Arabidopsis thaliana glutamate-cysteine ligase: functional properties, kinetic mechanism, and regulation of activity.

In plants, glutathione accumulates in response to different stress stimuli as a protective mechanism, but only limited biochemical information is available on the plant enzymes that synthesize glutathione. Glutamatecysteine ligase (GCL) catalyzes the first step in glutathione biosynthesis and plays an important role in regulating the intracellular redox environment. Because the putative Arabidopsis thaliana GCL (AtGCL) displays no significant homology to the GCL from bacteria and other eukaryotes, the identity of this protein as a GCL has been debated. We have purified AtGCL from an Escherichia coli expression system and demonstrated that the recombinant enzyme catalyzes the ATP-dependent formation of gamma-glutamylcysteine from glutamate (Km = 9.1 mm) and cysteine (Km = 2.7 mm). Glutathione feedback inhibits AtGCL (Ki approximately 1.0 mm). As with other GCL, buthionine sulfoximine and cystamine inactivate the Arabidopsis enzyme but with inactivation rates much slower than those of the mammalian, bacterial, and nematode enzymes. The slower inactivation rates observed with AtGCL suggest that the active site differs structurally from that of other GCL. Global fitting analysis of initial velocity data indicates that a random terreactant mechanism with a preferred binding order best describes the kinetic mechanism of AtGCL. Unlike the mammalian GCL, which consists of a catalytic subunit and a regulatory subunit, AtGCL functions and is regulated as a monomeric protein. In response to redox environment, AtGCL undergoes a reversible conformational change that modulates the enzymatic activity of the monomer. These results explain the reported posttranslational change in AtGCL activity in response to oxidative stress.

Regulation of the intracellular redox environment is critical in cellular physiology for influencing signaling pathways and cell fate in response to stress (1). In plants, as in other organisms, glutathione plays multiple roles as protection against various environmental stresses (2). As an antioxidant, glutathione quenches reactive oxygen species and is involved in the ascorbate-glutathione cycle that eliminates peroxide (2). Plants use glutathione for the detoxification of xenobiotics (3), herbicides (4), air pollutants such as sulfur dioxide and ozone (5,6), and heavy metals (7). Although glutathione accumulates in response to different stress stimuli in plants, the structural and kinetic properties of the plant enzymes responsible for its production remain biochemically uncharacterized.
Glutathione synthesis occurs in two ATP-dependent steps. In the first reaction, glutamate-cysteine ligase (GCL) 1 (EC 6.3.2.2) catalyzes formation of the dipeptide ␥-glutamylcysteine from cysteine and glutamate (Scheme I). Addition of glycine to the dipeptide occurs in a second reaction, catalyzed by glutathione synthetase. Of the two enzymes, GCL appears to be rate-limiting (8). Exposure to heavy metals increases the levels of GCL mRNA in Brassica juncea and activates transcription of both GCL and glutathione synthetase in Arabidopsis thaliana (9 -11). Overexpression of E. coli GCL in plants improves tolerance to cadmium and arsenic, demonstrating the importance of this enzyme in heavy metal protection (12,13).
Bioinformatic analysis of the GCL genes from multiple species suggests that these sequences group into three families (14), 1) sequences from the ␥-proteobacteria, such as Escherichia coli; 2) sequences from non-plant eukaryotes (mammals, Drosophila, and nematodes); and 3) sequences from plants (Arabidopsis) and ␣-proteobacteria (Rhizobium). Sequence comparisons within each family show similarity, but pairwise comparisons between groups display no statistically significant relationships (14). For example, A. thaliana GCL (AtGCL) shares less than 15% amino acid sequence identity with the members of other families (15). The differences among GCL sequences also reflect the functional properties of each family.
