Characterization and kinetic mechanism of mono- and

Eur. J. Biochem. 267, 5217±5226 (2000) q FEBS 2000

Characterization and kinetic mechanism of mono- and bifunctional ornithine acetyltransferases from thermophilic microorganisms FreÂdeÂric Marc1, Pierre Weigel1, Christianne Legrain2, Yann Almeras1, Marie Santrot1, Nicolas Glansdorff2,3 and Vehary Sakanyan1 1

FRE-CNRS 2230 Biocatalyse, Laboratoire de Biotechnologie, UniversiteÂ de Nantes, France; 2Jean-Marie Wiame Institute for Microbiological Research, Brussels, Belgium; 3Laboratory for Genetics and Microbiology, Vrije Universiteit, Brussels, Belgium

The argJ gene coding for N 2-acetyl-l-ornithine: l-glutamate N-acetyltransferase, the key enzyme involved in the acetyl cycle of l-arginine biosynthesis, has been cloned from thermophilic procaryotes: the archaeon Methanoccocus jannaschii, and the bacteria Thermotoga neapolitana and Bacillus stearothermophilus. Archaeal argJ only complements an Escherichia coli argE mutant (deficient in acetylornithinase, which catalyzes the fifth step in the linear biosynthetic pathway), whereas bacterial genes additionally complement an argA mutant (deficient in N-acetylglutamate synthetase, the first enzyme of the pathway). In keeping with these in vivo data the purified His-tagged ArgJ enzyme of M. jannaschii only catalyzes N 2-acetylornithine conversion to ornithine, whereas T. neapolitana and B. stearothermophilus ArgJ also catalyze the conversion of glutamate to N-acetylglutamate using acetylCoA as the acetyl donor. M. jannaschii ArgJ is therefore a monofunctional enzyme, whereas T. neapolitana and B. stearothermophilus encoded ArgJ are bifunctional. Kinetic data demonstrate that in all three thermophilic organisms ArgJ-mediated catalysis follows ping-pong bi±bi kinetic mechanism. Acetylated ArgJ intermediates were detected in semireactions using [14C]acetylCoA or [14C]N 2-acetyl-l-glutamate as acetyl donors. In this catalysis l-ornithine acts as an inhibitor; this amino acid therefore appears to be a key regulatory molecule in the acetyl cycle of l-arginine synthesis. Thermophilic ArgJ are synthesized as protein precursors undergoing internal cleavage to generate a and b subunits which appear to assemble to a2b2 heterotetramers in E. coli. The cleavage occurs between alanine and threonine residues within the highly conserved PXM-ATML motif detected in all available ArgJ sequences. Keywords: N-acetyltransferase; arginine biosynthesis; bifunctional enzymes; thermostability; protein processing.

The glutamate family of amino acids provides striking examples of the evolutionary strategies which allowed the emergence of related but independent biochemical pathways [1,2]. l-glutamate, via two consecutive intermediates, can undergo a spontaneous cyclization and conversion to l-proline, whereas acetylation prevents this cyclization and initiates the eight-step pathway for arginine synthesis (Fig. 1). Acetyl-group removal then produces ornithine which is converted to citrulline by condensation with carbamoylphosphate. At the fifth step, N-acetylornithine can be converted to ornithine by two completely different enzymes. Acetylornithinase (N 2 -acetyl-l-ornithine amidohydrolase EC 3.5.1.16), encoded by the argE gene in Escherichia coli, deacetylates N-acetylornithine. This linear pathway is functional in representatives of Enterobacteriaceae [3], in Myxococcus [4], in Vibrionaceae [5] and appears to be present in the thermophilic archaeon Sulfolobus [6]. Another enzyme, OATase (N 2-acetyl-l-ornithine: l-glutamate N-acetyltransferase; EC 2.3.1.35, encoded by the argJ gene) catalyses the transfer of Correspondence to V. Sakanyan, FRE-CNRS 2230 Biocatalyse, Laboratoire de Biotechnologie, FaculteÂ des Sciences et des Techniques, UniversiteÂ de Nantes, 2, rue de la HoussinieÁre, 44322 Nantes Cedex 3, France. Fax: 1 33 2 51 12 56 12, Tel.: 1 33 2 51 12 56 20, E-mail: [email protected] Abbreviations: PVDF, poly(vinylidene difluoride). Enzymes: ornithine acetyltransferase (OATase, EC 2.3.1.35), acetylglutamate synthase (AGSase, EC 2.3.1.1). (Received 13 March 2000, accepted 21 June 2000)

the acetyl-group from N-acetylornithine to l-glutamate. The enzymes of this cyclic pathway have been described in Corynebacterium glutamicum (formerly Micrococcus glutamicus [7]), yeasts [8], cyanobacteria [9], Pseudomonas aeruginosa [10], Neisseria gonorrhoeae [11], methanogens [12], representatives of Bacillus [13], Streptomyces coelicolor [14] and Thermus thermophilus [15]. Once this pathway is initiated from glutamate and acetylCoA by AGSase (acetylCoA: l-glutamate N-acetyltransferase; EC 2.3.1.1, encoded by the argA gene), the biosynthetic flow is maintained by recycling the acetyl-group, thereby leaving acetylCoA available for other intracellular processes. In pseudomonads and yeasts the OATase and AGSase activities are separable [10,16]. Moreover, in these organisms [17,18] (see also Saccharomyces and Pseudomonas Genome Databases) as well as in Neisseria [19,20] two separate argJ and argA genes have been identified. However, the energetically economical acetyl cycle of arginine biosynthesis can function with two different versions of ArgJ proteins. Biochemical and genetic studies showed that a single argJ-encoded N-acetyltransferase displays both AGSase and OATase activities in the Gram-positive bacterium Bacillus stearothermophilus; this enzyme can therefore fulfil both the first and the fifth catalytic steps of the pathway [13,21]. In contrast, in C. glutamicum [22] and T. thermophilus [15] ArgJ protein only displays OATase activity. Such a behaviour of microbial argJ encoded proteins has allowed us to group them into mono- and bifunctional enzymes [21,22]. Surprisingly, argA and argJ-encoded N-acetyltransferases displaying AGSase

5218 F. Marc et al. (Eur. J. Biochem. 267)

q FEBS 2000

Fig. 1. Arginine biosynthesis in microorganisms. The linear (E. coli is taken as a reference) and the alternative cyclic (in B. stearothermophilus, T. neapolitana and M. jannaschii) arginine biosynthesis pathways are shown. Amino acids which inhibit enzymatic steps are boxed.

