JOURNAL OF BACTERIOLOGY, Oct. 1967, P. 1007-1015Copyright © 1967 American Society for Microbiology

Vol. 94, No. 4-Printed in U.S.A.

Regulation of Glutamine SynthetaseV. Partial Purification and Properties of Glutamine Synthetase

from Bacillus licheniformisJERRY S. HUBBARD' AmN E. R. STADTMAN

Laboratory of Biochemistry, National Heart Institute, National Institutes of Health, Bethesda, Maryland 20014

Received for publication 29 June 1967

The glutamine synthetase of Bacillus licheniformis has been obtained at about15% purity. Sucrose gradient centrifugation gave a molecular weight value of ap-proximately 612,000. Both L- and D-glutamate can be utilized as substrates in thebiosynthetic reaction, although the L isomer was five times more active. The re-quirement for adenosine triphosphate (ATP) can be partially replaced by guanosineor inosine triphosphates, but not by cytidine or uridine triphosphates. The Mn ++was required for activity, and the requirement cannot be satisfied with Mg ++.Maximal activity of the biosynthetic reaction was observed when ATP and Mnwere present in equimolar amounts. An excess of either reactant gave less activity.However, other purine and pyrimidine nucleotides, when added in combinationwith ATP, can partially substitute for ATP in attaining the equimolar ratio ofnucleotide to Mn 11. A complex of ATP and Mn i' is the preferred form of sub-strate. The B. licheniformis enzyme catalyzes the glutamyl transfer reaction but ata much slower rate than the Escherichia coli glutamine synthetase. Either adenosinediphosphate (ADP) or ATP can activate the glutamotransferase, although ADP ismore active.

The amide and a-amino nitrogen atoms ofglutamine may be utilized in diverse pathwaysleading to the formation of a variety of metabo-lites (8, 12). Therefore, the synthesis of glutaminecatalyzed by glutamine synthetase represents acomplex metabolic branch-point and thus im-poses special requirements for an effective meansof control. In studies with the glutamine syn-thetase of Escherichia coli (13), evidence waspresented for a unique type of end-productcontrol, viz., cumulative feedback inhibition.When tested individually, eight end products ofglutamine metabolism cause limited degrees ofinhibition. However, they appear to act inde-pendently of each other, so that in mixturestheir effects are cumulative.A survey of 10 diverse microbial types (3)

revealed that the glutamine synthetase of eachorganism was unique with respect to the extentof inhibition caused by each of the various endproducts. Nevertheless, in eight of the organismstested, regulation of the glutamine synthetaseappears to involve a simple cumulative feedbackmechanism.

I Recipient of a Public Health Service PostdoctoralFellowship. Present address: Bioscience Section,Jet Propulsion Laboratory, California Institute ofTechnology, Pasadena.

On the other hand, two organisms, Bacilluscereus and B. licheniformis, exhibit atypicalpatterns of feedback regulation. For example,the glutamine synthetase of these organisms ismarkedly inhibited by glutamine; moreover,certain pairs of end products [viz., adenosine-5'-monophosphate (AMP) and glutamine orAMP and histidine] act synergistically to in-hibit glutamine synthetase activity. This studywas undertaken to characterize further theglutamine synthetase of B. licheniformis.

MATERIALS AND METHODSEnzyme assay. Glutamine synthetase activity was

measured both by the production of orthophosphate(Pi) in the biosynthetic reaction (equation 1) and bythe production of Sy-glutamylhydroxamate in theglutamyl transfer reaction catalyzed by the enzyme(equation 2).

glutamate + NH3 + ATP Mn.i glutamine +

ADP + Pi (1)

Mn++ ADPglutamine + NH20H ' I

arsenate

y-glutamylhydroxamate + NH, (2>

1007

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HUBBARD AND STADTMAN

The assay conditions were similar to those whichwere previously described (3, 12). The concentrationsof reactants employed in various experiments aregiven in the appropriate figures and tables. Netsynthesis of glutamine (equation 1) was establishedby paper chromatography of deproteinized incu-bation mixtures with 1-butanol-water-acetic acid(4:4:1, v/v) as the developing solvent. The RF valuesof glutamine and glutamic acid were 0.18 and 0.25,respectively. The formation of -y-glutamyl hydrox-amate (equation 2) was established by thin layerchromatography of deproteinized incubation mix-tures on silica gel F254 plates, with 1-butanol-water-acetic (4:1:1, v/v) and phenol-water (3:1, w/w) asthe solvent systems; y-glutamylhydroxamate hadRF values of 0.08 and 0.09, respectively. Ferricchloride reagent (10% in 0.2 N HCl) was used fordetecting hydroxamate spots.

Growth of cells. B. licheniformis ATCC 9945A wasgrown aerobically for 18 hr at 37 C in a medium ofthe following composition: K2HPO4, 10.6 g; KH2PO4,5.3 g; MgSO4.7H20, 0.2 g; K2SO4, 2.5 g; NaCl, 2.5 g;NaNO3, 8.5 g; MnCl2-2H20, 8 mg; glycerol, 4.2 g;and tap water, 1,000 ml. Cells grown with nitrate as arate-limiting source of nitrogen contained abouteight times more glutamine synthetase than cellsgrown with an equimolar level of ammonia. It haspreviously been reported that supplying E. coli (12)or Saccharomyces cerevisia (4) with a readily avail-able nitrogen source causes a repression of glutaminesynthetase.

