Mutational Analysis of the Gephyrin-Related Molybdenum Cofactor Biosynthetic Gene cnxE From the Lower Eukaryote Aspergillus nidulans

Corresponding author: James R. Kinghorn, University of St. Andrews, St. Andrews, Fife KY16 9TH, United Kingdom., jrk{at}st-andrews.ac.uk (E-mail)

Communicating editor: J. J. LOROS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We report the identification of a number of mutations that result in amino acid replacements (and their phenotypic characterization) in either the MogA-like domain or domains 2 and 3 of the MoeA-like region of the Aspergillus nidulans cnxE gene. These domains are functionally required since mutations that result in amino acid substitutions in any one domain lead to the loss or to a substantial reduction in all three identified molybdoenzyme activities (i.e., nitrate reductase, xanthine dehydrogenase, and nicotinate hydroxylase). Certain cnxE mutants that show partial growth with nitrate as the nitrogen source in contrast do not grow on hypoxanthine or nicotinate. Complementation between mutants carrying lesions in the MogA-like domain or the MoeA-like region, respectively, most likely occurs at the protein level. A homology model of CnxE based on the dimeric structure of E. coli MoeA is presented and the position of inactivating mutations (due to amino acid replacements) in the MoeA-like functional region of the CnxE protein is mapped to this model. Finally, the activity of nicotinate hydroxylase, unlike that of nitrate reductase and xanthine dehydrogenase, is not restored in cnxE mutants grown in the presence of excess molybdate.

THE molybdenum cofactor is an identical structural and functional component of most molybdoenzymes, catalysts that carry out key metabolic reactions necessary for sustaining the sulfur, nitrogen, and carbon cycles in organisms ranging from bacteria to human (RAJAGOPALAN 1996 and references therein). The first evidence for the biosynthetic pathway of this ubiquitous and ancient molecule came from genetic studies of nitrate assimilation in the lower eukaryote Aspergillus nidulans; these studies, by Pateman and Cove, of a number of cnx mutants resulted in the identification of five gene loci (COVE and PATEMAN 1963 ; COVE et al. 1964 ; PATEMAN et al. 1964 ). Mutant cnxE strains were found to be unique among cnx mutants in that they could be repaired for growth on minimal medium with nitrate or purines (such as hypoxanthine) as the sole sources of nitrogen by addition of relatively high concentrations (33 mM) of sodium molybdate (COVE et al. 1964 ; ARST et al. 1970 ). This phenotypic repair has been shown to be due to the partial restoration of nitrate reductase and xanthine dehydrogenase activities, respectively (ARST et al. 1970 ). Partial rectification of enzyme activity led to the proposal that the CnxE protein acted directly on molybdate when present at low concentrations to incorporate it into the cofactor. As the final step in the biosynthetic pathway, its catalytic activity became largely unnecessary in cells grown in the presence of excess molybdate concentrations, probably as a result of mass action.

Pioneering work on the chemical structure of the molybdenum cofactor and the gene/enzyme relationships implicated in the biosynthesis of the molybdenum cofactor, including mogA and moeA cistrons in Escherichia coli (discussed below), was provided by RAJAGOPALAN 1996 and colleagues. On the basis of the illuminating work of Rajagopalan, our A. nidulans studies have suggested the likely involvement of the cnx loci (including cnxE) in molybdenum cofactor biosynthesis (UNKLES et al. 1997 , UNKLES et al. 1999 ; APPLEYARD et al. 1998 ).

KAMDAR et al. 1994 isolated and sequenced the eukaryotic gene Cinnamon from Drosophila melanogaster. Their interesting comparison studies showed that the N-terminal and C-terminal sections of the fruit fly protein are similar in amino acid sequence to MogA and MoeA, respectively, which are single proteins implicated in the later stage of the molybdenum cofactor biosynthetic pathway in E. coli (RAJAGOPALAN 1996 ). Cinnamon is also highly similar in amino acid sequence to the rat protein Gephyrin (KAMDAR et al. 1994 ), which appears to have two functions: one anchoring neurotransmitter receptors to the cytoskeletal structures and the other involving the biosynthesis of the molybdenum cofactor (FENG et al. 1998 ). Recent studies have shown that A. nidulans CnxE is the ortholog of these eukaryotic fused proteins (MILLAR et al. 2001 ). Curiously, in the plant ortholog, Cnx1, the fused MogA-like and MoeA-like domains are in the opposite orientation to the fungal, mammalian, and fruit fly proteins (STALLMEYER et al. 1995 , STALLMEYER et al. 1999 ).

The precise biochemical role of E. coli MogA and MoeA (or the counterpart MogA- and MoeA-like fused proteins in eukaryotes) remains unclear. MogA and its eukaryotic counterpart domain have been implicated in molybdopterin binding while MoeA has also been shown to bind molybdopterin, albeit with a lower affinity (SCHWARZ et al. 1997 ; STALLMEYER et al. 1999 ; KUPER et al. 2000 ). However, no binding of molybdenum (in the form of molybdate) by either domain has been observed. Whatever the exact mechanisms of the later stages of the pathway, it is clear from the Aspergillus cnxE molybdate repairability growth test studies that CnxE (and inter alia, homologs) is needed only to synthesize the molybdenum cofactor when the concentration of molybdate available to the organism is low. Indeed, recent studies by LEIMKUHLER and RAJAGOPALAN 2001 have demonstrated directly, using an in vitro system, that MogA and MoeA reactions are largely dispensable in the presence of higher concentrations of molybdate.

Recent research has solved the crystal structure of several proteins involved in molybdenum cofactor biosynthesis, including MogA (LIU et al. 2000 ; SOLA et al. 2001 ) and MoeA (SCHRAG et al. 2001 ; XIANG et al. 2001 ). Although no definitive biochemical role is obtained from these reports, several important observations have been made. First, while MogA is composed of a single globular domain, MoeA consists of four discrete domains. In this article, therefore, we use the terms MogA-like domain and MoeA-like region (containing the four domains). Second, sequence and structural similarities between MogA and the third domain of MoeA suggest that they bind similar ligands and therefore have similar functions (SCHRAG et al. 2001 ). It has been suggested that the MogA-like domain and MoeA-like region in the eukaryotic protein may form a composite binding region (XIANG et al. 2001 ). High-resolution crystal structures provide the position of the highly conserved ThrThrGlyGlyThrGly motif, which is thought to be involved in molybdopterin binding.