activity do not appear homologous, which suggests a different origin [22]. Moreover, argA genes from E. coli and P. aeruginosa share only very limited similarity with the Neurospora crassa arg-14 gene encoding AGSase [23]. However, simply comparing sequences is not sufficient to distinguish between mono- and bifunctional enzymes [22]. New insight into geneenzyme relationships was gained by studying OATase from S. cerevisiae. The yeast protein is synthesized as a precursor which undergoes internal cleavage between alanine and threonine residues to form two subunits which then fold into a 57-kDa heterodimeric molecule [24]. The protein precursor appears to be targeted to mitochondria through its N-terminus [24]. The cloned S. cerevisiae OATase was shown to exhibit a weak additional AGSase activity in E. coli [25]. A thorough functional and structural analysis of OATases from different microbial sources will be needed to understand the kinetic mechanism of their action and the basis of their catalytic diversity. As a first step in this work we have examined two distantly related thermophilic bacteria (B. stearothermophilus and Thermotoga neapolitana) and an archaeon (Methanococcus jannaschii) as thermophilic organisms branch off early on the 16S rRNA-based microbial phylogeny [26] and therefore they might reflect ancestral events in the development of arginine

metabolism and corresponding regulatory mechanisms. We show that the internal processing is a phenomenon common to the bacterial and archaeal ArgJ protein precursors. We propose a kinetic mechanism for microbial ArgJ action. Our data indicate that in this mechanism l-ornithine plays a crucial role in regulating the flow of arginine precursors in thermophilic microorganisms.

E X P E R I M E N TA L P R O C E D U R E S Bacterial strains, plasmids and growth conditions E. coli strains JM109 [F 0 traD36 lacIq ZDM15 proAB] e142 (McrA2) D(lac-proAB) thi-1 gyrA96 endA1 hsdR17 (rK2 mK1) relA1 supE44 recA1 and BL21 F2 ompT hsdSB (rB2 mB2) gal dcm (DE3) pLysS were purchased from Stratagene and Novagen, respectively. Other E. coli strains XA4 (F2 argA nalA l2 ls hsdR), XS1D2R [F2 D (ppc-argE ) nalA rpoB l2 hsdR recA)], XA4::argE (F2 argA argE nalA l2 ls hsdR) and XC33 (F2 argC nalA rpoB l1 hsdR) were described previously [13]. Strains B. stearothermophilus NCIB8224 and T. neapolitana DSM5068 were used previously for cloning arg genes [27,28]. The argJ gene region from

q FEBS 2000

Mono- and bifunctional ornithine acetyltransferases (Eur. J. Biochem. 267) 5219

T. maritima genome [30]. In order to clone the corresponding argJ gene endonuclease restriction sites were created at its extremities by PCR and, by following digestion with the appropriate restriction enzymes, the amplified DNA was inserted into the chosen vector plasmid. The B. stearothermophilus and M. jannaschii argJ genes were fused to a C-terminal His-tag using the corresponding sequence from the pET21d(1) plasmid. Similarly, all three thermophilic argJ genes were fused to an N-terminal His-tag taken from pET19b. In the latter case the `overlapping extension' method [31] was applied to join up two independently amplified PCR products. The His-tagged argJ genes were inserted into the pACYC184 plasmid digested by EcoRV or EcoRV/BamHI (Fig. 2). The resulting plasmids were designated as pargJ-X/N for N-terminal, and pargJ-X/C for C-terminal fused proteins, in which `X' stands for the gene source, namely Bs for B. stearothermophilus, Tn for T. neapolitana and Mj for M. jannaschii. DNA sequencing was performed either manually with the dideoxynucleotide chain termination method [32] to verify the junctions created in recombinant DNA constructs or automatically for longer DNAs. The gene and protein sequences were analyzed by homology searches against available sequences in gene and protein banks using the tblastx program [33]. Design of oligonucleotide primers and protein sequence alignements were conducted with the macdnasis program, version 3.6 (Hitachi Software). Recombinant ArgJ purification

Fig. 2. Recombinant pACYC184 derivatives carrying argJ genes with C- or N-terminal His-tag. The strategy of construction of N-terminal His-tagged ArgJ is shown. A couple of A /B primers along with other couples, Bs1/Bs2 or Mj1/Mj2 or Tn1/ Tn2 primers were used in the first PCR step to create overlapping sequences. Then the DNA fragment, amplified by A /B primers, was mixed with each amplified argJ DNA fragment and used in the second PCR step only with primers A /Bs2 or A / Mj2 or A / Tn2.

M. jannaschii was purchased as a pUC71K plasmid (ATCC 624904) from The Institute for Genomic Research, Rockville, MD, USA. The pACYC184 plasmid was from our laboratory collection; pET19b, pET21d(1) and pET24d(1) were from Novagen. Bacteria were cultivated in rich LB or in M9 minimal media [29] supplemented with succinate 0.2% for the E. coli XS1D2 strain. When required, antibiotics were added at 50 mg´mL21 for ampicillin and 25 mg´mL21 for chloramphenicol or kanamycin. The capacity of microbial argJ genes to complement E. coli mutants argA (strain XA4) or argE (strain XS1D2) was tested in liquid or on solid selective minimal media devoided of l-arginine at 37 8C. The test was indicative of bi or monofunctional enzymes displaying AGSase and/or OATase activities. Complementation data were further verified by enzymatic assays (see below). DNA manipulations and recombinant plasmid constructions Arginine biosynthetic genes were cloned from T. neapolitana DSM5068 by PCR [28] assuming sequence similarity with the

