Regulation of Glutamine Synthetase Activity

EARL R. STADTMAN

[SECTION EDITOR: GEORGES COHEN]

Posted September 9, 2004

Laboratory of Biochemistry, National Heart, Lung, and Blood Institute, National Institutes of Health, Building 50, Room 2140, Bethesda, MD 20892-8012.

Phone: 301-496-4096, Fax: 301-496-0599, E-mail: This e-mail address is being protected from spambots. You need JavaScript enabled to view it

INTRODUCTION

The enzyme glutamine synthetase (GS) catalyzes the ATP-dependent conversion of glutamate and ammonia to glutamine (reaction 1).

ATP + glutamate + ammonia → ADP + P_i + glutamine (1)

The amide group of glutamine is a source of nitrogen for synthesis of the indole moiety of tryptophan, the imidazole group of histidine, the purine and pyrimidine moieties of adenylic and cytidylic acids, and the synthesis of carbamyl-phosphate, NAD⁺, and complex polysaccharides (via glucosamine-6-phosphate). In addition, transfer of the α-amino group of glutamine to various α-ketoacids leads to the formation of several amino acids. Furthermore, some organisms contain a glutamate synthetase that catalyzes transfer of the amide group of glutamine to α-ketoglutarate to form two equivalents of glutamate (reaction 2). When coupled with the well-known transamination reactions (reaction 3), this provides a pathway for the glutamine-dependent synthesis of almost all amino acids (reactions 2–4).

glutamine + α-ketoglutarate + NADPH + H⁺ → 2 glutamate + NADP⁺ (2)

glutamate + α-ketoacid → α-ketoglutarate + amino acid (3)

Sum:

glutamine + α-ketoacid + NADPH + H⁺ → glutamate + amino acid + NADP⁺ (4)

In view of its central role in nitrogen metabolism, it is not surprising that the activity of glutamine synthetase is subject to rigorous control. Detailed studies of the glutamine synthetase in Escherichia coli and other bacteria have shown that the activity of this enzyme is regulated by at least five different mechanisms: (i) cumulative feedback inhibition by multiple end products of glutamine metabolism, (ii) interconversion between taut and relaxed protein configurations in response to binding and dissociation of divalent cations at one of its two metal binding sites, (iii) dynamic interconversion of the enzyme between covalently modified (adenylylated) and unmodified forms by a novel bicyclic cascade system, (iv) repression and derepression of glutamine synthetase formation by cyclic phosphorylation and dephosphorylation of an RNA factor that governs transcription activities, and (v) regulation of glutamine synthetase turnover by the coupling of site-specific metal ion-catalyzed oxidation with proteolytic degradation of the enzyme.

ENZYME STRUCTURE

Based on ultracentrifugation (31), electron microscopic examination (2), x-ray crystallographic analysis (2, 45), and amino acid analysis (6), it has been established that the glutamine synthetases from gram-negative and some gram-positive bacteria are proteins of 620 to 640 kDa and are composed of 12 identical subunits arranged in two face-to-face hexagonal rings (Fig. 1), which are held together by hydrophobic and hydrogen-bonding reactions (2). Each subunit contains a "bifunnel-like"catalytic site, which is situated between the monomers (2, 45). The N terminus and the C terminus of each subunit are helical and the central channel of the dodecamer is lined by six four-stranded β-sheets (2). The dodecamer contains several exposed loops, one of which is a site for proteolysis (18) and ADP-ribosylation (23) and one that contains a specific tyrosyl residue that is the target for adenylylation and thereby plays a major role in the regulation of glutamine synthetase activity (see below). The enzyme contains two metal-binding sites (n1 and n2) to which Mg²⁺ or Mn²⁺ can bind. Binding at the n1 site leads to stabilization of the protein (7, 33) and facilitates the binding of glutamate to the catalytic site (12), whereas metal binding to the n2 site is implicated in phosphoryl transfer (13).

Fig. 1Ultramicroscopic views of unfixed glutamine synthetase molecules. Three characteristic orientations are shown. When the molecule is resting on a face, the subunits appear as a hexagonal ring (Top). Molecules viewed from the side appear as two layers of subunits seen as four spots when viewed exactly down a diameter between subunits (Middle) or, in general, as two layers of subunits (Bottom). Magnification, ×3,160,000. The results are thus consistent with results of ultracentrifugal and X-ray analyses showing that the enzyme is composed of 12 identical subunits and show that the subunits are arranged in two, superimposed hexagonal arrays.