Here, we report the results of a mutation experimental approach to study this eukaryotic complex fusion protein. Although mutants have been isolated previously in several eukaryotes (COVE and PATEMAN 1963 ; COVE et al. 1964 ; PATEMAN et al. 1964 ; KUPER et al. 2000 ; REISS et al. 2001 ), only one eukaryotic amino acid replacement mutant is thus far (i.e., the molybdate-repairable chl-6 mutant of Arabidopsis thaliana; SCHWARZ et al. 2000 ) identified at the nucleotide level with the residue change related to the phenotypic consequences of the resultant mutation within the natural eukaryotic host. We describe in this article studies of the phenotypic consequences of a series of cnxE mutations, identified at the amino acid level and within the natural host, A. nidulans, and the complementation patterns of these mutations within heterokaryons.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

A. nidulans strains and mutant isolation:
Standard wild-type (with regard to nitrogen metabolism) strains used for the isolation of mutants on the basis of chlorate resistance (see below) were (a) G1070, yellow yA2 with no other known markers or (b) G001, a biotin auxotroph, biA1. The puA2 putrescine auxotrophic strain (G071) was used to isolate mutants on the basis of selection for nitrate nonutilization. Standard Aspergillus growth media and handling techniques were as described before (CLUTTERBUCK 1974 ). The generation of mutations was carried out using the chemical mutagen 4-nitroquinoline-1 oxide (BAL et al. 1977 ). After mutagenesis, mutant selection was carried out on the basis of (i) resistance to 150 mM chlorate toxicity with 10 mM proline as the sole nitrogen source or (ii) nitrate nonutilization using the putrescine starvation method (COVE 1976A ,b and references therein). The subsequent characterization and identification of cnxE mutants were by growth tests according to those described by COVE 1976A , COVE 1976B . All cnxE mutants containing the putrescine auxotrophic marker (puA2) were outcrossed to provide cnxE putrescine auxotropic strains for characterization. For growth testing of molybdate repair with nicotinate or hypoxanthine, cnxE mutants were crossed to mutant crnA-niiA-niaD506 (abbreviated in the article to niaD506), a deletion extending through the nitrate assimilation gene cluster and therefore lacking nitrate reductase (as well as nitrite reductase and nitrate transport) activity (TOMSETT and COVE 1979 ). These cnxE niaD506 double-mutant strains were used to circumvent molybdate toxicity effects that occur when these two nitrogen sources are used in molybdate repair growth tests (ARST et al. 1970 ).

Isolation of a cnxE mutant in the mogA-like domain by in vitro mutagenesis:
pMON5, containing the entire cnxE gene in pUC19, was digested with EagI and SpeI. The 6-kb fragment was isolated, blunt-ended with Klenow, and ligated to give p36-130. This removed a 375-bp fragment encoding amino acid residues 36–130, but left the reading frame unchanged. The construct p36-130, linearized with XbaI, was transformed into cnxE13 and transformants selected by growth on nitrate as the sole nitrogen source. DNA was isolated from 17 transformants and subjected to PCR amplification using primers situated on either side of the deletion (5'-GTGCCTGAGGTGTCAAT and 5'-GATGATGATCATGTTGTGAC). All the transformants contained a 781-bp fragment expected for the wild-type cnxE. However, the amplification products of transformant T1 contained, in addition, a fragment of 406 bp expected from p36-130, suggesting that T1 was a heterokaryon. Since the conidia of A. nidulans are uninucleate, single colonies were isolated from a conidial suspension of T1, thus permitting isolation of homokaryotic strains. When tested for the ability to utilize nitrate as the sole nitrogen source, some of these colonies were unable to grow. PCR amplification of DNA from these colonies using the above primers showed that those that grew on nitrate contained the 781-bp fragment alone, while those that did not grow contained only the 406-bp fragment. A 406-bp fragment from colony C3 was sequenced and found to contain the expected deletion of nucleotides encoding residues 36–130, but leaving the CnxE start codon in frame with the rest of the protein. Finally, the entire cnxE coding region in C3 was sequenced following the finding that the only mutation present was the 375-bp deletion. This strain was designated cnxE3.

Mutant DNA sequence determination:
The DNA sequence of mutant cnxE genes was determined by automated sequencing following PCR amplification of genomic DNA as described before (UNKLES et al. 1999 ). The entire coding region was amplified in five overlapping sections using primers E1A and E1B (5'-GTGCCTGAGGTGTCAAT and 5'-AGTGACTTGTGTCGGGT, nucleotide positions -119–481 relative to the A of the start codon as 1; MILLAR et al. 2001 ), E2A and E2B (5'-GTTGAAGGTCACGCCTT and 5'-CAGCATATCCATCAACG,positions 371–987), E3A and E3B (5'-CGAGGTTCCGGTGAATA and 5'-CCCGTAGGCTTTGTTCC, positions 881–1575), E4A and E4B (5'-CTGAAATCATGGGGGAT and 5'-CTTTGGGAGACGCTGTA, positions 1494–2031), and E5A and E5B (5'-CCCTTGTAGCAGTAAC and 5'-GAATAAGTCTTTGGAGG, positions 1942–2336).

Molecular modeling of CnxE:
The homology model was constructed from the coordinates of E. coli MoeA (PDB code 1FC5; SCHRAG et al. 2001 ) using the homology module of INSIGHT (Molecular Simulations, San Diego) based on a multiple sequence alignment of the MoeA homologous domains of CnxE, Cnx1, Cinnamon, and Gephyrin with E. coli MoeA. After assignment of coordinates of equivalent residues, the model was energy minimized in CNS (BRUNGER et al. 1998 ).

