Acetylglutamate Kinase-Acetylglutamyl-phosphate Reductase Complex of Neurospora crassa EVIDENCE FOR TWO POLYPEPTIDES*

Mutations at the arg-6 locus in Neurospora crmsa are divided into two complementation groups (A and B) and a third noncomplementing group. There are many suppressible nonsense mutations among mutants in complementation group B and one in the noncomple- menting group; no nonsense mutations exist among mutants in complementation group the are of are separable an a of a a that the

between the three groups; suppressible mutations are found only in group B and the noncomplementing group (1,3). The latter's suppressibility suggests that a double deficiency can arise by virtue of a nonsense mutation and that suppression restores function to the domain in which the mutation lies and also relieves a translational polarity effect on the second function. The complementation and suppression data suggested that the arg-6 locus encodes a bifunctional protein. Similar findings have been reported for the argB-argC gene cluster of Saccharomyces cereuisiae (5, 6). However, subsequent characterization of the argB-argC gene products showed that the two activities were separable by either of two different purification steps.
Enzymological analysis of the polypeptide products of the similar N. crassa arg-6 locus was made in order to gain a better understanding of the organization and expression of the locus. Representatives of each of the three complementation groups of the arg-6 locus were screened for acetylglutamate kinase and acetylglutamyl-phosphate reductase activities. In addition, gel filtration and Cibacron dye chromatography were performed to determine whether acetylglutamate kinase and acetylglutamyl-phosphate reductase are separable enzymes, as they are in yeast. If so, these two eukaryotic systems could be used to study the mechanism whereby two proteins are produced from a single mRNA, e.g. mRNA processing, internal translation initiation, or post-translational processing.

EXPERIMENTAL PROCEDURES
Strains, Media, and Chemicals-The strains of N. crasa used in the experiments described here are shown in Table I. Strain LA16 is an arginine bradytroph having an altered ornithine carbamoyltransferase (EC 2.1.3.3.). The strain has a low arginine pool and is derepressed for the enzymes of arginine biosynthesis (7,8). The arginine auxotrophic strains have been previously described (1,3). The medium used for growth was Vogel's minimal medium N (9) containing 1.5% sucrose as the carbon source. Strains were maintained on solid medium containing 2% agar and supplemented with 0.  arg-6 CD174 A through two layers of cheesecloth, washed thoroughly with cold water, and then squeezed to remove excess water. The mycelia were then suspended in three times their wet weight of cold Hepes' extraction buffer (50 mM Hepes, pH 7.4, 1 mM EDTA, 0.2 mM dithiothreitol, 70 mM sucrose, 220 mM mannitol, 0.5 g/liter bovine serum albumin, 1 mM phenylmethylsulfonyl fluoride). The suspension was centrifuged at 3000 X g for 5 min. The supernatant was discarded, and the washed mycelia were used to isolate mitochondria by a modification of the method of Cramer et al. (10).
The wet mycelial pad was added to a mixing chamber (the Bead Beater, Biospec Products) one-third filled with glass beads (0.3-0.5 mm in diameter). The chamber was then filled with Hepes extraction buffer. The chamber was cooled with an ice-water jacket. The mycelia were broken by 3 X 30-5 pulses of mixing, with cooling between pulses to minimize heating. The homogenate was decanted from the glass beads, and mycelial debris was removed by centrifugation at 3000 X g for 5 min, followed by filtration through Miracloth (glass fiber filters may be substituted).
A crude organellar fraction was prepared by centrifugation at 30,000 X g for 20 min. The mitochondria were separated from other organelles on a sucrose step gradient. The gradient was prepared by layering 20 ml of 1.2 M sucrose in Hepes buffer (50 mM Hepes, pH 7.4, 1 mM EDTA, 0.2 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride) on top of 10 ml of 1.6 M sucrose in Hepes buffer. Approximately 6 ml of the organellar fraction, resuspended in 3 volumes of Hepes extraction buffer, was layered on the gradient. It was then centrifuged for 75 min at 70,000 X g in a Beckman SW 27 rotor. The mitochondria banded at the interface between the two layers. Fifteen ml of the top layer was removed with an aspirator and discarded. The band at the interface was collected with a Pasteur pipette. This suspension was diluted 1:3 with Hepes extraction buffer, and the purified mitochondria were collected by centrifugation at 30,000 X g for 20 min.
