Studies of the Mechanism of Anthranilate Synthase

The mechanism of anthranilate synthase reaction was studied using hydroxylamine as an inhibitor. Changes in the NH,OH concentration result in a concomitant alteration in the rates of synthesis of y-glutamylhydroxamate and anthranilate. However, the sum of the two products appears to be constant. Moreover, the degree of inhibition elicited by NHsOH at different concentrations is similar for the over-all reaction, i.e., formation of anthranilate and the partial reaction of glutamine hydrolysis also catalyzed by anthranilate synthase in the absence of Mg+f. y-Glutamylhydroxamate is hydrolyzed by anthranilate synthase and the slopes and V,,,, for its hydrolysis as well as that of glutamine are similar. y-Ethylglutamate is also hydrolyzed by anthranilate synthase and this reaction requires chorismate. Addition of NH20H yields y-glutamylhydroxamate. The V max for this substrate is again similar to that obtained with y-glutamylhydroxamate. These studies suggest the participation of an acyl-enzyme complex in the reaction catalyzed by anthranilate synthase.

i.e. formation of anthranilate from chorismate and glutamine, NHzOH also inhibits the glutaminase activity exhibited by anthranilate synthase in the absence of Mg++.
Chorismate exerts a positive cooperative effect on the glutaminase activity. CH,-NHOH, the methyl derivative of NH20H, is a more potent inhibitor than NH20H but the mechanism of its inhibition is different. CH,-NHOH deprives the enzymatic reaction of chorismate by cleaving it to form an adduct with the enolpyruvyl moiety of chorismate. The evidence obtained suggests that the structure of this adduct is (~-carboxy-cY,N-dimethylnitrone.
The formation of the adduct is dependent only on the presence of chorismate and anthranilate synthase.
CH3-NHOH does not form a derivative with glutamine, nor does it inhibit the glutaminase activity.
Anthranilate synthase, the first enzyme specific for the biosynthesis of tryptophan, catalyzes the formation of anthranilate from either chorismate and glutamine or chorismate and ammonia (Fig. 1). The enzyme exists as an aggregate with the next enzyme in the tryptophan pathway, anthranilate-5'-phosphoribosylpyrophosphate phosphoribosyltransferase (1,2). The aggregate is capable of utilizing either glutamine or ammonia as amino donor. The anthranilate synthase protein devoid of the transferase, termed Component I, can utilize only ammonia as amino donor (2). The formation of anthranilate from chorismate and an amino donor is a complex reaction requiring amination at carbon 2 (3), as well as elimination of a hydroxyl group and an enolpyruvyl group. Elimination of the enolpyruvyl group is accompanied by protonation to form pyruvate and this proton originates from water (4). In view of the complexity of the enzymatic reaction, efforts have been directed towards finding inhibitors which might allow the reaction to proceed in partial * This investigation WBS supported by a grant from the National Institutes of Health, United States Public Health Service. form only. Somerville and Elford (5) have shown that NH20H, and even more so, its derivative, CHS-NHOH, are inhibitors of anthranilate synthase, the inhibitions resulting in the formation of unidentified hydroxamates.
The fact that hydroxamate formation is completely dependent on chorismate, glutamine, and enzyme and is inhibited by the end product, tryptophan, suggests that hydroxylamine may function by interfering with the over-all reaction.
Zalkin and Kling (6), working with Component I of anthranilate synthase, observed the formation of the hydroxamate only in the presence of CHB-NHOH.
Moreover, it was dependent only on chorismate and enzyme, whereas the amino donors, ammonia or glutamine, had no effect. In view of these findings it was felt that the identification of the hydroxamate may shed light on the mechanism of anthranilate synthase reaction. It  The isolation of homogenous anthranilate synthase from Salmonella typhimurium has been described by us (7). Chorismic acid was isolated from the accumulation medium of Aerobacter aerogenes 62-l by the method described by Gibson and Gibson (8). Chorismate-U-*4C was isolated by adding to the accumulation medium uniformly labeled glucose ( (9) ; in a final volume of 200 ml. The two substrates and NADH were added in 4 aliquots at 20-min intervals. The reaction mixture was flushed with Hz gas and kept sealed. After 90 min at 37" the reaction mixture was cooled and the pH adiusted to 3.0 with 20 ml of 2 N HzS04. The precipitated proteins were removed by centrifugation and the chorismic acid was extracted with ether from the clear aqueous layer. The ether was removed and the crude chorismic acid was purified by adsorption on Dowex l-Cl-and elution with 1 M NH&l.
