Nội dung

Eur. J. Biochem. 270, 1014–1024 (2003)
FEBS 2003 doi:10.1046/j.1432-1033.2003.03477.x The N-acetylglutamate synthase/N-acetylglutamate kinase metabolon of Saccharomyces cerevisiae allows co-ordinated feedback regulation of the first two steps in arginine biosynthesis Katia Pauwels, Agnes Abadjieva, Pierre Hilven, Anna Stankiewicz and Marjolaine Crabeel Department of Genetics and Microbiology of the Vrije Universiteit Brussel, Brussels, Belgium In Saccharomyces cerevisiae, which uses the nonlinear pathway of arginine biosynthesis, the first two enzymes, N-acetylglutamate synthase (NAGS) and N-acetylglutamate kinase (NAGK), are controlled by feedback inhibition. We have previously shown that NAGS and NAGK associate in a complex, essential to synthase activity and protein level [Abadjieva, A., Pauwels, K., Hilven, P. & Crabeel, M. (2001) J. Biol. Chem. 276, 42869–42880]. The NAGKs of ascomycetes possess, in addition to the catalytic domain that is shared by all other NAGKs and whose structure has been determined, a C-terminal domain of unknown function and structure. Exploring the role of these two domains in the synthase/kinase interaction, we demonstrate that the ascomycete-specific domain is required to maintain synthase activity and protein level. Previous results had suggested a participation of the third enzyme of the pathway, N-acetylglutamylphosphate reductase, in the metabolon. Here, genetic analyses conducted in yeast at physiological level, or in a heterologous background, clearly demonstrate that the reductase is dispensable for synthase activity and protein level. Most importantly, we show that the arginine feedback regulation of the NAGS and NAGK enzymes is mutually interdependent. First, the kinase becomes less sensitive to arginine feedback inhibition in the absence of the synthase. Second, and as in Neurospora crassa, in a yeast kinase mutant resistant to arginine feedback inhibition, the synthase becomes feedback resistant concomitantly. We conclude that the NAGS/NAGK metabolon promotes the co-ordination of the catalytic activities and feedback regulation of the first two, flux controlling, enzymes of the arginine pathway. De novo arginine biosynthesis in plants and microorganisms occurs in eight biochemical steps starting from glutamate. In the fifth step of this pathway ornithine is generated from N-acetylornithine. Two different ornithine synthesis reactions can be distinguished. In the linear pathway, ornithine is generated through the hydrolysis of N-acetylornithine. In the cyclic pathway, the acetyl group of N-acetylornithine is transferred to glutamate, thereby regenerating N-acetylglutamate (Fig. 1). Because it avoids the acetyl-CoA consuming initial step, catalysed by N-acetylglutamate synthase (NAGS) (EC 2.3.1.1), the cyclic pathway is energetically more favourable. However, an organism, which regenerates N-acetylglutamate through ornithine synthesis, still requires the synthase in order to ensure a constant level of acetylated compounds during cell growth. Therefore an anaplerotic role is attributed to acetylglutamate synthase in organisms using the cyclic pathway of ornithine synthesis [1,2]. The linear pattern of ornithine synthesis is found in Escherichia coli and some other bacteria and archea [1–5]. The cyclic pattern is more widespread among the procaryotes [6–13], and it is observed in all investigated ascomycetes, including Candida utilis [14], Saccharomyces cerevisiae [15], Neurospora crassa [2], and in Chlamydomonas algae [16]. In the fungi, ornithine synthesis proceeds entirely in the mitochondria [17,18]. Control of the metabolic flux through a biosynthetic pathway usually occurs at the level of the first committed step and is often mediated by the end product of the pathway. This classical mechanism operates in organisms using the linear pathway of arginine synthesis: arginine exerts feedback inhibition on N-acetylglutamate synthase in E. coli and Salmonella typhimurium [19–21]. In pathways where acetylglutamate is regenerated, the second enzyme of arginine biosynthesis, N-acetylglutamate kinase (NAGK) (EC 2.7.2.8) becomes the main controlling step. Feedback inhibition of the kinase by arginine has been demonstrated in several bacteria [7,22,23]. Yet, metabolic control should occur on the production of acetylglutamate, regardless of its origin. Therefore, feedback inhibition on both the synthase and the kinase is believed to be general for organisms using Correspondence to M. Crabeel, Department of Genetics and Microbiology of the Vrije Universiteit Brussel, c/o CERIA-COOVI, Emile Gryson avenue 1, B-1070 Brussels, Belgium. Fax: + 32 2 526 72 73, Tel.: + 32 2 526 72 84, E-mail: mcrabeel@vub.ac.be Abbreviations: NAGS, N-acetylglutamate synthase; NAGK, N-acetylglutamate kinase; NAGPR, N-acetylglutamylphosphate reductase; CD, catalytic active domain; ASD, ascomycetes specific domain. Enzymes: N-acetylglutamate synthase (EC 2.3.1.1), N-acetylglutamate kinase (EC 2.7.2.8), N-acetylglutamylphosphate reductase (EC 1.2.1.38). (Received 25 November 2002, revised 14 January 2003, accepted 22 January 2003) Keywords: yeast; N-acetylglutamate synthase; N-acetylglutamate kinase; metabolon; co-ordinated feedback inhibition. 