Recombinant E. coli XS1D2 cells carrying the His-tagged argJ gene were cultivated in arginine-less minimal medium, harvested by centrifugation (5000 g, 4 8C, 15 min) and washed twice in Tris/HCl 0.1 m, pH 8.0. Cells were disrupted by sonication (15 min by pulsing 10 s at 19 kHz) in Na2HPO4 50 mm, NaCl 300 mm, imidazole 10 mm, pH 8.0 at 4 8C and centrifuged (18 000 g, 4 8C, 15 min). The supernatants obtained were treated at 65 8C for 30 min, recentrifuged and then purified by affinity chromatography on a Ni/nitriloacetic acid resin following manufacturer's instruction (Qiagen). The protocol used gave homogenous patterns of purified enzymes (more than 95% of purity). Molecular masses of purified recombinant ArgJ proteins were then estimated by gel permeation chromatography on a Sephacryl S200 gel (Pharmacia) using standard protein markers (Bio-Rad): thyroglobulin (670 kDa), bovine gamma globulin (158 kDa), chicken ovalbumin (44 kDa), equine myoglobin (17 kDa) and vitamin B12 (1.3 kDa). Protein electrophoresis, blotting and sequencing SDS/PAGE of proteins was carried out as described by Ausubel et al. [34]. Gels were stained with Coomassie brilliant blue R-250 (Bio-Rad). Molecular masses of proteins were determined with standard proteins: b phosphorylase (97.4 kDa), BSA (66.2 kDa), ovalbumin (45 kDa), carbonic anhydrase (31 kDa), trypsin inhibitor (21.5 kDa) and lysozyme (14.4 kDa) from Bio-Rad. Protein concentration was determined by the Bradford method [35] using BSA as standard. Immunological analysis of Western-blotted His-tagged proteins was performed using anti(H)4 primary antibodies against His-tag (Qiagen) and secondary antibodies (Bio-Rad) as described by Ausubel et al. [34]. Protein bands on ImmunoBlot poly(vinylidene difluoride) (PVDF) membranes (Bio-Rad) were visualized with the NBT-BCIP chromogenic reagent (Sigma).


q FEBS 2000

Fig. 3. Alignment of B. stearothermophilus (B. ste), M. jannaschii (M. jan) and T. neapolitana (T. nea) ArgJ amino-acid sequences. The arrow indicates the cleavage site determined for each of the three enzymes. The highly conserved amino acids around the cleavage site are shown in bold. M. jannaschii enzyme has additional LH residues within the supposed motif (in italic), in which P, A and T amino acids are conserved in all known 31 ArgJ sequences.

Proteins were also blotted on Sequi-Blot PVDF membranes (Bio-Rad) and then microsequencing was performed as described by Matsudaira [36] by a 477-A amino-acid analyzer (Applied Biosystems). Enzymatic assays We developed an HPLC-based method to quantitatively evaluate the AGSase or OATase catalysed formation of N-acetyl-lglutamate. OATase and AGSase activities were measured in a mixed buffer (Mes 0.1 m, Pipes 0.1 m, Tris 0.1 m, glycine 0.1 m and K2HPO4 0.1 m, pH range 5±11) containing l-glutamate (10 or 20 mm) and N2-acetyl-l-ornithine (20 mm) or acetylCoA (2.5 mm). The mixed buffer with a high ionic strength was chosen empirically in order to provide similar conditions for enzyme activities at different values of pH and high temperatures. After 2 min preincubation at the desired temperature the reaction was initiated by adding the purified enzyme (1±10 mg) or crude extracts (20±100 mg of total protein). The reactions were carried out for 5 min in a temperature range of

30 8C to 95 8C and then stopped by adding phosphoric acid 500 mm. After discarding the precipitated proteins the soluble fractions were analyzed on a Luna C18 column (Phenomenex) directly online with the HPLC system (Kontron) using phosphoric acid 0.1 m and methanol (90 : 10 v: v) as the mobile phase at a flow rate of 1 mL´min21. Detection of reaction products was carried out at 215 nm. AGSase activity was also measured by a radioassay method in presence of l-[U-14C]glutamic acid (10 mm, 50 mCi´mmol21, Amersham) and acetylCoA (0.5±2.5 mm). Labeled l-glutamate and N-acetyl-l-glutamate were separated on a cation exchange resin (AG50W, Bio-Rad) previously equilibrated with HCl 0.2 m. After elution with 3 mL of HCl 200 mm, the radiolabelled N-acetyl-l-glutamate fraction was counted in a scintillation mixture 2.67% 2,5-diphenyloxazol, 0.067% phenylen-bis (5 phenyloxazol) in 66% toluene, 33% triton X100 with a Beckman LS6500 scintillator. Temperature inactivation of recombinant B. stearothermophilus, T. neapolitana or M. jannaschii enzymes was determined at 75, 90 or 95 8C, respectively, for various periods of times. One enzyme unit is defined as the amount of enzyme producing 1 mmol of N-acetyl-l-glutamate per min. Kinetic studies

Fig. 4. Analysis of His-tagged ArgJ enzymes of B. stearothermophilus, M. jannaschii and T. neapolitana synthesized in E. coli XS1D2. (A) SDS/PAGE of native proteins eluted from the S200 column: lane 1, molecular mass markers, lanes 2, 3 and 4, respectively, ± N-terminal His-tagged ArgJ of B. stearothermophilus M. jannaschii and T. neapolitana. (B) Western-blot analysis of the proteins with anti(H)4 antibodies. The proteins were separated in duplicate SDS/PAGE.

OATase kinetic studies were carried out in presence of l-glutamate (5±25 mm) and N2-acetyl-l-ornithine (5±25 mm) or acetylCoA (0.5±2.5 mm). To evaluate the equilibrium of reaction, the amount of N-acetyl-l-glutamate was regularly estimated during 16±48 h in reaction media prepared with equimolar concentrations of N 2-acetyl-l-ornithine and l-glutamate (20 mm) or acetylCoA and l-glutamate (5 mm) or N-acetyl-l-glutamate and l-ornithine (20 mm) or N-acetyl-lglutamate and coenzymeA (5 mm). Product inhibition studies were carried out by adding l-ornithine or coenzymeA up to 10 mm. The data from kinetic studies were analyzed as described by Cleland [37]. Double-reciprocal plots using initial velocity vs. varied substrate concentrations were used to determine the mode of inhibition [38].

q FEBS 2000


Table 1. Kinetic parameters of ArgJ enzymes from thermophilic microorganisms. The data obtained with N-terminal His-tagged ArgJ enzymes are shown for OATase and AGSase activities. Both OATase and AGSase activities were found to undergo competitive substrate inhibition at unbalanced ratios of the acetyl-group donor and acceptor concentrations. Values in brackets were estimated at 10mM glutamate. The B. stearothermophilus N-terminal His-tagged ArgJ protein has Km 2.7 mm close to Km 2 mm obtained after removing His-tag by enterokinase, whereas Km for N-acetyl- l-ornithine was estimated as 16.3 mm with the C-terminal tagged enzyme. l-arginine (10 mm) had no inhibitory effect on enzyme activities. B. stearothermophilus

T. neapolitana

M. jannashii

OATase activity pH optimum T (8C) optimum Vm (U´mg protein21) Km N-acetyl-L-ornithine (mm) Km L-glutamate (mm) Ki L-ornithine (mm)

