INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Molybdoenzymes are ubiquitous and essential for almost all organisms, from bacteria to plants and animals. All molybdoenzymes (with the exception of nitrogenase) contain the molybdenum cofactor (Moco), which consists of a unique molybdopterin (MPT) complexing one Mo atom via a dithiolene group and which has the same principle structure in eubacteria, archaebacteria, and eukaryotes (29). The biosynthesis of this unique cofactor is complex and requires the multistep synthesis of the MPT moiety and the subsequent incorporation of Mo into MPT. Biosynthesis of Moco is best studied in Escherichia coli, and it was shown that the gene products of five distinct gene loci, moa, mob, mod, moe, and mog, are required (30). Two gene products of the moa locus, MoaA and MoaC, are suggested to be involved in the first steps of Moco biosynthesis, leading to a precursor molecule (precursor Z) of Moco. MPT synthase, encoded by moaD and moaE, catalyzes the conversion of precursor Z to MPT by introducing the sulfur which is finally needed to coordinate Mo. MoeB, one gene product of the moe operon, is required for activation of MPT synthase by sulfurylation (27). The physiological role of the second gene product of the moe operon, MoeA, is not yet established, although MoeA is suggested to be involved in activation of molybdenum by sulfurylation (8). Molybdenum is transported into the cell via a high-affinity molybdate transport system, encoded by the mod gene products (6). Molybdoenzyme activities in mogA mutants can be partially restored to 10 to 13% of the wild-type level by growing these mutants at high molybdate concentrations (43). Therefore, the mogA gene product, characterized as a molybdochelatase, is suspected to be involved in the last step of Moco formation, namely in the insertion of Mo into MPT (15). In E. coli, molybdopterin is further modified by covalent attachment of a GMP moiety to the terminal phosphate group of molybdopterin via a pyrophosphate link, to form the molybdopterin guanine dinucleotide (MGD) cofactor (13).

The phototrophic purple bacterium Rhodobacter capsulatus contains two well-characterized molybdoenzymes containing Moco, dimethyl sulfoxide (DMSO) reductase and xanthine dehydrogenase (XDH) (20, 37). In contrast to E. coli molybdoenzymes, all of which contain MGD, it was shown that R. capsulatus harbors molybdoenzymes coordinating different variants of Moco. The crystal structure of DMSO reductase from R. capsulatus revealed that the molybdenum atom is coordinated by two MGD cofactors (37); in contrast, XDH contains the MPT cofactor, the form of the cofactor present in all eukaryotic molybdoenzymes (20). Although the cofactor of these molybdenum enzymes is characterized in detail, some gene products involved in Moco biosynthesis in R. capsulatus remain to be identified. Nevertheless, Moco biosynthesis for R. capsulatus XDH and DMSO reductase is expected to proceed by a pathway similar to the Moco biosynthesis pathway in E. coli.

During the screening for R. capsulatus mutants unable to degrade xanthine, we obtained one mutant in which XDH activity could be restored to wild-type levels by high concentrations of molybdate (20), whereas DMSO reductase and nitrate reductase activities could not be restored under these conditions. The corresponding mutation was mapped within the moeA gene of R. capsulatus. In this report, we describe a detailed analysis of R. capsulatus MoeA and demonstrate that the pathways of MPT and MGD biosynthesis split at a step when molybdenum is added to the cofactor.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, media, and growth conditions. The bacterial strains and plasmids used in this study are listed in Table 1. R. capsulatus strains were grown in RCV medium as described by Klipp et al. (17). Methods for conjugational plasmid transfer between E. coli and R. capsulatus and the selection of mutants, anaerobic growth conditions, and antibiotic concentrations were as previously described (17, 18).

DNA biochemistry. DNA isolation, restriction enzyme analysis, agarose gel electrophoresis, and cloning procedures were performed by use of standard methods (35). Restriction endonucleases and T4 DNA ligase were purchased from Pharmacia or MBI Fermentas and used as recommended by the suppliers.