FEBS 2003 Fig. 1. Simplified scheme of the arginine biosynthesis pathway in S. cerevisiae. Step 1 is catalysed by N-acetylglutamate synthase (synthase), step 2 by N-acetylglutamate kinase (kinase), step 3 by N-acetylglutamylphosphate reductase (reductase), and step 5 by N-acetylornithine-glutamate acetyltransferase (acetyltransferase). Co-ordinated feedback regulation (Eur. J. Biochem. 270) 1015 cyclic ornithine synthesis. The feedback regulation of these first two steps in the arginine pathway has been clearly demonstrated in the bacterium Pseudomonas aeruginosa and in two ascomycetes: S. cerevisiae and N. crassa [24–28]. In the latter two organisms, the control of the first two steps of the arginine pathway includes an extra level of complexity. Beside its own structural gene (ARG2), N-acetylglutamate synthase activity also requires the yeast ARG5,6 gene (arg-6 in N. crassa). The ARG5,6 and arg-6 genes encode each a polyprotein precursor which is maturated in the mitochondrial matrix to N-acetylglutamate kinase and N-acetylglutamylphosphate reductase (NAGPR) (EC 1.2.1.38), catalysing, respectively, the second and third step of arginine biosynthesis [29,30]. This requirement of an extra gene for the synthase activity was first observed in N. crassa, where cells containing some nonsense mutants of the arg-6 gene displayed no detectable synthase activity, despite the presence of an intact synthase encoding gene (arg-14) [28,31]. An interaction between the synthase and the kinase of N. crassa was demonstrated by the yeast two-hybrid system (R. L. Weiss, S. K. Chae, J. Chung, C. McKinstry, M. Karaman and G. Turner, University of California, Los Angeles, CA, USA, personal communication). Similar data in yeast were independently obtained by our group [32]. An increase in synthase activity, expected to result from higher copy numbers of its structural gene ARG2, has only been observed with a parallel increase in the ARG5,6 gene copy number. The yeast synthase/kinase interaction was demonstrated by coimmunoprecipitation methods [32]. The physical participation of reductase, the second maturated gene product of ARG5,6, to the synthase/kinase complex, has not been proven so far. Hence, it is not clear whether synthase activity and protein level require reductase. However, the existence of mutations in the reductase-encoding domain of the N. crassa arg-6 gene, which affect synthase activity, suggests a possible role for the reductase [28,31]. Moreover, increasing the copy-number of a synthetic gene, only encoding the kinase domain of S. cerevisiae ARG5,6 gene, is not sufficient to increase the activity of yeast NAGS when coexpressed with high copy-number of ARG2 [32]. Another remarkable result, concerning the regulation of the first enzymes of the arginine pathway, has been reported by the team of R. L. Weiss. A series of ornithine-overproducing N. crassa mutants [33], were mapped to the N-terminus of N-acetylglutamate kinase and shown to bear F81L modifications. The data suggest that this single amino-acid modification of the kinase might result in the deregulation of the first two enzyme activities of the arginine pathway, leading to the hypothesis of a co-ordinated feedback control (R. L. Weiss, S. K. Chae, J. Chung, C. McKinstry, M. Karaman and G. Turner, University of California, Los Angeles, CA, USA, personal communication). The co-ordinated regulation of the first two enzymes of the arginine pathway in ascomycetes seems to correlate with some particular features of both the synthase and the kinase genes. The ascomycete N-acetylglutamate synthase encoding genes are conserved and appear evolutionarily not related to the gene family encoding N-acetylglutamate synthase in bacteria [32,34]. Recently, the murine and the human genes encoding N-acetylglutamate synthase were 
FEBS 2003 1016 K. Pauwels et al. (Eur. J. Biochem. 270) characterized and shown to pertain to the same family as the ascomycete synthase [35,36]. This apparent dual origin of the synthases is in sharp contrast with the common evolutionary relationship ascribed to all other genes involved in the arginine biosynthesis in different organisms. Amino-acid sequence alignments of known members of the N-acetylglutamate kinase family illustrate conservation over all three domains of living organisms (Bacteria, Archaea, and Eucarya) of a region corresponding to the E. coli NAGK, the only NAGK of known 3D structure [37], representing the catalytic NAGK domain. However, all the ascomycete N-acetylglutamate kinases characterized to date, namely those of S. cerevisiae, Schizosaccharomyces pombe, N. crassa, and Candida albicans, have two specific features: (a) they are encoded together with NAGPR as a bi-functional precursor protein that is processed into two distinct enzymes in the mitochondria, and (b) they possess an extra region of about 200 amino acids at their C-terminus, that we call the ascomycete-specific domain (ASD) [29,30]. It is tempting to speculate that the ascomycete-specific domain (ASD) of the kinase might play a role in formation of the synthase/kinase protein complex. This work investigates three important unsolved questions related to the structure and function of the yeast NAGS/NAGK metabolon. We analyse (a) the role of the reductase in the activity and protein level of the synthase, (b) the role of the ASD of the kinase in its interaction with the synthase, and (c) the significance of the yeast NAGS/ NAGK metabolon in terms of its co-ordinated feedback regulation by arginine. Experimental procedures All yeast strains were grown at 30 C on M.am medium, a minimal medium containing 0.02 M (NH4)2SO4, 3% glucose, vitamins, and trace minerals [41]. Where required, uracil, L-histidine or L-arginine was added to a concentration of 25 lgÆmL)1. Genes which are transcriptionally controlled by the GAL promoter were induced by growing cells on M.gal medium (containing 2% galactose as the carbon source) instead of the usual M.am medium (containing 3% glucose). Arginine starved cells were initially grown on medium containing arginine, centrifuged, washed with water and resuspended in M.am medium without arginine. Cells were starved for 3 h before harvesting them. E. coli. Strains XA4(argA–) and XC33(argC–) from the laboratory of S. Baumberg have been described previously [32]. Rosetta(DE3)(pRARE) is a commercial strain (Novagen) in which