Purification of glutaminie synthetase. The puri-fication procedure was similar to that used for thepurification of E. coli glutamine synthetase (12) butdiffers sufficiently in detail to warrant a description.Unless otherwise indicated, all steps were performedat 4 C. The buffer used throughout consisted of 0.01M imidazole and 0.01 M MnCl2, adjusted to pH 7.0with HCI. The extent of purification was followedwith the biosynthetic assay. Protein was determinedby the method of Lowry et al. (4).

Step 1. Freshly harvested cells were suspended inimidazole-Mn buffer and were ruptured by passagethrough a French press. Cell debris was removed bycentrifugation at 20,000 X g for 30 min.

Step 2. To the supernatant fluid from step 1 con-taining from 10 to 20 mg of protein per ml, 10% byvolume of a 10% solution of streptomycin sulfatewas added dropwise. After 15 min of equilibration,the mixture was centrifuged and the precipitatewas discarded.

Step 3. The precipitate formed between 45 and60% saturation of ammonium sulfate was collected.This pellet was resuspended in 10% of the originalvolume and was dialyzed overnight against 100volumes of buffer. A precipitate which formed duringthe dialysis was removed by centrifugation and wasdiscarded.

Step 4. The dialyzed extract was adjusted to pH4.6 with 1 M acetic acid and centrifuged, and thesupernatant fluid was discarded. The precipitate wasresuspended in 5% of the original volume, adjustedto pH 7 with 2 N KOH, triturated for 1 hr, and cen-trifuged.

Step 5. The supernatant fluid from step 4 wasadjusted to pH 5.7 with 1 M acetic acid, heated at60 C for 10 min, cooled, and centrifuged.

Step 6. To the supernatant fluid of step 5, acetonewas added to a final concentration of 45% (by vol-ume). The mixture was centrifuged, and the precipitatetriturated with buffer (2% of the original volume)for 1 hr and centrifuged.

Step 7. The supernatant fluid of step 6 was adjustedto pH 4.6 with acetic acid and centrifuged. Theprecipitate was suspended in buffer (1 to 2% oforiginal volume), adjusted to pH 7 with KOH,triturated for 30 min, and centrifuged. The super-natant fluid was dialyzed overnight against 1,000volumes of buffer. Enzyme preparations were storedat 4 C, and no loss in activity was detected over a2-month period.

Disc electrophoresis. The disc electrophoresis ofproteins was performed by the procedure of Ornsteinand Davis according to the instructions provided byCanalco Co., Silver Spring, Md. The 7% gel was usedand tris(hydroxymethyl)aminomethane Tris glycinewas the developing buffer. For enzyme assays on thepolyacrylamide gel, the developed electrophoreto-gram was added to a beaker containing the reactionmixture for the glutamyl transfer assay. After incu-bation, ferric chloride reagent was added, and a faintband of hydroxamate was detected in the area of thegel containing the glutamine synthetase.

Sucrose gradienit centrifugation. Sucrose densitygradients were prepared and sampled according to theprocedure of Martin and Ames (7). A buffer consist-ing of 0.01 M imidazole and 0.01 M MnCl2 (pH 7.0)was mixed with the sucrose solutions. Samples werecentrifuged for 8.5 hr at 35,500 rev/min in a Spincomodel L equipped with a Swinging bucket rotor(SW39).

Reagents. Adenosine-5'-triphosphate (ATP) diso-dium, cytidine-5'-triphosphate (CTP) disodium,uridine-5'-triphosphate (UTP) sodium salt, L-glu-tamine (Grade III), and L-glutamine acid-mono-hydroxamate were obtained from Sigma ChemicalCo., St. Louis, Mo. Adenosine-5'-diphosphate (ADP)sodium salt, inosine-5'-triphosphate (ITP) sodiumsalt, inosine-5'-diphosphate (IDP) sodium salt, andguanosine-5'-triphosphate (GTP) sodium salt wereobtained from Pabst Laboratories, Milwaukee, Wis.Pyruvate kinase (rabbit muscle) and phosphoenol-pyruvate (sodium salt) were obtained from C. F.Boehringer and Soehne GmbH, Mannheim, Germany.All other chemicals were reagent grade. Purified E. coliglutamine synthetase was generously provided byBennett Shapiro of this Institute.

RESULTS

Properties of B. licheniformis glutamine syn-thetase. The purification procedure gave enzymepreparations which catalyzed the biosyntheticreaction at a rate of 4,umoles of Pi liberated permin per mg of protein. This value is about 15 %,-of the specific activity reported for the glutaminesynthetase of E. coli (12). Disc gel electrophoresis

1008 J. BACTERIOL.

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

GLUTAMINE SYNTHETASE. V

of the B. licheniformis preparation showed thatthe band of glutamine synthetase activity cor-responded to about 15 to 30% of the total pro-tein present in the gel. The most successful puri-fication experiment gave a 126-fold purificationwith 29% recovery of activity. However, theextent of the purification could not be definitelydetermined, since it was subsequently found thatthe glutamine synthetase present in crude ex-tracts was not in a fully active state. For example,passage of crude extracts over Sephadex G-25gave a fourfold increase in specific activity. Apossible explanation for the increased activityis that inhibitors were removed by the Sephadextreatment. If so, then it might be assumed thatthe purification procedure would also result inthe removal of the inhibitors and a correspondingincrease in specific activity.