DNA sequencing. DNA sequence analysis was carried out with an Auto Read sequencing kit (Pharmacia) according to the protocol devised by Zimmermann et al. (47). Sequence data were obtained and processed by using the A.L.F. DNA sequencer (Pharmacia LKB) as instructed by the manufacturer by using standard fluorescence-labelled primers and appropriate subclones of the 7.3-kb EcoRI fragment (pWKR500). The Staden programs were used for editing and translating DNA sequences (41). For homology searches in sequence databases, the BLAST algorithms (basic local alignment search tool [1]) were used. Sequence alignments were performed with the CLUSTAL W program (44).

Construction of R. capsulatus moeA mutant strains. For the construction of R. capsulatus moeA interposon mutants, various wild-type fragments were cloned by standard methods (35) into mobilizable narrow-host-range vector plasmids. Suitable restriction sites were subsequently used to insert a gentamicin (9) or a kanamycin resistance gene (2). The resulting hybrid plasmids were mobilized from E. coli S17-1 into R. capsulatus by filter mating (23). Mutants were selected for the interposon-encoded resistance, and marker rescue was identified by loss of the vector-encoded resistance.

Construction of a moeA-lacZ fusion plasmid. To create a translational moeA-lacZ fusion, an EcoRI-PstI fragment (Fig. 1B) carrying the 5' part of R. capsulatus moeA was cloned into the polylinker of the broad-host-range lacZ fusion vector pPHU235 (10). The resulting replicative reporter plasmid was designated pWKR513.

-Galactosidase assays. To determine -galactosidase activities of R. capsulatus strains carrying the translational moeA-lacZ reporter plasmid pWKR513, the corresponding strains were grown in RCV medium supplemented with tetracycline (0.25 µg/ml). Ammonium (final concentration, 10 mM), serine (10 mM), hypoxanthine (1 mM), and molybdate (0 to 1 mM) were added to the growth medium as described in the text. Following growth in the respective media to late exponential phase, -galactosidase activities of R. capsulatus strains were determined by the sodium dodecyl sulfate-chloroform method (10, 25).

Enzyme assays. XDH activity was assayed as described by Leimkühler et al. (20), with NAD as the electron acceptor. Nitrate reductase activity was analyzed by the method described by Garrett and Nason (5), and DMSO reductase activity was measured as described by McEwan et al. (24), with dithionite-reduced methyl viologen as the electron donor.

Enzyme purification. To purify XDH from R. capsulatus KS36 and R507, a plasmid overexpressing the xdhABC genes was constructed. To uncouple xdh expression from its native promoter, the nifH promoter region derived from pNF3 (28) was cloned in front of xdhABC. In order to link the ATG start codons of nifH and xdhA, an NdeI restriction site was introduced into the ATG codon of xdhA by PCR mutagenesis. Subsequently, a SacI-HindIII fragment carrying the structural genes of XDH (xdhABC) expressed from the nifH promoter was cloned into the polylinker of the broad-host-range plasmid pPHU231 (31). The resulting hybrid plasmid was designated pSL157. This plasmid was introduced into the R. capsulatus moeA mutant strain R507 and into the corresponding parental strain KS36. Ten liters of the strains was grown under anaerobic, phototrophic conditions in 2-liter bottles containing RCV medium with 10 mM serine as the nitrogen source to induce the nifH promoter. Cells were harvested by centrifugation at late log phase after 40 h of growth, when the A₆₆₀ reached 0.8. Cells were washed with a buffer containing 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, and 2.5 mM dithiothreitol (buffer 1). Unless otherwise stated, all purification steps were carried out at 4°C and in buffer 1. Cell lysis was achieved by repeated passages through a French pressure cell (9,000 lb/in²); the suspension was centrifuged at 16,000 × g for 30 min, and the supernatant was used as crude extract for enzyme purification. The crude extract was centrifuged at 100,000 × g in an ultracentrifuge for 1 h. The resulting supernatant was brought to 45% ammonium sulfate saturation at 4°C, and the suspension was centrifuged (16,000 × g, 30 min). The pellet was dissolved in the minimal volume of buffer 1 and dialyzed against the same buffer to remove ammonium sulfate. After dialysis, the ammonium sulfate fraction was electrophoresed by means of the model 491 Prep Cell system from Bio-Rad in a gel column (diameter, 28 mm) consisting of 4 ml of stacking gel (4% acrylamide) and 10 ml of separating gel (4% acrylamide). Electrophoresis was carried out at 12 W of constant power at 4°C. Fractions were collected in buffer 1, and those containing XDH were pooled and concentrated by ultrafiltration under vacuum. With this method, inactive XDH from R. capsulatus moeA mutants as well as active XDH from the parental strain KS36 or moeA mutants grown in the presence of 1 mM molybdate could be purified with a yield of approximately 700 µg. Protein concentrations were determined as described by Smith et al. (40).