8 75 27 (13.3) 2.7 19.2 2.3

8 90 290 (66.7) 8.1 27.9 2.1

8 . 95 175 9.6 11.3 1.1

AGSase activity pH optimum Vm (U´mg protein21) Km AcCoA (mm)

8±8.5 (2.5) 0.9

7.5±8.0 (7.7) 2.4

Synthesis of

14

C-labelled N-acetyl-l-glutamic acid

Labelled on the glutamate moiety. One millimole of l-[U-14C] glutamic acid (250 mCi) was dissolved in 2 mL of H2O and chilled on ice then 1 mL of acetic anhydride (Aldrich) was added under extensive mechanical stirring and continued cooling at 5 8C for 6 h. An additional 0.2 mL of acetic anhydride was added and the solution was allowed to stand overnight at 5 8C under stirring. The residue obtained by evaporating the solution to complete dryness in vacuo was dissolved in 1.5 mL HCl 0.2 m, applied to an AG50W (BioRad) column, and eluted with 5 mL HCl 0.2 m. The eluate was evaporated in vacuo to dryness. The residue was extracted with 1 mL ethanol and evaporated again to dryness. Labelled on the acetyl moiety. l-glutamic acid (6 mmol) were dissolved in 2 mL H2O, chilled on ice, and 80 mL of unlabelled acetic anhydride was added. The mixture was introduced by depression in the vial containing [1± 14C]acetic anhydride (500 mCi). The solution was allowed to stand overnight at 5 8C under stirring and the product was purified as described above. Detection of acetylated ArgJ intermediates Purified B. stearothermophilus or T. neapolitana ArgJ enzymes or BSA (as a control) were incubated with only acetyl donors (semireaction), namely with [14C]acetylCoA (60 mCi´mmol21, Amersham) or with [14C]N-acetyl-l-glutamic acid or with N-acetyl-l-[14C]glutamic acid in the above mentioned mixed buffer at 37 8C for 30 min and then subjected to gel permeation chromatography on a Superose P12 column (Pharmacia). The purified protein fractions were collected and 14C-acetyl-labelled intermediates detected with a Beckman LS6500 scintillator.

R E S U LT S Cloning and sequence analysis of argJ genes from the three thermophilic organisms Previously it was shown that expression of B. stearothermophilus argJ above a critical level becomes deleterious for E. coli host cells [39,40]. In keeping with these observations we did not succeed in constructing plasmids expressing the

B. stearothermophilus argJ gene from a T7 promoter after IPTG induction in the E. coli BL21 (DE3) pLysS strain. Therefore, in order to minimize the level of B. stearothermophilus ArgJ synthesis in E. coli host cells we used a low copy pACYC184 plasmid as a vector. The resulting pargJ-Bs/N and pargJ-Bs/C plasmids in which the argJ gene was transcribed from the vector tet promoter (see Fig. 2) were found to be stable in E. coli XA4 and XS1D2. Furthermore, endonuclease restriction and sequence analysis did not reveal DNA rearrangements in these plasmids. Stable recombinant plasmids were also constructed by cloning the argJ gene from two other hyperthermophiles in pACYC184 (Fig. 2). The resulting pargJ-Tn/N plasmid, carrying the T. neapolitana gene, complemented both E. coli argA and argE mutant strains as well as the double argA argE mutant strain E. coli XA4::argE. In contrast, pargJ-Mj/N or pargJ-Mj/C plasmids, carrying the M. jannaschii gene, only complemented an E. coli argE mutant. The ArgJ proteins from T. neapolitana and M. jannaschii thus appear to be bi- and monofunctional, respectively. Amino-acid sequence comparison of the three thermophilic ArgJ proteins showed that the N-terminal portion of B. stearothermophilus ArgJ was 17 and 10 amino acids longer than its T. neapolitana and M. jannaschii homologues, respectively (Fig. 3). Amino-acid variations were spread all along the protein sequences, however, some distinct conserved amino acids were detected in several regions. In particular, the conserved ATML sequence resembles the region involved in the cleavage of the S. cerevisiae ArgJ precursor [24]. Thermophilic ArgJ proteins undergo cleavage processing Both N- and C-terminal His-tagged ArgJ proteins from the three thermophiles appeared as two clearly distinct bands of almost the same intensity (i.e. in equimolar proportions) on SDS/PAGE gel whether the extracts of the E. coli XS1D2 host cells were subjected to heat-treatment or to an additional gel permeation chromatography (Fig. 4A). Western-blot analysis of N-terminal His-tagged ArgJ proteins (see Fig. 4B) identified the fast migrating band as the a subunit of the protein precursor, the more slowly migrating C-terminal part was designated as b subunit. The amino-acid sequence of the N-terminal part of all three b subunits was shown to be TMLXFITT with


q FEBS 2000

Fig. 5. Thermal inactivation of the B. stearothermophilus at 75 8C (), M. jannaschii at 95 8C (X) and T. neapolitana at 90 8C (W) ArgJ enzymes. Inactivation of both AGSase and OATase activities was similar for bifunctional enzymes. The effect of preincubation of the B. stearothermophilus enzyme with acetyl-group donors at 75 8C on OATase activity is shown in the insert.

alanine for B. stearothermophilus and serine for T. neapolitana at the fourth position (the cysteine residue expected at this position for M. jannaschii ArgJ was not resolved). These data showed that the ArgJ protein precursor was cleaved just upstream from the threonine residue within the conserved ATML sequence (see Fig. 3). Molecular masses of a subunits processed from N or C-terminal tagged ArgJ precursors were determined by SDS/PAGE to be close to the values deduced from their amino-acid sequences (between 21 kDa and 24 kDa). However, molecular masses of b subunits were determined as 28±29 kDa, approximatively 4 kDa higher than the expected values deduced from their amino-acid sequences (see Fig. 4). Molecular masses of N-terminal His-tagged ArgJ proteins were determined by gel permeation chromatography as 98 kDa, 85 kDa and 90 kDa for B. stearothermophilus, T. neapolitana and M. jannaschii, respectively. Molecular masses of partially purified non-tagged OATases from B. stearothermophilus [21] and T. thermophilus (C. Legrain, unpublished data) were found to be close to 92 kDa and 86 kDa, respectively. From both deduced and experimental values for molecular masses of the processed subunits, the ArgJ protein from the three thermophilic microorganisms appears to be a heterotetramer constituted of two a and two b subunits. The appearance of a and b subunit bands processed from C-terminal His-tagged ArgJ proteins was always accompanied by two other weaker bands of approximately 46 kDa and 30 kDa on SDS/PAGE gel or in Western-blots in protein patterns obtained after elution from the Ni/nitriloacetic acid resin (data not shown). Effect of pH and temperature on enzyme activity Enzymatic assays confirmed that the B. stearothermophilus and T. neapolitana His-tagged ArgJ proteins possessed both OATase and AGSase activities. In contrast, the purified M. jannaschii enzyme exhibited only OATase acitivity. Even when the purified enzyme was incubated for 48 h at 37 8C with l-glutamate and acetylCoA no formation of N-acetylglutamate was observed. Besides, the very low AGSase activity present in extracts of E. coli XA4 cells carrying the pargJ-Mj/N or pargJ-Mj/C