The isolation and purification of p-bromophenylhydrazone of pyruvic acid and anthranilic acid from incubation mixtures have been described before (4). The hydroxamates were isolated from the incubation mixtures on Dowex l-acetate columns using pyridine-acetic acid buffers at the concentrations indicated in the legends. The molarity of the buffer refers to the concentration of pyridine.
Analytical Procedures-Anthranilic acid was determined by the calorimetric assay as previously described (7). Pyruvic acid was assayed spectrophotometrically in the presence of NADH and excess lactate dehydrogenase.
Glutamic acid was assayed with glutamate dehydrogenase and 3-acetylpyridine-NAD (10). The I80 content of the various compounds was analyzed by the method of Rittenberg and Ponticorvo (11). Hydroxamate formation was assayed with acidic FeC13 reagent as described by Somerville and Elford (5). Radioactivity measurements were carried out in a low background Nuclear Chicago gas flow counter.
Mass spectra were obtained by Miss Vinka Parmakovich on a RMU-6D Hitachi mass spectrometer.

EJect of XH20H and CH3-NHOH
Concentration on Hydroxamate Form&on-The formation of hydroxamate by anthranilate synthase was studied by varying the concentrations of NHzOH and CHa-NHOH ( Fig. 2A). At a concentration of 0.1 M, CHa-NHOH is twice as effective as NH20H, whereas very high concentrations of CH3-NHOH markedly reduces the formation of hydroxamate.
NH20H, on the other hand, behaves differently; increasing the concentration yields a biphasic curve. At high concentrations of NHzOH the formation of hydroxamate is probably nonenzymatic (12) and this may explain the biphasic curve. In view of these findings, the inhibitors were used at a final con- The reaction mixture contained Tris-HCl buffer, pH 8.2, 50 pmoles; glutamine, 5.0 pmoles; chorismate, 0.5 pmole; anthranilate synthase (specific activity 45)) 0.2 units; and neutralized NHzOH or CH,-NHOH at the indicated concentration in a final volume of 1.0 ml. The enzyme was added last. The incubation was carried out for 30 min at room temperature and was stopped by the addition of FeCla reagent (5). The color was allowed to develop for 10 min and the intensity was read at 500 rnp. B, effect of chorismate concentration on hydroxamate formation with CH,-NHOH.
The incubation mixture contained Tris-HCI buffer, pH 7.9,50 pmoles; EDTA, 0.1 Imole; anthranilate synthase, 0.015 mg; neutralized CHp-NHOH, 100 pmoles; and chorismate at the indicated concentration in a final volume of 0.5 ml. In the experiments with glutamine, it was present at a concentration of 2.5 Nmoles. centration of 0.2 M for further studies. The absorption spectra of the chromogenic material formed with FeC13 had no sharp maximum and was dependent on the nature of the hydroxylamine used for hydroxamate formation.
With NHzOH the absorption maximum was centered around 505 rnp while with CHI-NHOH it was around 520 mp.
pH Optimum and Substrate Requirements for Hydroxamate Formation-The optimum pH for hydroxamate formation with either NHzOH or with CHa-NHOH was 7.9; at this pH, the inhibition of anthranilate synthesis by CHs-NHOH was 42$& and 37% with glutamine and ammonia, respectively.
However, at the optimum pH for glutamine as amino donor (i.e. 7.4), the inhibition was 53% and, at the optimum pH for NH, as amino donor (i.e. 8.7), the inhibition was only 25% with CHa-NHOH. The inhibition by NHzOH at the respective pH optima was 29% for glutamine and 9% for NH3 as amino donor.
Somerville and Elford (5)   Attempts were therefore made to isolate the hydroxamate that mate and anthranilic acid after removal of the pyridine-acetic was produced by CHB-NHOH with chorismate and anthranilate acid buffer by lyophilization, and were found to be negative. synthase.
