JOURNAL OF BACTERIOLOGY, June 1978, p. 1046-1055

Vol. 134, No. 3

0021-9193/78/0134-1046$02.00/0 Copyright i 1978 American Society for Microbiology

Printed in U.S.A.

Mutations That Alter the Covalent Modification of Glutamine Synthetase in Salmonella typhimurium SHARON BANCROFT, SUE GOO RHEE,t CYNTHIA NEUMANN, AND SYDNEY KUSTU* Department of Bacteriology, University of California, Davis, California 95616 Received for publication 31 January 1978

ginD and ginE mutant strains of Salmonella typhimurium lack three of the four activities required for reversible covalent modification of glutamine synthetase (GS; EC 6.3.1.2). The ginD strains, which are unable to deadenylylate GS and therefore accumulate the adenylylated or less active form of the enzyme, were isolated as glutamine bradytrophs. They lack the activity of PIIA uridylyltransferase, one of the proteins required for deadenylylation of GS; in addition, they lack PIID uridylyl-removing activity. Mutations in ginD are suppressed by second-site mutations in ginE that eliminate the activity of GS adenylyltransferase (EC 2.7.7.42) and thus prevent adenylylation of GS. The ginD and glnE strains have one-third to one-half as much total GS as the wild-type strain when they are grown in a medium containing a high concentration of NH4'. The wildtype strain derepresses synthesis of GS fourfold in response to nitrogen limitation; ginD and ginE strains derepress synthesis of the enzyme fourfold and sevenfold, respectively. Thus, mutations that alter covalent modification of GS in Salmonella do not significantly affect derepression of its synthesis. The ginD gene lies at 7 min on the Salmonella chromosome and is 50% linked to pyrH by P22mediated transduction.

Biochemical studies of Stadtman (32, 36; reviewed in 16, 33), Holzer (14, 25), and their colleagues have demonstrated that glutamine synthetase (GS) (L-glutamate:ammonia ligase [ADP forming], EC 6.3.1.2) is reversibly covalently modified by adenylylation-deadenylylation in a variety of gram-negative bacteria and that the adenylylated form of the enzyme is catalytically less active than the unmodified form. To probe the significance of this covalent modification in vivo, we isolated mutant strains of Salnonella typhimurium that lack three of the four activities required for covalent modification of GS. We report here their preliminary characterization. Magasanik and his colleagues characterized similar strains of Klebsiella aerogenes (reviewed in 22). They reported that mutations that alter the covalent modification of GS (12, 19) also exert a large effect on synthesis of the enzyme (19, 28; F. Foor, K. A. Janssen, S. L. Streicher, and B. Magsnik, Fed. Proc. 34:514, 1975). They proposed that covalent modification of GS directly controls its synthesis (12, 19, 22). In contrast to findings reported for Klebsiella, we find that mutations that alter covalent modification of GS in Salnonella do not affect derepression of its synthesis in response to nitrogen

limitation. Since physiological control of the synthesis and covalent modification of GS is essentially the same among the enteric bacteria (4, 14, 15, 36), we think that the differences between our findings and those of Magasanik and his colleagues are probably not due to strain differences between Salmonella and Klebsiella. (A preliminary report of a portion of this work appeared previously [15].) MATERIALS AND METHODS Chemicals. Mutagen ICR191E (10, 27) was kindly donated by R. M. Peck, R. K. Preston, and H. J. Creech, Chemotherapy Laboratory of the Institute for Cancer Research, Philadelphia, Pa. Snake venom phosphodiesterase (EC 3.1.4.1; 1 mg/ml, 1.5 U/mg) was obtained from Boehringer Mannheim. Glutamine solutions (100 mM) were neutralized to pH 7, sterilized by filtration, and stored at 4°C. Media and growth of bacterial strains. Minimal medium refers to the salts mixture listed below, which was supplemented with glucose (final concentration, 0.4%), histidine (0.3 mM), and a nitrogen source: 1 g of K2S04, 17.7 g of K2HP04-3H20, 4.7 g of KH2PO4, 0.1 g of MgSO4 * 7H20, and 2.5 g of NaCl per liter. For a condition of excess nitrogen, 10 mM NH4Cl plus 3 mM glutamine was added; for nitrogen limitation, 3 mM glutamine was the sole nitrogen source (5). Minimal agar plates were made with the above salts mixture supplemented with the carbon and nitrogen sources indicated in the text or with medium E of Vogel and Bonner (37) supplemented with 0.4% glucose as carbon

t Address: Laboratory of Biochemistry, National Heart, Lung and Blood Institute, Bethesdc, MD 20014. 1046