plasmids was close to the background activity level detected in crude extracts of the plasmid-less strain. Transacetylation was found to occur efficiently in a pH range from 7 to 10 with the highest activity around pH 8 for all three OATases activities (Table 1). The highest rate for AGSases activities was found at pH 8.5 and 7.5 for B. stearothermophilus and T. neapolitana bifunctional ArgJ enzymes, respectively. It was shown previously that thermal inactivation of B. stearothermophilus ArgJ was similar for AGSase and OATase activities in E. coli crude extracts [21]. Purified His-tagged OATases from B. stearothermophilus and T. neapolitana exhibited their highest activities, respectively, at 75 8C and 90 8C (Table 1). The half-life of the T. neapolitana enzyme (for both AGSase and OATase activities) was found to be close to 9 h at 90 8C (Fig. 5). The B. stearothermophilus ArgJ enzyme's OATase activity dropped after three minutes incubation at 75 8C. However, the addition of acetyl donors 1 mm acetylCoA or 20 mm N2-acetyl l-ornithine (but not l-glutamate as the acetyl-group acceptor) prolonged the half-live of the enzyme nearly fourfold and sevenfold, respectively (see Fig. 5). M. jannaschii ArgJ activity increased with increasing temperature from 30 8C to 95 8C and kept up to 70% of its activity after 9 h incubation at 95 8C. Kinetics of thermophilic ArgJ enzyme mediated reactions N2-Acetyl-l-ornithine : l-glutamate N-acetyltransferase activity. The reaction catalyzed by both mono- and bifunctional ArgJ enzymes involves the couple N 2-acetyl l-ornithine : l-glutamate as substrate and the couple N-acetyl l-glutamate : l-ornithine as product. In such a bi-reactant system the catalytic reaction could occur randomly or be coordinated with respect to the binding and release of each participant [41]. To elucidate which mechanism directs enzyme catalysis, B. stearothermophilus, T. neapolitana and M. jannaschii OATase activities were measured at different concentrations of the substrates. For the three enzymes, 1/V vs. 1/S plots showed a family of parallel lines. A typical graph for B. stearothermophilus OATase activity is shown in Fig. 6. When both substrate concentrations were varied at a fixed ratio, secondary reciprocal plots were linear (data not shown).

q FEBS 2000


Fig. 6. Lineweaver & Burk reciprocal plots of the B. stearothermophilus N-terminal His-tagged ArgJ enzyme activity (OATase). The plots are representative of 1/V, where V is the ArgJ specific activity measured at initial velocity conditions, vs. 1/C, where C is N 2-acetyl-l-ornithine (A) or L-glutamate (B), respectively, in the presence of increasing concentrations of N 2-acetyl-l-ornithine or l-glutamate: 5 mm (X), 10 mm (W), 15 mm (), 20 mm (B), and 25 mm (A). Replots of 1/V-axis intercept vs. 1/C, where C is the N 2-acetyl l-ornithine or L-glutamate concentrations, are presented in inserts. Similar plots with parallel lines were obtained for M. jannaschii and T. neapolitana enzymes and kinetic data determined were similar irrespective of N- or C-terminal tagged protein versions (data not shown). Linear reciprocal replots were obtained at variable concentrations but at fixed ratios 1 : 2, 1 : 1 or 2 : 1 of N 2-acetyl l-ornithine and l-glutamate (data not shown). Table 2. Detection of acetylated ArgJ intermediates. ND, not detected.

Enzyme

Substrate

Mol acetyl residues bound per mol enzymea

B.stearothermophilus ArgJ [14C]acetylCoA [14C [acetyl-l-glutamate N-acetyl-l-[14C]glutamate

0.20 0.61 ND

T. neapolitana ArgJ

[14C]acetylCoA [14C]N-acetyl-l-glutamate N-acetyl-l-[14C]glutamate

0.09 0.59 ND

BSA

ND [14C]acetylCoA [14C]N-acetyl-l-glutamate ND N-acetyl-l-[14C]glutamate ND

a

Data are presented for a2û2 protein structure.

Product inhibition studies demonstrated that l-ornithine was a competitive inhibitor vs. l-glutamate and a mixed-type noncompetitive inhibitor vs. N 2-acetyl-l-ornithine (Fig. 7). The corresponding Kil-ornithine values were found to be similar and close to 2 mm for all three enzymes (see Table 1). Thus, the kinetic and inhibition data support a ping-pong BiBi mechanism for the OATase-mediated reaction. The affinities of the three purified OATases towards l-glutamate and N 2-acetyl-l-ornithine substrates were found similar for N-terminal and C-terminal His-tagged proteins, except for B. stearothermophilus ArgJ (see Table 1). Kinetic data showed that T. neapolitana OATase catalyzed the reaction about 10 times faster than B. stearothermophilus ArgJ at their respective optimal temperatures (see Table 1). The OATase transacetylation catalysis was found to be reversible, as the equilibrium of reaction corresponded to 40% of N-acetyl-l-glutamate


q FEBS 2000

l-glutamate as competitive substrate inhibition occurred in reactions performed at unbalanced substrate concentrations. The AGSase reaction initiated with acetylCoA as donor was found to be complete and irreversible under the conditions used, no acetylation of CoA was detected when the reaction mixture contained N-acetylglutamate and CoA in the presence of any of the bifunctional enzymes. Transacetylation was found to be less efficient from acetylCoA than from N 2-acetyl-l-ornithine: five and nine times lower for B. stearothermophilus and T. neapolitana enzymes, respectively (see Table 1). However, both bifunctional enzymes displayed almost three times higher the apparent substrate affinity towards acetylCoA than N 2-acetyll-ornithine. l-Ornithine (but not l-arginine) was found to inhibit AGSase activity whereas no inhibition of AGSase or OATase activities could be detected by CoA. Formation of acetylated-ArgJ intermediates. Considering the ordered ping-pong bi±bi behaviour [37] of ArgJ enzymes one would expect the formation of an acetylated protein intermediate during the transacetylation reaction. To confirm such a requirement the B. stearothermophilus or T. neapolitana bifunctional enzymes were incubated with radiolabelled acetylCoA or N-acetyl-l-glutamate as acetyl-group donors in the absence of acetyl-group acceptor. Radiolabelled protein intermediates could be recovered from reactions performed in the presence of each enzyme (but not BSA) and any of the two [14C]acetyllabelled donors used (Table 2). No radioactive protein was detected when the acetyl donor used was labelled at its glutamate moiety.