The most suitable method was found to be elution Only the fraction that was eluted with 0.2 M pyridine-acetic acid from a Dowex l-acetate column with increasing concentrations buffer reacted with FeC13 in the manner typical of hydroxamates, of pyridine-acetic acid buffer, pH 5.05. ,4 typical separation and this fraction contained 10 to 15% of the total radioactivity pattern of an incubation mixture inhibited by CHs-NHOH is placed on the column. This fraction was lyophilized to remove shown in Fig. 3. Uniformly labeled 14C-chorismate was used to the buffer and a mass spectral analysis was performed. The facilitate the identification of different fractions eluted from the major fragments are shown in Fig. 4. The highest molecular ion column.
All the radioactive fractions were analyzed for choris-had a mass of 117. The fragmentation ljattern suggests that the The ultraviolet absorption spectra of the enzymatic product had a peak at 240 rnp and the infrared spectra had a strong absorption at 1175 cm-l, a stretching frequency that is considered to be due to N+-O- (14). These spectral properties, characteristic of nitrones, are in accordance with the structure of a-carboxy-a-Ndimethylnitrone. Again, Dowex l-acetate column and pyridine-acetic acid buffers were used to separate the different components of the reaction mixture.
The fractions were counted and analyzed for hydroxamate with FeCL. The results of this experiment are shown in Fig. 5. Because the labeled chorismate has been crystallized only once, a control experiment in which ring-r4C-chorismate, unlabeled chorismate, and CH3-NHOH were incubated in the absence of enzyme and run through a column identical with the one used for the experiment. A small amount of radioactive impurity was eluted with both the experimental and control runs. With the incubation mixture in which enzyme was present, only Fractions 16 to 18 from the Dowex column contained all the FeCL-positive material and were not radioactive.
The radioactivity was eluted between 0.5 to 1 M pyridine-acetic acid and these fractions gave no color with FeCh. In control experiments none of the fractions gave a Fe&positive material. This experiment therefore provides further evidence that the hydroxamate formed with CHS-NHOH and chorismate in the anthranilate synthase reaction is an adduct of the side chain and not of the ring of chorismate.
Studies The rate of reduction of the nucleotide was followed spectrophotometrically by recording the increase in absorbance at 365 rnp. l-acetate column for separation and identification of the hydroxamate (Fig. 6). All fractions were analyzed for hydroxamate by reacting with FeCla. Only the fractions eluted with water gave a positive test with Fe& When 14C-chorismate was included in the reaction mixture, no radioactivity was associated with the water eluate, which gave a positive color reaction with Fe&.
This observation, as well as the finding that y-glutamylhydroxamate is not bound to the Dowex lacetate column, suggested that the hydroxamate formed in the reaction with NHzOH is probably y-glutamylhydroxamate. This was confirmed by employing 1Gglutamine in the reaction mixture.
The components of the incubation mixture consisting of enzyme, chorismate, 14C-glutamine, and NHzOH were placed on a Dowex 1 column and separated by elution with Hz0 followed by pyridine-acetic acid. Radioactivity was found to be located in the fractions eluted with Hz0 as well as in fractions eluted with 0.1 M pyridine-acetic acid. The latter was identified as glutamic acid. However, only the water eluate gave a color reaction with acidic FeCL. This fraction which contained the hydroxamate was further analyzed by paper chromatography in two solvent systems: (a) n-butyl alcohol, acetic acid-Hz0 (60: 15:25) and (b) ethanol-NHtOH-Hz0 (80:5:15). Authentic y-glutamylhydroxamate was included as a reference. A chromatogram from each solvent system was sprayed with FeC13 reagent and a second chromatogram from each solvent system was subjected to ninhydrin spray. In all cases the enzymatic product had the same mobility as authentic y-glutamylhydroxamate. Furthermore, the infrared spectra of the enzymatic product and of the authentic y-glutamylhydroxamate were identical in all respects.
Recently Nagano, Zalkin, and Henderson (13) have demonstrated glutaminase activity in a partially purified preparation of anthranilate synthase in the absence of Mg++ and in the presence of chorismate.
We have confirmed this observation with the homogenous anthranilate synthase used in the experiments outlined above (Fig. 7). Glutaminase activity can be demonstrated only in the presence of chorismate.