Of the three GCL families, the enzymes from the non-plant eukaryotes are the best studied. The mammalian and Drosophila GCL consist of a 70-kDa catalytic or heavy subunit and a 30-kDa regulatory or light subunit (16,17). The heavy subunit catalyzes the formation of ␥-glutamylcysteine and is inhibited by glutathione (18), whereas the light subunit increases the affinity of the enzyme for glutamate and decreases the inhibitory effect of glutathione (19). Interestingly, the GCL from Trypanosoma brucei and the mammalian catalytic subunit are related by 45% amino acid identity, but the T. brucei enzyme functions as a 77-kDa monomer with kinetic constants similar to the "activated" heterodimeric rat enzyme (20). From the ␥-proteobacteria family of GCL, the E. coli enzyme has been isolated and characterized as a functional 58-kDa monomeric protein (21).
The plant GCL are largely unexamined at the molecular level. Expression cloning isolated a cDNA from Arabidopsis that was unrelated to the mammalian, E. coli, or yeast GCL but complemented a GCL-deficient E. coli strain (14). However, doubts about the specificity of the assay used to measure GCL activity in cell lysates and the limited sequence similarity with other GCL have challenged the identity of the AtGCL clone (4,22). To characterize the biochemical properties of the putative AtGCL, recombinant enzyme was expressed in E. coli and purified to homogeneity. We demonstrate that the purified protein catalyzes the formation of ␥-glutamylcysteine. Our analysis shows that the Arabidopsis enzyme shares some functional properties with the GCL from other species but is regulated differently than the GCL from either bacteria or nonplant eukaryotes.

EXPERIMENTAL PROCEDURES
Materials-Integrated DNA Technologies, Inc. synthesized all oligonucleotides used in this study. The pGEM-T Easy vector was obtained from Promega. E. coli Rosetta (DE3) cells were from Novagen. Ni 2ϩnitrilotriacetic acid (NTA)-agarose was bought from Qiagen. Benzamidine-Sepharose and the HiLoad 26/60 Superdex-75 FPLC column were from Amersham Biosciences. All other reagents were purchased from Sigma-Aldrich and were of ACS reagent quality or better.
Cloning and Generation of Expression Vectors-AtGCL (GenBank Z29490) (15) was amplified by PCR from an Arabidopsis cDNA library using 5Ј-dCCATGGCATGGCGCTGCTGTCTCAAGCAGG-3Ј as the forward primer (NcoI site is underlined, the AtGCL start codon is in bold, and two E. coli optimized codons are in italic) and 5Ј-dTTATAGACAC-CTTTTGTTCACGTCCCATTTTC-3Ј as the reverse primer (the putative AtGCL stop codon is in bold). The 1.6-kb PCR product was subcloned into the pGEM-T Easy vector (Promega). Automated nucleotide sequencing confirmed the fidelity of the PCR product (Washington University Sequencing Facility, St. Louis, MO). The pHIS8-AtGCL expression vector was constructed by digesting pGEM-T-AtGCL with NcoI and NotI and then ligating the fragment into a NcoI/NotI-digested pHIS8 vector (23). An expression construct of AtGCL with a truncated N terminus (pHIS8-AtGCL⌬85) was generated by PCR using the appropriate oligonucleotides.
Expression in E. coli and Protein Purification-Expression constructs were transformed into E. coli Rosetta (DE3) cells. Transformed E. coli were grown at 37°C in Terrific broth containing 50 g ml Ϫ1 kanamycin and 34 g ml Ϫ1 chloramphenicol until A 600 nm ϳ0.8. After induction with 1 mM isopropyl 1-thio-␤-D-galactopyranoside, the cultures were grown at 18°C for 6 h. Cells were pelleted by centrifugation and resuspended in lysis buffer (50 mM Tris (pH 8.0), 500 mM NaCl, 20 mM imidazole, 5 mM MgCl 2 , 10% (v/v) glycerol, and 1% (v/v) Tween 20). After sonication and centrifugation, the supernatant was passed over a Ni 2ϩ -NTA column previously equilibrated with lysis buffer. His-tagged protein was eluted with elution buffer (wash buffer containing 250 mM imidazole). Incubation with thrombin during overnight dialysis at 4°C against wash buffer removed the His-tag. Dialyzed protein was reloaded on a Ni 2ϩ -NTA column, and the flow-through was depleted of thrombin using a benzamidine-Sepharose column. The flow-through of this step was dialyzed overnight against 30% (v/v) glycerol, 25 mM HEPES (pH 7.5), 5 mM MgCl 2 , and 100 mM NaCl then loaded onto a Superdex-75 FPLC column equilibrated in the same buffer without glycerol.
