Purification and kinetic analysis of the two recombinant arogenate dehydrogenase isoforms of Arabidopsis thaliana Pascal Rippert and Michel Matringe Laboratoire Mixte CNRS/INRA/Bayer CropScience (UMR 1932), Lyon, France The present study reports the first purification and kinetic characterization of two plant arogenate dehydrogenases (EC 1.3.1.43), an enzyme that catalyses the oxidative decarboxylation of arogenate into tyrosine in presence of NADP. The two Arabidopsis thaliana arogenate dehy- drogenases TyrAAT1 and TyrAAT2 were overproduced in Escherichia coli and purified to homogeneity. Bio- chemical comparison of the two forms revealed that at low substrate concentration TyrAAT1 is four times more efficient in catalyzing the arogenate dehydrogenase reac- tion than TyrAAT2. Moreover, TyrAAT2 presents a weak prephenate dehydrogenase activity whereas Tyr- AAT1 does not. The mechanism of the dehydrogenase reaction catalyzed by these two forms has been investi- gated using steady-state kinetics. For both enzymes, steady-state velocity patterns are consistent with a rapid equilibrium, random mechanism in which two dead-end complexes, E–NADPH–arogenate and E–NADP–tyro- sine, are formed. Keywords: Arabidopsis thaliana; tyrosine biosynthesis; arogenate dehydrogenase; isoforms; enzymatic properties. Archae, eubacteria, plants, and fungi are capable of synthesizing, de novo, the three aromatic amino acids phenylalanine, tyrosine, and tryptophan. The enzymes catalyzing these reactions are thus potentially useful targets for the development of new antibiotics, fungicides and herbicides. The broad-spectrum herbicide glyphosate that inhibits 5-enolpyruvyl shikimate 3-phosphate synthase is the best example of this potential [1]. For the biosynthesis of phenylalanine and tyrosine, alternative routes exist in nature (Fig. 1). In organisms such as Saccharomyces cerevisiae or Escherichia coli [2,3], phenylpyruvate and p-hydroxyphenylpyruvate serve as the direct precursors of phenylalanine and tyrosine, respectively. In most plant however, both amino acids are formed from a common precursor L -arogenate [4–7]. Furthermore, a widespread combination of the alternat- ive routes can be found. For example, in cyanobacteria, and some other microorganisms, both arogenate-to- tyrosine and phenylpyruvate-to-phenylalanine pathways exist [8–12]. In other bacteria, e.g. Pseudomonas aerugi- nosa and Zymomonas mobilis, the alternative pathways to phenylalanine and tyrosine coexist [13,14]. Whatever the synthetic route adopted, a fine tuning at the branch points is required to balance the flow of intermediates; a wide variety of control mechanisms have been reported, including multivalent allosteric feed back control by end products (reviewed in [15–17]). In most organisms, the aromatic amino acid biosynthetic pathway also plays a pivotal role in the production of precursors for a myriad of aromatic secondary metabolites engaged in a very diverse range of processes [18]. This opens the possibility to deregulate the pathway in favor of one of the three aromatic amino acids, by overexpressing a foreign enzyme with a completely different pattern of regulation. However, the existence of many different combinations for routing prephenate to phenylalanine or tyrosine means that the substrate specificity of a particular enzyme, and its susceptibility to feedback regulation by different metabolites is not predictable, and must be studied in detail. In a previous study [19], we reported on the identi- fication of two structural genes encoding plastidic Arabidopsis thaliana arogenate dehydrogenase named, respectively, tyrAAT1 (accession number AF434681), and tyrAAT2 (accession number AF434682), and the characterization of the recombinant protein TyrAAT1. This protein was structurally unusual in that a single polypeptide chain housed two highly similar domains. Our study revealed that both peptide domains sustained arogenate dehydrogenase activity with similar biochemi- cal characteristics. The second isoform, TyrAAT2, appeared later in the A. thaliana databases [20] and did not present these repeated peptide domains. The aim of the present study was to conduct a biochemical compar- ison of these two isoforms of arogenate dehydro- genase and a detailed kinetic analysis of their substrate specificity and reaction mechanisms to gain further insight into their respective roles in tyrosine metabolism. To date, reports of such mechanistic studies have been limited to a related bifunctional enzyme choris- mate mutase-prephenate dehydrogenase from bacteria [3,21,22]. Correspondence to M. Matringe, Laboratoire Mixte CNRS/INRA/ Bayer CropScience (UMR 1932), Bayer CropScience, 14–20 rue Pierre Baizet, 69263 Lyon cedex 9, France. Fax: + 33 472 85 22 97, Tel.: + 33 472 85 28 47, E-mail: michel.matringe@bayercropscience.com Abbreviations: TyrAAT1, Arabidopsis thaliana arogenate dehydrogenase isoform 1; TyrAAT2, Arabidopsis thaliana arogenate dehydrogenase isoform 2. (Received 7 June 2002, revised 2 August 2002, accepted 6 August 2002) Eur. J. Biochem. 269, 4753–4761 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03172.x MATERIALS AND METHODS Materials Isopropyl thio-b- D -galactoside was purchased from Bio- probe. Prephenate, NAD, NADP, benzamidine HCl, amino caproic acid were obtained from Sigma. Arogenate was synthesized enzymatically from prephen- ate and aspartate as described by Rippert & Matringe [19]. Arogenate was further purified from the reaction using the following steps as described by Jensen et al.