Uncoupling the Enzymatic and Autoprocessing Activities of Helicobacter pylori γ-Glutamyltranspeptidase*

γ-Glutamyltranspeptidase (γGT), a member of the N-terminal nucleophile hydrolase superfamily, initiates extracellular glutathione reclamation by cleaving the γ-glutamyl amide bond of the tripeptide. This protein is translated as an inactive proenzyme that undergoes autoprocessing to become an active enzyme. The resultant N terminus of the cleaved proenzyme serves as a nucleophile in amide bond hydrolysis. Helicobacter pylori γ-glutamyltranspeptidase (HpGT) was selected as a model system to study the mechanistic details of autoprocessing and amide bond hydrolysis. In contrast to previously reported γGT, large quantities of HpGT were expressed solubly in the inactive precursor form. The 60-kDa proenzyme was kinetically competent to form the mature 40- and 20-kDa subunits and exhibited maximal autoprocessing activity at neutral pH. The activated enzyme hydrolyzed the γ-glutamyl amide bond of several substrates with comparable rates, but exhibited limited transpeptidase activity relative to mammalian γGT. As with autoprocessing, maximal enzymatic activity was observed at neutral pH, with hydrolysis of the acyl-enzyme intermediate as the rate-limiting step. Coexpression of the 20- and 40-kDa subunits of HpGT uncoupled autoprocessing from enzymatic activity and resulted in a fully active heterotetramer with kinetic constants similar to those of the wild-type enzyme. The specific contributions of a conserved threonine residue (Thr380) to autoprocessing and hydrolase activities were examined by mutagenesis using both the standard and coexpression systems. The results of these studies indicate that the γ-methyl group of Thr380 orients the hydroxyl group of this conserved residue, which is required for both the processing and hydrolase reactions.

␥-Glutamyltranspeptidase (␥GT), a member of the N-terminal nucleophile hydrolase superfamily, initiates extracellular glutathione reclamation by cleaving the ␥-glutamyl amide bond of the tripeptide. This protein is translated as an inactive proenzyme that undergoes autoprocessing to become an active enzyme. The resultant N terminus of the cleaved proenzyme serves as a nucleophile in amide bond hydrolysis. Helicobacter pylori ␥-glutamyltranspeptidase (HpGT) was selected as a model system to study the mechanistic details of autoprocessing and amide bond hydrolysis. In contrast to previously reported ␥GT, large quantities of HpGT were expressed solubly in the inactive precursor form. The 60-kDa proenzyme was kinetically competent to form the mature 40-and 20-kDa subunits and exhibited maximal autoprocessing activity at neutral pH. The activated enzyme hydrolyzed the ␥-glutamyl amide bond of several substrates with comparable rates, but exhibited limited transpeptidase activity relative to mammalian ␥GT. As with autoprocessing, maximal enzymatic activity was observed at neutral pH, with hydrolysis of the acyl-enzyme intermediate as the rate-limiting step. Coexpression of the 20-and 40-kDa subunits of HpGT uncoupled autoprocessing from enzymatic activity and resulted in a fully active heterotetramer with kinetic constants similar to those of the wild-type enzyme. The specific contributions of a conserved threonine residue (Thr 380 ) to autoprocessing and hydrolase activities were examined by mutagenesis using both the standard and coexpression systems. The results of these studies indicate that the ␥-methyl group of Thr 380 orients the hydroxyl group of this conserved residue, which is required for both the processing and hydrolase reactions.
Helicobacter pylori is a Gram-negative bacterial pathogen that colonizes the gastric mucosa. Infection puts the individual at greater risk for developing gastritis, peptic ulcer disease, and gastric cancer (1,2). H. pylori ␥-glutamyltranspeptidase (Hp-GT) 3 is a glutathione-degrading enzyme that has been shown to be a virulence factor in infection (3,4). H. pylori lacking ␥-glutamyltranspeptidase has been shown to grow normally in vitro, but exhibits diminished growth rates within the gut in animal model systems. Bacterial loads of the HpGT-deficient strains are reduced by nearly 70% relative to the parental strain. Although not essential for colonization, HpGT clearly confers a growth advantage to the bacteria in vivo by mechanisms that remain unclear. HpGT has also been shown to up-regulate COX-2 and epidermal growth factor-related peptides in human gastric mucosal cells (5) and to induce apoptosis in human gastric epithelial cells (6). Both these activities are abolished by inactivation of the enzyme with mechanism-based inhibitors. Despite its demonstrated involvement in H. pylori colonization, persistence, and disease progression, biochemical characterizations of HpGT have been limited.