VoL. 134, 1978

MUTATIONS THAT ALTER ADENYLYLATION OF GS

Nutrient broth medium contained 8 g of nutrient broth (Difco) and 5 g of NaCl per liter. Cultures were incubated at 37°C with aeration. Bacterial

source.

growth was monitored turbidimetrically by measuring absorbance at 650 nm Isolation and genetic analysis of strains. All strains constructed for this work (Table 1) were derived from S. typhimurium strain LT2. P22 phage was (HT int) (30); this phage was obtained from G. Roberts, who introduced the int mutation. Procedures for isolating strains sensitive to phage P1 and for transductions with phage Plkc were those described by Enomoto and Stocker (11) and Ornellas and Stocker (26). Strains with ginD mutations were isolated by ICR or diethyl sulfate mutagenesis and penicillin counterselection for glutamine auxotrophs. They were identified on the basis of a slight growth defect on minimal plates and a more pronounced defect on nutrient broth plates; their growth on both media was stimulated by glutamine. Revertants and suppressors ofglnD strains were isolated by selecting for glutamine prototrophs (strins resistant to glutamate inhibition) on minimal plates (medium E) to which 10 mM glutamate was added. Since most glutamine prototrophs were found to contain second-site suppressor mutations, it was necessary to devise a rapid procedure for screening revertants. The basis for this screening was the observation that wild-type donor phage yielded many large transductant colonies in 24 h when crossed with a ginD recipient on medium E + 10 mM glutamate, whereas phage grown on a donor strain carrying a suppressor mutation (and glnD) gave rise to fewer and smaller colonies. Strains that appeared to be revertants (wild type) by the above criterion no longer contained a gin mutation linked to pyrH by P22mediated transduction; the ginD mutation could be recovered from strains carrying second-site suppressors.

The ginD mutation in the ginD ginE double mutants was repaired by P22-mediated transduction as follows: a donor (glnD+) carrying a TnlO element (20) 10% linked to ginD was used to transduce the double

Strain

TA831 SK101 SK103

SKill SK203 SK204 SK205

SK254

SK254

SK256 SK257 JL2979 JL1268 JL1272 JL1220 JL1271 Recipient. b Donor.

Genotype

mutants to tetracycline resistance, and resistant clones were tested for their phenotype on nutrient broth (see Fig. 2). Of 20 tetracycline-resistant clones derived from strain SK203, 18 had the phenotype of the recipient, as expected; 2 clones had the phenotype of the wild type. These clones (glnD ginE) no longer contained a gin mutation linked topyrH by P22-mediated transduction. However, they still lacked adenylyltransferase activity. Similar results were obtained starting with strain SK204 as the recipient. Preliminary conjugation mapping of the gInD gene was carried out as described previously (21). Linkage of ginD to pyrH and pan by P22- or P1-mediated transduction was determined by using phage grown on ginD strains as donor and scoring for the donor phenotype on medium E + 10 mM glutamate. (The pyrH strains, which are cold sensitive, were transduced to growth at 20°C.) GS assays. Cells were grown to late exponential or early stationary phase and suspended in the following buffer (0.25 g of cells + 1.5 ml of buffer): 20 mM

tris(hydroxymethyl)aminomethane-chloride (pH 8.0), 10 mM MgCl2, 1 mM MnCl2. They were disrupted at 18,000 lb/in2 in a French pressure cell, and debris was removed by centrifugation at 16,000 x g for 30 min. Total activity of GS and the approximate degree of adenylylation of the enzyme in crude extracts were determined by using the y-glutamyl transfer assay of Stadtman et al. (34) with the modifications described (21). In the presence of 0.3 mM Mn2", this assay is a measure of total activity independent of degree of adenylylation. In the presence of 0.3 mM Mn2+ + 60 mM Mg2+, activity of adenylylated subunits is in-

hibited. Thus, the ratio of activities +Mg2+/-Mg2+, which decreases as adenylylation increases, provides an estimate of degree of adenylylation. To remove adenylyl groups from GS, extracts were treated with snake venom phosphodiesterase, essentially as described previously (21); prior to treatment, the extracts were dialyzed against the same buffer in which cells were broken. Preparation of extracts for ATase, Pn, UTase,