Gel electrophoresis. Analytical polyacrylamide gel electrophoresis was carried out in a discontinuous gel system (19). For nondenaturating polyacrylamide gel electrophoresis, 4% acrylamide stacking gels and 6% acrylamide separating gels were used.

Moco analysis. To analyze Moco present in XDH, the purified protein was subjected to a procedure which converts the Moco to its oxidized fluorescent degradation product (form A) after boiling the enzyme at pH 2.5 in the presence of iodine, as originally described by Johnson et al. (12). After treatment with alkaline phosphatase, dephospho-form A was further purified and desalted on 0.5-ml QAE-Sephadex columns. The column was washed with 10 ml of water. Dephospho-form A was subsequently eluted with 10 mM acetic acid and immediately analyzed by high-performance liquid chromatography (HPLC). To obtain form A-GMP from R. capsulatus crude extracts, iodine treatment was carried out at room temperature and form A-GMP was eluted from the QAE-Sephadex column with 50 mM HCl by the method described by Joshi and Rajagopalan (14). HPLC was performed at room temperature with a Perkin-Elmer C₁₈ reverse-phase column as described by Schwarz et al. (38). Comparable amounts of proteins were used for these analyses. The presence of form A in the respective fractions was verified by its absorption spectrum with a maximum in absorbance at 380 nm.

Nucleotide sequence accession number. The nucleotide sequence of a 2,005-bp ClaI DNA fragment encompassing the moeA coding region has been submitted to the EMBL nucleotide sequence database under accession no. AJ238348.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Cloning and DNA sequence analysis of a Tn5-containing EcoRI fragment from the R. capsulatus mutant strain Xan-17. To identify R. capsulatus genes required for xanthine degradation, a random Tn5 mutagenesis was performed (20). Among 70,000 Tn5-induced R. capsulatus mutants, which were screened for the loss of ability to grow with xanthine as the sole nitrogen source, one mutant strain was observed (Xan-17) in which the xanthine-negative phenotype could be suppressed by high concentrations of molybdate (1 mM) in the medium (20). The phenotype of this mutant strain resembled the phenotype described for E. coli mutants defective in mogA, coding for a molybdochelatase (15). In E. coli mogA mutants, molybdoenzyme activities can be partially restored by growing these mutants at high molybdate concentrations. However, growth of these mutants at nearly toxic levels of molybdate (10 mM) resulted in only modest increases in nitrate reductase (43) and biotin sulfoxide reductase levels (3). In contrast, the ability of the R. capsulatus Tn5 mutant strain Xan-17 to grow with xanthine as the sole nitrogen source was restored to wild-type levels in medium supplemented with 1 mM molybdate (20). To identify the gene inactivated by the Tn5 insertion in R. capsulatus mutant Xan-17, a 7.3-kb Tn5-containing EcoRI fragment was cloned from chromosomal R. capsulatus DNA into pUC8. Figure 1A gives the physical and genetic map of the corresponding R. capsulatus gene region and the location of the Tn5 insertion. DNA sequence analysis of a 2,005-bp ClaI fragment (indicated in Fig. 1A) revealed the presence of one complete open reading frame (ORF) preceded by a typical ribosome binding site. This ORF encoded a putative protein of 414 amino acid residues with a deduced molecular mass of 42,589 Da. Surprisingly, the predicted amino acid sequence of this ORF showed a high degree of identity (35%) to the sequence of E. coli MoeA, a protein which is also involved in Moco biosynthesis. Downstream of R. capsulatus moeA, the 5' end of another ORF transcribed in the opposite direction was identified. This ORF showed a high degree of identity in its predicted 122 C-terminal amino acids to Azorhizobium caulinodans NtrY, a histidine kinase of a two-component regulatory system involved in nitrogen level control (26).