The reclamation of extracellular glutathione and its conjugates is initiated by ␥-glutamyltranspeptidase (␥GT). The enzyme cleaves the ␥-glutamyl amide bond to liberate cysteinylglycine, and the catalytic mechanism proceeds via a ␥-glutamyl-enzyme intermediate (7)(8)(9)(10)(11). The ␥-glutamyl group can be transferred to water (hydrolysis) or to an amino acid or short peptide (transpeptidation). Whereas mammalian ␥GTs are embedded in the plasma membrane by a single N-terminal transmembrane anchor and are heterologously glycosylated, bacterial homologs are soluble and localized to the periplasmic space. Overall, the ␥GTs are highly conserved, with mammalian and bacterial homologs often sharing Ͼ25% sequence identity.
A required post-translational modification is the maturation of the precursor protein. ␥GT is synthesized as a 60-kDa polypeptide, and cleavage of the proenzyme yields a heterodimer composed of a 40-and a 20-kDa subunit (see Scheme 1) (3,(12)(13)(14). Processing of ␥GT is thought to be an intramolecular autocatalytic event, and a mechanism has been proposed in which processing proceeds via a nitrogen 3 oxygen acyl shift, with a conserved threonine residue serving as the nucleophile (15). Based on its enzymatic function and autoprocessing activity, ␥GT has been classified as an Ntn (N-terminal nucleophile) hydrolase (15). Members of the Ntn hydrolase family contain an ␣␤␤␣-core structure, are autocatalytically processed to yield an active enzyme, and catalyze amide bond hydrolysis (16,17). The new N-terminal residue of the processed enzyme, typically a serine, threonine, or cysteine residue, then serves as a nucleophile in the catalytic mechanism. Although the general features of the function of HpGT can be inferred based on its classification as an Ntn hydrolase, many mechanistic details of the autoactivation and catalytic function of HpGT have not been addressed.
In this study, we isolated and biochemically characterized recombinant H. pylori ␥-glutamyltranspeptidase. In contrast to previous reports of ␥GT purified from other organisms, the solubly expressed protein was isolated primarily as the 60-kDa precursor. This allowed us an opportunity to examine the autoprocessing of the protein. The rate of maturation was maximal at neutral pH, as was the enzymatic activity of the mature ␣ 2 ␤ 2heterotetramer. We used biochemical measurements of enzymatic activity in conjunction with site-directed mutagenesis and a coexpression system to investigate the involvement of a conserved threonine residue in the processing and catalytic activities.

Generation of Wild-type and Mutant HpGT Expression
Constructs-The isolation of HpGT has been described (3), and its sequence corresponds to a predicted open reading frame within the H. pylori genome (KEGG Data Base entry HP1118) (18,19). Although encoded by a single gene, the protein was isolated as two polypeptides of ϳ40 and ϳ20 kDa. Primers were designed to amplify HpGT, excluding a 26-amino acid signal sequence that targets the enzyme to the periplasmic space (3). Using H. pylori genomic DNA (American Type Culture Collection) as a template, a 1.6-kb DNA fragment was amplified by PCR and inserted into a pET-28a expression vector (Novagen) incorporating a thrombin-cleavable N-terminal histidine tag. The resulting expression construct (pET-28/HpGT) was sequenced at the Genomics Core Facility of the University of Nebraska (Lincoln, NE) and confirmed to be identical to HP1118 excluding the signal sequence.
Expression constructs for individual subunits of the processed enzyme were generated. The gene sequence for the N-terminal 40-kDa subunit was amplified by PCR, incorporating SpeI and SalI restriction sites. The insert was digested with the appropriate restriction enzymes and ligated into the pET-28a expression vector digested with NheI and SalI. Similarly, the gene sequence for the C-terminal 20-kDa subunit was amplified by PCR, incorporating NdeI and SalI restriction sites. The resultant 0.6-kb PCR product was digested with the relevant enzymes and ligated into a similarly digested pET-24a expression vector (Novagen), thus incorporating a C-terminal histidine tag. A bicistronic construct was also generated to express the N-terminal 40-kDa and C-terminal 20-kDa subunits separately but concurrently. The full-length HpGT expression construct was used as a template, and the HpGT sequence corresponding to the 20-kDa subunit was amplified by PCR, incorporating NdeI and XhoI restriction sites. The product was digested with the relevant enzymes and ligated into a similarly digested the pETDuet expression vector (Novagen). The dual expression vector containing the 20-kDa coding sequence was amplified and isolated. Next, the coding sequence for the 40-kDa subunit was excised from the above 40-kDa expression vector using NcoI and SalI. The resultant 1-kb fragment was gel-purified and ligated into a similarly digested pET-Duet expression vector containing the 20-kDa subunit sequence. The completed construct (HpGT-Duet) contained an N-terminal histidine tag on the 40-kDa subunit and an N-terminal methionine on the 20-kDa subunit. Point muta-tions were generated using the QuikChange site-directed mutagenesis kit (Stratagene) following the manufacturer's protocol, and all constructs were verified by sequencing at the Genomics Core Facility of the University of Nebraska.