Source: R. C. Valentine et al. (40).

CUMULATIVE FEEDBACK INHIBITION

Glutamine synthetase activity in E. coli is subject to inhibition by seven different end products of glutamine metabolism, namely, by tryptophan, histidine, carbamyl-phosphate, CTP, AMP, glucose-6-phosphate, and NAD⁺, and also by serine, alanine, and glycine (42). The binding affinities of these end products to the enzyme are quite low (in the millimolar range). At a physiological concentration, each substance by itself causes only partial inhibition (10 to 50%) of the enzyme activity; however, the inhibition by combinations of any two inhibitors is greater than that of either one alone, and physiological concentrations of all seven inhibitors leads to over 90% inhibition of the enzyme activity. Based on this observation, it was proposed that glutamine synthetase is subject to cumulative feedback inhibition (42, 43). This behavior and detailed analysis of various combinations of inhibitors established that the total enzyme activity observed with various mixtures of inhibitors was equal to the products of the activities obtained when each inhibitor in the mixture was tested by itself. Based on this study, it was deduced that each of the inhibitors binds to a separate allosteric binding site on the enzyme (42). In the meantime, detailed crystallographic and kinetic analyses of a homogeneous glutamine synthetase preparation by Eisenberg and coworkers (21) have clearly established that serine, glycine, and alanine all bind to the glutamate-binding site of the enzyme and that AMP binds to the ATP-binding site on the enzyme (20). The discrepancy between the results of these more recent kinetic studies and those obtained earlier with an different, apparently homogeneous preparation of glutamine synthetase is not evident. However, the possibility that the discrepancy reflects differences in the states of adenylylation of the enzymes studied remains an open question. The earlier studies (42, 43) were carried out prior to the discovery that the activity of glutamine synthetase is strictly controlled by the adenylylation of a specific tyrosine residue in each subunit of the enzyme (discussed below), and the preparation used by Eisenberg and coworkers contained no adenylyl groups, whereas the preparation used in the earlier study likely contained both adenylylated and unadenylylated subunits. In any case, whatever the mechanism, it is clear that the enzyme is subject to cumulative inhibition by multiple end products of glutamine metabolism.

INTERCONVERSION BETWEEN TAUT, RELAXED, AND TIGHTENED CONFIGURATIONS

As noted above, each subunit of glutamine synthetase from gram-negative bacteria contains two divalent metal-binding sites (n1 and n2) that are capable of binding Mg²⁺ or Mn²⁺. As is illustrated in Fig. 2, upon removal of the divalent cations by treatment with chelating agents or by exhaustive dialysis, the native so-called taut form of GS is converted to a catalytically inactive relaxed configuration. This is accompanied by exposure of cysteine sulfhydryl groups to reactions with various sulfhydryl group reagents (33, 44) and by changes in the ultraviolet absorption spectrum of the protein, presumably due to transfer of tryptophan residues to a more hydrophobic environment (29) and to ionization of the hydroxyl groups of some tyrosine residues. Conversion of the taut to the relaxed configuration does not involve changes in the molecular weight or in the double-layer hexagonal arrangements of the subunits. However, in contrast to the taut enzyme, the relaxed enzyme undergoes complete subunit dissociation at pH 8.7 or upon exposure to low concentrations of urea. For review, see reference 39. Reassociation of the subunits, with restoration of catalytic activity, occurs spontaneously upon addition of divalent cations to the reaction mixtures or by shifts to lower pH values. Significantly, reassociation of the subunits with restoration of the catalytic activity is greatly facilitated by the E. coli chaperon in an ATP-dependent process (10). Whereas in vitro reassociation of the subunits leads to restoration of GS activity and to regeneration of the dodecameric structure and kinetic characteristics comparable with those of the taut enzyme. At high concentrations (1.0 mg/ml), the regenerated form of the enzyme (referred to as the tightened form) precipitates out of solution as a paracrystalline aggregate due to stacking of the dodecameric molecules to form long hexagonal tubes as shown in Fig. 3.

Fig. 2Schematic representation of the divalent metal cation-dependent interconversion of glutamine synthetase subunits between free, taut, relaxed, and tightened configurations.

Source: B. M. Shapiro and A. Ginsburg (31).

