Anthranilate Synthetase from Serratia marcescens

Anthranilate synthetase was purified from Serratia marcescens. The enzyme was homogeneous by the criterion of disc gel electrophoresis. Glutamineand NH,-dependent anthranilate synthetase activities were subject to end product inhibition by tryptophan. Inhibition by tryptophan exhibited positive cooperativity. In the absence of tryptophan, Michaelis-Menten kinetics were obtained for saturation by the two substrates, chorismate and glutamine. Positive cooperativity for chorismate was detected in the presence of tryptophan. These kinetic properties of anthranilate synthetase from S. marcescens were similar to those of the multifunctional anthranilate synthetase-anthranilate 5-phosphoribosylpyrophosphate phosphoribosyltransferase (PR transferase) aggregate from Salmonella typhimurium. A molecular weight of approximately 141,000 was estimated for the oligomeric anthranilate synthetase from S. marcestens. Nonidentical subunits of molecular weights approximately 60,000 and 21,000 were detected by disc gel electrophoresis in sodium dodecyl sulfate. The glutamine analogue, 6-diazo-5-oxo-L-norleucine (DON), and sulfhydryl reagents selectively inactivated glutamine-dependent anthranilate synthetase activity; NH3dependent activity was largely retained. 14C-Labeled DON and iodo[14C]acetamide were incorporated into the small subunit thus establishing the relationship of this subunit to anthranilate synthetase Component II from anthranilate synthetase-PR transferase of S. typhimurium. Approximately 2 moles of DON or iodoacetamide were incorporated per mole of anthranilate synthetase, suggesting that the oligomeric enzyme contains two glutamine-binding subunits (anthranilate synthetase Component II). From a consideration of the approximate size of the native enzyme and its subunits a composition of two chains each of anthranilate synthetase Components I and II is suggested. A glutaminase activity was detected. This activity may serve to transfer the amide of glutamine from anthranilate synthetase Component II to the catalytic site on Component I. Catalytic properties and structural features of anthranilate synthetase from S. marcescens were compared to those of trypsin-treated anthranilate synthetase-PR transferase from

The enzyme was homogeneous by the criterion of disc gel electrophoresis.
Glutamine-and NH,-dependent anthranilate synthetase activities were subject to end product inhibition by tryptophan. Inhibition by tryptophan exhibited positive cooperativity.
In the absence of tryptophan, Michaelis-Menten kinetics were obtained for saturation by the two substrates, chorismate and glutamine. Positive cooperativity for chorismate was detected in the presence of tryptophan.
These kinetic properties of anthranilate synthetase from S. marcescens were similar to those of the multifunctional anthranilate synthetase-anthranilate 5-phosphoribosylpyrophosphate phosphoribosyltransferase (PR transferase) aggregate from Salmonella typhimurium. A molecular weight of approximately 141,000 was estimated for the oligomeric anthranilate synthetase from S. marcestens.
Nonidentical subunits of molecular weights approximately 60,000 and 21,000 were detected by disc gel electrophoresis in sodium dodecyl sulfate.
On the basis of similarities between the two enzymes it is suggested that gene fusion could account for the two types of anthranilate synthetase.
Anthranilate synthetase catalyzes the first reaction of tryptophan synthesis in microorganisms.
As shown by Equations 1 and 2 either glutamine or NH3 may serve as amino donor.  (2)). Anthranilate synthetase Component I from these organisms catalyzes Reaction 2 but cannot utilize glutamine because the glutamine site is apparently on the anthranilate synthetase Component II subunit (4).
In Xerratia marcescens (5) as well as in Bacillus subtilis (6) and in species of Pseudomonas (7) anthranilate synthetase is not aggregated with other enzymes of the tryptophan pathway. It was therefore of interest to compare the enzyme from S. marcestens with the anthranilate synthetase-PR transferase from S. typhimurium.
It is the purpose of this paper to report the purification to homogeneity of anthranilate synthetase from S. marcescens. The enzyme is an oligomer of molecular weight approximately 141,000 and contains nonidentical subunits of molecular weights approximately 60,000 and 21,000. The smaller subunit functions as a glutamine-binding protein.
A possibly close relationship be- Enzyme PuriJication-ilnthranilate synthetase was purified from S. marcescens strain HY 150 (ATCC 27143), a tryptophan ausotroph lacking indoleglycerol-3-P synthetase (5). Cells were grown in 1 liter of media in 2-liter flasks with vigorous shaking for about 17 hours at 30" or 37". Each flask contained salts (as described in Reference 8, but lacking citrate), 10 g of acidhydrolyzed casein, 4 g of glucose, and 5 mg of tryptophan. Growth was started with 1 ml of an overnight culture grown in the same media.