Expression and Purification of HpGT-The pET-28/HpGT expression construct was used to transform Escherichia coli strain Rosetta(DE3)pLysS (Novagen). Cultures grown at 30°C to A 600 nm ϭ 0.6 -0.8 in 2ϫ YT medium containing 34 mg/liter chloramphenicol and 30 mg/liter kanamycin were induced for 4 h by the addition of isopropyl ␤-D-thiogalactopyranoside to a final concentration of 500 M. The cells were harvested by centrifugation and stored at Ϫ80°C. Cells were resuspended in lysis buffer (50 mM NaH 2 PO 4 (pH 8.0), 300 mM NaCl, and 10 mM imidazole), lysed by sonication, and centrifuged. The supernatant was treated with 0.35% polyethyleneimine at 4°C and centrifuged to separate precipitated nucleic acids from the protein-containing supernatant. The protein was purified by affinity chromatography using a nickel-chelating column (Novagen) following the manufacturer's protocol. Recombinant HpGT was concentrated, dialyzed against 20 mM Tris (pH 7.4), and stored at 4°C. Protein concentrations were estimated using a calculated extinction coefficient based on aromatic residue content (A 280 nm of 1 mg/ml solution ϭ 0.766) (20). The histidine tag used for affinity purification was removed by proteolytic cleavage with thrombin (1 unit of thrombin/mg of HpGT, incubated overnight at 18°C), which left an additional seven residues (GSHMASA) on the N terminus of HpGT. Following thrombin incubation, Ͼ50% of HpGT was found in the processed form. To ensure complete maturation of the enzyme, it was incubated at 37°C for 6 h. The sample was centrifuged to remove precipitated proteins, concentrated, and dialyzed against 20 mM Tris (pH 7.4).
A similar protocol was followed for each of the expression constructs with the following modifications. After ascertaining that the N-terminal histidine tag did not impact the processing or enzymatic activity of the wild-type enzyme, the thrombin cleavage step was omitted from all subsequent enzyme purifications. The HpGT-Duet construct confers resistance to ampicillin, and kanamycin was thus replaced with 100 mg/liter ampicillin. For HpGT generated using the pETDuet expression construct (HpGT-Duet), the enzyme was isolated in its mature form, and additional incubations were not required.
Kinetic Characterizations of HpGT-To determine the rate of processing and its effects on catalytic activity, unprocessed HpGT was purified as described above with the following modifications. After concentration and dialysis of the enzyme, it was flash-frozen and stored at Ϫ80°C to limit processing. It is important to note that Ͻ10% of the protein was cleaved after purification (see Fig. 1A, lane 1). The following day, HpGT (1 mg/ml) was thawed and incubated in 20 mM Tris (pH 8.0) at 37°C. At the indicated times, an aliquot was removed, assayed for hydrolase activity as described below, and denatured by boiling in SDS loading buffer for 5 min. Samples were analyzed by SDS-PAGE, and the gel was stained using GelCode Blue (Pierce). To quantify the degree of processing, a digital image of the gel was taken, and densitometric measurements were made on the 20-, 40-, and 60-kDa bands at each time point. For wildtype HpGT, bands at 30 h (see Fig. 1A, lane 8) were taken to represent complete cleavage, and the percent processing was calculated for the 40-kDa band at a given time point. The percent processing values were plotted versus time to obtain a processing rate (see Fig. 1B). A similar analysis was performed with the T380S mutant and required an extended incubation of 30 days. Representative plots from three or more determinations are shown. To examine the pH profile of autoprocessing, a citrate/phosphate buffer system of constant ionic strength (21) was used, and the percent processing was calculated for each band at a given time point and averaged.