Sedimentation on sucrose density gradients.Figure 1A illustrates results obtained whenenzyme preparations were subjected to sucrosegradient centrifugation. A single peak of ac-tivity was detected with the biosynthetic assay.A low level of glutamotransferase activity cor-responded to this peak, and the rate of glutamyl-hydroxamate formation was markedly enhancedwhen a high level of glutamate was included inthe reaction mixture. The significance of thestimulatory role of glutamate is considered in

the accompanying paper (3a). Figure lB presentsresults of a sucrose gradient centrifugation of amixture of B. licheniformis and E. coli glutaminesynthetases. The concentrations of these twoenzymes in a mixture can be measured by takingadvantage of the fact that the B. licheniformisenzyme is completely inhibited by mixtures ofglutamine and AMP, whereas the E. coli enzymeis inhibited only 20% by these compounds. Thus,the standard biosynthetic assay measures theactivities of both enzymes; however, in thepresence of added AMP and glutamine, thisassay measures the E. coli enzyme only. Alterna-tively, the E. coli enzyme can be specificallymeasured by the standard glutamyl transferassay since the B. licheniformis enzyme exhibitsalmost no glutamotransferase activity underthe assay conditions employed for this experi-ment. The inability of the B. licheniformis en-zyme to catalyze the transfer reaction is pre-sumably due to the marked inhibition exhibitedby glutamine.The distribution of enzyme activities as meas-

ured by these differential assay procedures isshown in Fig. lB. As noted above, the activitydetected when the biosynthetic assay was supple-mented with glutamine and AMP was attributableto the E. coli enzyme only; the smaller amountof this activity as compared to the total activity

l/icheniform/s GLUTAM NESYNTHETASE

240F A

>-200

L1JW c 160

,

U)-

-,0w E 80

E40

-J

V

0

I I B.osynthe/icAssoy

II

I , 6/u/atomy/ Tronsfer- Assoy (# glutamafe!

I 1/a/o G/tomy/

_ /"\,/ Tronsfer Assay1:4

MIXTURE OF S. licheniforml'sAND E.co/iGLUTAMINE SYN-

THETASESB.

BB8iosynthet/cAssoy

W, G/utomy/I \<Transfer Assay

jI 0'1IBiosynthet/ic

- Assay(*g/uNH2lA¶\tand AMPJ

8---2-28IX 84u

0 4 8 1 2 6 20 24 28 0 4 8 1 2 1 6 20 24 28GRADIENT FRACTION NUMBER ( bottom top )

FIG. 1. Sucrose gradient centrifugation of Bacillus licheniformis and Escherichia coli glutamine synthetases.(A) 1.8 mg of B. licheniiformis glutamine synthetase preparation. (B) 1.8 mg of B. licheniformis glutamine syn-thetase preparation plus 88 jig of purified E. coli glutamine synthetase. Biosynthetic assay: 0.1 M gluta-mate, 0.0075M NH4Cl, 0.0075M ATP, 0.007M MnCl2, and 0.022M imidazole. Glutamine and AMP, whenemployed, were added at 0.005 M levels. Glutamyl transfer assay: 0.05 M L-glutamine, 0.1 M NH20H, 0.02 Marsenate, 0.003 M MnC12, 0.0004M ADP, and 0.021 M imidazole. Glutamate, when employed, was added at alevel of0.06 M. The ordinate refers to glutamine synthetase activity (millimicromoles ofPiper 15 mill per 5 ,uliters),biosynthetic assay, or to glutamotransferase activity (micromoles of -y-glutamylhlydroxamate per 30 min per5 jlAiters X 1w2)-

2.4G)r--~ C:

2.0 3 -10 >

cn O

1.6 Z 'M0 zx (no -mn

1.2 3 m

- (I)

0.8 O3 >

C)

0.4.Q <_P

1009VOL. 94, 1967

IHA.

. ,rl

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HUBBARD AND STADTMAN

as measured in the standard biosynthetic assay

reflects the fact that the amount of E. coli en-

zyme added initially was less than the amount ofB. licheniformis enzyme activity added. It isapparent from the data in Fig. 1B that the E.coli enzyme had a slightly higher sedimentationcoefficient than did the enzyme from B. licheni-formis.A molecular weight of 680,000 has previously

been determined for the E. coli glutamine syn-thetase (12). If it is assumed that the B. licheni-formis and E. coli enzymes have approximatelythe same shapes and partial specific volumes,then the molecular weight of the B. licheniformisenzyme can be estimated by the equations ofMartin and Ames (7). Based on these assump-tions, the calculated molecular weight of theB. licheniformis glutamine synthetase is 612,000.