TABLE 1. Bacterial strains Parent

hisF645 glnD77 hisF645 glnD79 hisF645 glnD96 hisF645 glnE121 glnD96 hisF645 glnE122 glnD79 hisF645 gbE123 glnD79 hisF645 glnE121 hisF645 zag-209:.:Tn10 gLnE122 hisF645 zag-209.:TnlO

hisF6W5 glnD T hisF645 glnD+ cysCD59 cdd-7 udk-10 zag-209:.:TnlO pan-3

1047

TA831 TA831 TA831 SKill SK103 SK103

SK2O3,a JL2979b SK204,a JL2979b SK103 SK103

JL1268 pan-3galE pyrHi609 cysCD519 cdd-7 udk-10 pyrHi 609 cysCD519 cdd-7 udk-lOgalE JL1220

Source G. F.-L. Ames ICR mutagenesis ICR mutagenesis Diethyl sulfate mutagenesis ICR mutagenesis ICR mutagenesis

Spontaneous Transduction Transduction

Spontaneous Spontaneous C. F. Beck J. L. Ingraham J. L. Ingraham J. L. Ingraham J. L. Ingraham

1048

BANCROFT ET AL.

and UR assays. Cells were usuaily grown in medium E supplemented with 0.4% glucose as carbon source and with histidine and 3 mM glutamine. Extracts were prepared by suspending cells in the following buffer: 10 mM 2-methylimidazole (pH 7.6), 10 mM MgCl2, 0.1 mM ethylenediaminetetraacetic acid, 0.5 mM dithiothreitol. For adenylyltransferase (ATase; EC 2.7.7.42) assays, 0.25 g of cells was suspended in 1 ml of buffer. Extracts were prepared as described for GS assays. For PI,, uridylyltransferase (UTase), and uridylyl-removing (UR) assays, 1 g of cells was suspended in 2 ml of buffer. (In most cases cells had been frozen.) Cells were disrupted by sonic oscillation. Extracts were treated with 1% streptomycin sulfate, centrifuged, and dialyzed against the buffer in which cells were broken. Rationale for ATase, PIJA, PnD, UTase, and UR assays. Activities of ATase, PIIA, PIID, and UTase were determined spectrophotometrically at 260C. The assays are based on the ability to measure an increase in the activity of either the adenylylated or the unmodified form of GS, selectively, by using appropriate modifications of the y-glutamyl transfer assay. For ATase assays, SalmoneUa extracts were first incubated with glutamine and subsaturating amounts of unmodified GS from Escherichia coli (GSMao) (the subscript n'in GS;, which varies from 0 to 12, indicates the average number of adenylylated subunits per dodecamer); the increase in activity of adenylylated GS was then determined. (In the presence of glutamine, activation of ATase by the levels of PIIA present in crude extracts is negligible.) For PIIA and PIID assays, Salmonella fractions were incubated with subsaturating amounts of E. coli ATase and E. coli GSi-5 or GS12, respectively, and the increase in activity of adenylylated or unmodified GS, respectively, was then determined. For UTase assays, extracts were incubated-with partially purified E. coli PIIA, and the increase in activity of PIID was then determined as described above. Activities are expressed as AA1w0 (change in absorbance at 540 nm) per milligram of crude protein. (The AA140 is due to formation of the ferric chelate of -y-glutamylhydroxamate.) UR activity was assayed directly by following the release of [3H]UMP from PIID. Procedures for ATase assays. ATase activity was determined by incubating 10 A1 of crude extract (0.1 mg of protein) in a 50-1l reaction mixture that contained 24 mM 2-methylimidazole (pH 7.6), 36 mM MgCl2, 150 Ag of GSr.O, 6 mM ATP, and 24 mM glutamine. After 10 min of incubation, 1.0 ml of a final reaction mixture was added that contained 50 mM dimethylglutaric acid-triethylamine buffer (pH 6.8), 60 mM KCl, 0.4 mM MnCl2, 150 mM glutamine, 40 mM hydroxylamine, 20 mM arsenate, 0.8 mM ADP, and 60 mM alanine (before the addition of MnCl2 this reaction mixture was titrated to pH 6.4 with KOH). After 10 min of incubation, the reaction was stopped by addition of 1.0 ml of a stop mix that contained 6.6% FeCl3, 16% trichloroacetic acid, and 4 N HCL. Precipitate was removed by centrifugation, and the Am0 was determined. Procedures for Pn assays. The purification steps necessary to detect PnA activity in a ginD strain were described previously (15). For assays of Pi, activity in ginE strains, all of the PI, protein in the streptomycin