the pRARE plasmid over-expresses tRNAs for most rare E. coli codons. E. coli strains were grown at 37 C on rich medium supplemented with ampicillin (25 lgÆmL)1) and chloramphenicol (35 lgÆmL)1) where required. Cell cultures at a D600 of 0.600 were induced by addition of IPTG (2,5 mM) and overnight incubation at 30 C. Culture conditions for the spot tests: approximately 2 mL of cells at D600 of 0.250 grown on rich medium plus ampicilline, were harvested by centrifugation, washed and resuspended in minimal medium to a concentration of 1010 cellsÆmL)1. Drops of 10 lL of 10-fold serial dilutions (from 1010 cellsÆmL)1 to 105 cellsÆmL)1) were spotted on minimal medium with or without arginine (100 lgÆmL)1), and with or without IPTG (1 mM). Sets of four plates were incubated at 37 C, 30 C or 25 C. Strains and growth conditions Oligonucleotides S. cerevisiae. The wild-type strain of this laboratory is S1278b (Mat a). MG471 (Mat a, ura3–471) was directly derived from S1278b by M. Grenson, Universiteit Brussel, Belgium. The strains YeBR5 (Mat a, ura3–471, Darg5::genR), YeBR6 (Mat a, ura3–471, Darg6::genR, arg5–), and 14S31b (Mat a, ura3–, his3–) have been described previously [32]. The construction of strain SS1 (Mat a, ura3–471, Darg3), derived from MG471, has been described [38]. Strain KA44 (Mat a, ura3–, his3–, Darg2:: genR) and strain KA42 (Mat a, ura3–, his3–, Darg5,6::genR) are derived from 14S31b and were constructed using A. Wach’s method [39]: the genomic ARG2 ORF (KA44) or the genomic ARG5,6 ORF (KA42) were replaced by the kanMX4 cassette and the strains were selected on the basis of their geneticin resistance. PCR analysis confirmed the presence of the expected modification in those strains. SA2, derived from MG471, was constructed using the delitto perfetto system developed by Storici et al. [40]. The procedure allowed scarless removal of the NAGK encoding ARG6 region from the chromosomal ARG5,6 gene (deletion from amino acid 84–493 in the ORF encoding the kinase/ reductase precursor). The resulting ura3–, Darg6 mutant strain can be restored to prototrophy by plasmid pYB7, expressing ARG6 from a GAL promoter. This confirms that, as expected, SA2 expresses active NAGPR from the remaining ARG5 region of the ARG5,6 gene. BY4, BY5: [32], HP72 ¼ GTCTCACAACAACAATTGG CTGTGATCAAGGTG. HP73 ¼ CACCTTGATCACA GCTAATTGTTGTTGTGAGAC. HP79 ¼ CACACG ACTTCACAAAATTTTCAACTAATTTGTAACCTCT CCTGATCATAG. HP80 ¼ CTATGATCAGGAGAGG TTACAAATTAGTTGAAAATTTTGTGAAGTCGTG GTG. HP81 ¼ CACTAATTTGTAACCTCTCCTGAT AACCTCTCTTTTTGTGCTGATATTG. HP82 ¼ CAA TATCAGCACAAAAAGAGAGGTTATCAGGAGAG GTTACAAATTAGTG. AA29 ¼ CGTCAGACCATGG GGTGGAGGAGAATATTCGCGCATGAACTCAAG. K1 ¼ GGCCATGGTTTCATCTACTAACGGCTTTT CAG. K2 ¼ GGCCAAGCTTTCAACTACTTGCTGA TGAGTTGAGGGTAG. K4 ¼ GGCCTGCAGCTCAA GGCGCACTCCCGTTCTG. K8 ¼ GGCCTGCAGTCA ATGATGATGATGATGATGTGAAATATTTTTTTCA TTTTCCCAAC. K10 ¼ CCGGAAGCTTTCAGACAC CAATAATTTTATTTTCAGGG. K12 ¼ CCGGAAG CTTGTGAGCGGATAACAATTTCACACAGGAAAC AGACCATGCCTCGTCCCGAGGGAGTTAACACC. Plasmid constructs Table 1 gives an overview of the main features of the plasmids used in this work, including the new constructs. Plasmids pHP17, pHP21, and pHP22 (expressing the 
FEBS 2003 Co-ordinated feedback regulation (Eur. J. Biochem. 270) 1017 Table 1. Main features of the plasmids used in this work. Plasmids Cloning vector Origin of insert Nature of insert Expressed protein pYB2 pYB3 pYX213 (2l, URA3) pYX223 (2l, HIS3) S1278b S1278b PromoterGAL  ARG2 ORF-HAtag (32) PromoterGAL  ARG5,6 ORF (32) pYB7 pYB8 pYX223 (2l, HIS3) pYX223 (2l, HIS3) S1278b S1278b PromoterGAL  ARG6 (32) PromoterGAL  ARG5 (32) pHP17 pHP21 pYX223 (2l, HIS3) pYX223 (2l, HIS3) S1278b S1278b pHP22 pYX223 (2l, HIS3) S1278b p238 S288c pYK1 pYK7 pYK8 YCp50 (ARS-CEN, URA3) pTrc99a pTrc99a pTrc99a PromoterGAL  ARG5,6 ORF (F99L) PromoterGAL  ARG5,6 ORF (Damino acids 355–493) PromoterGAL  ARG5,6 ORF (Damino acids 85–347) GCN4 (4 uORFs untranslated)a WT NAGS-HA WT NAGK + WT NAGPR (amino acids 1–863) WT NAGK (amino acids 1–537) WT NAGPR (amino acids 1–38 + amino acids 494–863) FBR NAGK + WT NAGPR NAGK (DASD) + WT NAGPR pYK11 pTrc99a S1278b a S1278b S1278b S1278b PromoterTrc  PromoterTrc  PromoterTrc  PromoterTrc  PromoterTrc  PromoterTrc  ARG5 operon ARG6 ARG2 ARG2 ARG6 ARG2 ARG6 ORF-HIS6tag ORF-HIS6tag + ORF-HIS6tag + + NAGK (DCD) + WT NAGPR Constitutive expression of Gcn4p WT NAGK (amino acids 58 to 51) WT NAGS-HIS6 WT NAGS-HIS6 + WT NAGK (amino acids 58–513) WT NAGS-HIS6 + WT NAGK (amino acids 58–513) + WT NAGPR (amino acids 531–863) Gift of A. Hinnebusch, National Institute of Child Health and Human Development, Bethesda, MD, USA. ARG5,6 gene altered in its NAGK-encoding ARG6 region) were all constructed by recombinant PCR, using S1278b genomic DNA as template. Two overlapping fragments were generated in a first PCR amplification step, then selfannealed, elongated to duplex DNA, and amplified in a second PCR step using the two external oligonucleotide primers of the two oligonucleotide pairs of the first PCRs. These external primers are designed to add adequate restriction sites for classical cloning in the pYX223 vector (from R&D systems). The latter is a 2 micron-based yeast– E. coli shuttle vector, bearing HIS3 as selection marker, and in which the expression of the inserted genes is put under the control of a GAL promoter. The BY4/HP73 and HP72/ BY5 primer pairs were used to construct pHP17, BY4/ HP79, and HP80/BY5 for pHP21 and BY4/HP81 and HP82/BY5 for pHP22. Plasmids of the pYK series were all derived from the E. coli expression vector pTrc99a (Pharmacia) and contain different insertions, all obtained by PCR amplification. The inserted fragments allow the expression of the ORF under the transcriptional control of the IPTG-inducible strong bacterial trp-lac promoter and under the translational control of an appropriate Shine–Dalgarno sequence. Plasmids pYK1 expresses the ARG6 ORF, cloned as an NcoI–HindIII fragment amplified using K1 and K2 as primers and plasmid pYB3 as a template. Plasmid pYK7 expresses the ARG2 ORF, cloned as a NcoI–PstI fragment (primers AA29 and K8 and pYB2 as a template). With primer AA29, a tag of six histidine codons is fused in frame to the C-terminus of the ARG2 ORF for immunodetection of the enzyme. Plasmid pYK8 was obtained by inserting the ARG6 ORF and its trp-lac promoter (from position )115) as a PstI–HindIII fragment (primers K4/K2, tem- plate pYK1) into