Effect of substrate concentrations. Figure 2shows the effects of concentrations of NH4+,L-glutamate, and ATP on the activity of partiallypurified preparations of B. licheniformis gluta-mine synthetase. The kinetic responses to bothNH4+ and ATP are very complex and precludethe application of classical treatments to de-termine the apparent Km. The very narrow range,of optimal activity observed with NH4+ presentsa technical problem in measuring catalytic ac-

tivity under optimal conditions. The peak ofactivity occurs at such a low concentration ofNH4+ that the amount of substrate consumedduring normal assay conditions is a significantproportion of the total added. It is thereforedoubtful whether the amount of reaction in theincubation time of 7.5 min is a measure of initialrates, so the true optimal concentration for NH4+is uncertain. To avoid problems of changingcatalytic activity in a critical range of NH4+concentration, the concentration of NH4+ usedin the standard assay was 0.05 M; at this con-

-centration, the enzyme was significantly in-hibited, but its activity was relatively unin-fluenced by the small change in concentration ofNH4+ associated with the glutamine synthesis.The effect of ATP was equally complex and,

as will be shown below, was very dependentupon the concentration of MnCl2 and on theconcentration of other nucleoside triphosphates.The effect of glutamate concentration appearedto follow a more or less normal saturation func-tion; yet, as will be shown in the following paper(3a), the double reciprocal plots yielded bimodalcurves indicating a more complex function. Theconcentration of L-glutamate required for one-half maximal velocity was about 5 mM.D-Glutamic acid served as a substrate for the

biosynthetic reaction (Table 1). However, underotherwise comparable conditions, the rate of Pi

330Fr

al-r

CLL

E

ATP, mM

2 4 6 8 10

0 0.02 0.04 0.06 0.08

GLUTAMATE, NH4CI, M

12 14 16 18

0.10 0.12

FIG. 2. Effect of substrate concentration on glu-tamine synthetase activity. The reaction mixtures wereas follows: for the glutamate curve, 0.05 it NH4Cl,7.5 mi ATP, 22 mM imidazole (pH 7.0), 7 mi MnCl2,27 Mzg of glutamine synthetase preparation, and theindicated amounts of L-glutamate; for the ammoniumchloride curve, 0.05 i L-glutamate, 7.5 num ATP, 7 miMnCI2, 22 mt imidazole (pH. 7.0), 22 Mg of glutaminesynthetase, and the indicated amounts of NH4CI; forthe ATP curve, 0.05 mt NH4Cl, 0.1 M L-glutamate,22 mM imidazole (pH 7.0), 27,Mg of glutamine syn-thetase, and the indicated amounts of ATP. Totalvolume 0.3 ml, 37 C. Since different enzyme prepa-rations were used for the three experiments, all datawere normalized at the point corresponding to standardassay conditions, i.e., 0.05 m NH4CI, 0.05 it L-glU-tamnate, and 7.5 mM ATP.

TABLE 1. Substrate specificity of Bacilluslicheniformis glutamine synthetase

Substrate" Glutamine synthetaseSubstrate ~~~~~activityj'

None..................07.5 mM L-glutamic acid... 4317.5 mM D-glutamic acid... 887.5 mM L-aspartic acid.... 07.5 mM L-isoglutamine.... 0

vComplete reaction mixture: 50 mm NH4Cl,7.5 mM ATP, 7 mm MnCl2, 22 mm imidazole, andenzyme and substrate (as indicated).

b Expressed in millimicromoles of Pi per milli-gram per minute.

1010 J. BAcrERioL.

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

GLUTAMINE SYNTHETASE. V

liberation was five times greater with L-glutamicacid than with the D isomer. The biosyntheticreaction was not demonstrable when L-asparticacid or L-isoglutamine was substituted forL-glutamic acid. Table 2 shows that GTP andITP can partially substitute for ATP as thenucleotide substrate. The pyrimidine nucleotideswere ineffective as replacements for ATP.

Relationship of ATP and Mn+ . Unlike theglutamine synthetases from several other bac-teria, the enzyme from B. licheniformis showedan obligatory requirement for Mn++; Mg+ didnot satisfy the divalent cation requirement.The concentration of Mn+ required for maxi-mal activity varied directly with the nucleosidetriphosphate concentration (Fig. 3). With agiven concentration of Mn++, the catalyticactivity increased with increasing concentrationsof ATP up to a maximal value at which theamounts of ATP and Mn++ were about equal;further addition of ATP resulted in inhibition.

TABLE 2. Relationships of the ATP concentrationto the stimulation or inhibition of glutaminesynthetase by various purine and pyrimidine

nucleotidesa

Nucleotides added Glutamine Percentsynthetase activityactivity relative to

ATP Other nucleotide (Mmoles Pi) ATP control

mM 2.5 mM

0 ADP 00 CTP 00 GTP 100 UTP 00 ITP 30 IDP 00 ADP, CTP 0

2.5 97 1002.5 CTP 125 1292.5 GTP 126 1302.5 UTP 118 1222.5 ITP 117 1212.5 IDP 127 131

7.5 205 1007.5 CTP 155 767.5 GTP 130 647.5 UTP 145 717.5 ITP 132 657.5 IDP 191 93

a Reaction mixture: ATP as indicated, 2.5 mmpurine and pyrimidine nucleotides, 0.1 M gluta-mate, 0.05 M NH4C1, 7 mm MnCl2, and 22 mMimidazole and glutamine synthetase (27 ,ug ofprotein). Incubation time was 7.5 min. Glutaminesynthetase activity measured in biosyntheticassay system.