DISCUSSION

Fig. 7. Inhibition of the B. stearothermophilus N-terminal His-tagged ArgJ enzyme activity by l-ornithine (the OATase activity). The plots are representative of 1/V, where V is the ArgJ specific activity measured at initial velocity conditions, vs. 1/C, where C is N 2-acetyl-l-ornithine (A) or l-glutamate (B) concentrations in absence (X) and presence of increasing concentrations of l-ornithine 1.25 mm (W), 2.5 mm (), 5 mm (B) and 10 mm (A). Replot of slopes as a function of ornithine concentrations is presented in the insert. Similar plots were obtained for M. jannaschii and T. neapolitana ArgJ enzymes and kinetic data determined were similar irrespective of N- or C-terminal tagged protein versions (data not shown).

synthesized from N 2-acetyl-l-ornithine and l-glutamate or left from the reverse reaction, whatever the enzyme used. AcetylCoA : l-glutamate N-acetyltransferase activity. B. stearothermophilus and T. neapolitana ArgJ proteins also accept acetylCoA as an acetyl-group donor. Kinetic measurements for AGSase-mediated reactions were only performed at 10 mm

The arginine biosynthethic pathway starts by a transacetylation step involving acetylCoA and l-glutamate. In most procaryotes, a second transacetylation step recycles the acetyl group of N-acetylornithine on glutamate (see Fig. 1). In this study we show that argJ-encoded N-acetyltransferases from three evolutionary distant thermophilic microorganisms either catalyze only the acetyl group transfer from N-acetylornithine to glutamate (the OATase activity) or in addition, exhibit AGSase activity. Complementation and enzymatic data show that M. jannaschii ArgJ is a strictly monofunctional enzyme exhibiting only OATase activity. Contrary to the situation prevailing in S. cerevisiae [25] a residual AGSase activity cannot be detected with this archaeal enzyme, nor in crude extracts of recombinant E. coli argA mutant, nor with the purified enzyme. On the other hand, T. neapolitana possesses a bifunctional enzyme as already suggested previously for B. stearothermophilus [21] and confirmed here with purified enzyme preparations. The genomes of M. jannaschii [42] and T. maritima [30], closely related to T. neapolitana, have been completely sequenced. In M. jannaschii all arginine biosynthetic genes are scattered along the chromosome, whereas in T. maritima they are grouped in two separate argGHCJBD and purBA argF clusters. Both hyperthermophiles appear to lack the argA gene analogue. However, N-acetylglutamate synthase activity was detected in T. maritima [6], an observation that could be explained by ArgJ bifunctionality, though the existence of a AGSase cannot be excluded. Indeed, the origin of glutamate acetylation is far from clear. Enzymes which are not, or only distantly, related may catalyze this reaction, as for bifunctional OATases which are not homologous to ArgA proteins [22] or fungal and bacterial argA-encoded AGSases which share only

q FEBS 2000


weak similarity [23]. Furthermore, Vibrionaceae have recently been shown to harbour an extended argH gene able to complement E. coli argA mutants [5]. It is therefore possible that all AGSases do not have the same origin and that in different organisms different N-acetyltransferases were recruited to catalyze glutamate acetylation. Bacteria of the genus Thermotoga are, nevertheless, in principle able to synthesize arginine without an argA-encoded AGSase. The situation is more intriguing with M. jannaschii as the ArgJ protein of this organism appears unable to acetylate glutamate from the acetylCoA donor and an argA-like gene cannot be identified in the genome. Either this archaeon encodes an AGSase that cannot be identified by sequence similarity, or the Methanococcus protein displays in vivo an AGSase activity that was not detected under the conditions used. A systematic search for a gene able to complement an E. coli argA mutant could settle the question. ArgJ protein precursors from the three thermophilic microorganisms undergo cleavage between alanine and threonine residues to generate two a and b subunits in E. coli cytoplasm as found also for S. cerevisiae OATase [24]. The appearance of a 46-kDa polypeptide in C-terminal tagged protein patterns after purification by affinity chromatography indicates that a small portion of ArgJ remains unprocessed. The fact that the cleavage of both cytoplasmic ArgJ precursors takes place in the heterologous E. coli background, naturally devoid of OATase, or in yeast mitochondria between the same alanine and threonine residues present in a PXM-ATML motif, suggests that the maturation of active ArgJ is an autocatalytic rather than a protease-mediated process. Such an autocatalytic processing was shown to occur in bacterial glycosylasparaginase, where the nucleophilic side chain of threonine attacks the carbonyl side chain of the adjacent amino acid, generating N-O acyl shifted bond rearrangements with further liberation of two subunits from the single-chain precursor [43,44]. Anomalous migration was observed for both subunits of yeast OATase when comparing deduced molecular masses with experimental values [24,25]. We found that only b subunits of the three thermostable ArgJ enzymes studied show abnormal migration on SDS/PAGE. The reason for this is presently unknown but it could reflect postranslational modification of ArgJ. As compared to the yeast enzyme which forms an ab heterodimer [24], both bacterial and archaeal ArgJ proteins appear to assemble as functionally active a2b2 heterotetramers. This assumption should still be proved, however, a higher oligomerization order may be a prerequisite for higher thermostability as found in the case of thermostable ornithine carbamoyltransferase [45] and dihydrofolate reductase [46]. Whether or not thermophilic OATases also exhibit AGSase activity, they share a rather high sequence similarity all along their amino-acid sequences. The three ArgJ enzymes follow the same type of ordered ping-pong bi±bi kinetic mechanism via the formation of an acetylated-enzyme intermediate. As l-ornithine behaves as a competitive inhibitor with respect to l-glutamate, the latter appears to accept the acetyl-group only after releasing l-ornithine from the binding pocket. It is worth mentioning, that the absence of inhibition of AGSase or OATase activity by CoA does not rule out the ping-pong bi±bi kinetic mechanism. In fact, it indicates that the CoA-moiety of acetylCoA either does not enter into the binding pocket or that conformational change in a covalent enzyme intermediate prohibits the entrance for CoA but not for l-ornithine. The absence of acetylation of CoA from N-acetylglutamate supports this assumption in our interpretation of the kinetic data. Currently we use other approaches to address the question