Cells were harvested by centrifugation, washed with 0.05 M potassium phosphate, pH 7.4, and st,ored at -10" until used.
For a typical enzyme preparation 120 to 170 g of frozen cells were thawed in 4 ml per g of 0.1 M Buffer A (0.1 M potassium phosphate, pH 7.4, 0.2 mM dithiothreitol, and 0.1 mM EDTA) and were then subjected to sonic disruption with a Branson S-125 sonifier.
Power output and time were adjusted to yield an extract containing between 18 and 20 mg of protein per ml. The temperature was maintained below 10" during sonic disruption. The crude extract was obtained following centrifugation at 37,000 x g for 45 min. All succeeding steps were performed at 2-5".
The extract was adjusted to a protein concentration of 15 mg per ml. A freshly prepared 2v0 solution of protamine sulfate in 0.1 M potassium phosphate, pH 7.4, was added dropwise with stirring to yield a final concentration of 0.16 mg of protamine sulfate per mg of protein.
Stirring was continued for 15 min after the last addition.
The solution was centrifuged at 37,000 x g for 10 min and the sediment was discarded.
To t,he supernatant solution was added 230 mg of ammonium sulfate per ml. Stirring was continued for 15 min after the last addition and the solution was then centrifuged at 37,000 i< g for 10 min. The precipitate was dissolved in 0.05 M Buffer A (similar to 0.1 M Buffer A except for the concentration of potassium phosphate) and was dialyzed overnight against approximately 50 volumes of the same buffer solution.
After dialysis the protein solution was diluted with 0.05 M Buffer A to a protein concentration of 10 to 12 mg per ml and was applied to a column (3 x 50 cm) of DEAE-cellulose. The charged column was washed with 500 ml of 0.05 M Buffer A after which the enzyme was eluted with a linear gradient of KC1. The mixing chamber contained 1 liter of 0.05 M Buffer A and the reservoir had the same buffer solution plus 0.5 M potassium chloride.
Fractions containing at least 15 to 20% of the tot,a! activity in the peak fraction were pooled, and the enzyme was precipitated with 277 mg of ammonium sulfate per ml. Stirring was continued for 1 hour after the last addition of the salt. The enzyme was sedimented by centrifugation at 37,000 X g for 10 min and dissolved in 0.01 M Buffer B (0.01 M potassium phosphate, pH 7.4, and 0.2 mM dithiothreitol).
The enzyme solution was dialyzed overnight against 100 volumes of the same buffer solution and then applied to a column (2.2 x 22 cm) of ECTEOLA-cellulose, previously equilibrated with the same buffer solution.
The enzyme was eluted with about 50 to 75 ml of buffer solution and was concentrated and dialyzed against 0.01 M Buffer B as before. Any turbidity was removed by centrifugation, and 1.0 M potassium phosphate, pH 6.0, was added to bring the buffer concentration to 0.05 M. The enzyme solution was applied to a column (2.5 x 5.0 cm) of hydroxylapatite previously washed with 400 ml of 0.05 M potassium phosphate, pH 6.0, con-taming 0.2 mM dithiothreitol.
The enzyme was eluted from the column with this buffer solution, and fractions of constant specific activity were pooled.
The pooled enzyme could be stored at -10". An initial drop in specific activity was always observed but, thereafter, the activity was stable for several months.
A typical purification starting with 166 g of frozen cell paste is summarized in Table I. lwaterials--S. marcescens strain HY 150 was obtained from Dr. W. L. Belser. Unlabeled and 14C-labeled DON were preparations that were synthesized and used previously (4). Chorismic acid was made according to the procedure of Gibson (9). Iodo-[1-i4C]acetamide (52.3 mCi per mmole) was obtained from Amersham-Searle.
Radioisotopic purity ( > 95%) was verified by thin layer chromatography on silica gel with the use of benzene-dioxane-acetic acid (90 :25 :4) and methanol-acetone (50 : 50) solvent systems. Other chemicals were obtained from commercial suppliers and were used without further purification.
Disc Gel Electrophoresis-Electrophoresis of the native enzyme was performed by the method of Davis (10). Electrophoresis in SDS was according to the method of Shapiro, Vinuela, and Maize1 (11). For electrophoresis in urea, the method of Davis (10) was modified by inclusion of 8 M urea in the buffers and gels. Sample and spacer gels were not used. Sucrose, final concentration 20%, or a drop of glycerol was added to increase the density of enzyme solutions which were applied directly to the separating gel. COOmassie blue and acetic acid-methanol-water (12) were used for staining and destaining, respectively, in all experiments.