HPLC analysis of molybdenum cofactor precursor levels in cell-free extracts:
Shake flask cultures were grown for 12 hr at 30° and 250 rpm in liquid minimal medium (COVE 1966 ) containing 10 mM proline plus 10 mM nitrate as the sole nitrogen sources. Preparation of samples for HPLC was carried out using the following modification of the method reported by JOHNSON and RAJAGOPALAN 1987 . Mycelium (0.8 g), suspended in 3.5 ml of distilled water, was homogenized by sonication and centrifuged for 20 min at 20,000 x g. Cell-free extracts corresponding to 1 mg of protein were mixed with 125 µl of iodine solution (1% I₂, 2% KI, 0.25 M HCl) and water to a final volume of 1.125 ml. After 16 hr at room temperature, 138 µl 1% ascorbic acid and 0.5 ml 0.25 M Tris base were added and samples were cleared using a table-top centrifuge. To the supernatant, 13 µl 1 M MgCl₂ and 1 unit alkaline phosphatase dissolved in 85 µl water were added and incubated for 4 hr at 37° to carry out dephosphorylation. The samples were bound to 0.5 ml of QAE Sephadex (acetate form). After washing the columns with water, form A dephospho was eluted with 5 ml of 10 mM acetic acid and compound Z was eluted with 13 ml of 10 mM HCl. The eluates were stored at -70° for HPLC analysis. The samples were used without pH titration or a concentration step. HPLC buffer for form A dephospho analysis was 7% methanol, 50 mM ammonium acetate, and 5% methanol, 50 mM triethylammonium acetate, pH 7.0 for compound Z. Reverse phase (C18) HPLC of form A dephospho samples (1 ml) and compound Z (0.5 ml) samples was carried out with fluorescence detection set at 350/450 nm and 360/440 nm, respectively. Internal standards of form A dephospho and compound Z added during iodine oxidation showed an extraction efficiency of 95%, which significantly decreased when extracts corresponding to several milligrams of protein were applied to the QAE Sephadex columns. Form A dephospho is derived from all sources of molybdopterin, including the cofactor itself, molybdopterin bound to nonmolybdoenzymes, and free molybdopterin.

Nitrate reductase, xanthine dehydrogenase, and nicotinate hydroxylase activities in mutants and their repair by molybdenum:
Mycelial cells were grown in orbital shake flask culture at 25° according to previous methods (COVE 1966 ; ARST et al. 1970 ). Cultures grown on minimal medium with 5 mM urea as the sole source of nitrogen, with and without sodium molybdate (33 mM), were induced by the addition of 10 mM nitrate (for nitrate reductase activity) or 0.6 mM uric acid (for xanthine dehydrogenase) for 5 hr. For nicotinate hydroxylase assays, cnxE strains were first crossed to a mutant in the structural gene (hxA1) encoding xanthine dehydrogenase (SCAZZOCCHIO 1994 ) to generate cnxE hxA1 double mutants. The hxA1 mutation abolishes possible interfering xanthine dehydrogenase activity (SEALY-LEWIS et al. 1978 ). Cells of the double mutants were grown as before, except that for induction, the mycelium was washed in deionized water and suspended in fresh minimal medium containing 10 mM nicotinic acid as the sole nitrogen source and incubated for a further 5 hr. After induction, mycelia were harvested and stored at -70° until assay.

Cell-free extracts from 0.5 g of cells were prepared by homogenization using sonication with 3 ml of 100 mM phosphate buffer, pH 7.5 (for nitrate reductase) or 100 mM pyrophosphate buffer pH 9.4 (xanthine dehydrogenase and nicotinate hydroxylase) and centrifuged 20 min at 20,000 x g.

Nitrate reductase activity at 25° (nanomoles of nitrite produced per minute per milligram) was determined according to HECK and NINNEMANN 1995 . Xanthine dehydrogenase and nicotinate hydroxylase activities were determined using described procedures (SEALY-LEWIS et al. 1978 ; SCAZZOCCHIO 1994 ). Specific activity (nmol/min/mg) was calculated as the reduction of cytochrome c (from horse heart) using a millimolar extinction coefficient of 29.5. Xanthine dehydrogenase and nicotinate hydroxylase activities were determined once in each of three independent mycelial cultures, while nitrate reductase was determined in triplicate in each of three independent cultures.

Protein content was estimated by the Bradford method (Bio-Rad) with BSA as the standard.

E. coli strains, plasmids, and media:
Standard growth media and handling techniques for E. coli were employed. Standard procedures were used for propagation of plasmids and for subcloning and propagation of plasmids in E. coli strain DH5.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Generation and sequence analyses of in vivo cnxE mutants:
A total of 2000 mutants was isolated on the basis of resistance to chlorate toxicity or the nonnitrate utilization/putrescine starvation method using random chemical mutagenesis. From growth tests, 205 were observed to be cnx mutants, of which 25 were found to be cnxE mutants on the basis of complementation tests in heterokaryons (COVE and PATEMAN 1963 ; COVE et al. 1964 ; PATEMAN et al. 1964 ). Chlorate toxicity screening yielded 20 cnxE mutants while 5 mutants came from the nonnitrate utilization approach. The latter crop of mutants were all found to be chlorate resistant. In addition, 6 mutants (cnxE3, -12, -13, -14, -16, -22) originally isolated by Cove and Pateman were included in this study. Strain cnxE14 did not conform to the phenotype previously described, i.e., sensitivity to chlorate toxicity (ARST et al. 1970 ). Fig 1A summarizes the changes in the CnxE protein in representative cnxE mutants analyzed at the nucleotide level.

View larger version (18K): [in this window] [in a new window] [Download PPT slide]   Figure 1. The position of cnxE mutant changes within the MogA-like domain and MoeA-like region in the CnxE protein and genetic complementation data. (A) The thick line represents the 709 amino acid residues of the CnxE protein (MILLAR et al. 2001 ). The thin lines above show the extent of the MogA-like domain and the MoeA-like region while the dashed line indicates the position of the deletion in strain cnxE3. Arrows show the position of mutations with the amino acid change given in single letter code. Stops are indicated by an asterisk. (B) Results of pairwise mutant complementation tests in heterokaryons. +, vigorous growth of heterokaryons; -, no detectable growth. Hypoxanthine (5 mM) was chosen as the nitrogen source since individual cnxE mutants are completely unable to grow on this nitrogen source. Mutants (28) were divided into six complementation classes, A–F, as indicated in the rightmost column. (C) Establishment of a complementation map derived from testing pairwise heterokaryon combinations (FINCHAM 1966 , FINCHAM 1994 ).