The cleavage of the ␥-glutamyl group of the physiological substrate glutathione was assessed using a coupled assay system to detect the release of glutamate. HpGT (10 g/ml) in 0.1 M Tris (pH 8.0) was incubated with 20 mM glutathione at 37°C for 15 min. The reaction was stopped by boiling, and precipitated protein was removed by centrifugation. Parallel incubations were performed in the absence of HpGT to establish the nonenzymatic rate of hydrolysis. The amount of glutamate released was determined using a glutamate dehydrogenase assay following the conversion of glutamate to ␣-ketoglutarate concomitant with the production of NADH. Aliquots of the HpGT reaction mixtures were withdrawn, incubated at 37°C with glutamate dehydrogenase (0.1 mg/ml; Sigma) and NAD ϩ (2 mM), and allowed to come to equilibrium (Ͼ2 h). The absorbance at 340 nm was measured, and a standard curve for glutamate (0 -100 M) was determined. Similar experiments were performed using glutamine as the HpGT substrate.
For routine enzymatic characterizations, the substrate analog L-glutamic acid ␥-(4-nitroanilide) (GNA) was used (11,22). The release of 4-nitroaniline can be monitored continuously at 412 nm, and concentrations can be determined using the reported extinction coefficient of 8800 M Ϫ1 cm Ϫ1 . For standard assays, hydrolysis activity measurements were made in 0.1 M Tris (pH 8.0) containing 1 mM GNA at 25°C using a Cary 50 spectrophotometer. The pH profile of the enzymatic activity was assessed using a citrate/phosphate buffer system of constant ionic strength (21). To determine the apparent kinetic constants, HpGT activity was assayed at various concentrations of GNA ranging from 1 to 1000 M. The reaction exhibited saturation kinetics with respect to GNA, and the data were fit using the Michaelis-Menten equation to determine K m and V max values. To assess transpeptidation to an acceptor substrate, 20 mM glycylglycine was added to the assay mixture. Pre-steady-state kinetic studies were also performed. HpGT (17 M final concentration) was rapidly mixed with GNA (1 mM final concentration) at 10°C, and the absorbance at 412 nm was monitored using an Applied Photophysics stopped-flow apparatus (23).

Molecular Mass Determination of HpGT-Purified
HpGT was loaded onto a Superdex 200 HR 10/30 gel filtration column (Amersham Biosciences) and separated by fast protein liquid chromatography in 50 mM NaPO 4 (pH 7.0) containing 150 mM NaCl at a flow rate of 0.25 ml/min. Size determination was made by comparison with molecular mass standards (Amersham Biosciences) chromatographed under the same conditions. The molecular mass standards used were as follows: thyroglobulin, 699 kDa; ferritin, 416 kDa; catalase, 219 kDa; aldolase, 176 kDa; albumin, 67 kDa; ovalbumin, 47 kDa; chymotrypsinogen A, 20 kDa; and RNase A, 15 kDa. Dynamic light scattering experiments were performed on a DynaPro MSXTC instrument (Proterion Corp., Somerset, NJ). Protein samples (1 mg/ml) were centrifuged at 20,000 ϫ g for 15 min to remove particulates and loaded into the sample cuvette. Data were analyzed using the manufacturer's DYNAMICS software package, and the molecular mass was estimated based upon the hydrodynamic radius of the protein assuming a spherical model. Electrospray mass spectrometric analyses were performed at the Nebraska Redox Biology Center Metabolomics Core Facility of the University of Nebraska (Lincoln).

Protein Expression and
Purification-HpGT is expressed as an ϳ60-kDa precursor that undergoes autocatalytic processing to an ϳ40 kDa ␣-subunit and an ϳ20-kDa ␤-subunit (Scheme 1). HpGT has been implicated as a virulence factor in colonization (3,4), an inducer of apoptosis (6), and a modulator of COX-2 and epidermal growth factor-related peptides (5). However, the autoprocessing activity, enzymatic profile, and quaternary structure of HpGT have not been rigorously characterized. To examine these properties, we generated a prokaryotic expression vector containing the coding sequence of HpGT and overexpressed the protein in E. coli. HpGT was purified by affinity chromatography exploiting an engineered N-terminal hexahistidine tag, with typical yields of 50 mg of purified protein/liter of bacterial culture. The enzyme was Ͼ95% pure based on SDS-PAGE, and the vast majority of HpGT exhibited an apparent molecular mass of ϳ60 kDa (Fig. 1A, lane 1), corresponding to the inactive precursor. The precursor could be induced to catalytically cleave itself to produce the active heterodimer composed of an ϳ40and an ϳ20-kDa subunit (Fig.  1A, lane 8).