Adeasurement of Incorporation of [V]DOLIT and todo[14C]acetamide into Protein Subunits-Anthranilate
synthetnse was labeled with radioactive DON or iodoacetamide as described in the appropriate figure legends.
The enzymes were then dennt.ured with 1% SDS and 1 y0 mercaptoethanol or with 8 JI uretl and the subunits separated by SDS or urea gel electrophoresis.
The gels were stained, destained, and then exhaustively washed in acetic acid-methanol-water as reported previously (4). The gels were sliced, solubilized by treatment with HZOz, and counted for radioactivity (4).
Sucrose Gradient Centrifugation-The techniques and calculations described by Martin and Ames (13) were followed. Yeast alcohol dehydrogenase, sZo,ur o f 7.6 S and molecular weight 141,000 (14), was used as a reference protein.
In most esperiments the sedimentation profile for anthranilate synt'hetase was superimposable with that of yeast alcohol dehydrogcnase.
A unit of activity corresponds to the formation of 1 nmole of anthranilate per min under the conditions of assay. In all figures velocity, v, is expressed as enzyme units. Specific activity is expressed as units of activity per mg of protein.
Protein was determined by the method of Lowry as described by Layne (17)  a small portion of the dialyzed enzyme. The remaining enzyme was divided into two portions and dried at 105" to constant weight in stainless steel planchets.
An equal volume of the dialysis buffer was also dried and weighed.
In a duplicate experiment the drying and weighing were performed by Dr. C. T. Yeh of the Microanalytical Laboratory, Chemistry Department, Purdue University.
It was determined that 1 mg of protein according to the Lowry method corresponded to 0.87 mg of protein, dry weight.

Enzyme
Purification-The results of a typical purification are summarized in Table I. After storage for 1 day at 2" or at -10' the specific activity declined to 3400 to 4000 units per mg and remained constant for several months.
Following the hydroxylapatite purification step the enzyme was homogeneous by the criterion of disc gel electrophoresis.
Kinetics for Substrate Saturation and Inhbition by Tryptophan-Similar to other glutamine amidotransferases (18) and anthranilate synthetase enzymes from other species (l-3, 6, 7), the  of chorismate is summarized in Fig. 2. Absence of curvature precludes cooperative interactions for these substrates. Values of Km for chorismate of 18 PM and 15 pM were calculated from secondary plots for the reactions with glutamine and ammonium sulfate, respectively.
In the above experiments both activities were assayed in triethanolamine-chloride, pH 8.5. The Km for chorismate for glutamine-dependent anthranilate synthetase was 5-fold lower when determined with potassium phosphate, pH 7.4.
Both the glutamine-and NHB-dependent activities were completely inhibited by low concentrat,ions of tryptophan.
Concave upward curvature in plots of l/v against tryptophan concentration (Fig. 3) and slopes, n', greater than 1 in Hill plots (20) indicate positive cooperativity for tryptophan sites. For the experiment shown in Fig. 3, 50% inhibition of glutamine-dependent anthranilate synthetase activity required 7 PM tryptophan. The effect of tryptophan on substrate saturation of glutamine-dependent anthranilate synthetase activity is shown in Fig. 4. Concave upward curvature (Fig. 4A) is consistent with positive cooperativity for chorismate in the presence of tryptophan. End product inhibitior increased the chorismate concentration required for half-maximal velocity (S),., and also decreased the Vmax (Fig. 4A). Inhibition by tryptophan therefore may not be the result of simple competition for the chorismate site. Tryptophan inhibition was clearly noncompet8itive with respect to glutamine (Fig. 4B). Similar data were obtained for tryptophan inhibition of NHB-dependent anthranilate synthetase activity (not shown).
The data in Fig. 4A suggest that inhibition by tryptophan was not a resuIt of simple competition for the chorismate site. Evidence supporting a multisite model with distinct substrate and end product regulatory sites was provided by the data in Fig. 5. Lineweaver-Burk plots for chorismate saturation of glutaminedependent anthranilate synthetase performed in potassium phosphate buffer, pH 7.4, show deviations from linearity indicative of substrate inhibition.
One interpretation for substrate inhibition is that at high concentrations, chorismate interacts with the tryptophan regulatory site. The data in Fig. 5 show that a low concentration of tryptophan relieves substrate inhibition by chorismate.
Tryptophan may displace chorismate from a regulatory site.