One amino acid replacement, in mutant cnxE22, was found in the MogA-like domain of the CnxE protein (Fig 1), resulting in Gly79 being replaced with Asp. Nine mutations were found to result in amino acid substitutions within the MoeA-like region (Fig 1 and Fig 2): cnxE12 (Gly343 to Asp) and cnxE849 (Ala372 to Ser) within domain 2 and cnxE107, cnxE601, and cnx140 (all Gly434 to Asp), cnxE230 (Gly439 to Cys), cnxE160 (Gly515 to Arg), cnxE16 (Asp522 to Glu), and cnxE340 (Gly546 to Glu), all within domain 3. Four single-base-pair substitution mutations, which resulted in stop codons, were used in complementation studies: cnxE170 at residue 51, cnxE108 at 364, cnxE103 at 471, and cnxE250 at 663 (Fig 1). An additional 13 mutations resulted in stop codons: cnxE104 at residue 4, cnxE360 at 62, cnxE106 at 149, cnxE116 at 183, cnxE101 at 185, cnxE109 at 187, cnxE170 at 256, cnxE14 at 429, cnxE3 at 502, cnxE150 at 520, cnxE13 at 542, cnxE4 at 552, and cnxE190 at 598. Finally, only one substantial deletion was obtained with this mutagen: cnxE505 with a 10-bp deletion starting at residue 606. Surprisingly, three mutants appeared to be unchanged at the DNA level and are being examined further.

View larger version (71K): [in this window] [in a new window] [Download PPT slide]   Figure 2. Altered residues in the MoeA-like region of the A. nidulans CnxE protein. Mutant designations are shown beside the vertical lines indicating the alteration. Domains 2 and 3 of the E. coli MoeA protein (second line) are indicated in boldface and italic type, respectively. The amino acid sequences of eukaryotic orthologs, A. thaliana Cnx1, D. melanogaster Cinnamon, and Rattus norvegicus Gephyrin, are shown for comparison. Numbers to the right refer to residues in the respective proteins. Conserved residues are indicated below by an asterisk.

Isolation of a cnxE mutant in the MogA-like domain by in vitro mutagenesis:
Since only one in vivo mutation (i.e., cnxE22) has been found to reside in the MogA-like domain coding region (Fig 1) and since cnxE22 is not a complete loss-of-function mutation (see below), an in-frame deletion mutant was generated (see MATERIALS AND METHODS). This mutation, designated cnxE3, results in deletion of a section of the MogA-like domain from amino acid residues 36–130 to create a mutant devoid of MogA-like function.

Genetic complementation relationships:
Twenty-eight mutants were analyzed to identify complementation patterns. Mutants were co-inoculated in all pairwise combinations to form heterokaryons on hypoxanthine as the sole nitrogen source. From the results of complementation that restored growth approximating wild type on hypoxanthine (Fig 1), six complementation classes, A–F, were derived (Fig 1C). Mutants cnxE22 and cnxE3 belong to complementation classes A and D, respectively (encompassing the MogA-like domain; Fig 1A) whereas cnxE16, cnxE160, cnxE107, cnxE230, and cnxE340 form complementation class C (domain 3 of the MoeA-like region of CnxE; Fig 2). Mutant cnxE12, representing complementation class B, is located within domain 2 of the MoeA-like region (Fig 2). The largest complementation class, F, consists, not unexpectedly, of most of the chain termination mutants, including cnxE170 and cnxE108 (shown in Fig 1A), which did not show complementation with any other classes. Surprisingly perhaps, cnxE849, an amino acid substitution residing within domain 2 of the MoeA-like region, is included in class F. Finally, class E, as exemplified by the chain termination mutants cnxE250 and cnxE103, differs from class F in that the two class E mutants complement class A mutants (Fig 1B and Fig C).

Phenotype of mutants:
All cnxE mutants failed to grow on nitrate as the sole source of nitrogen (reflecting the absence of nitrate reductase activity) with the exception of mutants cnxE12 and cnxE16, which showed significant growth on nitrate, and are described below as "leaky" (Fig 3A, -Mo). No growth of mutants, including cnxE12 and cnxE16, was observed with hypoxanthine, indicating the loss of xanthine dehydrogenase activity (Fig 3B, -Mo). Testing on growth regimes that reflect the level of nicotinate hydroxylase activity with (i) nicotinate as the sole nitrogen (Fig 3C, -Mo) or (ii) hypoxanthine, allopurinol, nicotinamide containing minimal media (J. R. KINGHORN, unpublished results) revealed no growth on either medium by any of the cnx mutants, including cnxE12 and cnxE16.

View larger version (65K): [in this window] [in a new window] [Download PPT slide]   Figure 3. Growth and chlorate toxicity tests of cnxE mutants. Growth is shown in the presence (+Mo) or absence (-Mo) of 33 mM sodium molybdate on minimal medium containing the nitrogen sources as follows: (A) 10 mM nitrate: the mutants had no other markers associated with nitrogen metabolism other than the cnxE allele indicated. Mutants cnxE12 and cnxE16 show intermediate levels of growth between the wild type and a loss-of-function mutant such as cnxE3 (the narrow line of growth between cnxE12 and cnxE16 is due to heterokaryon formation and genetic complementation of the two alleles). (B) 5 mM hypoxanthine or (C) 10 mM nicotinate as the sole source of utilizable nitrogen: for growth tests on these nitrogen sources, all cnxE mutants were in a niaD506 genetic background (see MATERIALS AND METHODS). The niaD506 mutant grows as wild type on these nitrogen sources (TOMSETT and COVE 1979 ; J. R. KINGHORN, unpublished results). Protection against molybdate toxicity was afforded by this strain combination (i.e., in cnxE niaD506 double mutants) together with 10 mM nitrate being added to hypoxanthine or nicotinate minimal medium. For further details of nitrate protection against molybdate toxicity, see ARST et al. 1970 . (D) Resistance to toxicity of 150 mM chlorate with 10 mM proline as the sole source of nitrogen.