[28]. The arogenate solution in 1 m M potassium phosphate, pH 8.0 was passed through a Dowex-chloride column (1.6 · 30 cm, Sigma) at a flow rate of 2 mLÆmin )1 .The column was then washed with 100 mL of 1 m M potassium phosphate, pH 8.0. Arogenate is eluted from the Dowex- chloride column with a 100 mL of 0.5 M NH 4 HCO 3 . Arogenate eluted was then lyophilized, resuspended in 50 m M Hepps pH 8.6 and stored frozen at )20 °C. Bacterial strains The E. coli AT 2471 [23], tyrA relA spoT thi mutant was purchased from E. coli Genetic Stock Center. This mutant was lysogenized with the helper phage (kDE3) harboring a copy of the T7 RNA polymerase gene, using the kDE3 lysogenization kit from Novagen, according to manufac- turer’s instructions. The resulting E. coli AT 2471(kDE3) was used to express cDNAs cloned in the pET vector under the control of the T7 promoter. Electro-competent cells of this strain, prepared according to the method of Dower et al. [24], were transformed with the pET28-TyrAAT1 and pET21-TyrAAT2 constructs. Engineering of expression vector BLAST search using available prephenate and arogenate dehydrogenase protein sequences allowed us to identify two A. thaliana arogenate dehydrogenase genes (accession num- ber AF434681 and AF434682) [19]. The corresponding full- length cDNAs were obtained by PCR amplification of an Arabidopsis (var. Columbia) cDNA library constructed in pYes. The cloning of tyrAAT1 was previously described [19] under the name tyrAATc.TheNdeI–BamHI DNA frag- ments containing the coding sequence without the transit peptide [19] was cloned into the plasmid pET28 b(+) to place a 6·His tag at the NH 2 terminus of the polypeptide. According to the CHLOROP prediction [30] the putative mature protein started at position 36 from the first methionine [19]. The 5¢-oligonucleotide P1 (5¢-CACTAC TCACAATGCTACTCCATTTCTCTCCG-3¢ and the 3¢-oligonucleotide P2 (5¢-GCATAATCCAG GATCCC TTGTGATCTTAAGATG-3¢) were used for PCR ampli- fication of the full length tyrAAT2 cDNA. This 3¢-oligonu- cleotide is complementary to the beginning of the 3¢UTR, and introduced a BamHI restriction site (underlined). The full-length PCR fragment was cloned into the plasmid pPCR-Script (Stratagene). Constructs lacking the transit peptide sequence were produced according to the CHLOROP prediction [30] by replacing the arginine in position 36 by a methionine. We used P3 (5¢-CTCTTCGAATT CATATG ATCGACGCCGCCC-3¢) that introduced an ATG initi- ator and a NdeI restriction site (underlined) in position 108 from the first in frame ATG codon and P2. Then the resulting PCR NdeI–BamHI DNA fragment was cloned into the plasmid pET21 a(+), yielding the plasmid pET21- TyrAAT2. For both constructs, sequencing of the entire insert was carried out and was in complete agreement with the expected sequences. The theoretical masses of the products TyrAAT1 and TyrAAT2 were determined to be 68.5 kDa and 36 kDa, respectively. Expression and purification of recombinant TyrAAT1 The E. coli AT 2471(kDE3) cells transformed with pET28- TyrAAT1 were grown at 37 °C in Luria–Bertani medium supplemented with the appropriate antibiotics. When D 600 reached 0.6, 1 m M of isopropyl thio-b- D -galactoside was added to induce recombinant protein synthesis. The cells were further grown for 16 h at 28 °C, harvested, and centrifuged for 20 min at 4000 g. The pellet was resus- pended in buffer A (20 m M Tris/HCl, pH 7.9, 500 m M NaCl) containing 5 m M imidazole and sonicated with a Vibra-cell disrupter (Sonics and Materials, Danbury, CT, USA) (100 pulses every 3 s on power setting 5). The crude extract obtained was centrifuged for 20 min at 18 000 g. The soluble protein extract was applied to a Ni-nitrilo- triacetic acid agarose column (1.6 · 3 cm; Qiagen) column previously equilibrated with 30 mL of buffer A containing 5m M imidazole. The column was washed with buffer A containing 60 m M imidazole. The recombinant protein was eluted using buffer A containing 250 m M imidazole. Frac- tions containing the major arogenate dehydrogenase acti- vity were pooled and dialyzed against buffer A. Enzyme was stored at 4 °C without loss of activity during several months. Fig. 1. Biosynthesis pathway leading to tyrosine and phenylalanine. The alternative for the synthesis of tyrosine and phenylalanine from pre- phenate are represented. EPSP synthase, 5-enolpyruvyl shikimate 3-phosphate synthase. 4754 P. Rippert and M. Matringe (Eur. J. Biochem. 