plasmid pYK7. The artificial operon of plasmid pYK11 expresses a bi-cistronic ARG5/ARG6 mRNA under the control of the trp-lac promoter and was obtained by inserting a HindIII fragment (primers K10/K12, template pYK3), containing the reductase encoding region, into plasmid pYK8. Primer K12 has a 35-nucleotide 5¢ extension containing a ribosome site and an initiator codon. DNA sequencing The nucleotide sequence of the ARG5,6 gene, cloned in the plasmids pHP17, pHP21, pHP22 and pYK11, was determined. Beside the intended modification or deletion, these constructions, issued from independent PCRamplifications, share additionally the same 15 singlenucleotide differences with respect to the data base sequence. These S1278b-specific differences with respect to the ARG5,6 gene of strain S288c, used to establish the data base, result in only one amino-acid difference: the E803K modification in the region of the gene encoding NAGPR. Enzyme activity assays Acetylglutamate synthase. This enzyme activity was measured by a radioassay using L-[U-14C] glutamate and acetylCoA as substrates, as described previously [32]. Dependent on the experiment, 400 mL to 2 L of yeast cultures at D600  0.4 were required. Extracts were prepared using the French press. For E. coli experiments, cells of 100 mL cultures (induced overnight) were collected and extracts were obtained by ultrasonication. 
FEBS 2003 1018 K. Pauwels et al. (Eur. J. Biochem. 270) Acetylglutamate kinase. The assay used to measure NAGK activity has been described previously [18]. In total yeast extracts, this assay detects two distinct enzymatic reactions [18,26]. As the interfering activity is not inhibited by arginine (in contrast to the full inhibition of NAGK), a blank including 5 mM arginine was used by Jauniaux et al. to subtract the interfering activity [18]. Because we used arginine feedback resistant mutants in this work, we used adapted blanks containing 50 mM arginine. In some experiments, the blanks were reaction mixtures incubated without the substrate acetylglutamate. This explains the presence of a residual activity, resistant to arginine inhibition, in Fig. 5 (about 15% of the initial kinase activity). A kinase activity similar to that residual activity is measured in extracts of strain KA42 bearing a full deletion of the ARG5,6 gene. All NAGS and NAGK activities reported in this work are means of at least three independent experiments. Standard deviations generally did not exceed 15%. Western blots A standard chemiluminescence Western blotting protocol (Roche) was used to analyse the yeast NAGS expressed in E. coli from plasmids pYK7, pYK8, and pYK11. Equal amounts of total proteins of the different crude extracts were separated by SDS/PAGE on 12% gels, and then blotted on an ECL Hybond nitrocellulose membrane (Amersham Pharmacia Biotech) in transfer buffer [25 mM Tris, 192 mM glycine, 20% (v/v) methanol] using a Mini PROTEAN 3 blotting cell (Bio-Rad). Specific primary mouse anti-HIS Ig (Santa Cruz Biotechnology) (0.1 ngÆmL)1) and 40 UÆmL)1 peroxidase-labelled secondary antibody (Roche) were used to detect the tagged synthase protein. Chemiluminescence was monitored by autoradiography. Detection of Haemaglutinin (HA)-tagged NAGS, expressed by the pYB2 plasmid in yeast cells, was as described previously [32]. synthase, the synthase activity was measured in different mutants carrying deletions in relevant parts of the chromosomal ARG5,6 gene. Strain YeBR6 expresses neither the kinase nor the reductase, while only the kinase is expressed by strain YeBR5 [32]. A new strain, SA2 bears a deletion of the kinase-encoding domain of ARG5,6 and has its remaining reductase-encoding domain fused to the mitochondrial targeting peptide. The SS1 strain is used as the ARG5,6+ positive control. SS1 bears an ARG3 deletion rendering the control strain arginine-dependent, like the tested strains. SS1, SA2, YeBR5 and YeBR6 are all directly derived from MG471. In a crude extract of wild-type yeast, the physiological level of synthase activity is barely detectable. The detection becomes even more difficult for the strains requiring arginine for cell growth, presumably due to a tight binding of the feedback inhibitor. Moreover, adequate removal of the inhibiting arginine, by dialysis or repeated gel filtration, is limited by the lack of stability of the synthase. To overcome this difficulty, we choose to assay NAGS in extracts of arginine-starved cells (see strains and growth conditions). The arginine deprivation results in a Gcn4pmediated transcriptional activation of the ARG2 gene (K. Pauwels and M. Crabeel, unpublished results) and reduces the pool of the feedback inhibitor. Even higher levels of synthase activity were detected in strains bearing the p238 plasmid, due to a constitutive production of the Gcn4p transcriptional transactivator (Table 2). Synthase activity was assayed in crude extracts of arginine starved SS1, YeBR5, SA2 and YeBR6, with and without the plasmid p238 (Table 2). No synthase activity was detectable in absence of the kinase (SA2 and YeBR6 vs. SS1). In contrast, the absence of reductase did not affect considerably the synthase activity, though a small decrease was observed (YeBR5 vs. SS1). These data demonstrate that, at physiological level, the synthase activity requires the presence of the kinase, and that the additional presence of the reductase is dispensable. Results Activity and protein level of the yeast synthase expressed in E. coli, require the coexpression of the yeast kinase but not of the yeast reductase At physiological levels, the presence of N-acetylglutamyl phosphate reductase is dispensable to synthase activity In order to determine the influence of N-acetylglutamyl phosphate reductase on the activity of N-acetylglutamate The E. coli strain XA4 (argA–), which is defective in N-acetylglutamate synthase, cannot be restored to arginine prototrophy by a trp-lac-promoter-driven expression of the Table 2. Physiological levels of the N-acetylglutamate synthase in strains bearing different deletions in the ARG5,6 gene. Status of Strain Relevant genotype NAGK NAGPR NAGS activity (nmolÆmin)1Æmg)1 protein) SS1 SS1 (p238)a YeBR5 YeBR5 (p238)a SA2 SA2 (p238)a YeBR6 YeBR6 (p238)a ARG5,6, Darg3 + + Darg5 + – Darg6 – + Darg5,6 – – 2.2 9.36 1.5 7.3 <0.2b <0.2b <0.2b <0.2b a Plasmid p238 expresses Gcn4p constitutively; b below detection. 