Enzyme Assays-The activity of AtGCL was determined spectrophotometrically at 25°C by measuring the rate of ADP formation using a coupled assay with pyruvate kinase and lactate dehydrogenase. A standard reaction mixture (0.5 ml) contained 100 mM MOPSO (pH 7.0), 150 mM NaCl, 20 mM MgCl 2 , 10 mM cysteine, 20 mM sodium glutamate, 5 mM disodium ATP, 2 mM sodium phosphoenolpyruvate, 0.2 mM NADH, 5 units of type III rabbit muscle pyruvate kinase, and 10 units of type II rabbit muscle lactate dehydrogenase. Reactions were initiated by adding AtGCL (50 g). The rate of decrease in A 340 nm was followed using a Cary Bio300 spectrophotometer. Steady-state kinetic parameters were determined by initial velocity experiments. Measurements of the k cat and K m values for glutamate (1-80 mM) were made at 20 mM ATP and 20 mM cysteine. Kinetic constants for cysteine (0.1-20 mM) were measured at 20 mM ATP and 80 mM glutamate. For determination of the kinetic constants for ATP (0.25-20 mM), 20 mM cysteine and 80 mM glutamate were used. Kinetic parameters were calculated to fit untransformed data to v ϭ k cat [S]/(K m ϩ [S]) using Kaleidagraph (Synergy Software).
Inhibition and Inactivation Assays-For glutathione inhibition of AtGCL⌬85, initial velocities were determined spectrophotometrically using the standard assay system. Enzyme activity was determined after addition of glutathione (0 -5 mM) to assay solutions containing either varied glutamate (1-40 mM) or cysteine (0.5-20 mM). Global fitting analysis was used to simultaneously fit all data to the equation The time-dependent inactivation of AtGCL⌬85 by buthionine sulfoximine and cystamine was performed as follows. AtGCL⌬85 (125 g) was incubated (37°C) in 100 l of 0.1 M MOPSO (pH 7.0) and 20 mM MgCl 2 in the presence of either 0 -50 mM cystamine or 10 mM ATP and 0 -50 mM buthionine sulfoximine. All incubations were initiated by the addition of the inactivator. Aliquots (20 l) were withdrawn from the incubation mixture and diluted into the standard assay system and then the enzymatic activity remaining was determined. All inactivation experiments were monitored relative to a control sample without inactivator, which is set to 100% activity at each time point. Inactivation data were plotted as log (% initial enzyme activity) versus time. Semilog plots were fitted to the equation ϪdE/dt ϭ k [I], where the disappearance of enzyme activity over time is related to the concentration of inactivator (I), multiplied by k, a rate constant. This allowed a determination of the half-life for inactivation (t1 ⁄2 ) at each [I]. A Kitz-Wilson analysis of the data was used to generate the limiting constant for inactivation (k inact ) and K i by plotting t1 ⁄2 versus 1/[I] (24).
Analysis of the Kinetic Mechanism-Analysis of the kinetic mechanism of AtGCL⌬85 used global curve fitting (25,26). The reaction rates were measured as described above using a matrix of substrate concentrations (glutamate, 2-40 mM; ATP, 0.5-20 mM; cysteine, 1-20 mM). In this matrix, the rate measured for the concentration of one substrate is measured over the entire range of the other two substrates. SigmaPlot was used for curve fitting and modeling of the kinetic data to rapid equilibrium rate equations of the possible ter-reactant kinetic mechanisms (27).