Invariably, a lag of approximately 1 to 3 min was observed in the appearance of glutamic acid. This delay could not be abolished by preliminary incubation of chorismate and enzyme before the addition of glutamine or by using a very large excess of anthranilate synthase.
The inset in Fig. 7 demonstrates that, as the chorismate concentration is increased, the rate of glutaminase activity exhibited by anthranilate synthase does not follow linear kinetics.
The effect of NHzOH and CHI-NHOH on the glutaminase activity of anthranilate synthase is shown in Table II. Whereas NHzOH inhibits the glutaminase activity, CHI-NHOH had no effect on the rate or the extent of hydrolysis of glutamine.

DISCUSSION
The results presented demonstrate that although CHI-NHOH is a methyl derivative of NH20H, its mechanism of inhibition of anthranilate synthase is different from that of hydroxylamine. The formation of hydroxamate with NHzOH required the presence of enzyme, chorismate, and glutamine, whereas only enzyme and chorismate were needed for reaction with CH3-NHOH.
The evidence presented indicate that the reaction product is an adduct of pyruvate and CHB-NHOH, cY-carboxya: ,N-dimethylnitrone.
Nitrones are generally prepared by reaction of aldehydes and ketones with substituted hydroxylamines and have been extensively studied as useful intermediates in chemical syntheses (14,15). We have observed that concentrated solutions of cY-keto acids react with CHS-NHOH and FeC13 in strong acids to yield a chromogenic product with an absorption maximum around 500 mp. It should be stressed that even at high concentrations of chorismate the formation of Fe&positive material with chorismate and CHs-NHOH required the presence of anthranilate synthase. The formation of the adduct of pyruvate and CHS-NHOH suggests that a binding site of the enzyme to chorismate may exist through the enolpyruvate side chain. This is supported by the finding that pyruvate can partially substitute for chorismate in eliciting a glutaminase activity and in overcoming the inhibition of anthranilate synthase by tryptophan (13).
Moreover, when the anthranilate synthase reaction was carried out in DzO, the pyruvate formed not only contained close to 1 atom of deuterium in the methyl group, but also a small yet significant amount of -CHD2 species because the enzyme-chorismate complex underwent a limited exchange with water (4).
The formation of a hydroxamate with hydroxylamine required both chorismate and glutamine.
The identification of the reaction product with NHzOH as y-glutamylhydroxamate suggests that anthranilate synthase in the presence of chorismate can form an acylenzyme complex through the y-car-boxy1 of glutamic acid and can exhibit glutaminase activity. Evidence for such an activity has recently been provided by Nagano et al. (13). We have confirmed the finding with our homogenous anthranilate synthase preparation. Glutaminase activity can be demonstrated only in the presence of chorismate. Addition of NHtOH, to the reaction mixture containing enzyme, chorismate, and glutamine, results in a decrease of glutamic acid formed with the concomitant production of y-glutamylhydroxamate. Thus, hydroxylamine competes with water in the hydrolysis of the acyl enzyme.
Although chorismate is essential for demonstrating glutaminase activity of Mg ++ free-anthranilate synthase, increasing the chorismate concentration had a nonlinear effect on the glutaminase activity.
A similar positive cooperative effect could be shown to occur with anthranilate synthase only when studying the effect of chorismate concentration on the inhibition of the over-all reaction (i.e. formation of anthranilic acid) by tryptophan (7). Chorismate, in the absence of tryptophan, does not exhibit a homotropic effect normally observed with regulatory enzymes in which the allosteric ligand affects the affinity for the substrate.
On the basis of the present work the following scheme for the enzymatic reactions catalyzed by the anthranilate synthase can be advanced (Fig. 8). Anthranilate synthase reacts with chorismate to form enzyme-chorismate complex. This complex can combine with ammonia in the presence of Mg++ to give anthranilate, pyruvate, and enzyme or react with CH,-NHOH to yield the adduct shown in the figure.
The enzyme-chorismate complex also interacts with glutamine.
This new complex can undergo hydrolysis with Hz0 to give glutamate or give yglutamylhydroxamate with NHtOH or form anthranilate, pyruvate, and glutamate in the presence of Mg++.
Further studies to obtain direct evidence for the existence of an enzyme intermediate in anthranilate synthase reaction are in progress.