J. BACTERIOL.

supernatant (prepared as described above) was converted to the PIID form by the action of endogenous UTase; for this reaction 100 id of streptomycin supernatant was incubated in a 120-,ul reaction mixture containing 6.5 mM UTP, 38 mM a-ketoglutarate, and 7.5 mM ATP. After 15 or 40 min of incubation (the reaction was complete at 15 min), PIID activity was measured by incubating 10pl of the above mix (0.16 mg of protein) in a 50-pl reaction mixture that contained 70 mM 2-methylimidazole (pH 7.6), 21 mM MgCl2, 12 U of ATase (S. G. Rhee, M. Wittenberger, P. B. Chock, and E. R. Stadtman, submitted for publication), 21 pg of GSjio5, 1.4 mM ATP, 28 mM aketoglutarate, 21 mM K-PO4 (pH 7.0), 2 mM dithiothreitol, and 0.4 mM ethylenediaminetetraacetic acid. After 15 min of incubation, 1.0 ml of a final reaction mixture was added that contained 50 mM 2-methylimidazole (pH 8.0), 100 mM KCl, 0.4 mM MnCl2, 150 mM glutamine, 40 mM hydroxylamine, 40 mM potassium arsenate, and 0.1 mM ADP. After 5 min of incubation, the reaction was stopped, the precipitate was removed, and the Asw was determined. P>rcedures for UTase assays. UTase activity was measured by incubating 10 i1 of the dialyzed streptomycin supernatant (0.2 mg of protein) in a 30id reaction mixture that contained 50 mM 2-methylimidazole (pH 7.6), 100 mM KCl, 10 mM MgCl2, 0.009 U of PIIA (S. G. Rhee, C. Y. Huang, P. B. Chock, and E. R. Stadtman, submitted for publication), 1 mM UTP, 5 mM a-ketoglutarate, and 1 mM ATP. After 7 min of incubation, 70 ,l of a second reaction mixture for measuring PIID activity (see above) was added. After 4 min of incubation, 1.0 ml of a final reaction mixture for measuring the activity of unmodified glutamine synthetase was added. After 4 min of incubation, the reaction was stopped, the precipitate was removed, and the Aw^0 was determined. Prcdures for UR assays. UR activity was measured by incubating 10 pd of the dialyzed streptomycin supernatant in a 50-pl reaction mixture containing 50 mM 2-methylimidazole (pH 8.0), 100 mM KCI, 1 mM MnCl2, [3H]UMP-labeled P-,, (5.3 U, 24,600 cpm [Rhee, Chock, and Stadtman, submitted for publication]). After 30 min, 300 ll of water and 1 ml of 10% trichlofoacetic acid were added to terminate the reaction. The precipitated protein was removed by centrifugation, and the amount of radioactivity that had been released into the supernatant (0.5-ml sample) was determined.

RESULTS Rationale for the isolation of mutant strains. The isolation procedures were based on biochemical studies of the covalent modification of GS summarized in Fig. 1. As demonstrated by the elegant studies of Stadtman, Ginsburg, and their colleagues, several proteins are required for covalent modification of GS (reviewed in 16, 33). Both adenylylation and deadenylylation of the enzyme are catalyzed by ATase working in conjunction with regulatory protein P,1 (3, 31). When PI1 is in the unmodified form, the adenylylation reaction is favored (9). When P,, is in

VOL. 134, 1978

MUTATIONS THAT ALTER ADENYLYLATION OF GS

ADENY LYLATION ATP GS

UR

GS

\N,.ATa!! PIIA

PPj -.

GS-AMP

UTose

PrrD-UMP GS-AMP

ADP Pi DEADENY LYLATION FIG. 1. Proteins involved in covalent modification of GS. Adapted from Ginsburg and Stadtman (16). PIIA, Protein component that stimulates adenylylation of GS. PII, Uridylylated PnIA, which stirnulates deadenylylation of GS; GS-AMP is the less active form (16). Strains with a glnD mutation lack UTase and UR activities; glnE strains lack ATase activity.