Mass Spectrometry-An Applied Biosystems QSTAR XL hybrid quadrupole time-of-flight (TOF) mass spectrometry (MS) system equipped with a nanoelectrospray source (Protana XYZ manipulator) was used for an accurate molecular weight determination. The nanoelectrospray was generated from a PicoTip needle (New Objectives, Inc.) at 1200 volts. The sample flow rate was estimated to be 100 nl min Ϫ1 . The instrument m/z response was calibrated with standards from the manufacturer to provide molecular mass measurement accuracy of Ͻ30 ppm for proteins up to 50 kDa and of Ͻ5 ppm for lower mass peptides. For detection of protein complex species, data were acquired initially over the range m/z 900 -10000 and later over a narrow mass range with intense signals. The accumulation time was 1 s, for 300 cycles. The two declustering potential parameters and focusing potential, i.e. DP, DP2, and FP, were 100, 15, and 300, respectively. To maintain protein complexes in gas phase, the gas pressure in the collision cell was increased to 4.0 ϫ 10 5 Torr for ion cooling. For MS/MS analysis, data were acquired using the information-dependent acquisition feature in the Analyst QS software. TOF and tandem MS data were acquired over m/z ranges of 300 -2200 and 65-2000, respectively. Every spectrum was accumulated for 1 s and was followed by three-product ion spectra (each for 3 s). The DP, DP2, FP settings were 50, 10, and 200, respectively, and the collision energy was dependent on the m/z values of the ions.
For identification of the reaction product by electrospray ionization (ESI)-TOF MS, scaled-up reactions (5-ml volume, 1 mg of protein) were performed using standard assay conditions. Reactions were quenched by an addition of 5% (v/v) 5-sulfosalicyclic acid. After centrifugation, the supernatant was evaporated to dryness and dissolved in water. Enzymatically synthesized ␥-glutamylcysteine was compared with an authentic standard (Sigma-Aldrich).

Expression and Purification of AtGCL-
The reported nucleotide sequence (15) was used to design oligonucleotides for amplifying AtGCL from an Arabidopsis library. Several clones were identical to each other but had a single base pair deletion (T1470) compared with the published sequence that shifts the reading frame at the C terminus. The sequence obtained was identical to other GenBank TM entries for AtGCL, including that from genome sequencing (NP_194041). The C-terminal amino acid sequence of our AtGCL clone was 70% identical to other plant GCL sequences, suggesting that this is the correct sequence. Because the N-terminal region of AtGCL encodes a chloroplast transit signal (10), we generated a version of AtGCL (At-GCL⌬85) lacking the localization sequence based on comparison with the cytosolic GCL from Zea mays (28).
AtGCL and AtGCL⌬85 were overexpressed and purified by Ni 2ϩ -affinity and size-exclusion chromatography (Fig. 1). SDS-PAGE analysis of the purified proteins showed that AtGCL and AtGCL⌬85 migrated with molecular masses of 58 and 50 kDa, respectively, corresponding to their predicted masses. Fulllength AtGCL was primarily insoluble, but ϳ1 mg of pure soluble protein was obtained from an 8-liter growth. Removal of the plastid target sequence improved protein solubility, as purification of AtGCL⌬85 yielded 5 mg of pure protein liter Ϫ1 of culture.
AtGCL catalyzed the ligation of glutamate to cysteine in the presence of ATP with a specific activity of 120 nmol min Ϫ1 mg protein Ϫ1 . AtGCL⌬85 catalyzed the same reaction with a similar specific activity. Because the removal of the localization sequence improved solubility without altering enzymatic activity, we used AtGCL⌬85 for subsequent analysis. Substitution of ␣-aminobutyrate for cysteine in these reactions resulted in 20-fold reductions in specific activity even when high concentrations (up to 100 mM) were used. Therefore, cysteine was employed as a substrate for all further enzyme assays.