269) Ó FEBS 2002 Expression and purification of recombinant TyrAAT2 The E. coli AT 2471(DE3) strain transformed with pET21- TyrAAT2 were grown at 37 °C in Luria–Bertani medium supplemented with the appropriate antibiotics. When D 600 reached 0.6, 1 m M of isopropyl thio-b- D -galactoside was added to induce recombinant protein synthesis. The cells were further grown for 16 h at 28 °C, harvested, and centrifuged for 20 min at 4000 g. The pellet was resuspended in buffer B containing 50 m M Tris/HCl, pH 7.5, 1 m M EDTA, 1 m M dithiothreitol, 1 m M benzamidine HCl, 5m M amino caproic acid, and sonicated with a Vibra-cell disrupter (Sonics and Materials) (100 pulses every 3 s on power setting 5). The crude extract obtained was centrifuged for 20 min at 18 000 g. The soluble protein extract was subjected to ammo- nium sulfate fractionation by addition of solid (NH 4 ) 2 SO 4 (20% saturation) at 4 °C. After 20 min of stirring, the mixture was centrifuged at 40 000 g for 20 min and the supernatant was brought to 40% saturation with solid (NH 4 ) 2 SO 4 at 4 °C. The precipitate was recovered by centrifugation and resuspended in a minimum volume of buffer A. The resulting protein extract was applied to a Hiload Superdex S200 (3.2 · 60 cm, Pharmacia), column connected to an FPLC system (Pharmacia) previously equilibrated in 200 mL of buffer B. Fractions enriched with arogenate dehydrogenase were pooled and the enzyme was stored at 4 °C without loss of activity during several month. Protein concentration Total protein concentration was determined with the Bio- Rad protein assay with c-globulin as the standard, as described by Bradford [25]. The concentration of purified arogenate dehydrogenase was also determined by measur- ing the absorbance at 205 nm [26]. Electrophoretic analyses of proteins For the analysis of the protein purification, polypeptides were separated by SDS/PAGE containing 12% (w/v) acrylamide as detailed by Chua [27] and visualized by staining with Coomassie Brilliant Blue R250. Enzyme activities To monitor protein purification, prephenate and arogen- ate dehydrogenase activities were assayed according to Bonner & Jensen [29] by following at 25 °Ctheformation of NADH or NADPH at 340 nm in a buffer containing 50 m M Tris/HCl, pH 7.5, 300 l M prephenate or arogen- ate, and 1 m M NAD or NADP in a total volume of 200 lL. Enzyme activity is expressed in UÆmg )1 ,were1Uis defined as 1 lmol NAD(P)H formed per min. Data analysis Kinetic data were analyzed with KALEIDAGRAPH program (Abelbeck Software) providing an iterative fit to the appropriate equation by using a nonlinear curve-fitting method. Hyperbolic curve were fitted to the following equation: v ¼ V app m :½S K app m þ½S ð1Þ And sigmoidal curve were fitted to the following equation: v ¼ V app m :½S n K app m þ½S n ð2Þ where V app m and K app m are the apparent V m and K m values for one substrate at different fixed value for the concentration of the second substrate and n is the Hill coefficient. RESULTS Two forms of arogenate dehydrogenase were previously identified in the A. thaliana databanks, tyrAAT1 (accession number AF434681), and tyrAAT2 (accession number AF434682) [19]. tyrAAT1 was previously described under the name tyrAATc [19]. According to the CHLOROP predic- tion program [30], the two forms present a putative plastidic transit peptide sequence. This finding is in good agreement with a plastidic localization of all other enzyme activities involved in aromatic amino acids biosynthesis (reviewed in [16,17]). We have reported previously that TyrAAT1 is comprised of two highly similar peptide domains Tyr- AAT1.1 and TyrAAT1.2 (Fig. 2A). A biochemical analy- sis carried out on crude protein extracts revealed that both domains possess arogenate dehydrogenase activity [19]. The second isoform, TyrAAT2, does not contain repeated domains (Fig. 2A). Sequence comparison between TyrAAT2 and each of the two peptide domains of TyrAAT1 revealed more than 50% identity in the two cases (Fig. 2B). Purification of TyrAAT1 and TyrAAT2 The two recombinant isoforms of A. thaliana arogenate dehydrogenase TyrAAT1 and TyrAAT2 lacking their putative plastidic transit peptide were overproduced in the E. coli mutant strain AT 2471(DE3) cells devoid of endogenous prephenate dehydrogenase activity [23]. Both isoforms were purified to near homogeneity (more than 95% of purity) as assessed by denaturing PAGE and visualization by Coomassie Blue staining. As documented in Fig. 3, in denaturing condition both purified proteins have the expected molecular mass deduced from their coding sequences, i.e. 66 vs. 68 kDa and 38 vs. 37 kDa for TyrAAT1 and TyrAAT2, respectively. Recombinant TyrAAT1 was never recovered in large amounts in the soluble protein fraction [19], we thus decided to carry out a purification of a recombinant His-tagged TyrAAT1 in a one-step procedure via Ni-nitrilotriacetic acid affinity chro- matography. This purification process resulted in a 700-fold enhancement of the specific activity (Table 1). About 1.4 mg of enzyme was routinely obtained from 1 g of soluble proteins. TyrAAT2 was expressed without its transit peptide as described in Materials and methods. The resulting protein was successfully overproduced in AT 2471(DE3) cells and was recovered mainly as soluble protein. Recombinant TyrAAT2 arogenate dehydrogenase was excluded from a gel filtration S200 column (Fig. 3B). Therefore, a two step purification strategy was adopted for Ó FEBS 2002 Arogenate dehydrogenases of A. thaliana (Eur. J. Biochem. 