FEBS 2003 yeast synthase [32]. In the present study we analysed the influence of the additional expression of the yeast kinase, and of the yeast kinase and reductase together. Three plasmids, derived from pTrc99A, are designed to express yeast synthase (pYK7), yeast synthase and kinase (pYK8) or yeast synthase, kinase and reductase (pYK11). These new constructions, including the empty vector pTrc99a, are transformed in strain XA4. SDS/PAGE/Coomassie Blue analysis and kinase activity assays confirmed that XA4 (pYK8) and XA4(pYK11) are over-expressing functional kinase protein (data not shown). Beside the kinase protein, XA4 (pYK11) expresses the reductase protein, however, in lower amounts. The functionality of the reductase, encoded by pYK11, was verified by complementation of the reductase deficient E. coli strain XC33 (argC–) (data not shown). First, all four plasmids were tested for their efficiency to complement the argA– deficiency of the XA4 strain, using spot tests of serial dilutions incubated at 37 C. Under noninducing conditions (Fig. 2A), pYK8 and pYK11 (both expressing the kinase protein) allow growth of the argininedeficient mutant in the absence of arginine. On the other hand, plasmids pYK7 and the empty vector pTrc99a (both lacking the yeast kinase ORF) were completely unable to complement the mutation. These data demonstrate that the presence of the kinase is essential to yeast synthase activity while the additional presence of the reductase (pYK11 vs. pYK8) does not improve complementation. The observation that complementation is even slightly lower in the presence of the reductase, could be due to a lower copy number of pYK11, which is larger than pYK8. Unexpectedly, expression of pYK8 and pYK11 under induced conditions did not improve the efficiency of complementation of the argA– XA4 strain (Fig. 2C). In contrast, it appeared to be toxic to the cell. This cell toxicity was demonstrated by the severe growth handicap observed when ITPG and arginine were supplemented together to the medium (Fig. 2D vs. 2C). On the other hand, the absence of growth of strain XA4(pYK7) under inducing conditions Co-ordinated feedback regulation (Eur. J. Biochem. 270) 1019 (Fig. 2C) demonstrates the incapacity of the plasmid that bears only the synthase gene to complement the arginine deficiency, rather than revealing cell toxicity. This is shown by a similar behaviour in growth of XA4(pTrc99A) and XA4(pYK7), in all conditions used. Same series of spot test were also realized with plate incubations at 30 C and 25 C. These milder temperatures did not allow any growth of XA4(pYK7) in the absence of arginine, but complementation of the synthase deficiency by pYK8 and pYK11 slightly and gradually improved with decreasing incubation temperatures (data not shown). In a second step, synthase specific activity was determined in XA4 strains bearing one of the four plasmids mentioned above, following an overnight induction at 30 C with 2.5 mM IPTG. Table 3 summarizes the results. No synthase activity was detected for XA4(pYK7) (expressing only the yeast synthase) and high activity was measured for XA4(pYK8) (coexpressing synthase and kinase). Compared to XA4(pYK8), XA4(pYK11), which additionally expresses the reductase, showed a slight decrease in synthase activity, which can presumably be ascribed to a lower plasmid copy number. To test whether the absence of synthase activity is the result of low levels of NAGS protein, an immunoWestern Table 3. Yeast N-acetylglutamate synthase specific activity in the XA4 (argA–) E. coli background. Plasmid Yeast enzymes expressed NAGS activity after IPTG induction (nmolÆmin)1Æmg)1 protein) pTrc99A pYK7 pYK8 pYK11 none NAGS NAGS + NAGK NAGS + NAGK + NAGPR <0.2a <0.2a 44 30 a Below detection. Fig. 2. Spot growth tests of the E. coli strain XA4(argA–) transformed with various plasmids as indicated. In each row, from left to right, 10 lL of 10-fold serial dilutions of a cell suspension (going from 1010 cellsÆmL)1 to 105 cellsÆmL)1) were spotted, either under noninducing conditions, without arginine (A) and with arginine (B); or under inducing conditions, without arginine (C) and with arginine (D). Plates were incubated at 37 C. 