Kinetic Analysis and Identification of the Reaction Product-Steady-state kinetic parameters (k cat and K m ) for glutamate, cysteine, and ATP were determined for AtGCL⌬85 (Table I). The k cat and K m values were comparable with those reported for GCL purified from tobacco cell suspensions (29). The reaction rate of recombinant purified AtGCL⌬85 was 50-fold higher than the tobacco enzyme likely because of the greater purity of these samples. Compared with other characterized GCL (17, 19 -21, 30 -33), AtGCL displays a K m value for glutamate 3-10fold higher and a K m value for cysteine 0.5-10-fold different.
The reaction product of AtGCL⌬85 was analyzed by ESI-TOF mass spectrometry (Fig. 2). A TOF-MS survey scan showed a major component of m/z ϭ 251.0894. The mass matches that of authentic ␥-glutamylcysteine (MW 251.0841). Fragmentation of the precursor ion generated two major product ions corresponding to ␥-glutamylcysteine cleaved at the peptide bond. The data demonstrated that the purified AtGCL catalyzed the production of ␥-glutamylcysteine and is a GCL, even though it displays a low sequence homology with the GCL from other species.
Buthionine sulfoximine (34) and cystamine (35) inactivate GCL. GCL catalyzes the ATP-dependent phosphorylation of buthionine sulfoximine to form a ␥-glutamylphosphate intermediate that mimics the transition state bound at the active site (34). Cystamine is a thiol-specific inactivator of GCL that targets a cysteine within the active site (35).
Kinetic Mechanism of AtGCL-To determine the kinetic mechanism of AtGCL, we obtained a complete matrix of kinetic data for the three substrates (25)(26)(27). Six families of data were generated in which one of the ligands is maintained at a saturating concentration whereas the other two substrates are varied. Kinetic data were first analyzed as double-reciprocal plots (Fig. 5). Because the plots for each family of data converged, potential ping-pong kinetic mechanisms for AtGCL were eliminated. Next, the initial velocity data were globally fit to the 16 possible rapid equilibrium rate equations describing a terreactant system (27), including random, ordered, and partially  ordered/random mechanisms. Simultaneous fitting of the initial velocity data provides a more robust analysis than replotting reciprocal data parameters (37). Based on global fits to the initial velocity data, the best results were obtained using the equation for a random terreactant system (Equation 1), where K ATP , K Cys , and K Glu are the equilibrium dissociation constants for the binding of substrate with the free enzyme and, ␣, ␤, and ␥ are the interactions factors between cysteine and glutamate, ATP and glutamate, and ATP and cysteine, respectively. These factors indicate how the dissociation con  (Fig. 6). Table II summarizes the fitted parameters and the calculated substrate dependences.
The K m values reported in Table I are in agreement with the results from global fitting of the initial velocity data (Table II). Because binding of substrates influences the binding of other substrates in a random kinetic mechanism, the K m values (Table I) will not equal the dissociation constants calculated from fitting the data (27). Instead, the K m values approximate a combination of the dissociation constant and interaction factors. For example, the K m for cysteine (Table I) is similar to ␣␥K Cys calculated from fits of the initial velocity data (Table II).