View larger version (13K): [in this window] [in a new window]   FIG. 1.   Physical and genetic maps of the R. capsulatus moeA gene region. (A) The localizations of ORFs are given by arrows carrying their respective gene designations. The vertical arrow indicates the location of the Tn5 insertion in R. capsulatus mutant strain Xan-17. The 2,005-bp ClaI fragment sequenced in this study is marked by a heavy line, and a stem-loop structure located between moeA and ntrY is indicated. Below the map, the locations of interposon insertions are shown. The direction of transcription of interposon resistance genes are symbolized by arrows in boxes (gentamicin resistance gene, white arrow; kanamycin resistance gene, black arrow), indicating polar and nonpolar insertions. The ability of the corresponding R. capsulatus mutant strains to grow with xanthine as the sole nitrogen source is indicated by a plus or minus. (B) Construction of a translational moeA-lacZ fusion. In plasmid pWKR513, an EcoRI-PstI fragment was fused to the reporter gene lacZ located on the broad-host-range vector plasmid pPHU235. Abbreviations: B, BamHI; C, ClaI; E, EcoRI; P, PstI; X, XhoI.

Mutational analysis of the R. capsulatus moeA gene region. To confirm that the molybdate-reparable phenotype of Tn5 mutant strain Xan-17 was indeed linked to the Tn5 insertion and not to secondary mutations, appropriate mutant strains carrying kanamycin interposon insertions (R509I and R509II) were constructed (Fig. 1A). The Xan phenotype of these interposon mutant strains was suppressible by the addition of 1 mM molybdate to the medium, confirming that the Tn5 insertion was indeed responsible for this phenotype (data not shown).

R. capsulatus MoeA is homologous to the E. coli Moco biosynthesis proteins MoeA and MogA and to the eukaryotic proteins Cnx1, Cinnamon, and Gephyrin. The amino acid sequence alignment of R. capsulatus MoeA with E. coli MoeA is shown in Fig. 2A. The two proteins are similar in length (414 and 411 amino acids, respectively) and display 35% identity and 69% similarity over the entire length of the proteins. In addition, weak similarities of both MoeA proteins to E. coli MogA were identified (15% identity, 46% similarity). This alignment demonstrates that MoeA contains an internal MogA-like domain, which corresponds to amino acids 184 to 362 in R. capsulatus MoeA and amino acids 181 to 357 in E. coli MoeA. Another E. coli protein homologous to MogA and MoeA is encoded by moaB. The moaB gene in E. coli is located downstream of moaA, but the specific function of MoaB for Moco biosynthesis is unknown, and no mutants defective in moaB have been described yet (32). Furthermore, a moaB-like gene is absent in other organisms.

View larger version (51K): [in this window] [in a new window]   FIG. 2.   Alignment of the deduced amino acid sequence of R. capsulatus MoeA with E. coli MoeA and MogA. (A) Amino acid sequences were aligned for maximum matching by use of the CLUSTAL W program (44). The C-terminal ends of polypeptides are marked by black dots. Identical amino acids are boxed, and in addition, the similarity of amino acids between E. coli MogA, R. capsulatus MoeA, and E. coli MoeA is emphasized by shading. (B) Schematic overview of different bacterial and eukaryotic proteins showing similarities to E. coli MoeA and MogA. The E. coli MogA protein and MogA-like domains found in other proteins are indicated by dark gray shading. Protein domains showing similarities to E. coli MoeA are indicated by light gray shading.