FEBS 2003 1020 K. Pauwels et al. (Eur. J. Biochem. 270) Fig. 3. NAGS detection by immunoWestern blot analysis of total protein extracts of E. coli strain transformed with plasmid pYK7 expressing His6-tagged yeast N-acetylglutamate synthase (NAGS), pYK8 expressing His6-tagged NAGS and N-acetylglutamate kinase (NAGK) or pYK11 expressing His6-tagged NAGS, NAGK and N-acetylglutamyl phosphate reductase. Plasmid pTrc99a is the corresponding empty cloning vector. MM, molecular mass markers. The arrow indicates the protein band corresponding to NAGS. blot analysis was performed, comparing equal amounts of total proteins from crude extracts of the four type of transformants (Fig. 3). The synthase protein was detected by its His6 tag in extracts of the strains bearing pYK8 or pYK11, but not in extracts of a strain bearing pYK7. A small difference in protein concentration was observed between pYK8 and pYK11 in some blots, ascribed to a lower plasmid copy number. Thus, protein concentration data correspond to the data of the growth assays and of the activity measurements. Therefore, unless kinase is coexpressed, yeast synthase appears unstable, both in a heterologous bacterial background (present data) and in an yeast homologous context [32]. This suggests an intrinsic instability of the synthase protein. Furthermore, in a heterologous bacterial background, the supplementary presence of the yeast reductase, in addition to the yeast kinase, does not enhance synthase activity or levels. The ascomycete-specific domain of N-acetylglutamate kinase is required to maintain N-acetylglutamate synthase activity and protein level Yeast N-acetylglutamate kinase consists of two distinguishable domains. The N-terminal domain is conserved in both eucaryotes and procaryotes and is therefore inferred to be the catalytic active domain (CD). The C-terminal domain is specific to ascomycetes (ASD). It extends from about amino acid 348 to a residue located between amino acid 510 and 540, the region in which the kinase/reductase precursor is maturated [29]. We addressed the question whether the two kinase domains are needed to observe synthase activity and stability. By inference, this would indicate a role for each domain in the association of the NAGS/NAGK in a complex. For this experiment, new high copy number plasmids were derived from pYB3, each lacking one of the kinase domains. Plasmid pYB3 encodes the full length ARG5,6 gene, plasmid pHP21 is truncated for the ascomycete specific domain of the kinase (Daa355–493) and plasmid pHP22 is truncated for the catalytic domain of the kinase (Daa85–347). The functionality of the kinase protein encoded by those plasmids was assessed by transforming the plasmids in the Darg5,6 genetic background of strain KA42 and measuring kinase activity. As expected KA42(pHP22) lacks any kinase activity while KA42(pHP21) keeps more than 50% of the wild-type kinase activity as compared to KA42(pYB3) (data not shown), implying that the ASD-truncated kinase is stably expressed. We then analysed the synthase activity and protein level when the synthase protein was coexpressed with one of the truncated kinases. Therefore, pYX223, pYB3, pHP21 and pHP22 were cotransformed in 14S31b with pYB2, which is a GAL promoter-driven plasmid, over-expressing the synthase fused to a C-terminal haemaglutinin (HA)-tag. The first two combinations served as a negative and a positive control, respectively. The host strain 14S31b has a ARG2, ARG5,6 genetic background circumventing the need to add arginine in the growth medium, but explaining the low background synthase activity and protein level of the negative control. Table 4 summarizes the values of the synthase specific activity measured under galactose promoter inducing growth conditions. As expected, the coexpression of wildtype kinase and synthase resulted in high synthase activity. The combined expression of synthase with each of the truncated kinase proteins, however, showed no increase in the synthase activity compared to the negative control. This suggests that complex formation does not occur, resulting in synthase protein instability. Alternatively, a non-productive but stable association can be the cause of this inactivity. To assess the synthase protein level, an immunoWestern blotting was performed on crude extracts of several strains, as shown in Fig. 4A. No synthase was detectable in any of the samples, except for 14S31b(pYB2 + pYB3), which was used as a positive control. Only when the gels were deliberately overloaded, did synthase become detectable in the extracts from strains bearing pHP21 and pHP22, yet in amounts comparable to the basal level produced in the negative control (Fig. 4B). The results demonstrate that the ascomycete-specific domain of the kinase is required for accumulation of the synthase. However, if this domain is assumed to be Table 4. N-acetylglutamate synthase specific activity in strains coexpressing promoter GAL-driven ARG2 and ARG5,6 genes: effect of domain deletions in the N-acetylglutamate kinase. Status of NAGK Strain 14S31b 14S31b 14S31b 14S31b a (pYB2 (pYB2 (pYB2 (pYB2 + + + + pYX223) pYB3) pHP21) pHP22) CD, catalytic domain; b CDa ASDb NAGS activity (nmolÆmin)1Æmg)1 protein) – + + – – + – + 14 206 10 16 ASD, ascomycete specific domain. 
FEBS 2003 Co-ordinated feedback regulation (Eur. J. Biochem. 270) 1021 M. Karaman and G. Turner, University of California, Los Angeles, CA, USA, personal communication). Alignment of the amino-acid sequences of S. cerevisiae, S. pombe, C. albicans, and N. crassa kinases shows the phenylalanine 81 of N. crassa to be conserved in ascomycete kinases. It corresponds to the phenylalanine 99 in S. cerevisiae. We constructed the yeast kinase ARG5,6 F99L mutant in a vector with a GAL promoter, yielding plasmid pHP17. Plasmid pYB2, encoding the yeast synthase, was cotransformed with pHP17 in the strain YeBR6. YeBR6 (pYB2 + pYB3), over-expressing both the wild-type kinase and synthase, was used as a reference strain. The transformants were grown on galactose medium and N-acetylglutamate kinase activity in cell extracts was assayed in the presence of increasing arginine concentrations. Figure 5A compares the arginine inhibition curves of the wild-type and F99L mutant kinases. The arginine concentration required to inhibit 50% of the activity of the wild-type kinase (I0.5) is 0.1 mM, a value that is comparable to an I0.5 of 0.05 mM Fig. 4. ImmunoWestern blot detection of N-acetylglutamate synthase in total protein extracts of yeast strain 14S31b bearing plasmid pairs as indicated above the lanes. pYB2 expresses a haemaglutinin-tagged NAGS, pYB3 expresses the wild-type NAGK/NAGPR, pHP21 and pHP22 are derived from pYB3 and, respectively, lack the ascomycetespecific domain and the catalytic domain of the kinase encoding region, pHP17, also derived from pYB3, bears the F99L modification in NAGK. pYX213 and pYX223 are the empty cloning vectors. (A) Equal amounts of total protein were loaded in lanes 1–4, and double that amount in lanes 6–9. (B) All lanes contain equal amounts of total protein. (C) Lanes 1 and 2 contain 7.5 lg total protein, lanes 3 and 4 contain 15 lg, and lanes 5 and 6 contain 30 lg. MM is the molecular mass standard. expressed and to be stable in the CD truncated kinase, these data indicate that the ASD is insufficient to maintain the activity of the synthase. The kinase F99L mutant leads to arginine feedback resistance of both the kinase and the synthase R. L. Weiss and coworkers found that the F81L modification in the N-acetylglutamate kinase of N. crassa renders the enzyme resistant to arginine feedback inhibition (R. L. Weiss, S. K. Chae, J. Chung, C. McKinstry, Fig. 5. Feedback inhibition by arginine of yeast N-acetylglutamate (A) kinase and (B) synthase activities in extracts of strain YeBR6 (pYB2 + pYB3) expressing NAGS, NAGK and NAGPR (d) and strain YeBR6(pYB2 + pHP17) expressing NAGS, mutant F99L NAGK and NAGPR (s), after growth on galactose medium. The insert shows the effect of arginine at higher concentrations, (A) up to 100 mM, (B) up to 10 mM. The arginine-resistant residual activity in A is due to a distinct enzymatic activity not encoded by ARG6. 