269) 4755 this enzyme and consisted of precipitation in a 20–40% ammonium sulfate followed by chromatography on a S200 gel filtration column (Fig. 3A, Table 1). This purification resulted in a 12-fold enhancement in specific activity. About 17 mg of enzyme was routinely obtained from 1 g of soluble protein by this procedure. Size exclusion chromatography using Superdex S200 revealed that TyrAAT1 eluted with a mobility consistent with that of a monomeric protein of 67 kDa (Fig. 3B), and in total agreement with the mass estimated by denaturing PAGE(i.e.66kDa,Fig.3A).Incontrast,TyrAAT2eluted in the void volume of the S200 column even in presence of 300 m M NaCl (Fig. 3B). A similar behavior was found with the TyrAAT2 arogenate activity found in crude protein extract of cultured cells from A. thaliana. This finding, Fig. 2. Schematic diagram of the amino acid sequence of the two A. thaliana arogenate dehydrogenases TyrAAT1 and TyrAAT2 deduced from their respective coding sequences (A) and alignment of amino acid sequence of the two protein domains of TyrAAT1 (TyrAAT1.1, and TyrAAT1.2), and TyrAAT2 using the CLUSTALW program (B). (A) Putative transit peptide cleavage site determined using CHLOROP pro- gram are indicated (i.e. Val36 for TyrAAT1 and Arg 36 for TyrAAT2). (B) Conserved residues are noted by *, and high similarity residues are noted by colons. Fig. 3. Purification for the two A. thaliana arogenate dehydrogenases TyrAAT1 (left) and TyrAAT2 (right) monitored by SDS/PAGE (A) and behavior of TyrAAT1 and TyrAAT2 by chromatography on Superdex 16/60 S200 (B). (A) Polypeptides were separated on SDS/ PAGE gels and stained with Coomassie brillant blue R-250. Lane 1: crude soluble extract (15 lg) of E. coli AT 2471(DE3) cells trans- formed with pET28-TyrAAT1. Lane 2: Fraction containing TyrAAT1 eluted from a Ni-nitrilotriacetic acid agarose column by a 250-m M imidazole buffer (1 lg). Lane 3: crude soluble extract (15 lg) of E. coli AT 2471(DE3) cells overproducing TyrAAT2. Lane 4: ammonium sulfate 20–40% fractionation of crude soluble extract of E. coli AT 2471(DE3) cells overproducing TyrAAT2 (10 lg). Lane 5: Fraction containing TyrAAT2 eluted from a S200 gel filtration column (2 lg). The minor polypeptide of 72 kDa resolved in lane 5 correspond to a dimeric form of TyrAAT2. This was determined by Western blot analysis. MW: molecular mass markers. (B) A 0.5-mL sample of purified TyrAAT1 or TyrAAT2 (1 mg) were applied to the S200 column and eluted at 1.5 mLÆmin )1 in a buffer containing 50 m M Tris/HCl, pH 7.5. The straight line represents the profile of activity obtained with TyrAAT1 and the dashed line the profile of activity obtained with TyrAAT2. TyrAAT2 eluted in the void volume of the S200. TyrAAT1 eluted with a Ve/Vo of about 1.7 which correspond to an apparent mass of 67 kDa as determined by comparison with mobility of standard protein (insert panel). Standards used (1 mg) were thyroglobulin (669 kDa), apoferritin (443 kDa), b-amylase (200 kDa), alcohol dehydrogenase (150 kDa) and bovine serum albumin (66 kDa). 4756 P. Rippert and M. Matringe (Eur. J. Biochem. 269) Ó FEBS 2002 together with the result from SDS/PAGE, indicates that native and recombinant TyrAAT2 form oligomers with a molecular mass greater than 600 kDa. Kinetic analyses The two enzymes are strictly NADP-dependent. Both forms obeyed Michaelis–Menten behavior at saturating concen- trations of the second substrate. The presence of a His-tag at the NH 2 terminus of TyrAAT1 did not modify its kinetics properties as its K m values for NADP and arogenate were very similar to the values previously determined with the unpurified recombinant protein devoid of His-tag [19]. Michaelis–Menten constants for arogenate and NADP are presented in Table 2. K m values for NADP were found to be 10.2 l M and 14.3 l M for TyrAAT1 and TyrAAT2, respect- ively. They are in the same order of magnitude as that previously reported for unpurified plant arogenate dehy- drogenases [5,6,19,31]. The K m values for arogenate were 52.6 and 84.2 l M for TyrAAT1 and TyrAAT2, respectively (Table 2). The K m values for arogenate reported in the literature for other plant enzymes differ significantly. They range from 67 to 340 l M [4–6,19,31]. Values for the maximal velocity (V m ) obtained for purified TyrAAT1 and TyrAAT2 were 142 and 73 UÆmg )1 , respectively. They are higher than those reported for prephenate dehydrogenase [3,21], and are of the same order of magnitude as that reported for the arogenate dehydrog- enase purified from Phenylobacterium immobile [12], and from Synechocystis (Rippert & Matringe, in preparation). The V m of TyrAAT1 is twice that of TyrAAT2. If we consider that the two active sites of TyrAAT1 are equivalent, values for the turn over number (k cat )were determined as 84 s )1 for one domain of TyrAAT1 and 37 s )1 for one TyrAAT2 monomer. Values for the catalytic efficiency (k cat /K m ) for arogenate were estimated at 1.6 · 10 6 M )1 Æs )1 and 0.44 · 10 6 M )1 Æs )1 for TyrAAT1 and TyrAAT2, respectively (Table 2). The two isoforms differ also with respect to their capacity to use prephenate as an alternative substrate. TyrAAT1 was incapable of cata- lyzing the transformation of prephenate to p-hydroxyphe- nylpyruvate regardless of the concentration of prephenate tested(i.e.upto1m M ). However, prephenate at high concentrations could bind to the active site as it behaves as a competitive inhibitor with respect to arogenate (not shown). Its apparent K i value was estimated at 4.2 m M (Table 3). In contrast, the second isoform TyrAAT2 exhibits weak prephenate dehydrogenase activity. From our data, its K m and k cat values could be estimated as 17 m M and 3.4 s )1 , respectively, and the catalytic efficiency for prephenate (k cat /K m )was2· 10 2 M )1 Æs )1 (Table 2). Therefore, at nonsaturating concentration of prephenate (< 17 m M ), the reaction is 2000 times less efficient in catalyzing the