Molybdoenzyme activities in R. capsulatus moeA mutant strains. As described previously (20), the inability of R. capsulatus moeA mutants to grow with xanthine as the sole nitrogen source was restored by the addition of 1 mM molybdate to the medium. This phenotype resembles the phenotype of E. coli mogA mutants. In contrast, even in the presence of high molybdate concentrations, no molybdoenzyme activity was observed in E. coli moeA mutants (11). This significant difference in the phenotype of moeA mutants from E. coli and R. capsulatus might be explained by the fact that E. coli harbors only MGD-containing molybdoenzymes whereas XDH from R. capsulatus was shown to contain the MPT cofactor (20). To study the role of R. capsulatus MoeA in biosynthesis of the MGD cofactor, the activity of DMSO reductase and nitrate reductase, both enzymes binding the MGD cofactor, was tested in R. capsulatus moeA mutants. To test nitrate reductase activity, plasmid pFR400 carrying the genes encoding the periplasmic nitrate reductase of Rhodobacter sphaeroides (31) was introduced into R. capsulatus. As shown in Table 2, the activities of XDH, DMSO reductase, and nitrate reductase were analyzed in the R. capsulatus moeA mutant strain R507 and compared to those of the corresponding parental strain KS36 (carrying a nifHDK deletion). Enzyme activities were determined as described in Material and Methods after 2 days of growth in medium containing either 1 µM or 1 mM molybdate. The parental strain KS36 contained active XDH regardless of the molybdate concentration, whereas the moeA mutant strain R507 required high molybdate concentrations in the growth medium to produce active XDH. In contrast, the MGD cofactor-containing enzymes nitrate reductase and DMSO reductase remained inactive in the R. capsulatus moeA mutant independent of the molybdenum concentration in the medium. In the parental strain KS36, DMSO reductase activity was 2.5-fold higher when cells were grown with 1 mM instead of 1 µM molybdate. A similar dependence of DMSO reductase activity on the molybdenum concentration of the medium has been observed by Bonnet and McEwan (2a). In conclusion, like E. coli MoeA, R. capsulatus MoeA is absolutely required for the biosynthesis of the MGD cofactor, whereas in contrast, the biosynthesis of the MPT cofactor of XDH was restored by the addition of 1 mM molybdate to the medium in a moeA mutant.

Purification of XDH from moeA mutant strains. To determine the reason for inactive XDH in moeA mutants grown under low molybdate concentrations, XDH was purified from an R. capsulatus strain defective in moeA. In order to purify inactive XDH, a plasmid overexpressing xdhABC in R. capsulatus was constructed (see Materials and Methods). Plasmid pSL157 carries the structural genes for XDH (xdhAB) and xdhC required for XDH activity (21) under the control of the nifH promoter, which is inducible under nitrogen-limiting conditions. Plasmid pSL157 was introduced into the R. capsulatus moeA mutant strain R507 and into the corresponding parental strain KS36. The resulting strains were cultured under anaerobic, phototrophic conditions in RCV minimal medium containing serine as the nitrogen source to induce expression of plasmid-encoded XDH. Inactive as well as active XDH were enriched from cells grown either without the addition of molybdate or with 1 mM molybdate by preparative gel electrophoresis, as shown in Fig. 3 (Materials and Methods). Inactive XDH purified from R. capsulatus R507 grown without the addition of 1 mM molybdate has a reduced electrophoretic mobility compared to active XDH from R. capsulatus KS36 or R507 grown in the presence of 1 mM molybdate. As already described by Leimkühler and Klipp (21), inactive XDH isolated from an xdhC mutant strain, which was shown to be devoid of the MPT cofactor, also revealed a reduced electrophoretic mobility following polyacrylamide gel electrophoresis. Therefore, it could be assumed that inactive XDH from moeA mutants might also be devoid of the MPT cofactor (see below).

View larger version (69K): [in this window] [in a new window]   FIG. 3.   Analysis of active and inactive XDH from R. capsulatus by polyacrylamide gel electrophoresis. Purified XDH was electrophoresed under nondenaturing conditions in 6% polyacrylamide gels and stained for protein with Coomassie brilliant blue. Arrows indicate different electrophoretic mobilities of active (lane 1 and 2) and inactive (lane 3) XDH. Lane 1, 2 µg of active XDH purified from R. capsulatus KS36; lane 2, 2 µg of active XDH purified from R. capsulatus R507 (moeA) grown in the presence of 1 mM molybdate; lane 3, 2 µg of inactive XDH purified from R. capsulatus R507 (moeA) grown without molybdate supplementation.