Redox Sensitivity of AtGCL-Size-exclusion chromatography of full-length AtGCL (58 kDa) and AtGCL⌬85 (50 kDa) during their purification showed that each protein migrated as a 60 -65-kDa species. However, incubation of AtGCL⌬85 with DLdithiothreitol (DTT) shifted the elution volume to a corresponding molecular mass of ϳ30 kDa (Fig. 7A). SDS-PAGE analysis of peaks A (60 kDa) and B (30 kDa) from the size-exclusion column in buffer containing 1 mM DTT showed that AtGCL⌬85 (50 kDa) was present in both peaks (Fig. 7B). Similar sizeexclusion chromatography and SDS-PAGE results were ob-served for full-length AtGCL. Assays for GCL activity demonstrated that peaks A and B had specific activities of 130 and 15 nmol min Ϫ1 mg protein Ϫ1 , respectively. Size-exclusion chromatography after dialysis in the presence of either 10 mM ␤-mercaptoethanol or 10 mM glutathione had the same effect as DTT. The change in elution profile of AtGCL in response to reducing environment is reversible (Fig. 7C). After incubation in the presence of 10 mM DTT, AtGCL⌬85 elutes from the size-exclusion column at a volume corresponding to 30 kDa with a specific activity of 15 nmol min Ϫ1 mg protein Ϫ1 . Removal of DTT shifts the elution profile back to a 60-kDa species with a specific activity of 120 nmol min Ϫ1 mg protein Ϫ1 . After incubation with DTT, AtGCL⌬85 was incubated with iodoacetamide to block free thiol groups on the protein (Fig. 7C, inset). Following dialysis, size-exclusion chromatography showed that the protein primarily eluted as the 30-kDa species indicating that blocking of free thiols prevents formation of the active monomer form. These results suggest that the plant GCL are regulated differently than non-plant eukaryotic GCL (16,17).
ESI-MS Analysis of Oligomerization State-Two models can explain the change in elution profile from the size-exclusion column in non-reducing and reducing environments. One possibility is that the active form of AtGCL is a dimeric protein that dissociates into monomers in response to reducing agents. In this model, the difference between the expected and observed molecular masses under reducing conditions suggests that AtGCL adopts a rod-like shape, because spherically shaped proteins migrate more quickly through the column matrix. Alternatively, AtGCL is a monomeric protein that adopts either a spherical (Fig. 7A, peak A) or a rod-like (Fig. 7A, peak  B) shape in response to redox environment.
ESI-TOF mass spectrometry was used to determine the oligomerization state of AtGCL⌬85 (38,39). Under non-reducing conditions, a monomer signal was observed for the 60-kDa species purified by size-exclusion chromatography (Fig. 8, A  and B). The observed molecular mass (50,110 Da) is consistent with the calculated mass (50,113 Da) of AtGCL⌬85. However, a small amount of dimeric species was observed at a higher m/z range (Fig. 8B, inset). The abundance of each species in the sample is best represented by the intensity of each species on the ESI mass spectra, not the intensity on the deconvolution spectra. After reduction with 10 mM DTT, only a monomeric species (50,111 Da) was observed (Fig. 8, C and D). ESI-TOF MS experiments using the 30-kDa species purified by sizeexclusion chromatography showed that only the monomeric 50-kDa species was present in the absence and presence of DTT (not shown). Because similar results were obtained in ammonium acetate buffer (not shown), the presence of an intermolecular interaction can be eliminated. The ESI-TOF MS conditions for determining the oligomerization state were optimized using bovine serum albumin and yeast enolase (38,39). DISCUSSION AtGCL has remained functionally uncharacterized (15) and the lack of sequence homology with other GCL and concerns with the original activity assays have caused controversy over the identity of this protein (4,22). To resolve this issue and to provide a biochemical analysis of AtGCL, we have expressed  7. Sensitivity of AtGCL to reductants. A, size-exclusion chromatography of AtGCL⌬85 in the absence and presence of DTT. Before injection on a Superdex-75 26/60 FPLC column, purified protein was dialyzed overnight at 4°C in buffer (25 mM HEPES (pH 7.5), 100 mM NaCl, and 5 mM MgCl 2 ) containing 0 (orange), 1 (black), or 10 (red) mM DTT. Peak A corresponds to 60 kDa, and peak B corresponds to 30 kDa using gel-filtration molecular weight standards. B, SDS-PAGE analysis of the fractions corresponding to peaks A and B obtained using At-GCL⌬85 incubated with 1 mM DTT. Arrowheads correspond to molecular mass markers as indicated. C, reversible dissociation of At-GCL⌬85. Purified protein was dialyzed for 4 h at 4°C in 25 mM HEPES (pH 7.5), 100 mM NaCl, 5 mM MgCl 2 , and 10 mM DTT. Approximately half of the sample was injected on a Superdex-75 26/60 FPLC column (black). The remaining sample was dialyzed overnight at 4°C in buffer without DTT and then loaded onto the same column (red). Inset, iodoacetamide blocking of free thiols. After treatment with 10 mM DTT, AtGCL was incubated with 100 mM iodoacetamide. The elution from the Superdex-75 column is shown. Axis are the same as C. and purified the enzyme using an E. coli expression system. Our results demonstrate that AtGCL shares functional properties with the GCL of other families but is regulated differently than the GCL of other non-plant eukaryotes or bacteria.