Effect of pH on Enzyme Activity and Inhibition by Tryptophan-The pH optima for enzyme activity and inhibition by tryptophan are shown in Fig. 6 NHs-dependent reaction was at pH 8.5 or higher. Inhibition of both activities by 10 pM tryptophan was maximal at high pH. Molecular Weight and Subunit Composition-The molecular weight of anthranilate synthetase estimated by sucrose gradient centrifugation was 141,000 relative to yeast alcohol dehydrogenase. In seven experiments the range was 132,000 to 145,000. Nonidentical polypeptide chains were detected by disc gel electrophoresis in urea or in SDS. The molecular weights of the polypeptide chains were estimated to be approximately 60,000 and 21,000 by SDS gel electrophoresis.
Evidence supporting a subunit composition of two polypeptide chains of molecular weight approximately 60,000 plus two chains of molecular weight approximately 21,000 was obtained from ligand-binding studies (see below) and from intramolecular crosslinking.
Cross-linking with dimethyl suberimidate (21) was used as one approach to distinguish between anthranilate synthetase oligomers containing one or two small polypeptide chains. Most of the eight species that should result from random intramolecular cross-linking of a tetramer contaning subunits of two sizes were detected by SDS gel electrophoresis.
Significantly, a protein chain of molecular weight approximately 40,000 was found following treatment with dimethyl suberimidate. Formation of this species was independent of protein concentration and therefore suggested that oligomeric anthranilate synthetase might contain two protein chains of molecular weight approximately 21,000. Further evidence for a tetramer containing two large and two small subunits is presented below.

Synthetase
Activity-Evidence for a glutamine site distinct from the site for NH3 was obtained by selective inactivation of glutamine-dependent anthranilate synthetase activity. Such inactivation was obtained by treatment with the glutamine analogue DON or with sulfhydryl reagents.
The data in Fig. 7 show apparent first order kinetics for inactivation of glutamine-dependent anthranilate synthetase activity by 3.3 pM DON and 2.5 mM iodoacetamide.
Little or no inactivation of NHI-dependent anthranilate synthetase was obtained.
Similar but less reproducible results were obtained with p-mercuribenzoate and DTNB. For these experiments the half-times for inactivation by 3.3 PM DON and 2.5 rnitf iodoacetamide were 1. 5 (Fig. 8). No radioactivity over the background level was detected in the region of the gel containing the large polypeptide chain. Similar results were obtained with iodo[14C]acetamide except that in some experiments small amounts of radioactivity were incorporated into the large polypeptide chain.
The kinetics for incorporation of iodo[W]acetamide into the polypeptide chain of molecular weight approximately 21,000 are shown in Fig. 9. It is apparent that incorporation of radioactiv- The foregoing experiments while clearly establishing anthmnilate synthetase Component II as the glutamine-binding protein were not satisfactory for determining the number of sites alkylated by DON or iodoacetamide.
Quantitative recovery of radioactive material and enzyme was difficult following SDS gel electrophoresis.
The number of DON sites for anthranilate synthetase was estimated by the kinetic method of Xagano, Zalkin, and Henderson (4). A typical experiment is shown in Fig. 11. In this method glutamine-dependent anthranilate synthetase activity was determined in reaction mixtures containing different molar ratios of enzyme to DON.
The amount of DON required to achieve complete inactivation was obtained by extrapolation and as shown in Fig. 11 was 2.2 moles of DOK per mole of enzyme.
The number of sites on anthranilate synthetase Component II that were specifically required for utilization of glutamine and were alkylated by iodoacetamide was determined. Anthranilate synthetase was treated with iodo[14C]acetamide until 90% inactivation of glutamine-dependent activity. Under such conditions 100% of the NHs-dependent anthranilate synthetase activity was retained.
The reaction mixture was applied t.o a column of Sephadex G-25 in order to separate alkylated enzyme from unreacted iodo[l*C]acetamide.
The recovered enzyme fraction was assayed for protein and counted for radioactivity.
The number of sites calculated according to this procedure was 1.8 per enzyme.
Inactivation of glutamine-dependent anthranilate synthet.ase activity by sulfhydryl reagents and incorporat.ion of iodoacetamide into the enzyme indicates sulfhydryl groups are essential for utilization of glutamine. It has also been shown that DON alkylates cysteine residues of other glutamine amidotransferases (4,24,25). Titration with DTNB was used to confirm that specific inactivation of X. marcescens glutamine-dependent anthranilate synthetase by DON and iodoacetamide resulted from alkylation of cysteine groups.
The data in Table III show the number of titratable sulfhydryl groups in the native and alkylated enzyme samples. Following inactivation of glutamine-dependent anthranilate synthetase by DON approximately 1.5 cysteine groups per enzyme were blocked and were unavailable for titration wit.h DTNB.
Interestingly, these residues were largely inaccessible to