Fluorescence analysis of MPT derivatives isolated from active and inactive XDH purified from moeA mutant strains. XDH in moeA mutants might be inactive, because (i) XDH lacks the MPT cofactor or (ii) enzyme-bound MPT lacks the molybdenum atom, which both might result in a reduced electrophoretic mobility of XDH. The first possibility indicates that only molybdenum-containing MPT (Mo-MPT) is incorporated into apo-XDH, and the second possibility suggests that apo-XDH would stay in an appropriate "open" conformation until the molybdenum atom is incorporated. To discriminate between these two possibilities, fluorescence derivatives of the MPT cofactor from active and inactive XDH isolated from moeA mutants were analyzed by the method described by Johnson et al. (12). Moco was released from active and inactive XDH by heat treatment followed by a treatment with acidic iodine, which converts MPT to its oxidized fluorescent degradation product, form A. By this method, Mo-MPT as well as MPT lacking molybdenum can be converted into form A. As described in Materials and Methods, dephospho-form A was purified on QAE-Sephadex columns and analyzed by HPLC. As shown in Fig. 4A and B, equal amounts of dephospho-form A were obtained from XDH isolated either from the parental strain KS36 or from the moeA mutant R507 grown in the presence of 1 mM molybdate. This result is consistent with the observation that XDH activity can be restored to approximate wild-type levels in moeA mutants grown in medium containing 1 mM molybdate (Table 2). In contrast, no form A was detected in XDH isolated from the moeA mutant strain R507 grown without molybdate supplementation (Fig. 4C). In accordance with the observed reduced electrophoretic mobility of inactive XDH from this strain, these data indicate that XDH in a moeA mutant does not contain MPT, suggesting that only molybdenum-bound MPT is incorporated into apo-XDH.

View larger version (13K): [in this window] [in a new window]   FIG. 4.   Analysis of fluorescent derivatives of the MPT cofactor of R. capsulatus XDH. HPLC elution profiles of dephospho-form A isolated from Mocos of XDH from different R. capsulatus strains. (A) R. capsulatus KS36 (parental strains); (B) R. capsulatus R507 (moeA), grown in the presence of 1 mM molybdate; and (C) R. capsulatus R507 (moeA), grown without molybdate supplementation. Mocos were converted into form A by oxidation with iodine at 95°C. Dephospho-form A was eluted from a QAE-Sephadex column with 10 mM acetic acid and analyzed by HPLC. Fluorescence was monitored with excitation at 370 nm and emission at 450 nm. Each protein was used for the analysis at a concentration of 0.8 µM.

Purification and cofactor analysis of XDH from an R. capsulatus modA mutant strain grown in the absence of molybdate. To corroborate the hypothesis that only molybdenum-containing MPT is incorporated into XDH, a cofactor analysis of inactive XDH isolated from a strain starved for molybdenum was carried out. R. capsulatus strains containing mutations in the high-affinity molybdate transport system grown on molybdate-free medium were shown to contain an inactive XDH because of the lack of molybdenum in the cell and thus were unable to grow with xanthine as the sole nitrogen source (22a). To purify XDH from a strain containing a mutation in the modA gene (R438I), coding for the periplasmic binding protein of the R. capsulatus high-affinity molybdate transport system (46), the xdhABC-overexpressing plasmid pSL157 was introduced into the mutant strain R438I. The resulting strain was cultured under anaerobic, phototrophic conditions on molybdate-free medium (see Material and Methods) containing serine as the nitrogen source to induce the overexpression of xdhABC genes. XDH was purified by ion-exchange chromatography and preparative gel electrophoresis as described in Materials and Methods. As shown in Fig. 5, inactive XDH isolated from the modA mutant also showed a reduced electrophoretic mobility in native polyacrylamide gels relative to that of active XDH purified from the R. capsulatus wild type.