FEBS 2003 1022 K. Pauwels et al. (Eur. J. Biochem. 270) mentioned by Hilger [26]. It is also comparable to the I0.5 of 0.075 mM determined for the wild-type kinase of N. crassa [27]. In the absence of arginine the kinase specific activity in the extract from yeasts carrying the F99L mutation was only one half of the activity of wild-type yeast. It remains susceptible to feedback inhibition by arginine, but 100 times higher arginine concentration is required to reach 50% inhibition (I0.5 of 10 mM). Synthase activity and its arginine sensitivity were also assayed using the same extracts (Fig. 5B). In the presence of the wild-type kinase, 0.015 mM arginine is required to reach I0.5 of the synthase. This value corresponds with the I0.5-value of 0.02 mM published by Wipf and Leisinger [25]. It is noticeable that the yeast synthase is 10 times more sensitive to feedback inhibition by arginine than the N. crassa synthase (50% inhibition at 0.16 mM [28]). When coexpressed with the F99L mutant kinase, the synthase behaves quite differently than when coexpressed with the wild-type kinase. The synthase specific activity is reduced fivefold and, like the mutant kinase, the enzyme becomes much less sensitive to arginine feedback inhibition (I0.5 of 0.75 mM). The reduction in activity is not a consequence of a loss in synthase protein, as immunoWestern blots revealed equal amounts of the haemaglutinin-tagged synthase in both transformants (Fig. 4C). The increased resistance to feedback inhibition of the wild-type synthase, resulting from the presence of the feedback-resistant F99L kinase, suggests a mechanism of co-ordinated feedback regulation of the synthase and the kinase. Arginine feedback inhibition of N-acetylglutamate kinase is altered in the absence of N-acetylglutamate synthase protein Proper feedback inhibition of the synthase appears to require an association with a feedback-sensitive kinase. To find out whether the opposite is also true, we studied arginine feedback inhibition of the kinase in the presence and absence of the synthase protein. KA44, a strain derived from 14S31b (ura3–, his3–) and lacking the ARG2 ORF, was constructed and cotransformed with (pYB2 + pYB3) and (pYX213 + pYB3). Because of the growth requirement of the latter transformants, cells were grown on galactose medium supplemented with arginine. Kinase specific activity in extracts was assayed in the presence of increasing arginine concentrations. The results are presented in Fig. 6. In extracts of KA44(pYB2 + pYB3), the wild-type kinase proved to be sensitive to arginine inhibition. The inhibition curve displays a normal hyperbolic shape and the I0.5-value of 0.26 mM arginine in the illustrated experiment (Fig. 6) is comparable to the I0.5 of 0,1 mM measured with YeBR6(pYB2 + pYB3) extracts (Fig. 5A). In fact, three similar, less detailed experiments (data not shown), display an I0.5-value closer to 0.1 mM. Interestingly, the apparent affinity of the kinase for the feedback inhibitor is markedly lower when the synthase is absent (I0.5-value of 1.5 mM). In addition, the inhibition curve becomes reproducibly sigmoidal. These data show that the kinase requires an interaction with the synthase for its normal arginine feedback inhibition. Furthermore, the values of the specific activity of the kinase were reproducibly two times higher in extracts of Fig. 6. Feedback inhibition by arginine of yeast N-acetylglutamate kinase activity in extracts of Darg2 strain KA44(pYB2 + pYB3) (d) and KA44(pYX213 + pYB3) (s), after growth on galactose medium supplemented with arginine. KA44(pYX213 + pYB3) compared to those of KA44 (pYB2 + pYB3), suggesting that the association with the synthase inhibits partially the kinase activity. Discussion In our previous work, we showed that synthase forms a complex with the kinase, an association essential to synthase activity and synthase protein accumulation. In contrast, no physical interaction could be demonstrated conclusively between synthase (or kinase) and reductase despite the fact that some data suggested a role of the reductase for synthase activity [32]. To investigate further the role of the reductase in the metabolon, we have now followed two new genetic approaches. First, the activity of the synthase, expressed from its natural locus, was measured in yeast deletion mutants lacking relevant parts of the chromosomal ARG5,6 gene. Second, synthase activities and protein accumulation were monitored by over-expressing yeast genes in the heterologous E. coli background. Both approaches led to the same conclusions, namely that synthase activity is strictly dependent on the presence of the kinase and essentially independent of the reductase. This dispensibility of the reductase for synthase activity, is in contrast with the previous results obtained in a context of over-expression in yeast, which had indicated an apparent requirement of the reductase [32]. One (unexplored) hypothesis that might explain the discrepancy, is that the bulk of over-expressed kinase is inefficiently targeted to the yeast mitochondria in the absence of reductase. The present data in E. coli further show that no synthase protein is detectable in the absence of kinase, a situation similar to the one observed previously in yeast. We attribute this drastic reduction in steady state concentration of the protein to an instability of the yeast synthase when not associated to the kinase. Because it is also observed in the heterologous E. coli background, this apparent instability is likely to be an intrinsic feature of the protein, rather than to result from of a yeast specific degradation process. Alternatively, the uncomplexed synthase might present structural features rendering it susceptible to proteolytic degradation 