All cnxE single mutants grew substantially better on nitrate when molybdate was present in the medium over a range of concentrations (1, 10, 20, and 33 mM; Fig 3A, +Mo). Restoration of growth was observed, but to a lesser extent relative to nitrate, on hypoxanthine with molybdate supplementation up to 10 mM. Above this concentration, wild-type growth is reduced due to molybdate toxicity as previously observed (ARST et al. 1970 ). To overcome this problem of toxicity exhibited at higher molybdenum concentrations, mutants in domain cnxE22, -3, -12, -16, or -849 were each combined with the deletion allele niaD506 (lacking nitrate reductase activity; see MATERIALS AND METHODS) to yield cnxE niaD506 double mutants, which no longer are subject to molybdate toxicity due to protection afforded by nitrate added to the medium but not available for use as a nitrogen source by the niaD506 mutant (ARST et al. 1970 ). cnxE niaD506 double mutants do not show growth on hypoxanthine/nitrate (in molybdate unsupplemented minimal media) due to the lack of xanthine dehydrogenase activity (J. R. KINGHORN, unpublished results). However, substantial restoration of growth of all five cnxE niaD506 double mutants examined was observed on hypoxanthine/nitrate supplemented with excess (20 or 33 mM) molybdate concentrations (Fig 3B, +Mo). In contrast, no repair of the ability to grow on nicotinate with molybdate supplementation was exhibited by any of the single cnxE mutants (J. R. KINGHORN, unpublished results). This lack of phenotypic repair was confirmed when it was observed that cnxE22, -3, -12, -16 , and -849 double mutants (again each cnxE mutation being in combination with niaD506) failed to grow on nicotinate/nitrate minimal medium containing molybdate concentrations of up to 33 mM (Fig 3C, +Mo).

Finally, all cnxE mutants isolated in this and previous studies (COVE and PATEMAN 1963 ; PATEMAN et al. 1964 ; ARST et al. 1970 ), such as cnxE3 and cnxE849 (Fig 3D), are highly resistance to chlorate toxicity, with the exception of mutants cnxE12 and cnxE16, which are sensitive to chlorate toxicity with the three sole nitrogen sources tested (i.e., arginine, proline, or uric acid) and cnxE22, which showed a low/intermediate level of resistance.

Nitrate reductase, xanthine dehydrogenase, and nicotinate hydroxylase activities in mutants and their repair by molybdenum:
cnxE strains, mutant in the MogA-like domain (cnxE22, cnxE3), domain 2 (cnxE12), or domain 3 (cnxE16, -849, -230) of the MoeA-like region, were examined for the restoration of enzyme activities in cells supplemented with 33 mM sodium molybdate (Table 1). All mutant strains examined showed a similar increase in nitrate reductase or xanthine dehydrogenase on supplementation. In contrast, no increase in nicotinate hydroxylase activity was observed in any mutant cells grown in the presence of high molybdate concentrations. It is noteworthy that the values obtained for the wild type in cells grown in the presence of molybdate are lower for nitrate reductase and nicotinate hydroxylase. This is probably connected with the fact that high concentrations of molybdate reduce the growth rate (see above), although it is not clear to us why this is not also the case for xanthine dehydrogenase.

HPLC analysis of molybdenum cofactor precursor Z and molybdopterin levels in cell-free Aspergillus extracts:
All cnxE mutants assayed appeared to synthesize similar levels of molybdopterin as the wild type in cells growing on nitrate/proline as the sole nitrogen sources (Table 2). In contrast, precursor Z concentrations were found to be significantly higher than wild-type concentrations in the cnxE mutants examined.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The MogA-like domain and the MoeA-like region of CnxE are both required for activity:
A number of randomly generated mutations within the cnxE gene result in amino acid replacements in the MogA-like domain as well as in domains 2 and 3 (identified from the crystal structure; see below) of the MoeA-like region. No discernible phenotypic differences could be observed between mutants in either the MogA-like domain or in domains 2 or 3 of the MoeA-like region of CnxE. First, loss-of-function mutants in either MogA-like domain or domains 2 or 3 of the MoeA-like region lead to the complete abolition of all three A. nidulans molybdoenzyme activities (i.e., nitrate reductase, xanthine dehydrogenase, and nicotinate hydroxylase), resulting in the inability to use nitrate, hypoxanthine, or nicotinate, respectively, as sole sources of nitrogen. Second, nitrate reductase and hypoxanthine dehydrogenase activities of all mutants are repaired to similar levels by molybdate supplementation (as judged by growth tests or direct assay), but not nicotinate hydroxylase (see below). Third, pathway intermediate levels in all the mutants are higher (precursor Z) than or similar to those of wild type (molybdopterin). The MogA-like domain and domain 3 of the MoeA-like region are structurally similar (SCHRAG et al. 2001 ; XIANG et al. 2001 ) and so they may have similar roles in molybdenum cofactor synthesis and processing. Nevertheless, the characteristics of the mutants indicate that the fused MogA-like domain and MoeA-like region of the eukaryotic protein are both required for CnxE function.

Genetic complementation occurs between mutants in different domains of CnxE:
Complementation tests were carried out between pairwise combinations of mutants. The results support the independent functional nature of the MogA-like domain and MoeA-like region of CnxE since mutants in the MogA-like domain, i.e., cnxE22 and cnxE3, clearly complement mutants (cnx-E16, -160, -230, -340) within MoeA-like domain 3, which is similar in sequence and structure to the MogA domain. Also, complementation provides evidence that the CnxE protein is multimeric in structure, in agreement with the physical data obtained from crystallography studies of the trimeric MogA-like N-terminal domain gephyrin (SOLA et al. 2001 ). Moreover, mutant cnxE3 (in which most of the MogA-like domain is deleted) will complement cnxE16 (or cnxE160, -230, -340), suggesting that the MoeA-like domain can function when expressed with the MogA-like domain in trans. This would indicate that the complete, although mutant, CnxE16 (or CnxE160, CnxE230, CnxE340) polypeptide can aggregate with the truncated CnxE3 peptide to yield a functional hybrid CnxE protein, consisting of both MogA- and MoeA-like partial activities.