Biochemical Evidence of GCL Function-Using purified recombinant protein, we have shown that AtGCL is a genuine GCL. The steady-state kinetic parameters of AtGCL⌬85 were similar to those reported for GCL purified from tobacco cell suspensions (29), indicating that the GCL from Arabidopsis and tobacco are functionally similar. The steady-state kinetic values of AtGCL (Table I) are most similar to those of the isolated catalytic subunit of Drosophila GCL (K m Glu ϭ 2.9 mM, K m Cys ϭ 5.5 mM, V max ϭ 244 nmol min Ϫ1 mg protein Ϫ1 ) (16). In addition, ESI-TOF MS analysis confirms that ␥-glutamylcysteine is the reaction product of AtGCL.
Although AtGCL is functionally similar to other GCL, inactivation studies suggest structural differences at the active site of AtGCL compared with members of the bacterial and nonplant eukaryotic GCL. The inactivation of AtGCL by both buthionine sulfoximine and cystamine occurs with slower k inact rates and higher K i values than observed for the mammalian, nematode, or bacterial GCL (20,(33)(34)(35)(36). Importantly, the inactivation specificity of AtGCL compared with those of GCL from T. brucei (20), Ascaris suum (32), and Onchocerca volvulus (33) suggests that the design of anti-nematode compounds is possible. Although these nematodes are human parasites, the generation of inhibitors against nematodes that attack crop and non-crop plants can also be envisioned (40).
Kinetic Mechanism of AtGCL-The kinetic mechanism of GCL has been described as ping-pong (31,41), ordered A (ATP), random BC (42,43), or random ter-reactant (25) mechanisms. The kinetic data generated for AtGCL are not consistent with a ping-pong mechanism. In addition, evidence for the formation of different reaction intermediates does not support a pingpong kinetic mechanism in the GCL of other organisms (36). The initial velocity line patterns for AtGCL are best fit to a random ter-reactant kinetic mechanism; however, the interaction factors indicate a preferred binding order. Earlier studies (42, 43) describing a partially ordered mechanism for rat GCL were performed before robust computational methods for si-multaneous fitting of kinetic data were available. As noted by Brekken and Phillips (25), a global fitting analysis of initial velocity data can distinguish between partially ordered and random mechanisms, which would be difficult to perform using reciprocal plotting methods alone. In summary, AtGCL and the GCL from T. brucei and human all share a random a terreactant kinetic mechanism displaying a preferred order of substrate addition (25).
The interaction factors of AtGCL for a random mechanism (Table II) indicate that some substrate binding sites exhibit positive interactions with other ligand sites. For example, the binding of either ATP or glutamate increases the binding affinity for the other substrate 2.5-fold (␤ ϭ 0.38). AtGCL also displays a positive interaction between the glutamate and cysteine binding sites (␣ ϭ 0.063) with binding of either ligand increasing binding affinity for the other 16-fold. Similar cooperativity was described for T. brucei GCL. With ␣-aminobutyrate as a substrate, binding of either ATP or glutamate increased the binding affinity for the other 18-fold (25). Interestingly, the T. brucei GCL showed a negative interaction between the glutamate and ␣-aminobutyrate binding sites with a 6-fold decrease in binding affinity between these ligands. The differences between the Arabidopsis and T. brucei enzymes likely result from using different substrates for the kinetic analysis or may also reflect potential structural variations at the active site. For both AtGCL and the T. brucei enzyme, ATP and cysteine or ␣-aminobutyrate does not significantly alter the binding energies of each other.
