzymes. Thus, a pool of insertion-competent Moco may be stored and provided on demand. In eukaryotes, the first MoBP named Moco carrier protein (MCP) was identified in the green alga C. reinhardtii (50). The protein is able to bind and protect Moco against oxidation, and the atomic structure shows that it forms a homotetramer capable of holding four molecules of Moco (51). Without any denaturing procedure, the subsequent transfer of Moco from the carrier protein to aponitrate reductase (NR) from Neurospora crassa was possible. It is unknown, however, whether MCP is also able to donate Moco to molybdenum enzymes other than NR. Preliminary data suggest that molybdenum is bound in a trioxo-coordinated form in MCP. However, a complex structure of MCP with Moco is still unavailable.
In the higher plant A. thaliana, a family of eight MCP-related proteins was identified, all of which can bind Moco (52). Their biochemical characterization showed reversible Moco-binding properties, however, with overall lower affinities. Therefore, these MoBPs are not good candidates to serve as Moco storage proteins. Rather, they seem to be involved in the cellular distribution of Moco because they were found to undergo protein-protein interactions with both the Moco donor protein Cnx1 and the Moco acceptor protein NR (Fig. 3). This observation does not exclude a direct transfer of Moco from the donor protein Cnx1 to the molybdenum enzyme, which has been shown in vitro.
Insertion of Moco into molybdenum enzymes is still not understood. All crystal structures of molybdenum enzymes demonstrate that Moco is buried deeply within the holoenzymes (12). Hence, it follows that Moco needs to be incorporated prior to or during completion of folding and dimerization of the apoprotein monomers. In bacteria, a complex of proteins synthesizing the last steps of Moco biosynthesis donates the mature cofactor to apoenzymes assisted by enzyme-specific chaperones. Nearly every bacterial molybdenum enzyme has a private chaperone available (4). However, no eukaryotic Moco chaperones have so far been identified.

Final Sulfuration of Moco
Two different molybdenum enzyme families are known in eukaryotes (5): (i) the sulfite oxidase family, to which also NR and the mitochondrial amidoxime-reducing component belong, and the (ii) xanthine oxidase family, to which also aldehyde oxidase belongs. In bacteria, a third class of molybdenum enzymes is known in which two MPT equivalents coordinate one molybdenum atom (4). It is assumed that the rare eukaryotic molybdenum enzymes pyridoxal oxidase (53) and nicotinate hydroxylase (54) represent specific isoforms of aldehyde oxidase. In contrast to the sulfite oxidase family, the members of the xanthine oxidase family require a final step of maturation prior to or after insertion of Moco. In addition to the dithiolene sulfurs of the pterin moiety and two oxo groups, the molybdenum atom of Moco needs the addition of a terminal inorganic sulfur to provide enzyme activity to these enzymes (55). This final step is catalyzed by the Moco sulfurase protein (ABA3 in plants and HMCS in humans) (Fig. 3). ABA3 is a homodimeric two-domain protein (56), with its N-terminal domain sharing structural and functional homologies with bacterial cysteine desulfurases, thereby being more similar to SufS than to NifS or IscS. In a pyridoxal phosphate-dependent manner, the N-terminal domain of ABA3 decomposes L-cysteine to yield alanine and elemental sulfur (57), the latter being bound as a persulfide to a highly conserved cysteine residue of ABA3. The C-terminal domain of ABA3 shares a significant degree of similarity with the newly discovered mitochondrial amidoxime-reducing component proteins and was shown to bind sulfurated Moco, which receives the terminal sulfur via an intramolecular persulfide relay from the N-terminal domain (58,59). It is likely that subsequent to Moco sulfuration, ABA3 exchanges non-sulfurated for sulfurated Moco, thus activating its target molybdenum enzyme.

Moco Deficiency and Therapy in Humans
Human Moco deficiency is a rare recessive hereditary disorder, which ultimately results in the death of affected patients. Lack of Moco biosynthesis results in the pleiotropic loss of all molybdenum-dependent enzyme activities (2). Symptoms develop shortly after birth, when the infant's metabolism starts to operate, and toxic metabolites (mainly sulfite, which is formed upon degradation of sulfur-containing amino acids) accumulate within the body (60). Two-thirds of patients have a defect in step 1 of Moco biosynthesis (formation of cPMP). The first human exposure to cPMP treatment has been reported recently (61), where a patient was diagnosed on day 6 of life, and experimental treatment was started on day 36. Within days, all biomarkers returned to almost normal readings and stayed constant.