whether the mono and bifunctionality of ArgJ is related to the accessibility of various donors to the same catalytic site, or to the presence of two different catalytic domains in the same protein. The argA gene encoded AGSase is feed-back inhibited by arginine in P. aeruginosa and E. coli [10±47], whereas both AGSase and OATase activities of bifunctional ArgJ from B. stearothermophilus and T. neapolitana are only inhibited by ornithine ([21] and this study). Therefore, l-ornithine (but not l-arginine) appears to be a key regulatory molecule for the first half of the arginine pathway metabolic flow in these thermophiles (see Fig. 1). Given this inhibitory role of ornithine and the existence of an ArgR regulatory protein in B. stearothermophilus [48,49] and in T. neapolitana [28], one may assume that the functional state of the whole arginine biosynthesis pathway is controlled both at the transcriptional and enzymatic levels in these thermophilic bacteria.

ACKNOWLEDGEMENTS We are indebted to M. Lecocq (UniversiteÂ de Nantes, France) and M. Demarez (Vrije Universiteit, Brussels, Belgium) for participation in some experiments and to D. Gigot (Vrije Universiteit, Brussels, Belgium) for synthesis of labelled N-acetyl-l-glutamate. We are grateful to D. Levitsky (UniversiteÂ de Nantes, France) for advice in performing Western-blot experiments and H. Rabesona (INRA, Nantes, France) for amino-acid sequencing of ArgJ proteins. This work was supported by a grant from the ReÂgion des Pays de la Loire (Contrat de Plan Etat-ReÂgion) and the Tournesol Programme (Collaborations Franco-Belges, CommunauteÂ Flamande, Dossier 99042) for mutual visits. F. M. is a postgraduate student supported by the French MinisteÁre de l'Enseignement SupeÂrieur et de la Recherche.

REFERENCES 1. Cunin, R., Glansdorff, N., Pierard, A. & Stalon, V. (1986) Biosynthesis and metabolism of arginine in bacteria. Microbiol. Rev. 50, 314±352. 2. Davis, R.H. (1986) Compartmental and regulatory mechanisms in the arginine pathways of Neurospora crassa and Saccharomyces cerevisiae. Microbiol. Rev. 50, 280±313. 3. Vogel, H.J. & MacLellan, W.L. (1970) Acetylornithinase (E. coli). Methods Enzymol. 17A, 265±269. 4. Harris, B.Z. & Singer, M. (1998) Identification and characterisation of the Myxococcus xanthus argE gene. J. Bacteriol. 180, 6412±6414. 5. Xu, Y., Liang, Z., Legrain, C., Ruger, H.L. & 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. Van de Casteele, M., Demarez, M., Legrain, C., Glansdorff, N. & Pierard, A. (1990) Pathways of arginine biosynthesis in extreme thermophilic archaeo- and eubacteria. J. Gen. Microbiol. 136, 1177±1183. 7. Udaka, S. & Kinoshita, S. (1958) Studies on l-ornithine fermentation. I. The biosynthetic pathway of l-ornithine in Micrococcus glutamicus. J. Gen. Appl. Microbiol. 4, 272±282. 8. De Deken, R.H. (1962) BiosyntheÁse de l'arginine chez la levure. 1. Le sort de la N a-acetylornithine. Biochim. Biophys. Acta 78, 606±616. 9. Hoare, D.S. & Hoare, S.L. (1966) Feedback regulation of arginine biosynthesis in blue-green algae and photosynthetic bacteria. J. Bacteriol. 92, 375±379. 10. Haas, D., Kurer, V. & Leisinger, T. (1972) N-acetylglutamate synthase of Pseudomonas aeruginosa. An assay in vitro and feedback inhibition by arginine. Eur. J. Biochem. 31, 290±295. 11. Shinners, E.N. & Catlin, B.W. (1978) Arginine biosynthesis in Neisseria gonorrhoeae: enzymes catalyzing the formation of ornithine and citrulline. J. Bacteriol. 136, 131±135. 12. Meile, L. & Leisinger, T. (1984) Enzymes of arginine biosynthesis in methanogenic bacteria. Experientia 40, 899±900.