reaction with prephenate than with arogenate (0.44 · 10 6 M )1 Æs )1 vs. 2 · 10 2 M )1 Æs )1 ) (Table 2). This prephenate dehydrogenase activity could not be attributed to an endogenous E. coli activity that would coelute with arogenate dehydrogenase as the recombinant TyrAAT2 protein was overproduced in the E. coli strain AT 2471, devoid of prephenate dehydrogenase activity [23]. As expected for an alternate substrate, prephenate at high concentrations exhibited competitive inhibition with respect to arogenate (not shown). An apparent K i value for prephenate was determined to % 2.4 m M (Table 3). Finally, Table 2. Kinetic properties of the two purified A. thaliana arogenate dehydrogenases TyrAAT1 and TyrAAT2. KALEIDAGRAPH program was used to analyze data. Hyperbolic curves were fitted to Eqn (1). To determine K m NADP, plots of the velocity as a function of increasing concentrations of NADP (8–1000 l M ) at a fixed concentration of arogenate (300 l M ) was used. To determine K m arogenate, plots of the velocity as a function of increasing concentrations of arogenate (0–300 l M ) at a fixed concentration of NADP (1000 l M )wereused.K m values were determined using Eqn (1) with a nonlinear curve-fitting of plots. TyrAAT1 TyrAAT2 K m NADP (l M ) 10.2 ± 1.4 14.7 ± 1.1 K m arogenate (l M ) 52.6 ± 4.3 84.2 ± 5.6 k cat arogenate (s )1 ) 84.2 ± 3.4 37.3 ± 2.6 k cat /K m arogenate ( M )1 Æs )1 ) (1.6 ± 0.3) · 10 6 (4.4 ± 0.4) · 10 5 K m prephenate (l M ) – 17000 ± 7200 k cat prephenate (s )1 ) – 3.4 ± 0.7 k cat /K m prephenate ( M )1 Æs )1 ) – (2 ± 0.7) · 10 2 Table 1. Purification of the two A. t haliana arogenate dehydrogenases TyrAAT1 and TyrAAT2. NTA, nitrilotriacetic acid. Total protein (mg) Total activity (U) Specific activity (UÆmg protein )1 ) Recovery (%) Purification -fold TyrAAT1 Crude extract 750 159 0.21 100 1 Ni-NTA pool 1 142 142 90 700 TyrAAT2 Crude extract 1749 9290 5.31 100 1 (NH 4 ) 2 SO 4 755 7310 9.68 79 1.8 S200 pool 30 2190 73 23.5 11.7 Ó FEBS 2002 Arogenate dehydrogenases of A. thaliana (Eur. J. Biochem. 269) 4757 the arogenate activities of both isoforms were totally insensitive towards concentration up to 1 m M of p-hydroxy- phenylpyruvate, phenylalanine or tryptophan, and highly sensitive towards inhibition by tyrosine, the product of the reaction (see below). Kinetic mechanism Steady-state velocity studies in the absence of product. The kinetic mechanism of both enzymes was investigated by varying the concentration of one substrate at fixed concentrations of the second substrate. For both enzymes, initial velocity patterns obtained by varying the concentra- tions of NADP (4–500 l M ) at fixed concentrations of arogenate (90–250 l M ) were linear and intersected to the left of the vertical axis (Fig. 4B,D). Similar patterns were obtained by varying the concentrations of arogenate (10–250 l M ) at fixed concentrations of NADP (50–100 l M ) (data not shown). Thus, under these condi- tions, the reaction conforms to a simple sequential mech- anism for both isoforms. However, when NADP is used at low concentrations (between 10 and 30 l M ), with arogenate as the variable substrate sigmoı ¨ dal, curves were obtained (Fig. 4A,C), indicating positive kinetic cooperativity with respect to the binding of arogenate for both enzymes. When the data in Fig. 4A and C were fitted to the Hill equation, Hill coefficient values of 1.8 and 1.6 were obtained for TyrAAT1 and TyrAAT2, respectively. When NADP was the variable substrate, in both cases, double reciprocal plots were linear (Fig. 4B,D). Thus there is no kinetic coopera- tivity for the interaction of NADP with the enzyme. Table 3. Values for the apparent inhibition constant of the two arogenate dehydrogenases TyrAAT1 and TyrAAT2 by product and substrate analogues. The concentrations of each inhibitor and fixed or varied substrate used are given in the table. ND, not determined; C, competitive; NC, noncompetitive; M, mixed type inhibition. Values for the apparent inhibition were obtained using equations (a), (b) and (c) for competitive, noncompetitive and mixed inhibition, respectively. K m and V m values used in equation were directly determined from hyperbolic curve or sigmoidal curve using Eqns (1) and (2). (a) K i ¼ V max :½I V m À V app m (b) K i ¼ K m :½I K app m À K m and (c) K i ¼ ½I Slope 1 ½s Á V max K max À V app max Inhibitor Substrate Type of inhibitionVaried Fixed Apparent K i (l M ) NADPH (0, 50, 100 l M ) Arogenate NADP (20 l M ) NC TyrAAT1 53.9 ± 11 TyrAAT2 ND NADP Arogenate (200 l M ) C TyrAAT1 58.8 ± 6 TyrAAT2 ND Tyrosine (0, 50, 100 l M ) Arogenate NADP (50 l M ) C TyrAAT1 8.1 ± 0.3 TyrAAT2 7.5 ± 0.4 NADP Arogenate (70 l M ) NC TyrAAT1 14.2 ± 0.4 TyrAAT2 16.6 ± 3.4 AMP (0, 10, 20 l M ) Arogenate NADP (50 l M ) M TyrAAT1 25 900 ± 900 TyrAAT2 17 800 ± 6200 NADP Arogenate (200 l M ) C TyrAAT1 115 000 ± 9000 TyrAAT2 73 900 ± 11 000 Prephenate (0, 1, 2 m M ) Arogenate NADP (100 l M ) C TyrAAT1 4200 ± 800 TyrAAT2 2400 ± 400 NADP Arogenate (70 l M ) M TyrAAT1 20 200 ± 1100 TyrAAT2 22 500 ± 2500 cis-Aconitate (0, 10, 20 l M ) Arogenate NADP (100 l M ) C TyrAAT1 27 700 ± 5100 TyrAAT2 5800 ± 800 NADP Arogenate (200 l M ) M TyrAAT1 42 900 ± 9600 TyrAAT2 25 300 ± 2000 Fig. 4. Effects of substrate concentration on the velocity of the two arogenate dehydrogenases TyrAAT1 (upper panels), and TyrAAT2 (lower panels). (A) and (C) Plots of the velocity as a function of increasing concentrations of arogenate (0–250 l M )atfixedconcen- trations of NADP (10, 20 and 30 l M ). Sigmoidal curves were fitted to Eqn (2) using KALEIDAGRAPH program. (B) and (D) Lineweaver–Burk representations of velocity as a function of increasing concentrations of NADP (8–1000 l M ) at fixed concentrations of arogenate (20, 30 and 40 l M ). 