Autocatalytic Processing of HpGT-To characterize the autoprocessing of the enzyme, we monitored the conversion of precursor HpGT to its mature form as a function of time ( Fig. 1). At the indicated times, an aliquot of HpGT was removed and denatured. Samples were analyzed by SDS-PAGE (Fig. 1A), and the percent processing was plotted versus time (Fig. 1B). Under the indicated conditions, the processing of HpGT exhibited t1 ⁄ 2 ϭ 1.73 Ϯ 0.22 h (Table 1). Catalytic activity was also plotted as a function of processing (Fig. 1B, inset). Hydrolysis activity was strongly dependent on autocatalytic processing of the enzyme, exhibiting a nearly 1:1 relationship. Extrapolation to the completely unprocessed enzyme suggested that the uncleaved enzyme exhibited ϳ7% activity. However, the uncertainty in the densitometric measurements may underestimate the extent of cleavage in the early time points. Subsequent studies of HpGT mutants (discussed below) with diminished processing activities indicated that processing is required for enzymatic activity. Thus, it is likely that the unprocessed wild-type enzyme is also completely inactive.
The maturation of HpGT was further characterized to gain insights into the autoprocessing reaction. To verify that matu- SCHEME 1 Characterization of H. pylori ␥-Glutamyltranspeptidase ration of HpGT is an intramolecular event, the rates of HpGT cleavage were determined over a range of protein concentrations (0.1-10 mg/ml). Comparable processing rates were observed at each protein concentration. Similarly, the addition of mature HpGT to unprocessed protein at various ratios (1:10, 1:1, and 10:1) did not impact the rate of precursor maturation (data not shown). These observations strongly support an autocatalytic intramolecular maturation mechanism.
The pH dependence of HpGT processing was also examined and found to be most efficient in the neutral pH range (data not shown). Because of the inherent uncertainty in the densitometric measurements and prolonged incubation times, considerable variability in rates was observed. Most notably, reliable rates for processing could not be determined below pH 5.0 and above pH 9.0, as incomplete maturation of the enzyme was observed. At these extreme pH values, HpGT was prone to aggregation as judged by dynamic light scattering experiments.
Determination of the Oligomeric State of HpGT-To examine the quaternary structure of HpGT, the apparent molecular mass of the native enzyme was determined by gel filtration (Fig.  2). Compared with molecular mass standards, mature HpGT exhibited an estimated molecular mass of ϳ90 kDa, suggesting that the enzyme is either an extended heterodimer (␣␤, 60 kDa) or a compact heterotetramer (␣ 2 ␤ 2 , 120 kDa). To differentiate between these two possibilities, dynamic light scattering experiments were performed. Measurements of mature HpGT indicated a molecular mass between 109 and 122 kDa, suggesting that the enzyme is an ␣ 2 ␤ 2 -heterotetramer. Although SDS-PAGE analysis suggested equivalent concentrations of 20-and 40-kDa subunits, the possibility of an unequal combination of ␣and ␤-subunits in the mature enzyme, such as ␣ 1 ␤ 2 or ␣ 2 ␤ 1 , was excluded by studies with the unprocessed precursor and an HpGT mutant incapable of forming the ␣and ␤-subunits (T380A; discussed below). Both of these proteins ran nearly identically to mature HpGT on the gel filtration column, and both had comparable molecular masses (110 -120 kDa) as judged by dynamic light scattering experiments.
Kinetic Characterization of HpGT-A coupled assay system was used to detect the release of glutamate from potential physiological substrates. HpGT exhibited a specific activity of 3.59 Ϯ 0.17 mol of glutathione hydrolyzed per min/mg of enzyme. Similar experiments were performed using glutamine as the substrate, and a comparable rate of hydrolysis (2.77 Ϯ 0.04 mol of glutamine hydrolyzed per min/mg of enzyme) was observed. Cleavage of the ␥-glutamyl bond in both glutathione and glutamine was verified by mass spectrometry. The coupled assay system is cumbersome, and therefore, most studies of ␥GT have employed the substrate analog GNA. To obtain kinetic constants (K m and V max ) for the artificial substrate, the steady-state release of 4-nitroaniline was monitored by its absorbance at 412 nm. Measurements to determine the dependence of the reaction on substrate concentration were done using purified HpGT incubated with increasing concentrations of GNA. Saturation kinetics were observed, and data were fit to the Michaelis-Menten equation to obtain K m and V max for the enzyme-catalyzed reaction ( Table 1). The apparent K m for GNA was 12.5 Ϯ 1.2 M, and the apparent V max was 5.81 Ϯ 0.13 mol/min/mg of enzyme.