View larger version (43K): [in this window] [in a new window]   FIG. 5.   Analysis of R. capsulatus XDH isolated from the wild type and a modA mutant strain by polyacrylamide gel electrophoresis. Purified XDH was electrophoresed under nondenaturing conditions on 6% polyacrylamide gels and stained for protein with Coomassie brilliant blue. Arrows indicate different electrophoretic mobilities of active (lane 1) and inactive (lane 2) XDH. Lane 1, 3 µg of active XDH purified from R. capsulatus wild type; lane 2, 5 µg of inactive XDH purified from R. capsulatus R438I (modA), grown in the absence of molybdate.

View larger version (15K): [in this window] [in a new window]   FIG. 6.   Analysis of fluorescent derivatives of the MPT cofactor obtained from XDH of the R. capsulatus wild type and a modA mutant strain. HPLC elution profiles of dephospho-form A isolated from the MPT cofactor of XDH from the R. capsulatus wild type (A) (20) and R. capsulatus R438I (modA), grown in the absence of molybdate (B). The MPT cofactor was converted into form A and analyzed as described in the legend to Fig. 4. Equal amounts of enzyme (0.25 µM) were used for each assay.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

DNA sequence and mutational analyses indicated that the R. capsulatus moeA gene forms a monocistronic operon, and moeA is cotranscribed with neither moeB, as shown for E. coli, nor with any other gene involved in Moco biosynthesis. The R. capsulatus moeA gene product shows high similarities to MoeA of E. coli and other organisms. An internal domain was identified within MoeA, exhibiting similarities to E. coli MogA. This observation indicates a similar function of MoeA and MogA for the biosynthesis of the Moco. In addition, the corresponding proteins in eukaryotes were found to form a fusion protein consisting of a C- or N-terminal MogA domain and a C- or N-terminal MoeA domain (Fig. 2B). The separated MogA domain as well as the separated MoeA domain from the Cnx1 protein of A. thaliana were shown to bind MPT in vitro; however, the MogA domain binds MPT with a higher affinity than the MoeA domain does (38). This observation indicates that the internal MogA-like domain in MoeA might be responsible for MPT binding.

R. capsulatus MoeA seems to fulfill a similar role for the biosynthesis of the MPT cofactor as does MogA for the biosynthesis of the MGD cofactor in E. coli. The phenotype of corresponding mutants can be suppressed by high concentrations of molybdate. However, XDH activity in R. capsulatus moeA mutants was restored to almost wild-type levels by the addition of 1 mM molybdate. In contrast, molybdoenzyme activities in E. coli mogA mutants were only partially restored (10 to 13% of wild-type levels), even at molybdate concentrations of 10 mM (43). R. capsulatus MoeA as well as E. coli MoeA seems to be essential for the biosynthesis of the MGD cofactor, since MGD-containing enzymes were shown to be completely inactive in moeA mutants from both organisms. These data indicate that MoeA is involved in the synthesis of both MPT and MGD, but the pathways might branch at the step of molybdenum insertion. Since a mogA mutant is not yet available in R. capsulatus, it could only be speculated that R. capsulatus MogA might fulfill a role similar to that of E. coli MogA, and thus only the biosynthesis of MGD-containing enzymes might be influenced in an R. capsulatus mogA mutant whereas MPT-containing XDH might not be hurt.

Analysis of fluorescent derivatives of the MPT cofactor in XDH revealed that inactive XDH isolated from moeA mutants is devoid of MPT, although enzyme activity could be restored to wild-type levels when these mutants were grown on high molybdate concentrations. Therefore, the spontaneous insertion of molybdenum into the cofactor, without the help of MoeA, seems to occur prior to cofactor insertion into XDH. Additionally, in XDH purified from a modA mutant grown in the absence of molybdate, only minor amounts of MPT were found. These data support the observation that molybdenum chelation to MPT, with or without the help of MoeA, precedes cofactor insertion into XDH. Taking into account all results available to date, the putative pathways of MPT and MGD cofactor biosynthesis in R. capsulatus are summarized in the model shown in Fig. 7.