E 200 -/D

1/,'~ ~ ~ ~ \,

0 t S,v ..1 , .

E 0~~~~/-

o 2.5 5 7.5 10 12.5 15

ATP concentrotion, mM

FIG. 3. Interrelationship of ATP and MnC12 forthe activation of glutamine synthetase. Reactionmixtures contained 0.1 M glutamate, 0.05 M NH4Cl0.022 M imidazole, and 27,uAg of glutamine synthetase.ATP and MnCI2 were addled as indicated. Temperature,37 C.

As the level of Mn++ was increased, so was theconcentration of ATP required to obtain maxi-mal activity.

These results suggest that an ATP-Mn++complex is the active substrate and that relativeexcesses of either Mn++ or ATP inhibit thereaction. Especially noteworthy from the stand-point of cellular regulation is the fact that,whereas the concentration of ATP required foroptimal activity is equal to the concentrationof Mn++, the maximal activity of the enzyme inthe presence of stoichiometric amounts of ATPand Mn /iis independent of theMn0 - concen-tration. As a consequence, the concentration ofATP required for the same maximal activityis subject to wide variation depending upon theavailability of Mn++, and, as will be shown inthe following section, it is dependent also onthe presence of other nucleoside triphosphates.

Effect of CTP. Previous studies (3) showedthat, under standard assay conditions, CTPis an inhibitor of the B. icheniformis glutaminesynthetase. Subsequent studies, however, showthat the effect of CTP is extremely complex andis depnentat upon the relative concentration ofATP and Mn++. In fact, under certain conditionsCTP actually activates the enzyme. These complexrelations are illustrated in Fig. 4. The solidbars on the graph show the typical response ofcatalytic activity to increasing concentrations ofATP at a fixed level of 7 mM MnI1. As was noted

1011VOL. 94, 1967

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HUBBARD AND STADTMAN

240 S ATP+ 5mM ClP

200

W0160

20

4-80

40

1.5 2.5 5.0 7.5 10.0ATP CONCENTRATION (mM )

FIG. 4. Influence of various mixtures of ATP andCTP on glutamine synthetase activity. Reactionmixtures contained 0.1 M glutamate, 0.05 M NH4Cl,0.007 M MnCl2, 0.022 M imidazole, and glutaminesynthetase (27 ,ug of protein). ATP and CTP wereadded as indicated. Incubation time was 7.5 min at37 C.

before (Fig. 3), optimal activity was attainedwhen the level of ATP was almost equal to thelevel of Mn++ added, i.e., at 7.5 mm. The openbars in Fig. 4 show the effects of increasing ATPconcentration when the same assay mixture wassupplemented with 2.5 mm CTP. Under theseconditions, the concentration of ATP requiredfor optimal activity was reduced from 7.5 to5 mM, i.e., by an amount that is equivalent tothe amount of CTP added. Similarly, as shownby the cross-hatched bars in Fig. 4, in the pres-ence of 5 mm CTP, the level of ATP requiredfor optimal activity was 2.5 mm. Thus, as inthe preceding instance, the amount of ATPrequired for optimal activity was reduced by anamount equivalent to the amount of CTP added.In other words, maximal activity of the enzymewas attained when the total amount of nucleotide(ATP + CTP) was equivalent to the concen-tration of Mn++ present; when the nucleosidetriphosphate concentration was greater than theMn++ concentration, inhibition similar to thatobtained with excesses of ATP alone occurred.The effects of CTP could be readily explained

if CTP could replace ATP as a substrate forglutamine synthetase. This, however, is not thecase. As is shown in Table 2, CTP was completelyinactive when tested by itself. Moreover, itfailed to support glutamine synthesis in thepresence of ADP. This precludes the possibilitythat the stimulation observed by CTP at lowATP to Mn++ ratio was due to the presence of acontaminating enzyme capable of catalyzingthe regeneration of ATP from ADP producedin the biosynthetic reaction. Other data in Table

2 show that the effect of CTP was not specific.Similar results were obtained also with UTP,GTP, ITP, and IDP. Thus, in the presence of7.0 mm Mn++, mixtures containing 2.5 mm ATPand 2.5 mm of any one of the above four nucleo-tides were more active than 2.5 mm ATP alone.However, when tested in the presence of 7.5mM ATP, each of the four nucleotides wasinhibitory.The complex interrelationship of ATP and