Fig. 3Ultramicroscopic view of paracrystalline aggregates of tightened glutamine synthetase.

Source: R. C. Valentine et al. (40).

INTERCONVERSION OF GLUTAMINE SYNTHETASE BETWEEN ADENYLYLATED AND UNADENYLYLATED FORMS

The most remarkable mode of GS regulation involves the attachment of an adenylyl group to a particular tyrosine residue in one or more of the enzyme’s 12 subunits. This is achieved by the coupling of two protein nucleotidylation processes (illustrated in Fig. 4). In one process, a 50-kDa regulatory protein, P_II, composed of three identical subunits, undergoes cyclic interconversion between uridylylated and unmodified forms (1, 3); whereas, in the other process GS undergoes cyclic interconversion between adenylylated and unmodified forms (16, 35). As illustrated in Fig. 4, uridylylation of the P_II protein is catalyzed at the uridylylation site (UT_u) of the bifunctional enzyme, uridylyltransferase (UT), and involves the α-ketoglutarate-dependent attachment of the uridylic acid moiety of UTP, in phosphodiester linkage, to the hydroxyl group of a particular tyrosyl residue in one or more of the three identical subunits of the P_II protein (reaction 5).

Fig. 4The bicyclic cascade that regulates glutamine synthetase activity. Interrelationship between the cyclic interconversion of the regulatory protein between uridylylated [P _II(UMP) _n] and unuridylylated (P _II) forms (upper cycle), and the cyclic interconversion of glutamine synthetase (GS) between adenylylated [GS(AMP) _n] and unadenylylated forms (lower cycle), and the reciprocal control of these interconversions by l-glutamine (Gln) and α-ketoglutarate (α-KG). AT _aand AT _ddenote the adenylylation and deadenylylation sites of the adenylyltransferase (AT), respectively; UT _uand UT _ddenote the uridylylation and deuridylylation sites of uridylyltransferase, respectively. The values of nrefer to the average number (0 to 3) of P _IIsubunits of UT molecules that are uridylylated (upper cycle) or the number (0 to 12) of the GS subunits that are adenylated (lower cycle).

Source: S. G. Rhee et al. (26).

P_II + nUTP P_II(UMP)_n + nPP_i (5)

This reaction is opposed by the glutamine-dependent hydrolytic cleavage of the UMP-P_II phosphodiester bond, which is catalyzed at the deuridylylation site (UT_d) of the uridylyltransferase (reaction 6).

P_II(UMP)_n + nH₂O P_II + nUMP (6)

Because UT is composed of three identical subunits, one or all of which can be uridylylated, the value of nin reactions 5 and 6 will vary from 0 to 3 depending on the intracellular concentrations of α-ketoglutarate, glutamine, and other metabolites (see below). In an analogous manner, adenylylation of GS involves the glutamine-stimulated attachment in phosphodiester linkage of the adenylic acid moiety of ATP to the hydroxyl group of a unique tyrosine residue in each subunit of the enzyme (reaction 7).

GS + nATP GS(AMP)_n + nPP_i (7)

This is catalyzed at the adenylylation site (AT_a) of a bifunctional adenylyltransferase (AT) and is opposed by the α-ketoglutarate-dependent phosphorolytic cleavage of the phosphodiester bond, which is catalyzed at the deadenylylation site (AT_d) of AT to yield ADP and unmodified GS (reaction 8).

GS(AMP)_n + nP_i GS + nADP (8)

The singular importance of reactions 7 and 8 in the regulation of GS activity is highlighted by the fact that GS is composed of 12 identical subunits and that the adenylylation of any one or more of the GS subunits leads to loss of their activities. Therefore, the specific catalytic activity of a GS molecule is directly proportional to the number (n) of its subunits that are adenylylated. Moreover, because the uridylylation and deuridylylation ofP_II are catalyzed at separate sites on the same uridylyl transferase, and the adenylylation and deadenylation of GS are catalyzed at separate sites on the same adenylyl transterase, it follows that both cyclic interconversions must be rigorously controlled to avoid senseless decomposition of UTP and ATP. Such control is achieved by the reciprocal effects of α-ketoglutarate and glutamine on the nucleotidylation and denucleotidylation reactions in both cycles, as shown in Fig. 4. Tight coupling of the two cycles is because adenylylation of a GS subunit at the AT_asite of AT absolutely depends on the binding of the unmodified form of P_II at that site, and also because the deadenylylation of a GS subunit absolutely depends on the binding of P_II(UMP) at the AT_dsite of AT.