4758 P. Rippert and M. Matringe (Eur. J. Biochem. 269) Ó FEBS 2002 Steady-state velocity studies in the presence of product and product or substrate analogues. Patterns of inhibition by products and substrate or product analogues were constructed to determine if the reactions catalyzed by TyrAAT1 and TyrAAT2 conform to a rapid equilibrium random or steady-state ordered kinetic mechanism. Inhibi- tion by NADPH was monitored in the presence of 20 l M NADP as no inhibition was observed at saturating concentrations of NADP. At 20 l M of NADP double reciprocal plots were concave up due to the positive cooperativity for the interaction of arogenate with TyrAAT1. Replots of 1/v vs. 1/s 2 were used to linearize the curves (Fig. 5Ai) [32]. Inhibition by NADPH was linear noncompetitive with respect to arogenate, and competitive with respect to NADP (Fig. 5Ai,ii). K i values for NADPH were found to be about 58 l M (Table 3). For TyrAAT2, nonlinear plots were also observed when varying arogenate at a fixed concentration of NADP. However, NADPH did not inhibit the reaction at the concentrations tested (not shown), indicating that TyrAAT2 was less sensitive towards inhibition by NADPH than TyrAAT1. Lineweaver–Burk plots revealed that for both enzymes, tyrosine causes linear competitive inhibition with respect to arogenate (Fig. 5Bi,iii), and noncompetitive inhibition with respect to NADP (Fig. 5Bii,Biv). K i values for tyrosine were found to be about 8 and 7 l M for TyrAAT1 and TyrAAT2, respectively (Table 3). In order to confirm the kinetic mechanism, the inhi- bitory effect of AMP and cis-aconitate as potential analogues of NADP and arogenate, respectively [21] were also tested. Lineweaver–Burk plots revealed that AMP acts as a dead-end inhibitor, giving rise to inhibition that is linear competitive with respect to NADP, and linear mixed-type inhibition with respect to arogenate. cis-Aconi- tate exerts a competitive inhibition with respect to arogenate and mixed-type competitive with respect to NADP (not shown). As reported above, for both isoforms prephenate acts as a competitive inhibitor, with respect to arogenate, and noncompetitive with respect to NADP (not shown). DISCUSSION The shikimate pathway plays a pivotal role in providing the cell with three aromatic amino acids tyrosine, phenyl- alanine and tryptophan, and most of the aromatic secondary metabolites [18]. In certain circumstances, for example in woody plants, at least 30% of the carbon fixed during photosynthesis is incorporated by this pathway for the synthesis of lignin via phenylalanine [33]. Detailed kinetics analysis of branch point enzymes of this pathway is thus absolutely required for understanding the partition of the carbon flux between the different end products. The present study is the first to report the purification and detailed kinetic characterization of A. thaliana arogenate dehydrogenase. Both isoforms exhibit initial velocity patterns in the absence and presence of products and dead end inhibitors consistent with a rapid equilib- rium random kinetic mechanism and the formation of two dead-end complexes, enzyme–NADP–tyrosine and enzyme–NADPH–arogenate. Conclusions about the order of product release are limited because it was not possible to determine the inhibition patterns with bicarbonate (CO 2 ). A similar kinetic mechanism has also been reported for the NAD-dependent prephenate dehydrogenase activity of the bifunctional enzyme chorismate mutase-prephenate dehydrogenase from Aerobacter aerogenes and E. coli [3,21]. The two monofunctional plant enzymes also exhibit positive kinetic cooperativity in the binding of arogenate, paralleling those findings for the interaction of prephenate with chorismate mutase-prephenate dehydrogenase [35]. NADP increases the apparent affinity of the plant enzymes for arogenate but arogenate did not alter the affinity of the enzymes for NADP. Both A. thaliana arogenate dehydro- genases were very sensitive toward inhibition by the product of the reaction (tyrosine); in fact they exhibit a higher affinity for tyrosine, than for their arogenate substrate. This strong feedback inhibition arises from the necessity of the plant cell, to regulate the flux of arogenate between the synthesis of tyrosine and that of phenylalanine. Indeed, the monofunctional plants arogenate dehydrogenase and E. coli chorismate mutase-prephenate dehydrogenase are at the branch point of the synthesis of tyrosine and Fig. 5. Product inhibition of the two arogenate dehydrogenases Tyr- AAT1 and TyrAAT2. (A) (i) plot representing 1/v function of 1/[aro- genate] 2 (4–400 l M ) at fixed concentrations of NADP (20 l M )andat fixed concentrations of NADPH (0, 50 and 100 l M ). (A) (ii) Line- weaver–Burk representation of velocity as a function of increasing concentrations of NADP (3–1000 l M ) at fixed concentrations of aro- genate (200 l M ) and at fixed concentrations of NADPH (0, 50 and 100 l M ). (B) (i) and (iii) Lineweaver–Burk representations of velocity as a function of increasing concentrations of arogenate (5–400 l M )at fixed concentrations of NADP (100 l M ) and at fixed concentrations of tyrosine (0, 50 and 100 l M ). B (ii) and (iv) Lineweaver–Burk repre- sentations of velocity as a function of increasing concentrations of NADP (5–1000 l M ) at fixed concentrations of arogenate (70 l M )and at fixed concentrations of tyrosine (0, 5 and 10 l M ). Ó FEBS 2002 Arogenate dehydrogenases of A. thaliana (Eur. J. Biochem. 