CTP is further illustrated by Fig. 5, which showsdouble reciprocal plots of the reaction velocityagainst ATP concentration at a fixed concentra-tion of Mn++ (7 mM). In the plot with ATPalone, which is described by the solid line inFig. 5, a multiphasic curve is obtained with aminimum (corresponding to maximum velocity)occurring at an ATP concentration (7.5 mM)that is nearly equivalent to the concentration ofMn++ added. On the other hand, when thevarious incubation mixtures were supplementedwith CTP to yield a total nucleoside triphosphate(CTP + ATP) concentration of 7.5 mi, thedouble reciprocal plot gave a straight line fromwhich the apparent Km for ATP was calculatedto be 9.4 X 10-4 M. It appears from these dataand those presented in Fig. 3 and 4 that a one-to-one complex of ATP and Mn++ (ATP-Mn)is the catalytically active substrate, and thatexcesses of either Mn++ or ATP are inhibitory,perhaps because they produce secondary com-plexes that are less active or because they com-pete with ATP-Mn for the catalytic site. If theseassumptions are correct, the activating effect of

2-

0 100 200 300 400 500 600 700

[ATP]

FIG. 5. Double reciprocal plot of ATP concen-tration versus reaction velocity in the presence andabsence of CTP. Reaction mixture: 0.1 M glutamate,0.05 m NH4Cl, 0.007 m MnCl2, 0.022 M imidazole,27 ,tg ofglutamine synthetase, and ATP (as indicated).Incubation time was 7.5 min at 37 C. CTP was addedat a level such that the combined ATP and CTP con-centrations were equal to 0.0075 m.

1012 J. BACTERIOL.

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

GLUTAMINE SYNTHETASE. V

CTP might be attributable to its capacity tocomplex with Mn+ and thereby neutralize theinhibitory action of this cation. On the otherhand, the inhibitory effect of CTP, observedwhen the total nucleoside triphosphate levelexceeds the Mn++ level, could be explainedeither by a direct competition of CTP for ATP-Mnat the catalytic site or by its ability to competewith ATP for Mn++. This would lead to adecrease in the level of ATP-Mn and conse-quently with the formation of free ATP, which inturn could compete with ATP-Mn at the catalyticsite. Obviously, other, more complicated inter-pretations that invoke the existence of multiplenucleotide binding sites (activation, catalytic,and inhibition sites) could also account for theobserved effects.

Effect of CTP and divalent cations on theE. coli glutamine synthetase. In contrast to theB. licheniformis enzyme, the glutamine syn-thetase from E. coli can be activated by Mg+,and, at saturating concentrations of Mg++ (50mM), the activity is inhibited by CTP (14). Inview of the unusual response of the B. licheni-formis enzyme to Mnl and nucleoside triphos-phates, it seemed worthwhile to reinvestigatethe interrelationship between divalent cationand CTP inhibition for the E. coli enzyme. Datashowing the effect of ATP concentration at 5 mmMn+ and at 50 mm Mg++ are summarized inFig. 6. The data confirm the earlier studies in

300

D 200

C

C7

0 2 4 6 8 10

ATP, mM

12 14 16 18

FIG. 6. Effect of divalent cations and ATP concen-

tration on the activity of Escherichia coli glutaminesynthetase. In the experiment with Mn++, the reactionmixture contained 5 mm MnCI2, 50 mM NH4Cl, 0.1

L-glutamate, 22 mM imidazole (pH 7.0), 1.32 ,Agof pure E. coli glutamine synthetase. Total volwne 0.3ml. For the experiment with Mg++, the conditionswere the same except that 50 mM MgCl2 replaced theMnCl2 and 1.06 jig of enzyme was used. The finalvolume was 0.6 ml; temperature, 37 C.

that, when Mg++ was the activating cation,there was a more or less normal saturationresponse to ATP concentration over the range0 to 10 mm. Substrate inhibition was observed atconcentrations of ATP greater than 10 mM.It is evident that neither the saturation kineticsnor the ATP inhibition is related in a simpleway to the ratio of ATP to Mg++ since the latterwas always present in excess. Moreover, as canbe seen in Table 3, when Mg++ was the cation,CTP inhibited the enzyme at both high and lowconcentrations of ATP. These results are inaccord with the earlier conclusion (14) thatCTP is a specific feedback inhibitor of the en-zyme under these conditions. On the other hand,when Mn+ was the divalent cation, the effectof nucleoside triphosphates was very similarto that observed with the B. licheniformis en-zyme. Thus, as shown in Fig. 6 the ATP ac-tivity curve shows a sharp maximum at thepoint where the concentrations of ATP and Mn++are nearly equivalent (e.g., 6 mM). The importanceof the nucleoside triphosphate to Mn++ ratio isfurther illustrated by the data of Table 3 whichshow that, when the concentration of ATP wasless than the Mn++ concentration, the additionof CTP resulted in activation, so long as theCTP + ATP concentration was not greaterthan the Mn++ concentration; however, when theATP concentration was equal to or greater thanthat of the Mn++, further addition of CTP onlycaused inhibition. The situation is thus identical

TABLE 3. Effect of divalent cations on the responseof Escherichia coli glutamine synthetase

to CTP and ATPa

Divalent cation ATP added CTP added Activity,

mM mm

Mg++ 1.25 0 2281.25 2.5 1857.5 0 2927.5 2.5 2150 2.5 0

Mn++ 2.5 0 1142.5 2.5 1602.5 5.0 2627.5 0 2557.5 2.5 2380 0 10

2.5

a Reaction mixtures contained 0.1 M L-gluta-mate, 0.05 M NH4Cl, 22 mM imidazole buffer (pH7.0), 2,g of pure E. coli glutamine synthetase,and, when indicated, 50 mM MgCl2 or 7 mm MnCl2.Total volume, 0.3 ml. Incubation, 37 C, 10 min.

bExpressed in millimicromoles of Pi per 10 min.