A structural model for CnxE and the position of mutations therein:
The availability of crystal structures for orthologs MogA (LIU et al. 2000 ; SOLA et al. 2001 ) and MoeA (SCHRAG et al. 2001 ; XIANG et al. 2001 ) has aided our interpretation of the effects of mutations in CnxE. Only one amino acid substitution mutation affecting the MogA-like domain of CnxE was observed. Gly79 is the last Gly in the conserved GlyGlyThrGly motif contained within a loop that forms the base of a cavity proposed to be the site of molybdopterin binding (SOLA et al. 2001 ). Replacement of this residue with Asp in mutant cnxE22 most likely disrupts the loop directly affecting the active site conformation. A homology model of CnxE was built from the dimeric structure of E. coli MoeA (PDB code 1FC5; SCHRAG et al. 2001 ). Seven amino acid substitution mutations in the MoeA-like domain (Fig 2) were mapped on the CnxE model (Fig 4). Most of these mutations are replacements of residues containing sidechains for Gly and most are located along the border of the putative active site cleft formed between the two monomers of the dimer, providing supportive genetic evidence for the importance of this cleft in the function of the protein. Additionally, all the amino acid substitution mutations except cnxE849 are located in regions of high sequence conservation among similar proteins from different species.

View larger version (34K): [in this window] [in a new window] [Download PPT slide]   Figure 4. Structural model of the dimer of the CnxE MoeA-like region. Domain 3 of monomer A is shown in cyan, domain 4 in magenta, domain 1 in blue, and domain 2 in red. Domain 2 of monomer B is light green and the remainder of monomer B is in dark green. The orange spheres mark the locations of the mutations. The TTGG signature sequence is marked in red within monomer A domain 3. Insertions in the A. nidulans amino acid sequence relative to that of E. coli were not modeled and so certain loops that should be present in A. nidulans are not shown. One such insertion involves residue 372 (mutation E849). Since this residue was not in the model, the adjacent residue 373 was marked, indicating the approximate location of the mutation. This model was made using MOLSCRIPT (KRAULIS 1991 ) and Raster3D (MERRITT and MURPHY 1994 ; MERRITT and BACON 1997 ).

Four mutations lie within the third domain of the MoeA-like region of the protein as predicted from the model. The mutation in strain cnxE160, resulting in Gly515 being replaced with Arg, involves the Gly residue located in the putative active site. Introduction of the long Arg sidechain at residue 515 disrupts the geometry of the ThrThrGlyGly loop and acidic cluster that are conserved in MoeA and MogA homologs. The backbone conformation observed for Gly515 would be highly strained upon introduction of an Arg residue and some local conformational changes would be likely. The Arg sidechain also disrupts the electrostatic potential in the region of the conserved acidic cluster. In mutant cnxE230, a Cys replacement of Gly439 introduces two extra atoms into the area of the putative ligand-binding site. This residue is adjacent to Gly515 of the conserved ThrThrGlyGly signature motif and the presence of a sidechain probably interferes sterically with ligand binding. The change of Asp to Glu at residue position 522 in mutant strain cnxE16 extends a fourth negative charge into the acidic cluster and may introduce both electrostatic and steric influences on ligand binding. However, in contrast to the other domain 3 mutants (i.e., cnxE107, -160, and -230), mutant cnxE16 possesses sufficient nitrate reductase activity to allow limited growth on nitrate (i.e., leaky), indicating that this conservative replacement permits low levels of the molybdenum cofactor to be synthesized. The cnxE340 mutation results in Gly546 being replaced by Glu and this change introduces a negative charge into a hydrophobic cluster. The backbone conformation of this residue is readily adopted by glycine, but is strained for other residues and will undoubtedly be altered as a result of the mutation. These local conformational changes will alter the shape of the cleft and could influence ligand binding.

The cnxE12 mutation results in change of Gly343, a highly conserved residue located in the putative ligand-binding cleft, to Asp. Unlike the four previously mentioned mutations in domain 3, Gly343 is located in domain 2 and on the opposite side of the putative active site cleft. The replacement of Gly343 by Asp introduces a negative charge that neutralizes the positive charge of Arg482, suggesting both electrostatic and steric influences on ligand binding. However, these effects probably do not cause complete inactivation of the protein, as strain cnxE12 is leaky (similar to mutant cnxE16), showing limited growth with nitrate due to low levels of nitrate reductase activity. Not unexpectedly perhaps, the counterpart mutant of Arabidopsis (chl-6) also possesses considerable levels of nitrate reductase (SCHWARZ et al. 2000 ). The model, moreover, may explain why cnxE12 is the only mutant within the MoeA-like region that can complement other missence mutants within the same region; i.e., it may explain why it appears in a complementation group by itself. This could be because the cnxE12 mutation lies in a different domain from the other mutations and so aggregation of monomers with mutations in different domains could produce an active protein by a mechanism of conformational correction as proposed by FINCHAM 1966 , FINCHAM 1994 . This mechanism proposes that the packing constraints imposed on monomers within a hybrid oligomer can induce the correct conformation of a faulty polypeptide. Alternatively, domains 2 and 3 may have different and distinct functions that could complement in trans within the oligomeric protein. Isolation and analysis of more mutants within these domains, in combination with structural studies of mutant proteins, would be necessary to resolve these possibilities.

Of the final two mutants modeled, Gly434, which is located in a ß-strand in the core of domain 3, is replaced by Asp in cnxE107, introducing a charged sidechain into a hydrophobic environment. This mutation is likely to alter the folding of the protein, at least locally, and may make the protein unstable. The Ala 372-to-Ser change in cnxE849 is located in domain 2, far from the putative active site cleft. There is no clear indication from the modeling as to why this mutation inactivates the protein. However, the fact that this mutant is unable to complement any other cnxE mutant (although complementation with other cnx mutants is normal) suggests that it may affect the overall oligomerization of the protein.