the modified or uridylylated form, the deadenylylation reaction is favored (9, 24). Covalent modification of PI, is catalyzed by the enzyme UTase, and removal of uridylyl groups is catalyzed by a UR activity (24). Co-purification of UTase and UR suggested that attachment and removal of uridylyl groups might be two activities of the same protein (1). Thus, either three or four proteins (ATase, PI,, UTase, and UR enzyme) are required for covalent modification of GS. Biochemical evidence suggested that strains lacking UTase activity would accumulate the adenylylated or less active form of GS and, thus, that they might require glutamine for growth. Consistent with this, such strains (glnD) were isolated as glutamine bradytrophs. The evidence suggested that strains lacking ATase activity would be unable to modify GS initially and, thus, that the loss of ATase activity should suppress the loss of UTase activity. Indeed, strains lacking ATase activity (glnE) were isolated among suppressors of glnD strains. Isolation of ginD, ginD glnE, and ginE strains. Five strains that lack UTase activity (glnD) were isolated as glutamine bradytrophs from ICR- and diethyl sulfate-mutagenized cultures of strain TA831 (hisF645) as described previously (15). (In all, twenty-three independent glutamine-requiring strains were characterized.) To isolate strains lacking ATase activity, strain SK1ll, which contains a diethyl sulfateinduced glnD mutation, was mutagenized with the frameshift mutagen ICR191E (2), and glu-

1049

tamine prototrophs (strains resistant to glutamate inhibition) were selected as described in Materials and Methods. This procedure was designed to enrich for strains with mutations that had suppressed the glnD mutation by causing complete loss of function of another gene product (2). Twelve glutamine prototrophs were tested directly for loss of ATase activity in crude extracts, and one such strain (gInD glnE) was found. After this strain was tested for a phenotype that could be used to identify additional glnD glnE strains on petri plates (see below), strains with spontaneous or ICR-induced glnE mutations were isolated from the glnD strain SK103. Based on phenotype (confirmed by enzyme assays in four cases), about a third (18/54) of spontaneous glutamine prototrophs (strains resistant to glutamate inhibition) derived from strain SK103 contained a glnE mutation. To characterize glnE mutations in a wild-type background, the glnD mutation in two gInD glnE strains, SK203 and SK204, was repaired as described in Materials and Methods to yield strains SK254 and SK255, respectively. Phenotype of strains with ginD and ginE mutations. The doubling times of strains with glnD and gInE mutations in minimal medium are summarized in Table 2. Three glnD strains, SK101, SK103, and SKill, had doubling times of about 60 min, whereas the parent strain, TA831, had a doubling time of 45 min. If the medium was supplemented with glutamine, glnD strains grew as well as wild type (doubling times of about 47 min). The glnD glnE strains SK203 and SK204 had doubling times of 50 or TABLE 2. Doubling times of strains with mutations affecting the covalent modification of GS Doubling time (min) Strain and relevant genotype 10 mM NH4a 10 mM NH4+a +5mMglutamine TA831 (wild type) 45 45 SK101 (gLnD77) 60 47e SK103 (ginD79) 60b 47e SKIll (gbtD96) 65 52c 52 SK203 (glnD96ginEl21) 45 SK204 (glnD79glnE122) 49 45 SK254 (glnE121) 45 43 SK255 (glnEl22) 45 45 50 SK256 (glnD+) (spontaneous 45 revertant of SK103) Strains were grown in minimal medium with the nitrogen sources indicated and glucose as carbon source. The growth temperature was 37°C. 'In other experiments the doubling time of ginD strains was as long as .83 min. In other experiments the doubling time of ginD strains was as long as 56 min. a

c

1050 BANCROFT ET AL.

J. BACTERIOL.

45 min in minimal medium without or with Glutamine prototrophs (strains resistant to gluadded glutamine, respectively. Like the wild tamate inhibition) including ginD ginE strains type, the ginE strains SK254 and SK255 had a were isolated from the glnD strains on these doubling time of 45 min in medium without or media. The doubling times of a ginD strain with added glutamine. (SK203), a ginD ginE strain (SK204), and wild Other phenotypic characteristics of strains type in minimal medium containing 10 mM gluwith ginD and ginE mutations were studied on tamate were 80, 60, and 47 min, respectively. agar plates; in general, their physiological basis On nutrient broth plates ginD strains grow is not yet understood. These phenotypic prop- poorly (Fig. 2). (There is little glutamine in erties have been used to differentiate between autoclaved nutrient broth, and amino acids inginD, ginD ginE, ginE, and wild-type strains. hibit ginD strains.) The glnD ginE strains grow Assignments of genotype made on the basis of better than ginD strains but not as well as wild these phenotypic properties were confinned by type. On nutrient broth plates the ginE strains enzyme assays and genetic analysis. Since ginE are indistinguishable from wild type. When numutations prevent adenylylation of GS, they trient broth plates are supplemented with glushould eliminate the effect of other mutations tamine, all strains grow as well as the wild type. altering covalent modification; hence, it was ex- The phenotype on nutrient broth plates was pected that ginD ginE and ginE strains would used to distinguish glnDglnE strains from other have the same phenotypic characteristics. In classes of glutamine prototrophs (strains resistseveral instances (see below) they do not; again, ant to glutamate inhibition) derived from ginD the physiological basis for differences in pheno- strains; most other prototrophs grew as well as type between these strains is not understood. wild type. It was also used to distinguish between On minimal plates ginD strains grow at near glnD glnE and ginE strains during construction wild-type rate; they grow slowly on plates con- of ginE single mutants. taining glutamate (final concentration, 10 mM) Proteins required for covalent modificaor Casamino Acids (final concentration, 0.15%). tion of GS. Assays of GS activity in ginD strains