1013VOL. 94, 1967

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

HUBBARD AND STADTMAN

to that observed with the B. licheniformis en-

zyme. Examination of Fig. 6 shows that theslopes of the ATP-activity curves for concentra-tions greater than 9 mm ATP are essentially thesame whether the divalent cation is Mn+ or

Mg++. This suggests that the inhibition by ex-cess ATP is unrelated to the nature or concen-tration of divalent cation and is therefore notdue to competition between free ATP and anATP-metal complex for the catalytic site, aswas previously proposed. Perhaps there is aspecific allosteric inhibitor site for ATP in ad-dition to the substrate site.

Glutamotrantferase activity. Leonard et al. (5)reported that cell-free extracts of B. licheniformiscatalyze the formation of y-glutamylhydroxamatein a reaction mixture containing L-glutamine,NH20H, MnSO4, ATP, and arsenate (or ortho-phosphate). These authors could not demon-strate the glutamyl transfer reaction when ADPwas substituted for ATP. The above are incontrast to findings with mammalian and E. coliglutamine synthetases (9, 12) where ADP ismore active than ATP as the activator for theglutamyl transfer reaction. In the present study,we were able to demonstrate the glutamyl trans-fer reaction with ADP (Fig. 7) and to a lesser

4.0

Plus 0.1 M Glutomate

3.0

0

Li

O 2.0

1.0

2 3 4 5 6 7 8

ADP CONCENTRATION (mM)

FIG. 7. Relationship of ADP concentration toglutamotransferase activity. Reaction mixture con-tained 0.05 M L-glutamine, 0.1 X NH20H, 0.02 marsenate, 3.0 mm MnCI2, 21 mm imidazole (pH 7.0),and 0.15 mg of glutamine synthetase. ADP and glu-tamate were added as indicated. Incubation time was30 min at 37 C.

extent with ATP (Table 4). Maximal activity wasfound with 2 mm ADP. It is of interest that thisADP concentration is approximately equal tothe 3 mm level of MnCl2 employed. This indicatesthat a complex of ADP and Mn+ is an "activeform" for the glutamotransferase similar to theproposed ATP-Mn+ complex for the biosyn-thetic reaction. With ATP as the activator, thehighest level of activity was detected with 10 mmATP which was the highest concentration tested.The possibility was considered that the latteractivity was due to the presence of small quanti-ties of ADP in the ATP solution. However, thisappeared not to be the case since preincubationof the reaction mixture with pyruvate kinase andphosphoenolpyruvate had little effect on theATP-dependent activity (Table 4). This treat-ment would convert contaminating quantitiesof ADP to ATP. Preincubation of the reactionmixture which contained ADP with the pyruvatekinase and phosphoenolpyruvate markedly re-duced the glutamotransferase activity.The discrepancy between our results and those

of Leonard et al. (5) is not understood. Theirfailure to observe the glutamyl transfer reactionwith ADP could be explained if their enzymepreparations contained an active adenylatekinase. In the presence of this enzyme, ADPwould be partly converted to AMP which, inthe presence of glutamine, is a very potent in-hibitor of the B. licheniformis glutamine syn-thetase (3).

DIscussIoNThe complex response of glutamine synthetase

activity to nucleoside triphosphates and Mn++TABLE 4. Glutamyl transfer reaction with ADP

and ATP as the nucleotide activators

Additionse Amt (umoles) of gluNHOHper 30 min

None................. 00.4 mM ADP............. 1.00.4 mM ADP plus2mMPEP.1.02PK ................... 1.122mMPEP andPK..... 0.04

10 mMATP.............. 0.1610 mM ATP plus 2 mmPEP and PK........... 0.14

v Reaction mixture: 0.05 M L-glutamine, 0.1 MNH20H, 0.02 M arsenate, 3 mm MnCl2, 21 mMimidazole, and 0.15 mg of glutamine synthetase.PEP = phosphoenolpyruvate; pK = pyruvatekinase (25 ,ug of protein). Prior to the introduc-tion of glutamine synthetase, the otherwise com-plete reaction mixture was preincubated for 15min.