Chromatography of Mitochondrial Matrix Proteins-Purified mitochondria were suspended in 3 pellet volumes of phosphate buffer (50 mM potassium phosphate, pH 7.5, 0.1 mM EDTA, 0.2 mM dithiothreitol, 10 mM acetylglutamate, 1 mM MgC12, 1 mM phenylmethylsulfonyl fluoride) and sonicated (3 X 10-5 pulses) to lyse the mitochondria. Mitochondrial membrane fragments were removed by centrifugation at 30,000 X g for 40 min. The resulting preparation of mitochondrial matrix proteins was used for all chromatographic separations (see below).
The mitochondrial matrix preparation was fractionated by gel filtration chromatography using Sephadex G-200. The column (2.6 X 65 cm) was equilibrated with phosphate buffer. A sample of the matrix proteins was applied to the column and subsequently eluted with phosphate buffer. Fractions were assayed for acetylglutamate kinase and acetylglutamyl-phosphate reductase.
Blue A Dyematrex Chromatography-A sample of mitochondrial ' The abbreviation used is: Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid. matrix proteins was partially purified by ammonium sulfate fractionation. The fraction precipitating between 30 and 65% saturation was resuspended in start buffer (10 mM potassium phosphate, pH 7.5,0.5 mM EDTA, 0.2 mM dithiothreitol, 5 mM acetylglutamate, 5 mM MgCl,, 1 mM phenylmethylsulfonyl fluoride), and 0.1-ml aliquots were desalted on four Penefsky columns (11). These desalted fractions were pooled and applied to a 9 X 32-mm Blue A Dyematrex column equilibrated with start buffer. Following a 30-min no-flow equilibration period to allow binding, the column was washed with 10 ml of start buffer. Bound proteins were then eluted with 10 ml of start buffer containing 1 mM NADP' followed with 10 ml of 100 mM potassium phosphate, pH 8.0, containing 5 mM acetylglutamate, 0.2 mM dithiothreitol, 10 mM EDTA, 1 mM phenylmethylsulfonyl fluoride. The effluents were assayed for acetylglutamate kinase and acetylglutamyl-phosphate reductase.
Synthesis of N-Acetylglutamate 5-Semialdehyde-N-Acetylglutamate 5-semialdehyde was synthesized enzymatically from acetylornithine and a-ketoglutarate by a method based on that of Vogel and McLellan (12) and modified as described below. The enzyme, acetylornithine transaminase, was isolated from E. coli strain MA13 (argE25, argR63), derepressed for arginine biosynthesis and lacking acetylornithinase.
Cells were grown at 37 "C for 24 h in M9 medium (13) and harvested using a Sharples Model T-1 centrifuge. Cells were resuspended in phosphate buffer (0.1 M potassium phosphate, pH 7.0,0.2 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride) to give approximately 0.15 g of cells/ml. Cells were disrupted in the Bead Beater (see above) by 5 X 1-min pulses. The crude extract was clarified by centrifugation at 20,000 X g for 30 min, and the supernatant was fractionated by ammonium sulfate precipitation. Acetylornithine transaminase was precipitated between 30 and 40% ammonium sulfate. The precipitate was resuspended in phosphate buffer and desalted by passage through a Sephadex G-25 column (2.6 X 44 cm). The pooled fractions containing activity could be frozen and stored at -20 "C or used directly for the enzymatic synthesis of acetylglutamate 5-semialdehyde.
Acetylglutamate 5-semialdehyde was synthesized in a reaction mixture containing 0.14 M a-ketoglutarate, 0.12 M acetylornithine, 0.8 mM pyridoxal phosphate (all in 0.1 M potassium phosphate, pH 7.0 (total volume of 5 ml)), and 5 ml of desalted enzyme. The pH of the reaction mixture was adjusted to 7.0 just prior to the addition of the enzyme. The reaction was carried out for 18-24 h at 30 "C and was stopped by boiling for 3 min. The precipitated protein was removed by centrifugation, and the supernatant was acidified with 0.4 ml of 12 N HCl. One-ml aliquots were applied to AG50W-X8 ion exchange columns (H' form, 0.7 X 12 cm), which were equilibrated with 0.1 N HC1. Acetylglutamate 5-semialdehyde was eluted with 10 ml of H20. Fractions were evaporated and pooled, and the pH was adjusted to 7.0 with 50% NaOH. The yield was 40-50%.