269) 4759 phenylalanine and they are both inhibited by tyrosine [5,6,36]. However, in the case of E. coli enzyme tyrosine binds to an allosteric site and an active tyrosine–enzyme– prephenate complex could be formed [37]. Under these conditions, complete inhibition by tyrosine would never be reached even at high tyrosine concentrations. In contrast, the plant enzymes could be completely inhibited by tyrosine as tyrosine is a competitive inhibitor with respect to arogenate. When it is required, this allows the plant to utilize the majority of the arogenate formed for the synthesis of phenylalanine. In Synechocystis and Actinomyces mis- souriensis, arogenate dehydrogenase is not at branch point between tyrosine and phenylalanine as phenylalanine is synthesized via phenylpyruvate. Accordingly, these enzyme activities are not regulated by tyrosine [10,38] (P. Rippert & M. Matringe, unpublished results). The comparison of the kinetics properties of these two isoforms revealed that the peculiar structure of TyrAAT1, i.e. its repeated peptide domains, confers to this enzyme a catalytic efficiency (k cat /K m ) four times greater than that of TyrAAT2 (Table 2). This may have important physiologi- cal significance in the partition of the flux of arogenate into tyrosine or phenylalanine because arogenate dehydrogen- ases are branch point enzymes in competition with arogen- ate dehydratases for arogenate (Fig. 1). The lower substrate specificity of TyrAAT2, probably explains its lower effi- ciency in catalyzing the arogenate dehydrogenase reaction. The high K m for prephenate and the low specific activity of this prephenate activity, indicate that it is a side reaction with no physiological significance. Interestingly, a BLAST search allowed us to identify a gene fragment from Brassica oleracea (accession number BH495674), encoding a putative arogenate dehydrogenase with two highly similar peptide domains, indicating that this peculiar structure is not restricted to A. thaliana, and may also be present in other plant species. Immunological analyses, which are presently under investigation in our laboratory, will help us to address this issue. The presence of two plastidic isoforms with different kinetic behavior raised the question of their respective physiological roles. The presence of several plastidic isoforms is a general feature for the enzymes of the aromatic amino acid pathway. Indeed, the complete sequence of the A. thaliana genome [20] revealed two plastidic isoforms for 5-enolpyruvyl shikimate 3-phosphate synthase and chorismate mutase, three for 2-ceto-3-deoxy- D -arabino-heptulosonate 7-phosphate synthase and shiki- mate kinase, and not less than six for arogenate dehydratase. Analysis of the patterns of expression of the two A. thaliana arogenate dehydrogenases in different organs and in response to different environmental condi- tions by real-time PCR will help us to address the question of their physiological roles. These studies are underway in our laboratory. ACKNOWLEDGEMENTS We are grateful to Roland Douce, Claude Alban, Gilles Curien and Renaud Dumas for critical reading of the manuscript. This work was supported by Centre National de la Recherche Scientifique, by the Institut National de la Recherche Agronomique and by Aventis CropScience. REFERENCES 1. Steinrucken, H.C. & Amrhein, N. (1980) The herbicide glyphosate is a potent inhibitor of 5-enolpyruvyl-shikimic acid-3-phosphate synthase. Biochem. Biophys. Res. Commu. 94, 1207–1212. 2. Lingens, F., Goebel, W. & Uesseler, H. (1967) Regulation der aromatischen aminosaˆ uren in Saccharomyces cerevisiae. Eur. J. Biochem. 1, 363–374. 3. Sampathkumar, P. & Morrisson, J.F. (1982) Chorismate mutase- prephenate dehydrogenase from Escherichia coli.Purificationand properties of the bifunctional enzyme. Biochim. Biophys. Acta 702, 204–211. 4. Byng, G., Whitaker, R., Flick, C. & Jensen, R.A. (1981) Enzymology of 1-tyrosine biosynthesis in corn (Zea Mays). Phytochemistry 6, 1289–1292. 5. Gaines, C.G., Gyng, G.S., Whitaker, R.J. & Jensen, R.A. (1982) 1-tyrosine regulation and biosynthesis via arogenate dehy- drogenase in suspension-cultured cells of Nicotiana silvestris. Speg. et Comes. Planta 156, 233–240. 6. Connelly, J.A. & Conn, E.E. (1986) Tyrosine biosynthesis in Sorghum bicolor: isolation and regulatory properties of arogenate dehydrogenase. Z. Naturforsch 41c, 69–78. 7. Bonner, C.A., Fisher, R.S., Schmidt, R.R., Miller, P.W. & Jensen, R.A. (1995) Distinct enzymes of aromatic amino acid biosynthesis that are highly conserved in land plants are present in the chloro- phyte alga Chlorella sorokiniana. Plant Cell Physiol. 36, 1013–1022. 8. Abou-Zeid, A., Euverink, G.J.W., Hessels, G.I., Jensen, R.A. & Dijkhuizen, L. (1995) Biosynthesis of 1-phenylalanine and 1-tyro- sine in the actinomycete Amycolatopsis methanolica. Appl. Env. Microbiol. 