FEBS 2003 in general. In any case, the lack of synthase protein in the absence of kinase has been observed with expression systems using totally different promoters and translation initiation signals. Therefore, the hypothesis that it results from an effect on transcription or translation can be reasonably excluded. The data in hand today do not allow to tell if the lack of synthase activity in the absence of kinase fully correlates with the physical disappearance of the enzyme, or if inactive free synthase can subsist. However, as discussed below, the kinetic properties of the synthase are likely to be modulated by its association with the kinase. The role of the catalytic and the ascomycete specific domain of NAGK in complex formation with NAGS was tested in an indirect way. Our approach is based on the knowledge that synthase activity and protein levels are dependent upon the enzyme association with the kinase (32 and new data above). Therefore, capacity for complex formation was deduced from measurements of overexpressed synthase activity and from estimations of the concentrations of over-expressed synthase protein. Present data showed that deletions of the catalytic domain (CD) or the ascomycete specific domain (ASD) of the kinase both result in the loss of synthase activity and stability. As the ASD-truncated kinase is shown to be stable and active, it implies that the ASD of the kinase is necessary for a productive association with the synthase. The presence of CD-truncated kinase in yeast extracts could not be demonstrated (neither over-expressed truncated kinases nor the wild-type are detectable by SDS/PAGE/Coomassie Blue staining analysis), but if it is assumed to be stable, then the data show that the ASD is not sufficient for association with the kinase. Previous data revealing highly reduced amounts of synthase when coexpressed with N-terminally His10-tagged kinase [32], support the hypothesis that the ASD of the kinase does not suffice for complex formation. By analogy with the F81L substitution of N. crassa kinase, which renders it feedback resistant, a new mutant of the yeast kinase (F99L substitution) was constructed. Our results show that the yeast mutant kinase is feedback resistant as well. In comparison to the wild-type yeast kinase, 100 times more arginine is required to reach halfinhibition of the F99L yeast mutant kinase. Our results illustrate further that feedback regulation of the wild-type yeast synthase is strongly dependent upon the presence of a normally regulated kinase. In the presence of the wild-type kinase, the synthase is fully inhibited by 0.1 mM arginine, while 10 mM arginine is required to inhibit completely the synthase activity when the partner is a feedback-resistant mutant kinase. Moreover, the kinetic properties of the synthase appear dependent upon its association with the kinase. Indeed, in the context of the mutated kinase, the synthase specific activity was reduced by 80% while the amount of enzyme remained unchanged. Contrasting with the strict requirement of NAGK for NAGS activity, the absence of NAGS increased the activity of NAGK by approximately twofold, possibly reflecting inhibition of NAGK in the NAGS/NAGK complex. The presence of NAGS had also the effect of rendering hyperbolic the inhibition of the kinase by arginine, whereas in absence of NAGS the inhibition was sigmoidal and Co-ordinated feedback regulation (Eur. J. Biochem. 270) 1023 exhibited an increased I0.5-value, strongly suggesting that more than one site for arginine has to be occupied to inhibit the kinase. Although the present results demonstrate quite different apparent affinities for arginine of the kinase and the synthase, the data do not allow to decide if specific inhibitory sites for arginine exist in the two enzymes or if only the kinase possesses a binding site for the inhibitor. In any case, the mutual influence of each enzyme on the other concerning its susceptibility to arginine inhibition suggest the existence of either a crosstalk between the inhibitory sites of the two enzymes, or an intermoleculair transmission of an inhibitory signal from a binding site on the kinase to the catalytic site of the synthase. Both are in agreement with the hypothesis of co-ordinated feedback regulation of NAGS and NAGK in yeast, as proposed initially for N. crassa (R. L. Weiss, S. K. Chae, J. Chung, C. McKinstry, M. Karaman and G. Turner, University of California, Los Angeles, CA, USA, personal communication). Acknowledgements K. P. is the recipient of a Specialization Grant from the IWT (Vlaams Instituut voor de bevordering van het wetenschappelijk-technologisch onderzoek in de industrie). We thank J. P. Ten Have for his help with the figures and tables. References 1. Cunin, R., Glansdorff, N., Pierard, A. & Stalon, V. (1986) Biosynthesis and metabolism of arginine in bacteria. Microbiol. Rev. 50, 193–225. 2. Davis, R. (1986) Compartmental and regulatory mechanisms in the arginine pathways of Neurospora crassa and Saccharomyces cerevisiae. Micr. Rev. 50, 280–313. 3. Harris, B. & Singer, M. (1998) Identification and characterization of the Myxococcus xanthus argE gene. J. Bacteriol. 180, 6412– 6414. 4. Van de Casteele, M., Demarez, M., Legrain, C., Glansdorff, N. & Piérard, A. (1990) Pathways of arginine biosynthesis in extreme thermophilic archaeo- and eubacteria. J. Gen. Microbiol. 136, 1177–1183. 5. Xu, Y., Liang, Z., Legrain, C., Ruger, H. & Glansdorff, N. (2000) Evolution of arginine biosynthesis in the bacterial domain: novel gene-enzyme relationships from psychrophilic Moritella strains (Vibrionaceae) and evolutionary significance of N-alpha-acetyl ornithinase. J. Bacteriol. 182, 1609–1615. 6. Udaka, S. & Kinoshita, S. (1958) Studies on 1-ornithine fermentation. I. The biosynthetic pathway of 1-ornithine in Micrococcus glutamicus. J. General Appl. Microbiol. 4, 272–282. 7. Hoare, D. & Hoare, S. (1966) Feedback regulation of arginine biosynthesis in blue-green algae. J. Bacteriol. 92, 375–379. 8. Haas, D., Kurer, V. & Leisinger, T. (1972) N-acetylglutamate synthetase of Pseudomonas aeruginosa. An assay in vitro and feedback inhibition by arginine. Eur. J. Biochem. 31, 290–295. 9. Shinners, E. & Catlin, B. (1978) Arginine biosynthesis in Neisseria gonorrhoeae: enzymes catalyzing the formation of ornithine and citrulline. J. Bacteriol. 136, 131–135. 10. Meile, L. & Leisinger, T. (1984) Enzymes of arginine biosynthesis in methanogenic bacteria. Experientia 40, 899–900. 11. Sakanyan, V., Kochikyan, A., Mett, I., Legrain, C., Charlier, D., Piérard, A. & Glansdorff, N. (1992) A re-examination of the pathway of ornithine biosynthesis in a thermophilic and two mesophilic Bacillus species. J. Gen. Microbiol. 138, 125–130.