The amount of N-acetylglutamate 5-semialdehyde was determined as follows (14). An aliquot of the sample was diluted to 0.5 ml with water. The acetyl group was removed by boiling for 30 min following the addition of 0.3 ml of 6 N HCl. The reaction mixture was cooled and stopped by the addition of 1.0 ml of 3.6 M sodium acetate. The product, glutamate 5-semialdehyde, was converted to a colored product by reaction with 0.2 ml of 33 mM o-aminobenzaldehyde for 15 min at room temperature. An absorbance of 0.1 at 440 nm corresponds to 86 nmol of glutamate 5-semialdehyde (14). The concentration of the remaining sample was adjusted to the desired concentration (0.01 M) with water and was stored in aliquots at -20 "C.
Enzyme Assays-Acetylglutamate kinase was assayed using a radioactive assay as previously described (2). One unit of enzyme is defined as the amount required to produce 1 pmol of product in 1 min at 30 "C. Acetylglutamyl-phosphate reductase was assayed by following the increase in fluorescence as NADP' was converted to NADPH (12). The reaction mixture contained 0.1 M glycine, pH 9.3, 1.33 mM acetylglutamate 5-semialdehyde, 25 mM potassium phosphate, pH 9.3, 0.67 mM NADP+ (freshly prepared) and 1.0 ml of appropriately diluted enzyme in a final volume of 3.0 ml. The reaction was initiated by the addition of enzyme, and the increase in fluorescence was followed using a Gilson Spectra/Glo filter fluorometer (excitation filter, 330-380 nm, and emission filter, 430-600 nm). The reactions were carried out at 25 "C. The activity is expressed as the change in fluorescence/min.
Determination of Native Molecular Weights-A 7.5 X 300-mm LKB TSK-3000 high pressure liquid chromatography gel filtration column was used to determine the native molecular weights of acetylglutamate kinase and acetylglutamyl-phosphate reductase. The column was equilibrated with running buffer (30 mM potassium phosphate, pH 7.0, 5 mM acetylglutamate, 2 mM MgCl,, 0.2 mM dithiothreitol, 0.5 mM EDTA, 100 mM potassium chloride). The following molecular weight protein standards were used to calibrate the column: thyroglobulin (670,000), ferritin (440,000), catalase (232,000), immunoglobulin G (150,000), lactate dehydrogenase from rabbit muscle (140,000), ovalbumin (45,000), and myoglobin (17,500). The protein standards were 5 mg/ml in running buffer, and 20-30 pg of each standard was applied to the column in separate trials. The flow rate was 1 ml/min, and chromatography was monitored by absorbance at 280 nm. The void volume of the column was determined to be 6.55 ml using a 5pg sample of blue dextran. The molecular weight was plotted versus Kav. The K, was determined using the following formula: K , = (V, where V, is elution volume of the protein, Vo is void volume, and V, is bed volume of the column. Partially purified preparations of acetylglutamate kinase and acetylglutamyl-phosphate reductase were applied to the column in separate trials under conditions described for the standards. One-ml fractions were collected and assayed for activity as described above. Determinution of Subunit Molecular Weights-Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed according to Laemmli (15), using 11 X 14-cm running gels (10% total acrylamide, 0.26% N,N'-methylenebisacrylamide or 12.5% total acrylamide, 0.34% N,N'-methylenebisacrylamide) and 2 X 14-cm stacking gels (3.75% total acrylamide). Partially purified preparations of acetylglutamate kinase and acetylglutamyl-phosphate reductase were mixed with sample buffer to give final concentrations of 1% sodium dodecyl sulfate, 30 mM Tris, pH 6.8, 5% glycerol, 2.5% 2-mercaptoethanol, and 0.0005% bromphenol blue and boiled for 3 min. Molecular weight standards (66,000-30,000) obtained from Sigma were treated in the same manner. The gels were stained with Coomassie Brilliant Blue R or silver-stained according to Goldman et al. (16).