41, 1298–1302. 9. Fazel, A.M. & Jensen, R.A. (1979) Obligatory biosynthesis of 1-tyrosine via the pretyrosine branchlet in coryneform bacteria. J. Bacteriol. 138, 805–815. 10. Hall, G.C., Flick, M.B., Gherna, R.L. & Jensen, R.A. (1982) Biochemical diversity for biosynthesis of aromatic aminoacids among cyanobacteria. J. Bacteriol. 149, 65–78. 11. Keller, B., Keller, E. & Lingens, F. (1985) Arogenate dehy- drogenase from Streptomyces phaeochromogenes.Purificationand properties. Biol. Chem. Hoppe Seyler 366 (11), 1063–1066. 12. Mayer, E., Waldner-Sander, S., Keller, B., Keller, E. & Lingens, F. (1985) Purification of arogenate dehydrogenase from Phenylo- bacterium immobile. FEBS Lett. 792, 208–212. 13. Patel, N., Pierson, D.L. & Jensen, R.A. (1977) Dual enzymatic routes to 1-tyrosine and 1-phenylalanine via pretyrosine in Pseu- domonas aeruginosa. J. Biol. Chem. 252, 5839–5846. 14. Zhao,G.,Xia,T.,Ingram,L.O.&Jensen,R.A.(1993)Anallo- sterically insensitive class of cyclohexadienyl dehydrogenase from Zymomonas mobilis. Eur. J. Biochem. 212, 157–165. 15. Hermann, K.M. & Weaver, L.M. (1999) The shikimate pathway. Annu. Rev. Plant Physiol. Plant Mol. Biol. 50, 473–503. 16. Schmid, J. & Amrhein, N. (1995) Molecular organization of the shikimate pathway in higher plants. Phytochemistry 39 (4), 737– 749. 17. Schmid, J. & Amrhein, N. (1999) The shikimate pathway. In Plant Amino Acids (Singh, B., ed.), pp. 147–169. Marcel Dekker, New York. 18. Wink, M. (1999) Plant secondary metabolites from higher plants: biochemistry, function and biotechnology. InBiochemistry of Plant Secondary Metabolism, Anual Plant Reviews (Wink, M., ed.), pp. 1–16. Sheffield Academics, Sheffield. 19. Rippert, P. & Matringe, M. (2002) Molecular and biochemical characterisation of an A. thaliana arogenate dehydrogenase with two highly similar and active domains. Plant Mol. Biol. 48, 361–368. 20. Arabidopsis Genome Initiative (2000) Analyses of the genome sequence of the flowering plant Arabidopsis thaliana. Nature 408, 796–815. 4760 P. Rippert and M. Matringe (Eur. J. Biochem. 269) Ó FEBS 2002 21. Heyde, E. & Morrison, J.F. (1978) Kinetic studies on the reactions catalysed by chorismate mutase-prephenate dehydrogenase from Aerobacter aerogenes. Biochemistry 17, 1573–1580. 22. Turnbull, J., Cleland, W.W. & Morrison, J.F. (1990) Chorismate mutase-prephenate dehydrogenase from Escherichia coli.1. Kinetic characterization of the dehydrogenase reaction by use of alternative substrates. Biochemistry 29, 10245–10254. 23. Taylor, A.L. & Trotter, C.D. (1967) Revised linkage map of. Escherichia C-Oli. Bacteriol. Rev. 31, 332–353. 24. Dower, W.J., Miller, J.F. & Ragsdate, C.W. (1988) High efficiency transformation of E. coli by high voltage electroporation. Nucleic Acids Res. 16, 6127–6145. 25. Bradford, M.M. (1976) A rapid and sensitive method for the quantification of microgram quantities of proteins uti- lizing the principle of protein-dye binding. Anal. Biochem. 72, 248– 254. 26. Scopes, R.K. (1974) Measurement of protein by spectrophoto- metry at 205 nm. Anal. Biochem. 59, 277–282. 27. Chua, N.H. (1980) Electrophoresis analysis of chloroplast pro- teins. Methods Enzymol. 69, 434–436. 28. Jensen, R.A., Zamir, L.S.T., Pierre, M., Patel, N. & Pierson, D.L. (1977) Isolation and preparation of pretyrosine, accumulated as a dead-end metabolite by Neurospora crassa. J. Of Bacteriol. 132, 896–903. 29. Bonner,C.A.&Jensen,R.A.(1987)Arogenatedehydrogenase. Meth Enzymol. 142, 488–494. 30. Emanuelsson, O., Nielsen, H. & Von Heijne, G. (1999) ChloroP, a neutral network-based method for predicting chloroplast transit peptides and their cleavage sites. Protein Sci. 8, 978–984. 31. Rubin, J.L. & Jensen, R.A. (1979) Enzymology of 1-tyrosine in Mung Bean (Vigna radiata [L.] Wilczek). Plant Physiol. 64, 727– 734. 32. Segel, I.H. (1993) Enzyme Kinetics, Wiley & Sons, Inc., New York. 33. Higuchi, T. (1985) Biosynthesis and Biodegradation of Wood Components, Academic Press, Orlando. 34. Siehl, D.L. (1999) The biosynthesis of tryptophan, tyrosine, and phenylalanine from chorismate. In Plant Amino Acids (Singh, B., ed.), pp. 171–204. Marcel Dekker, New York. 35. Christopherson, R.I. & Morrison, J.F. (1985) Chorismate mutase- prephenate dehydrogenase from Escherichia coli: positive cooperativity with substrates and inhibitors. Biochemistry 24, 1116–1121. 36. Koch, G.L., Shaw, D.C. & Gibson, F. (1971) The purification and characterisation of chorismate mutase-prephenate dehydrogenase from Escherichia coli K12. Biochim. Biophys. Acta 229, 795–804. 37. Turnbull, J., Morrison, J.F. & Cleland, W.W., (1991) Kinetic studies on chorismate mutase-prephenate dehydrogenase from Escherichia coli: models for the feedback inhibition of prephenate dehydrogenase by L -tyrosine. Biochemistry 30, 7783–7788. 38. Hund, H.K., Ba ¨ r, G. & Lingens, F. (1989) Purification and properties of arogenate dehydrogenase of Actinoplanes mis- souriensis. Z. Naturforsch. 44c, 797–801. Ó FEBS 2002 Arogenate dehydrogenases of A. thaliana (Eur. J. Biochem. 269) 4761 . Purification and kinetic analysis of the two recombinant arogenate dehydrogenase isoforms of Arabidopsis thaliana Pascal Rippert and Michel Matringe Laboratoire. compar- ison of these two isoforms of arogenate dehydro- genase and a detailed kinetic analysis of their substrate specificity and reaction mechanisms to gain further insight