The K m and V max values for the hydrolysis of GNA as catalyzed by HpGT are in good agreement with those reported for the human enzyme (Table 1) (24). However, there are considerable differences in transpeptidation. For human ␥GT, the addition of 20 mM glycylglycine as an acceptor substrate resulted in a Ͼ100-fold rate increase in the release of 4-nitroaniline. Furthermore, the apparent K m for GNA increased substantially (Table 1), presumably because the donor-and acceptor-binding sites are at least partially overlapping, i.e. the acceptor site is likely coincident with the cysteinylglycine portion of the glutathione-binding site (9,25). In contrast, HpGT was only marginally more active as a transpeptidase than a hydrolase, and the apparent K m for GNA was relatively unaffected by the presence of 20 mM glycylglycine (Table 1). HpGT did not exhibit saturation kinetics with respect to glycylglycine, whereas the human enzyme has a reported K m for glycylglycine of 2.5 mM. Furthermore, even at 100 mM glycylglycine, HpGT At each time point, a sample was removed and denatured by boiling in SDS loading buffer. Samples were run on a 12% polyacrylamide gel and stained. B, the extent of processing was determined using densitometric measurements of the 40-kDa band and plotted as a function of time. The processing reaction exhibited t1 ⁄2 ϭ 1.73 Ϯ 0.22 h in this representative characterization. Inset, the catalytic activity was plotted as a function of enzyme processing. A nearly 1:1 correlation between processing and activity was observed.
exhibited a modest ϳ2-fold rate increase relative to the hydrolase activity. Mass spectrometric analysis indicated that transpeptidation did occur, as a peak corresponding to the molecular mass of glutamylglycylglycine was observed. However, these data strongly indicate that HpGT is not an effective transpeptidase when using the standard acceptor glycylglycine.
The pH dependence of HpGT activity was also examined. At a given pH, steady-state kinetic constants were determined. No significant pH dependence for the K m values was observed (data not shown). However, a bell-shaped pH profile was observed for enzymatic activity (Fig. 3), suggesting that at least two ionizable groups contribute to catalysis. Similar to the pH dependence of autoprocessing, HpGT enzymatic activity with GNA as the substrate was greatest at neutral pH.
For mammalian ␥GT, the rate-limiting step in catalysis was shown to be the breakdown of the acyl-enzyme intermediate (23). To ascertain whether this is also the case for HpGT, presteady-state kinetic studies were initiated. HpGT was rapidly mixed with GNA, and the absorbance at 412 nm was monitored using an Applied Photophysics stopped-flow apparatus. A burst phase proportional to the enzyme concentration was observed, followed by a slower steady-state rate (Fig. 4). The predicted absorbance change for a single turnover (ϳ0.15 absorbance units) agreed closely with the observed burst phase.
The slower rate of 3.54 mol/min/mg of enzyme compared favorably with our steady-state kinetic measurement of 5.81 Ϯ 0.13 mol/min/mg of enzyme. This is consistent with the formation of an acyl-enzyme intermediate, the breakdown of which is rate-limiting overall.
Site-directed Mutagenesis of Thr 380 -Thr 380 is the new N-terminal residue of the ϳ20-kDa ␤-subunit of the mature enzyme. To assess the contributions of Thr 380 to autoprocessing and enzymatic activities, two point mutations were made at this position. Substitution with alanine (T380A) resulted in a protein that could not process even after 30 days at 37°C and that was devoid of enzymatic activity ( Table 2). Substitution with serine (T380S) resulted in an enzyme that was dramatically impaired with respect to processing. Prolonged incubation (Ͼ30 days) at 37°C led to complete maturation of the T380S mutant (Fig. 5A, lane 9) with a half-life of 5.75 Ϯ 0.82 days (Fig. 5B and Table 1). Thus, removal of the ␥-methyl group of Thr 380 resulted in a nearly 80-fold reduction in the rate of processing relative to the wild-type enzyme. The added bulk of the threonine side chain is likely required for proper orientation of the nucleophilic hydroxyl group in the processing reaction or possibly helps position other active-site residues needed for maturation.