Nicotinate hydroxylase is not repairable by molybdate supplementation:
An interesting feature of the enzyme activity restoration studies is that, unlike nitrate reductase and xanthine dehydrogenase, nicotinate hydroxylase activity is not repaired by molybdate supplementation of the growth medium. The reason for this lack of enzyme restoration is unclear to us. One possibility may be that nicotinate hydroxylase requires for activity a dinucleotide derivative of the molybdenum cofactor such as is commonly found in prokaryotes (RAJAGOPALAN 1996 ). A second possibility is that nicotinate hydroxylase may have a lower affinity for the molybdenum cofactor than xanthine dehydrogenase or nitrate reductase, as suggested before (MACDONALD and COVE 1974 ; ARST 1997 ). Another possible explanation for the lack of nicotinate hydroxylase repairability requires molybdopterin to be chelated with molybdenum before the resulting molybdenum cofactor insertion into molybdoenzymes. Failure to repair nicotinate hydroxylase by molybdate supplementation could be due simply to lack of accessibility or hindrance in this enzyme in the absence of CnxE.

	ACKNOWLEDGMENTS

We thank B. Tomsett and C. Scazzocchio for the strains. L. J. Millar was a recipient of a Biotechnology and Biological Sciences Research Council Postgraduate Studentship. J. R. Kinghorn acknowledges funding from the Biotechnology and Biological Sciences Research Council and a travel award from the Royal Society (London) to visit Australia.

Manuscript received December 7, 2001; Accepted for publication February 26, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

APPLEYARD, M. V., J. SLOAN, G. J. KANA'N, I. S. HECK, and J. R. KINGHORN et al., 1998  The Aspergillus nidulans cnxF gene and its involvement in molybdopterin biosynthesis. Molecular characterization and analysis of in vivo generated mutants. J. Biol. Chem. 273:14869-14876[Abstract/Free Full Text].

ARST, H. N., JR., 1997  Cosying up to MoCo. Microbiology 143:1037.

ARST, H. N., JR., D. W. MACDONALD, and D. J. COVE, 1970  Molybdate metabolism in Aspergillus nidulans. I. Mutations affecting nitrate reductase and xanthine dehydrogenase. Mol. Gen. Genet. 108:129-145[Medline].

BAL, J., E. M. KAJTANIAK, and N. J. PIENIAZEK, 1977  4-Nitroquinolone-1-oxide: a good mutagen for Aspergillus nidulans.. Mutat. Res. 56:153-156.

BRÜNGER, A. T., P. D. ADAMS, G. M. CLORE, W. L. DELANO, and P. GROS et al., 1998  Crystallography and NMR system: a new software suite for macromolecular structure determination. Acta Crystallogr. Sect. D Biol. Crystallogr. 54:905-921[Medline].

CLUTTERBUCK, A. J., 1974 Aspergillus nidulans, pp. 447–510 in Handbook of Genetics, Vol. 1, edited by R. C. KING. Plenum Press, New York.

COVE, D. J., 1966  The induction and repression of nitrate reductase in the fungus Aspergillus nidulans.. Biochem. Biophys. Acta 113:51-56[Medline].

COVE, D. J., 1976a  Chlorate toxicity in Aspergillus nidulans: the selection and characterisation of chlorate resistant mutants. Heredity 36:191-203[Medline].

COVE, D. J., 1976b  Chlorate toxicity in Aspergillus nidulans: studies of mutants altered in nitrate assimilation. Mol. Gen. Genet. 146:147-159[Medline].

COVE, D. J. and J. A. PATEMAN, 1963  Independently segregating loci concerned with nitrate reductase activity in Aspergillus nidulans.. Nature 198:262-263[Medline].

COVE, D. J., J. A. PATEMAN, and B. M. REVER, 1964  Genetic control of nitrate reduction in Aspergillus nidulans.. Heredity 19:529.

FENG, G., H. TINTRUP, J. KIRSCH, M. C. NICHOL, and J. KUHSE et al., 1998  Dual requirement for gephyrin in glycine receptor clustering and molybdoenzyme activity. Science 282:1321-1324[Abstract/Free Full Text].

FINCHAM, J. R. S., 1966 Genetic Complementation. W. A. Benjamin, New York.

FINCHAM, J. R. S., 1994 Genetic Analysis. Blackwell Science, London.

HECK, I. S. and H. NINNEMANN, 1995  Molybdenum cofactor biosynthesis in Neurospora crassa: biochemical characterisation of pleiotropic molybdoenzyme mutants nit-7, nit-8, nit-9A, B and C. Photochem. Photobiol. 61:54-60[Medline].

JOHNSON, M. E. and K. V. RAJAGOPALAN, 1987  Involvement of chlA, E, M and N loci in Escherichia coli molybdopterin biosynthesis. J. Bacteriol. 169:117-125[Abstract/Free Full Text].

KAMDAR, K. P., M. E. SHELTON, and V. FINNERTY, 1994  The Drosophila molybdenum cofactor gene cinnamon is homologous to three Escherichia coli cofactor proteins and to the rat protein Gephyrin. Genetics 137:791-801[Abstract].

KRAULIS, P. J., 1991  MOLSCRIPT: a program to produce both detailed and schematic plots of protein structures. J. Appl. Crystallogr. 24:946-950.

KUPER, J., T. PALMER, R. R. MENDEL, and G. SCHWARZ, 2000  Mutations in the molybdenum cofactor biosynthetic protein Cnx1G from Arabidopsis thaliana define functions for molybdopterin binding, molybdenum insertion, and molybdenum cofactor stabilization. Proc. Natl. Acad. Sci. USA 97:6475-6480[Abstract/Free Full Text].

LEIMKÜHLER, S. and K. V. RAJAGOPALAN, 2001  In vitro incorporation of nascent molybdenum cofactor into human sulfite oxidase. J. Biol. Chem. 276:1837-1844[Abstract/Free Full Text].