Regulation of AtGCL Activity-In mammals and bacteria, glutathione regulates GCL as a feedback inhibitor (8,21). This is a fundamental mechanism for controlling intracellular glutathione levels in response to the redox environment (2). Here we have shown that glutathione also inhibits AtGCL. However, our experimental results also suggest a mechanism for the regulation of GCL activity in plants that differs from other eukaryotes.
The GCL from mammals (18,19), Drosophila (17), and A. suum (32) consist of catalytic and regulatory subunits. The reversible association of these two subunits in response to redox environment modulates GCL activity (18,19). Fraser  Fig. 7A, peak A (60 kDa) from size-exclusion chromatography in an aqueous 1% acetic acid and 50% methanol solution. The ϩ30 marks the peak with that number of charges on the ion. B, result of deconvoluting the ESI spectra over the m/z range 45,000 -110,000. Inset, ESI m/z spectrum of peak A (60 kDa) at a higher m/z range corresponding to the dimeric species. Numbers on the peaks indicate the number of charges on the ion. C, ESI m/z spectrum of peak A (60 kDa) from size-exclusion chromatography in an aqueous 1% acetic acid and 50% methanol solution after reduction with 10 mM DTT. The ϩ30 marks the peak with that number of charges on the ion. D, result of deconvoluting the ESI spectra over the m/z range 45,000 -110,000. amu, atomic mass units. et al. (44) demonstrated the importance of intersubunit disulfide bonds by the mutagenesis of cysteines in the regulatory subunit of the Drosophila GCL. Likewise, mutagenesis studies of the human GCL catalytic subunit showed a loss of redox sensitivity when a conserved cysteine in the C-terminal region of human GCL was mutated, implicating this residue in forming a disulfide linkage with the regulatory subunit (45). Interestingly, this cysteine is not found in the plant GCL, because it is located beyond the C terminus of the plant enzymes.
Our results indicate that AtGCL is catalytically active as a monomeric protein that is sensitive to reducing agents. Hell and Bergmann (29) reported that treatment of the enzyme with DTT resulted in a loss of activity and suggested that the protein dissociated into 30-kDa subunits but never showed evidence of dissociation. We observed the same shift in the elution profile of AtGCL from a size-exclusion column following the addition of reductant; however, ESI-MS analysis before and after treatment with DTT indicates that the protein remains monomeric in both conditions. 2 Although some dimer species are observed under non-reducing conditions, this species accounts for one-thirtieth of the total protein. Similar results were reported for the T. brucei GCL using analytical ultracentrifugation and likely represent a nonspecific interaction between monomers (46). These experiments suggest that monomeric AtGCL undergoes a conformational change in response to redox environment, modulating enzymatic activity.
Because this change is mediated by reducing agents, an intramolecular disulfide bond is likely involved. When the cellular environment is reducing (i.e. high glutathione levels), the disulfide bond is disrupted and AtGCL adopts a rod-like shape with attenuated activity. Thus, regulation of AtGCL would not require association with a regulatory subunit, as in the mammalian system, for activation. This model for controlling AtGCL activity is also consistent with observations that oxidative stimulus induces glutathione synthesis (47,48). May et al. (48) noted that hydrogen peroxide treatment of a GCL-deficient yeast strain complemented with AtGCL yielded higher glutathione levels without increased gene expression and suggested that posttranslational modification of GCL occurred. Activation of GCL by a change in redox environment would also explain these results. Similar redox signal-induced conformational switching by disulfide formation of intracellular proteins has been described for other proteins, including Yap1 (49), thioredoxins (50), and OxyR (51).