ADVANTAGES OF CYCLIC CASCADES IN REGULATION

A theoretical analysis of the bicyclic cascade system shown in Fig. 4 demonstrated that such cascades are endowed with unique characteristics that make them unusually effective for the regulation of key enzymes in metabolism (4, 36). Thus, (i) cyclic cascades are capable of signal amplification, i.e., they can provoke large changes in the steady-state level of covalent modification of a target enzyme at concentrations of a primary allosteric effector that is orders of magnitude lower than the dissociation constant of the effector-converter enzyme complex. (ii) The amplitude of the response to saturating concentrations of one particular allosteric effector can be preset to values ranging from 0 to 100% of the theoretical maximum by variations in the concentrations of other effectors in the cascade. (iii) Cyclic cascades can serve as rate amplifiers, i.e., they can elicit a shift in the steady-state level of covalent modification in the millisecond time range. (iv) Cyclic cascades are able to elicit either positive or negative cooperativity in their kinetic responses to increases in the concentration of a given effector. And, finally, (v) cyclic cascades can serve as metabolic integration systems. By means of multisite interactions of the convertor enzymes with allosteric effectors and enzyme substrates, cyclic cascades are able to monitor continuously the intracellular concentrations of a multitude of metabolites. This leads automatically to shifts in the steady-state distribution of covalently modified and unmodified forms of the target enzyme and, hence, to changes in its catalytic activity commensurate with metabolic demand.

VERIFICATION OF THE STEADY-STATE CONCEPT

Forty substances (metabolites and metal ions) that have been found to serve directly or indirectly as either substrates or allosteric effectors of one or more of the enzymes in the bicyclic cascade (Fig. 4) are listed in Table 1 (9, 27, 37). Of these, α-ketoglutarate and glutamine are unquestionably the most important in the cascade regulation of GS activity. Nevertheless, the cascade theory predicts that the steady-state level of GS adenylylation and therefore its catalytic activity is determined by the combined effects of all metabolites that affect the kinetic parameters of one or more of the enzymes in the cascade. This prediction was verified by a study (38) in which a reaction mixture containing the cascade enzymes UT, AT, and GS, the P_II regulatory protein, and arbitrary concentrations of six different metabolites that collectively yielded a steady state in which about 50% of the 12 GS subunits were adenylylated (i.e., n= 6.0). It was then shown that a change in the concentration of any one of these effectors led to a rapid shift in the steady-state level of adenylylation of GS (26, 27, 38). Consistent with the cascade model, little or no adenylylation occurred in the absence of glutamine, and almost all of the subunits were adenylylated (n = 11.0) when α-ketoglutarate was omitted from the mixture (Fig. 5).

Fig. 5Variations in the steady-state level of glutamine synthetase adenylylation in response to changes in the effector concentrations.The heavy line (filled squares) shows how the number ( n) of GS adenylylated subunits that were formed when a reaction mixture containing 95 μg of GS, 20 mMMgCl ₂, 20mM orthophosphate (Pi), 1.0 mM ATP, 1.0 mM UTP, 15 mM α-ketoglutarate, 0.3 mM glutamine, and partially purified preparations of P _II, AT, and UT were incubated at pH 7.2, 37°C, for the times indicated. The other curves illustrate the effects of changing the concentration of only one of the various metabolites, as indicated.

Source: E. R. Stadtman et al. (27), with minor modifications.