1014 J. BAcTERioL.

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from

GLUTAMINE SYNTHETASE. V

concentration provides the basis of a uniquemechanism of enzyme regulation that may be ofbiological significance. As was noted earlier,the concentration of ATP required for optimalactivity is equal to the concentration of Mn l;however, the maximal activity obtained withstoichiometric amounts of ATP and Mn+ isindependent of the Mn concentration over a

considerable range of Mn+ concentrations.It is evident therefore that full expression of thebiosynthetic capacity of the enzyme can beachieved at different levels of ATP or of Mn++so long as they are present in stoichiometricamounts. This provides for considerable flexi-bility in the regulation of glutamine syntheticactivity, since factors that tend to decrease theconcentration of Mnt will coincidentallyreduce the concentration of ATP necessary forfull catalytic activity. Superimposed on this basicregulatory mechanism are the effects of othernucleoside triphosphates which may eitheractivate or inhibit the catalytic activity de-pending upon the ratio of ATP to Mn 1. Thuswhen the ratio of ATP to Mn+ is equalto or greater than 1.0, the addition of othernucleoside triphosphates will cause inhibition ofthe enzyme; however, when the ratio of ATPto Mn++ is less than 1.0, activation of the en-zyme is obtained with increasing amounts ofother nucleoside triphosphates up to the pointat which the total concentration of ATP plusother triphosphate is equal to the concentrationof Mn+ . Then, higher concentrations of thetriphosphates will cause inhibition. We see thatthe level of the Mn+ is a critical determinantin establishing the sensitivity of the enzyme toATP and nucleoside triphosphate concentra-tions, whereas the ratio of ATP to Mn+ de-termines the susceptibility of the enzyme toactivation or inhibition by the nucleoside tri-phosphates.

Other instances have been reported in which aone-to-one complex of nucleotide and divalentcation is the substrate for enzymes (10, 11).Moreover, the importance of divalent cationsin the effector responses of other enzymes, viz.,isocitrate dehydrogenase (1) and phospho-fructokinase (2), has been noted previously.The significance of such effects in the overallcellular regulation of enzyme activity probablydeserves more attention.

LITERATuRE CITED

1. ATKINSON, D. E., J. A. HATHAWAY, AND E. C.SMITH. 1964. Kinetic order of the yeast diphos-

phopyridine nucleotide isocitrate dehydro-genase reaction and a model for the reaction.J. Biol. Chem. 240:2682-2690.

2. ATKINSON, D. E., AND G. M. WALTON. 1964.Kinetics of regulatory enzymes. Escherichiacoli phosphofructokinase. J. Biol Chem.240:757-763.

3. HUBBARD, J. S., AND E. R. STADTMAN. 1967.Regulation of glutamine synthesis. II. Patternsof feedback inhibition in microorganisms. J.Bacteriol. 93:1045-1055.

3a. HUBBARD, J. S., AND E. R. STADTMAN. 1967.Regulation of glutamine synthetase. VI. Inter-actions of inhibitors for Bacillus licheniformissynthetase. J. Bacteriol. 94:1016-1024.

4. KOHLHAW, G., W. DRAGART, AND H. HOLZER.1965. Parallel-repression der synthese vonglutamin-synthetase und DPN-abhangiger glu-tamat-dehydrogenase in hefe. Biochem. Z.341:224-238.

5. LEONARD, G. G., R. D. HOUSEWRIGHT, AND C. B.THORNE. 1962. Effects of metal ions on theoptical specificity of glutamine synthetase andglutamyl transferase of Bacillus licheniformis.Biochim. Biophys. Acta 62:432-434.

6. LOWRY, 0. H., N. J. ROSEBROUGH, A. L. FARR,AND R. J. RANDALL. 1951. Protein measure-ment with the Folin phenol reagent. J. Biol.Chem. 193:262-275.

7. MARTIN, R. G., AND B. N. AMES. 1961. A methodfor determining the sedimentation behaviorof enzymes: application to protein mixtures.J. Biol. Chem. 236:1372-1379.

8. MEISTER, A. 1962. Amino group transfer (survey),p. 193-217. In P. D. Boyer, H. Lardy, andK. Myrback [ed.], The enzymes, vol. 6. Aca-demic Press, Inc., New York.

9. MEISTER, A. 1962. Glutamine synthesis, p. 443-468.In P. D. BOYER, H. LARDY, AND K. MYRBACK[ed.], The enzymes, vol. 6. Academic Press,Inc. New York.

10. MELCHIOR, J. B. 1965. The role of metal ions inthe pyruvic kinase reaction. Biochemistry4:1518-1525.

11. MILDVAN, A. S., ANM M. COHN. 1966. Kineticand magnetic resonance studies of the pyruvatekinase reaction. J. Biol. Chem. 241:1178-1193.

12. WOOLFOLK, C. A., B. SHAPIRO, AND E. R. STADT-MAN. 1966. Regulation of glutamine synthetase.I. Purification and properties of glutaminesynthetase from Escherichia coli. Arch. Bio-chem. Biophys. 116:177-192.

13. WOOLFOLK, C. A., AND E. R. STADTMAN. 1964.Cumulative feedback inhibition in the mutipleend product regulation of glutamine synthetasein Escherichia coli. Biochem. Biophys. Res.Commun. 17:313-319.

14. WOOLFOLK, C. A., AND E. R. STADTMAN. 1967.Regulation of glutamine synthetase. III.Cumulative feedback inhibition of glutaminesynthetase from Escherichia coli. Arch. Bio-chem. Biophys. 118:736-755.

1015VOL. 94, 1967

on March 29, 2019 by guest

http://jb.asm.org/

Dow

nloaded from