5226 F. Marc et al. (Eur. J. Biochem. 267) 13. Sakanyan, V., Kochikyan, A., Mett, I., Legrain, C., Charlier, D., Pierard, A. & Glansdorff, N. (1992) A re-examination of the pathway for ornithine biosynthesis in a thermophilic and two mesophilic Bacillus species. J. Gen. Microbiol. 138, 125±130. 14. Hindle, Z., Callis, R., Dowden, S., Rudd, B.A. & Baumberg, S. (1994) Cloning and expression in Escherichia coli of a Streptomyces coelicolor A3(2)argCJB gene cluster. Microbiology 140, 311±320. 15. Baetens, M., Legrain, C., Boyen, A. & Glansdorff, N. (1998) Genes and enzymes of the acetyl cycle of arginine biosynthesis in the extreme thermophilic bacterium Thermus thermophilus. Microbiology 144, 479±492. 16. Wipe, B. & Leisinger, T. (1979) Regulation of activity and synthesis of acetylglutamate synthase from Saccharomyces cerevisiae. J. Bacteriol. 140, 874±880. 17. Messenguy, F. (1979) Concerted repression of the synthesis of the arginine biosynthetic enzymes by aminoacids: a comparison between the regulatory mechanisms controlling aminoacid biosynthesis in bacteria and in yeast. Mol. Gen. Genet. 169, 85±95. 18. Dharmsthiti, S. & Krishnapillai, V. (1993) DNA sequence conservation at the gene level in a conserved chromosomal segment in two Pseudomonas species. J. Genet. 72, 1±14. 19. Picard, F.J. & Dillon, J.R. (1989) Cloning and organization of seven arginine biosynthesis genes from Neisseria gonorrhoeae. J. Bacteriol. 171, 1644±1651. 20. Martin, P.R. & Mulks, M.H. (1992) Sequence analysis and complementation studies of the argJ encoding ornithine acetyl transferase from Neisseiria gonorrhoeae. J. Bacteriol. 174, 2694±2701. 21. Sakanyan, V., Charlier, D., Legrain, C., Kochikyan, A., Mett, I., Pierard, A. & Glansdorff, N. (1993) Primary structure, partial purification and regulation of key enzymes of the acetyl cycle of aginine biosynthesis in Bacillus stearothermophilus: dual function of ornithine acetyltransferase. J. Gen. Microbiol. 139, 393±402. 22. Sakanyan, V., Petrosyan, P., Lecocq, M., Boyen, A., Legrain, C., Demarez, M., Hallet, J.N. & Glansdorff, N. (1996) Genes and enzymes of the acetyl cycle of arginine biosynthesis in Corynebacterium glutamicum: enzyme evolution in the early steps of the arginine pathway. Microbiology 142, 99±108. 23. Yu, Y.G., Turner, G.E. & Weiss, R.L. (1996) Acetylglutamate synthase from Neurospora crassa: structure and regulation of expression. Mol. Microbiol. 22, 545±554. 24. Liu, Y.S., Van Heeswijck, R., Hoj, P. & Hoogenraad, N. (1995) Purification and characterization of OATase from Saccharomyces cerevisiae. Eur. J. Biochem. 228, 291±296. 25. Crabeel, M., Abadjieva, A., Hilven, P., Desimpelaere, J. & Soetens, O. (1997) Characterization of the Saccharomyces cerevisiae ARG7 gene encoding ornithine acetyltransferase, an enzyme also endowed with acetylglutamate synthase activity. Eur. J. Biochem. 250, 232±241. 26. Woese, C. (1998) The universal ancestor. Proc. Natl Acad. Sci. USA 95, 6854±6859. 27. Sakanyan, V.A., Hovsepyan, A.S., Mett, I.L., Kochikyan, A.V. & Petrosyan, P.K. (1990) Molecular cloning and structural-functional analysis of arginine biosynthesis genes of the thermophilic bacterium Bacillus stearothermophilius. Genetika (Russia) 26, 1915±1925. 28. Dimova, D., Weigel, P., Takahashi, M., Marc, F., Van Duyne, G.D. & Sakanyan, V. (2000) Thermostability, oligomerization and DNA binding properties of the ArgR regulatory protein from hyperthermophilic bacterium, Thermotoga neapolitana. Mol. Gen. Genet. 263, 119±130. 29. Miller, J.H. (1992) A short course in bacterial genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, USA. 30. Nelson, K.E., Clayton, R.A., Gill, S.R., Gwinn, M.L., Dodson, R.J., Haft, D.H. et al. (1999) Evidence for lateral gene transfer between

q FEBS 2000

31. 32. 33.

34. 35. 36. 37. 38. 39. 40.

41. 42. 43. 44. 45.

46.

47. 48.

49.

Archaea and Bacteria from genome sequence of Thermotoga maritima. Nature 399, 323±329. Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K. & Pease, L.R. (1989) Site directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77, 51±59. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing with chain terminating inhibitors. Proc. Natl Acad. Sci. USA 74, 5463±5467. Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W. & Lipman, D.J. (1997) Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389±3402. Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A. & Struhl, K. (1993). Current Protocols in Molecular Biology. John Wiley & Sons, New York, USA. Bradford, M.M. (1976) A rapid sensitive method for the quantitation of microgram quantities of protein utilising the principle of protein-dye binding. Anal Biochem. 72, 248±254. Matsudaira, P. (1987) Sequence from picomole quantities of proteins electroblotted onto polyvinylidene difluoride membranes. J. Biol. Chem. 262, 10035±10038. Cleland, W.W. (1973) Derivation of rate equations for multisite pingpong mechanisms with ping-pong reactions at one or more sites. J. Biol. Chem. 248, 8353±8355. Cleland, W.W. (1979) Substrate inhibition. Methods Enzymol. 63, 500±513. Savchenko, A. (1996) Etude geÂneÂtique de la reÂgion de controle de l'opeÂron arginine de la bacteÂrie thermophile Bacillus stearothermophilus. PhD Thesis. UniversiteÂ de Nantes, Nantes, France. Savchenko, A., Weigel, P., Dimova, D., Lecocq, M. & Sakanyan, V. (1998) The Bacillus stearothermophilus argCJBD operon harbours a strong promoter as evaluated in Escherichia coli cells. Gene 212, 167±177. Segel, I.H. (1975). Enzyme Kinetics: Behavior and Analysis of Rapid Equilibrium and Steady-State Enzyme Systems. John Wiley & Sons. New York, USA. Bult, C.J., White, O., Olsen, G.J., Zhou, L., Fleischmann, R.D., Sutton, G.G. et al. (1996) Complete genome sequence of the methanogenic archaeon, Methanococcus jannaschii. Science 273, 1058±1073. Liu, Y., Guan, C. & Aronson, N.N. Jr (1998) Site-directed mutagenesis of essential residues involved in the mechanism of bacterial glycosylasparaginase. J. Biol. Chem. 273, 9688±9694. Xu, Q., Buckley, D., Guan, C. & Guo, H.C. (1999) Structural insights into the mechanism of intramolecular proteolysis. Cell 98, 651±661. Villeret, V., Clantin, B., Tricot, C., Legrain, C., Roovers, M., Stalon, V., Glansdorff, N. & Van Beeumen, J. (1998) The crystal structure of Pyrococcus furiosus ornithine carbamoyltransferase reveals a key role for oligomerization in enzyme stability at extremely high temperatures. Proc. Natl Acad. Sci. USA 95, 2801±2806. Wilquet, V., Gaspar, J.A., Van de Lande, M., Van de Casteele, M., Legrain, C., Meiering, E.M. & Glansdorff, N. (1998) Purification and characterization of recombinant Thermotoga maritima dihydrofolate reductase. Eur. J. Biochem. 255, 628±637. Leisinger, T. & Haas, D. (1975) N-Acetylglutamate synthase of Escherichia coli regulation of synthesis and activity by arginine. J. Biol. Chem. 250, 1690±1693. Dion, M., Charlier, D., Wang, H., Gigot, D., Savchenko, A., Hallet, J.N., Glansdorff, N. & Sakanyan, V. (1997) The highly thermostable arginine repressor of Bacillus stearothermophilus: gene cloning and repressor±operator interactions. Mol. Microbiol. 25, 385±398. Ni, J.P., Sakanyan, V., Charlier, D., Glansdorff, N. & Van Duyne, G.D. (1999) Structure of the arginine repressor from Bacillus stearothermophilus. Nat. Struct. Biol. 6, 427±432.