I

;, !f

i,

%,

-;[ I7(.,14 ( t

SI. .'jjl IJI ?.,;

with

Adi"D (11"k.'

"InP"

::;"

-i-ii1tric-pit agw'.

VOL. 134, 1978

MUTATIONS THAT ALTER ADENYLYLATION OF GS

(see below) indicated that the enzyme was highly adenylylated under all growth conditions. This suggested that ginD strains lacked the activity of a protein required to deadenylylate GS. As shown in Table 3, crude extracts of the ginD strains SK101, SK103, and SK1ll lacked UTase activity; they had normal ATase activity. Extracts of the ginD ginE strains SK203, SK204, and SK205 (derived from SKill and SK103) lacked ATase activity as well as UTase activity (Table 3). ATase activity of the wild-type strain TA831 increased nonlinearly with increasing protein (data not shown). However, ATase activity in an extract of strain SK203 remained undetectable over a range of protein concentrations. Extracts of the ginE strains SK254 and SK255 (derived from SK203 and SK204, respectively) lacked ATase activity but had normal UTase activity. Appropriate mixing experiments indicated that mutant extracts lacking ATase or UTase activity did not inhibit the ATase or UTase activity of a wild-type extract (data not shown). Since co-purification of UTase and UR had suggested that they might be two activities of the same protein (1), ginD strains were assayed for UR activity. The three that were assayed had only 2 to 12% of wild-type UR activity (Table 3). Strains in which the ginD mutation was repaired by P22-mediated transduction (SK254 and SK255) or in which it was spontaneously reverted (SK256 and SK257) had 80 to 95% of wild-type UTase activity and 90 to 110% of UR activity. Our previous report (15) that strain SK103 (ginD) had normal UR activity was in error due to interference with the spectrophotometric assay for UR activity in crude extracts of this strain (Rhee, Chock, and Stadt-

1051

submitted for publication). Representative ginD (SK103) and ginE strains (SK254 and SK255) had normal amounts of PI, protein (Table 3). As expected, only the PIIA form of the PI, protein was present in the ginD strain SK103 (15). GS. GS activity was measured by the (nonphysiological) -y-glutamyl transfer assay of Stadtman et al. (34), which allows an estimation of total activity, independent of degree of adenylylation, and of degree of adenylylation (see Materials and Methods). Table 4 summarizes the GS activities of strains containing ginD and ginE mutations. Extracts of ginD strains (SK101, SK103, and SKill) grown under repressing conditions (nitrogen excess) had approximately 50% of the total activity of wild type. Qualitative immunodiffusion tests with antibody against purified GS indicated that low activity in these strains was correlated with low antigen (data not shown). Under nitrogen excess conditions, GS from both wild-type and ginD strains was highly adenylylated (ratio of activities +Mg2+/-Mg2+ of about 0.2). To confirm that low ratios of activity +Mg2+/-Mg2+ represented adenylylated enzyme, extracts were treated with snake venom phosphodiesterase, which removes adenylyl groups from GS (36); it was demonstrated that the ratio of activities +Mg2+/-Mg2e increased to _1.2, a ratio characteristic of unadenylylated enzyme (data not shown). Extracts of ginD ginE strains (SK203, SK204, and SK205) had very low GS activity, 20 to 40% that of wild type. Extracts of ginE strains (SK254 and SK255) had about 40% of wild-type GS activity. As expected, GS from both ginD ginE and ginE strains was unadenylylated (ratio of activities +Mg2e/-Mg2+ : 1.2). man,

TABLE 3. Activities ofproteins required for covalent modification of GS Activity

Strain and relevant genotype

(% of wild type)

ATase UTase UR P,, 00 10 100a 100a looa TA831 (wild type) 113