LIU, M. T. W., M. M. WUEBBENS, K. V. RAJAGOPALAN, and H. SCHINDELIN, 2000  Crystal structure of the gephyrin-related molybdenum cofactor biosynthesis protein MogA from Escherichia coli.. J. Biol. Chem. 275:1814-1822[Abstract/Free Full Text].

MACDONALD, D. W. and D. J. COVE, 1974  Studies on temperature-sensitive mutants affecting the assimilatory nitrate reductase of Aspergillus nidulans.. Eur. J. Biochem. 47:107-110[Medline].

MERRITT, E. A. and D. B. BACON, 1997  Methods Enzymol. 297:505-524.

MERRITT, E. A. and M. E. P. MURPHY, 1994  Raster3D version 2.0: a program for photorealistic graphics. Acta Crystallogr. Sect. D Biol. Crystallogr. 50:869-873[Medline].

MILLAR, L. J., I. S. HECK, J. SLOAN, G. J. M. KANA'N, and J. R. KINGHORN et al., 2001  Deletion of cnxE encoding the gephyrin-like protein involved in the final stages of molybdenum cofactor biosynthesis in Aspergillus nidulans. Mol. Genet. Genomics 266:445-453[Medline].

PATEMAN, J. A., D. J. COVE, B. M. REVER, and D. B. ROBERTS, 1964  A common cofactor for nitrate reductase and xanthine dehydrogenase which also regulates the synthesis of nitrate reductase. Nature 201:58-60[Medline].

RAJAGOPALAN, K. V., 1996 Biosynthesis of the molybdenum cofactor, pp. 674–679 in Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, edited by F. C. NEIDHART, R. CURTISS, C. A. GROSS, J. L. INGRAHAM, E. C. C. LIN et al. ASM Press, Washington, DC.

REISS, J., S. GROSS-HARDT, E. CHRISTENSEN, P. SCHMIDT, and R. R. MENDEL et al., 2001  A mutation in the gene for the neurotransmitter receptor-clustering protein gephyrin causes a novel form of molybdenum cofactor deficiency. Am. J. Hum. Genet. 68:208-213[Medline].

SCAZZOCCHIO, C., 1994 The purine degradation pathway, genetics, biochemistry and regulation, pp. 221–257 in Aspergillus: 50 Years On, edited by S. D. MARTINELLI and J. R. KINGHORN. Elsevier Press, Amsterdam.

SCHRAG, J. D., W. HUANG, J. SIVARAMAN, C. SMITH, and J. PLAMONDON et al., 2001  The crystal structure of Escherichia coli MoeA, a protein from the molybdopterin synthesis pathway. J. Mol. Biol. 310:419-431[Medline].

SCHWARZ, G., D. H. BOXER, and R. R. MENDEL, 1997  Molybdenum cofactor biosynthesis. The plant protein Cnx1 binds molybdopterin with high affinity. J. Biol. Chem. 272:26811-26814[Abstract/Free Full Text].

SCHWARZ, G., J. SCHULZE, F. BITTNER, T. EILERS, and J. KUPER et al., 2000  The molybdenum cofactor biosynthetic protein Cnx1 complements molybdate-repairable mutants, transfers molybdenum to the metal binding pterin, and is associated with the cytoskeleton. Plant Cell 12:2455-2471[Abstract/Free Full Text].

SEALY-LEWIS, H. M., C. SCAZZOCCHIO, and S. LEE, 1978  A mutation defective in the xanthine alternative pathway of Aspergillus nidulans: its use to investigate the specificity of uaY mediated induction. Mol. Gen. Genet. 164:303-308[Medline].

SOLA, M., M. KNEUSSEL, I. S. HECK, H. BETZ, and W. WEISSENHORN, 2001  X-ray crystal structure of the trimeric N-terminal domain of gephyrin. J. Biol. Chem. 276:25294-25301[Abstract/Free Full Text].

STALLMEYER, B., A. NERLICH, J. SCHIEMANN, H. BRINKMANN, and R. R. MENDEL, 1995  Molybdenum cofactor biosynthesis: the Arabidopsis thaliana cDNA cnx1 encodes a multifunctional two-domain protein homologous to a mammalian neuroprotein, the insect protein Cinnamon and three Escherichia coli proteins. Plant J. 8:101-112.

STALLMEYER, B., G. SCHWARZ, J. SCHULZE, A. NERLICH, and J. REISS et al., 1999  The neurotransmitter receptor-anchoring protein gephyrin reconstitutes molybdenum cofactor biosynthesis in bacteria, plants, and mammalian cells. Proc. Natl. Acad. Sci. USA 96:1333-1338[Abstract/Free Full Text].

TOMSETT, A. B. and D. J. COVE, 1979  Deletion mapping of the niiA niaD gene region of Aspergillus nidulans.. Genet. Res. 34:19-32[Medline].

UNKLES, S. E., J. SMITH, G. J. M. KANA'N, L. J. MILLAR, and I. S. HECK et al., 1997  The Aspergillus nidulans cnxABC locus is a single gene encoding two catalytic domains required for synthesis of precursor Z, an intermediate in molybdenum cofactor biosynthesis. J. Biol. Chem. 272:28381-28389[Abstract/Free Full Text].

UNKLES, S. E., I. S. HECK, M. V. APPLEYARD, and J. R. KINGHORN, 1999  Eukaryotic molybdopterin synthase. Biochemical and molecular studies of Aspergillus nidulans cnxG and cnxH mutants. J. Biol. Chem. 274:19286-19293[Abstract/Free Full Text].

XIANG, S., J. NICHOLS, K. V. RAJAGOPALAN, and H. SCHINDELIN, 2001  The crystal structure of Escherichia coli MoeA and its relationship to the multifunctional protein gephyrin. Structure 9:299-310[Medline].

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Heck, I. S. Articles by Unkles, S. E. Search for Related Content PubMed PubMed Citation Articles by Heck, I. S. Articles by Unkles, S. E.