Actiuities in Mutant Strains-
The genetic characterization of the arg-6 locus suggested that the locus encodes both acetylglutamate kinase and acetylglutamyl-phosphate reductase. Representative members of the three complementation groups were assayed for each of these enzymes. As shown in Table 11, members of complementation group A had little or no acetylglutamate kinase activity but contained elevated activities for acetylglutamyl-phosphate reductase. Members belonging to complementation group B had acetylglutamate kinase activity but lacked acetylglutamyl-phosphate reductase. Acetylglutamate kinase activity varied from barely detectable (strains LA59 and LA305) to approximately 50% (LA302) of that observed in the wild-type strain. Mutations belonging to the noncomplementing group had little or no kinase activity and lacked reductase activity.
Separation of the Kinase and Reductase-Since the arg-6 locus encodes both the kinase and the reductase, it was of interest to determine whether these activities reside on a single bifunctional polypeptide or whether they represent separable enzymes as is the case in S. cereuisiue (6). Several different approaches were used to address this question. Mitochondria were freed of vacuoles by extensive purification in order to remove contaminating protease activity. In addition, purification was carried out in the presence of a protease inhibitor. A crude mitochondrial matrix fraction was prepared, and aliquots were subjected to gel filtration chromatography, Dyematrex chromatography, and ammonium sulfate fractionation.
As shown in Fig. 1, gel filtration chromatography on Seph- Enzyme activities in various strains of N. crassa All strains were grown in Vogel's minimal medium supplemented with 0.1 g/liter arginine. Mitochondria were purified from the indicated strains, lysed, and used to assay for the activities of acetylglutamate kinase and acetylglutamyl-phosphate reductase.  Two additional approaches were used in order to further substantiate the conclusion that the two enzyme activities reside on distinct polypeptides. First, the kinase and reductase activities were found to precipitate in distinct but overlapping fractions upon ammonium sulfate fractionation (data not shown). In the second approach, a crude extract was concentrated by ammonium sulfate precipitation and desalted on Penefsky columns, and the enzymes were then separated by specific stepwise elution from a Blue A Dyematrex column as shown in Table 111. Elution with 1 mM NADP+ resulted in recovery of 92% of the reductase activity. Subsequent elution with 10 mM EDTA yielded greater than 100% of the kinase activity applied to the column. This is probably due to inhibition of the kinase activity in the crude extract by residual sulfate, which is subsequently removed during chromatography on the Dyematrex column.
Determination of Molecular Weight-In order to get additional information on the structure of the arg-6-encoded protein(s), we determined the native and subunit molecular weights. Native molecular weights were determined by gel filtration chromatography on TSK-3000. Acetylglutamylphosphate reductase and acetylglutamate kinase activities migrated with molecular weights corresponding to 93,000 and 400,000, respectively (Fig. 2). Partially purified preparations of acetylglutamate kinase and acetylglutamyl-phosphate reductase were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and several bands were evident in both sample lanes (data not shown). Based on several trials using both 10 and 12.5% polyacrylamide gels, the subunit molecular weight of acetylglutamate kinase was determined to be between 49,000 and 52,000. This is consistent with an octameric structure. Acetylglutamyl-phosphate reductase had an apparent subunit molecular weight between 38,000 and 46,000, consistent with a dimeric structure. The purification, further characterization of these enzymes, and determination of the subunit molecular weights of the purified proteins will be presented elsewhere?

DISCUSSION
The arg-6 locus had previously been shown to be genetically complex (1). The absence of kinase and reductase activities in several strains carrying mutations at the arg-6 locus in&cates that it codes for both enzymes (Table 11). The results of the biochemical characterization of 40 mutations at the arg-6 locus are consistent with the expectations based on the genetic data. All the mutants in complementation group A lack acetylglutamate kinase activity, those in group B lack acetylglutamyl-phosphate reductase activity, and those in the noncomplementing group lack both activities.
The arg-6 alleles were also screened for suppressibility by a nonsense suppressor (3). The results, summarized in Table  IV   strong translational polarity was observed in the case of the noncomplementing class mutants, where both of the activities are absent as a result of a single mutation (both are restored in the presence of a nonsense suppressor). These results suggested that the arg-6 locus encodes a bifunctional protein, with acetylglutamyl-phosphate reductase being translationdistal.