Table 1Direct and indirect effectors of glutamine synthetase activity

REGULATION OF GLUTAMINE SYNTHETASE CONCENTRATION

In addition to their roles in the regulation of GS activity, α-ketoglutarate and glutamine have a central role in the regulation of biosynthesis of glutamine synthetase. The biosynthesis of GS is subject to repression and de-repression by glutamine and α-ketoglutarate, respectively. When grown in a nitrogen-deficient medium, the intracellular level of GS in E. coli is about 10 times greater than when the bacterium is grown on a medium containing high concentrations of glutamine and a good supply of glucose. The mechanisms that underlie the repression and de-repression of GS synthesis in response to nitrogen availability were elucidated by the pioneering studies of Magasanik (22, 24, 41) and Kustu (14, 15, 17). They showed that transcription of the structural gene for GS is under control of two closely linked genes glnL and glnG. The glnL gene product (NtrB) is a protein kinase that undergoes the ATP-dependent autophosphorylation of one of its histidine residues. This is followed by transfer of the phosphoryl group to the carboxyl group of an aspartyl residue in the glnG gene product (NR_I). Upon phosphorylation, NR_I is converted from an inactive form to a form (NR_I-P) that is able to activate the RNA polymerase for transcription of the GS structural gene. However, this process is opposed by theunmodified form ofP_II which promotes dephosphorylation of the NR_I-P. This and the fact that the uridylylated form of P_II has no effect on the dephosphorylation of NR_I-P is the basis of a novel mechanism for the coordinate regulation of GS biosynthesis and GS catalytic activity (illustrated in Fig. 6). It follows that, by means of allosteric interactions, the uridylyltransferase senses the intracellular concentrations of glutamine and α-ketoglutarate and thereby specifies the P_II/P_II(UMP)_n ratio, which dictates the activity of GS (state of adenylylation) on the one hand and the rate of GS biosynthesis (state of NR_Iphosphorylation) on the other.

Fig. 6The tricyclic cascade system that regulates the activity and synthesis of glutamine synthetase. Interrelationship between the uridylylation cycle, the adenylylation cycle, and the phosphorylation cycle that regulates glutamine synthetase transcription. The reciprocal controls of these interconversions by l-glutamine (Gln) and α-ketoglutarate (α-KG) are shown: + indicates stimulation; – indicates inhibition. Abbreviations: GS, glutamine synthetase; P _II, regulatory protein; AT _a, and AT _d, adenylylation and deadenylylation sites, respectively, on the bifunctional adenylyltransferase; UT _dand UT _u, deuridylylation (uridylyl-removing) and uridylation sites, respectively, on the bifunctional uridylyltransferase. NR _I, glnGproduct (also known as NTRC); NR _IIKand NR _IIP, glnLproduct (also known as NTRB) catalyzing phosphorylation and dephosphorylation of NR _I, respectively.

Source: S. G. Rhee et al. (27).

REGULATION OF GLUTAMINE SYNTHETASE DEGRADATION

The intracellular turnover of GS is a two-step process: (i) The enzyme is inactivated by a metal ion-catalyzed mixed-function oxidation (MFO) system comprising Fe(III), NADH or NADPH oxidases, and oxygen (11). (ii) The oxidatively modified enzyme is then rapidly degraded by a protease system exhibiting high specificity for degradation of the oxidized enzyme (29). Substantial evidence shows that the MFO-catalyzed oxidation of GS involves reduction of Fe³⁺ to Fe²⁺, the reduction of O₂ to H₂O₂, followed by the interaction of Fe²⁺ with H₂O₂ at a metal-binding site on the enzyme to form a reactive oxygen radical species (35) that oxidizes a histidine residue and an arginine residue at the metal-binding site to form 2-oxo-histidine (19) and glutamic semialdehyde (5) derivatives, respectively. The two-step mechanism of GS degradation (Fig. 7) is further supported by the demonstration that cells contain proteases that degrade oxidized forms of GS but have little ability to degrade the native form of the enzyme (28, 30). Finally, from the standpoint of cellular regulation, it is significant that the GS substrates, ATP, and glutamate inhibit the oxidation of unadenylylated GS but stimulate the oxidation of the adenylylated enzyme (25). This is reasonable since, so long as there is a demand for glutamine, GS will be present in the unadenylylated (catalytically active) form and will be protected from irreversible oxidative inactivation and degradation provided that there is an ample supply of ATP and glutamate; however, in the absence of these substrates, GS is nonfunctional and can be degraded to obtain amino acids for other functions. Furthermore, under conditions where the supplies of ATP and glutamate are not limiting and the production of glutamine exceeds the demand, GS is no longer needed, then it will be converted to the catalytically inactive adenylylated form that is not under protection of ATP and glutamate. No longer needed, it can then be processed for degradation by the MFO system.

Fig. 7Two-step process involved in the regulation of glutamine synthetase turnover. In step 1, H ₂O ₂and Fe ²⁺generated in the mixed-function oxidation (MFO) system react at divalent metal-binding sites on GS leading to an oxidized form of the enzyme (GS _ox) in which an arginine residue is converted to a carbonyl derivative. In step 2, the GS _oxis degraded by intracellular neutral proteases.