View larger version (30K): [in this window] [in a new window]   FIG. 7.   Model for Moco biosynthesis in R. capsulatus. Moco biosynthesis in R. capsulatus is predicted to proceed in a manner similar to that of Moco biosynthesis in E. coli. The molybdopterin molecule, still lacking molybdenum, is synthesized by the gene products of moaABC, moaDE, and moeB (30). For the biosynthesis of the MPT cofactor, MoeA is involved in molybdenum insertion into MPT. The role of MogA for MPT biosynthesis in R. capsulatus is yet unknown. XdhC is required to insert the MPT cofactor into XDH (21). In contrast, the biosynthesis of the MGD cofactor requires additional proteins. MogA is probably involved in molybdenum insertion, whereas MoeA is essential for MGD biosynthesis; the addition of the guanosine dinucleotide to MPT is catalyzed by MobA, the MGD synthase in R. capsulatus (22).

In comparison to Moco biosynthesis in E. coli (reviewed in reference 30), Moco biosynthesis in R. capsulatus probably proceeds in a similar manner, involving a series of reactions catalyzed by the moa, mod, and moe gene products (Fig. 7). At the stage when MPT is still lacking the molybdenum atom, biosynthesis of the MPT cofactor for XDH and the biosynthesis of the MGD cofactor for DMSO reductase and nitrate reductase seem to split in R. capsulatus. For the biosynthesis of the MPT cofactor, molybdenum is incorporated into MPT by the MoeA protein. However, MoeA is not essential for molybdenum insertion, because the phenotype of moeA mutants can be suppressed by high concentrations of molybdate. The role of the MogA protein in the biosynthesis of the MPT cofactor has not been examined in R. capsulatus, because a mogA mutant is not yet available. Only the molybdenum-containing MPT cofactor is incorporated into apo-XDH, which is already assembled as an ₂₂ tetramer containing flavin adenine dinucleotide and 2[Fe-2S] clusters (21). Additionally, the XdhC protein is required for the insertion of the MPT cofactor into XDH (Fig. 7) (21).

In contrast to the MPT cofactor biosynthesis, the biosynthesis of the MGD cofactor requires additional proteins. Assuming a situation comparable to E. coli, a putative R. capsulatus MogA protein might be involved in molybdenum insertion into the MGD cofactor. The attachment of the guanosine moiety to MPT requires the MobA protein, an MPT guanine dinucleotide synthase, which has recently been identified in R. capsulatus (22). In contrast to the biosynthesis of the MPT cofactor for XDH, MoeA is absolutely essential for the biosynthesis of the MGD cofactor. Additionally, the analysis of fluorescent derivatives present in crude extracts of R. capsulatus moeA mutants revealed that these mutants are able to synthesize MPT but not MGD (22a). Therefore, it is possible that the nucleotide attachment to MPT occurs only after insertion of molybdenum into the cofactor. This hypothesis is consistent with the data obtained for CO dehydrogenase from Hydrogenophaga pseudoflava, for which the biosynthesis of MPT was independent of molybdenum, but molybdenum was strictly required for the conversion of MPT to molybdopterin cytosine dinucleotide (MCD) (7). Additionally, Hänzelmann and Meyer (7) reported that only in the presence of molybdenum was Mo-MCD inserted into CO dehydrogenase. Comparable data were obtained for MGD insertion into DMSO reductase and nitrate reductase in E. coli for which it was shown that both metal chelation and nucleotide addition preceded cofactor insertion (33, 34). Therefore, it seems to be a general mechanism for all molybdoenzymes that Moco insertion into the target enzyme occurs only after molybdenum chelation to the cofactor.

In conclusion, the current model of MPT and MGD cofactor biosynthesis outlined in Fig. 7 indicates that the pathways of these two kinds of molybdenum cofactors divide in R. capsulatus at the stage of molybdopterin. The insertion of molybdenum and in the case of MGD cofactors the addition of the guanine nucleotide occur later and separately. However, for MPT-containing enzymes as well as for enzymes harboring MGD cofactors, the insertion of the respective cofactor into the appropriate target enzyme seems to proceed only when Moco biosynthesis is completed.