Multifunctional loci in which several enzymatic activities are associated with a single polypeptide chain have been reported in several eukaryotic organisms (17,18). One advantage of a bifunctional protein would be that the activities which synthesize and utilize the labile intermediate, acetylglutamyl phosphate, are on the same protein. However, the kinase and reductase activities were readily separable as judged by several criteria (Figs. 1 and 2 and Table 111). This is also true in S. cerevisiue (6). Even the addition to all buffers of a protease inhibitor, phenylmethylsulfonyl fluoride (which established the existence of a pentafunctional protein of the arom complex in N. crmsa (14)), failed to provide any evidence for a large bifunctional protein in the mitochondrial matrix.
Three models of the gene-protein relationship can be proposed based on these results. First, the arg-6 locus in fact comprises two loci with corresponding mRNAs and polypeptides. If this were the case, the noncomplementing group, lacking both activities, would be rare, and none would be expected to be a nonsense mutation. Neither expectation is borne out. Second, the locus is transcribed into a single mRNA transcript which is processed into two, each encoding a separate enzyme. Alternatively, a single mRNA could have two independent initiation sites for protein synthesis, one for each enzyme. These models would not be compatible with the large number of noncomplementing mutants (a nonsense mutant would not be expected to have a polar effect), and polycistronic mRNAs have thus far been found only in prokaryotic organisms. Third, the locus encodes a single mRNA which is translated into a single polypeptide. The latter is then cleaved to yield the two physically separable enzymes. In this model, many mutations of the noncomplementing class would be expected and some of these would be suppressible nonsense mutations. No complementing nonsense mutations would be expected in the translation-proximal region because their polarity would preclude complementation activity for the distal function, Both of the expectations of the one mRNA/one protein model are realized there are many noncomplementing mutants, one a nonsense type, and there are no nonsense mutants in the admittedly small sample of group A mutants. In this model, the group B (reductase-) mutations (among which are many nonsense mutants) are presumed to be translationdistal. The failure to detect a larger molecular weight precursor form carrying both activities is not inconsistent with this model for several reasons. Such a precursor might be inactive and cleavage may be concomitant with import into mitochondria or it might be present in small quantities and/or unstable. There is accumulating evidence that proteins destined for the mitochondrion are synthesized in the cytoplasm usually as larger molecular weight precursors (19, 20) and/or in an inactive form (21). These precursors are post-translationally processed, and subsequently imported following binding to a putative mitochondrial receptor (reviewed in Ref. 22).
The observation that mutants lacking acetylglutamyl-phosphate reductase (complementation group B) had decreased acetylglutamate kinase activity can also be explained by the proposed model. The mutation might interfere with normal processing and/or transport or an incorrectly processed form may be less active. Alternatively, even though the acetylglutamate kinase and acetylglutamyl-phosphate reductase reside on different polypeptides, they are likely to be associated in vivo to ensure efficient utilization of the labile intermediate, acetylglutamyl phosphate. Therefore, in the absence of functional acetylglutamyl-phosphate reductase, acetylglutamate kinase may be less stable or less active.
The determination of native and subunit molecular weights of acetylglutamate kinase and acetylglutamyl-phosphate re-ductase further supports the hypothesis that these activities reside on distinct polypeptides. Acetylglutamate kinase appears to be an octameric protein, while acetylglutamyl-phosphate reductase is a dimer. The proposed model that these proteins arise from a single precursor implies that they are synthesized stoichiometrically. If the kinase and the reductase are associated in uivo, these data suggest that four acetylglutamyl-phosphate reductase dimers are associated with every acetylglutamate kinase octamer. This association, if it exists, is not conserved in vitro in cell extracts, since the activities were never found to co-migrate during chromatography on a variety of resins. Further work is necessary before definitive conclusions about the structure and/or assembly of these proteins in vivo can be made.
A direct test of the proposed model requires that a single precursor polypeptide containing acetylglutamate kinase and acetylglutamyl-phosphate reductase be detected. If this is not found to be the case, then translation of the two enzymes may reveal a novel mechanism in eukaryotic protein synthesis, mRNA metabolism, or protein-protein interaction. The more extensive analysis of the yeast argBC locus has yielded the same picture of the two enzymes' genetic determination and the same hypotheses of gene-enzyme relationship. Whatever mechanism of gene action prevails, it is probably common to the two organisms.