FIGURE 2. Determination of the oligomeric state of HpGT. Recombinant
HpGT (1 mg/ml) was subjected to gel filtration chromatography, and the elution profile was compared with those of proteins with known molecular masses. Inset, the elution time of the standards was plotted as a function of molecular mass (f), and the equation for the line was used to determine the mass of HpGT (E). HpGT has an estimated molecular mass of ϳ90 kDa, suggesting that the mature enzyme is either an extended heterodimer (60 kDa) or a compact heterotetramer (120 kDa). Subsequent dynamic light scattering experiments are consistent with the latter as discussed under "Results." FIGURE 3. pH profile of the HpGT-catalyzed hydrolysis of the artificial substrate GNA. At the indicated pH values, kinetic constants for the hydrolysis of GNA were determined. The maximal enzymatic activity at a saturating substrate concentration was determined in triplicate and plotted as a function of pH. HpGT was most active at neutral pH and likely has at least two ionizable groups that contribute to catalysis.  (24). In these studies, saturation kinetics were observed with glycylglycine as the varied substrate with an apparent K m of 2.5 Ϯ 0.3 mM. b Saturation kinetics were not observed with glycylglycine as the varied substrate, and therefore, measurements were made using 20 mM glycylglycine as reported for the human enzyme.

Characterization of H. pylori ␥-Glutamyltranspeptidase
The ␥-methyl group of Thr 380 is also required for efficient enzymatic hydrolysis of GNA. Analysis of the kinetic constants for processed T380S indicated significant perturbations relative to the wild-type enzyme (Table 1). Overall, the rate of catalysis was reduced by ϳ12-fold, whereas the apparent K m for GNA was increased by ϳ2.5-fold. This reduced activity may result from the prolonged incubation (Ͼ30 days at 37°C) required for maturation. However, the native enzyme exhibited a Ͻ5% loss of activity over the same 30-day period required for maturation of the T380S mutant. Furthermore, the T380S mutant did not precipitate during the maturation process, and the mature T380S protein had a similar tryptophan fluorescence profile compared with the wild-type enzyme (data not shown), suggesting that the overall fold and intrinsic stability of the mutant were uncompromised. Although the ␥-methyl group of Thr 380 was required for optimal catalytic activity, its removal did not result in a dramatic loss of activity ( Table 1). As opposed to the wild-type protein, completely unprocessed T380S could be obtained. Analysis of enzymatic activity as a function of processing (Fig. 5B, inset) indicated that unprocessed T380S was devoid of enzymatic activity. In conjunction with similar observations with wild-type HpGT (Fig. 1B, inset), these data suggest that processing is required for HpGT activity.
Generation of a Coexpression Construct-Using the standard expression system, it is difficult to assess the precise role of Thr 380 in catalysis. As indicated above, the hydroxyl group of a serine residue could adequately substitute for the conserved threonine residue in the hydrolysis of GNA. Substitution with alanine resulted in a protein incapable of maturation (Fig. 6,  lane 3) and devoid of enzymatic activity. However, it was unclear whether the T380A mutant lacked enzymatic activity because it could not autoprocess or because the hydroxyl group of Thr 380 was required for catalytic activity. To address whether the hydroxyl group of Thr 380 is required for autoprocessing, enzymatic activity, or both, we generated a system in which the catalytic activity of HpGT could be uncoupled from its processing.
We created constructs to express the 40-kDa ␣-subunit and the 20-kDa ␤-subunit of mature HpGT independently in a bacterial expression system. Both the 40-and 20-kDa subunits were insoluble when expressed separately in E. coli over a wide range of conditions, suggesting that both subunits must be present for proper folding. Therefore, to express both subunits in the same cell, the gene sequences of the 40-kDa ␣-subunit and the 20-kDa ␤-subunit were subcloned into the pETDuet expression vector. In this construct, the 40-kDa subunit contained an N-terminal histidine tag, and the 20-kDa subunit had only the initiator methionine added (HpGT-Duet). The HpGT-Duet protein was purified by immobilized metal affinity chromatography with yields of 15-25 mg of HpGT/liter of bacterial culture. Similar molar ratios of ␣to ␤-subunits were observed HpGT was rapidly mixed with GNA, and the absorbance at 412 nm was monitored using a stopped-flow apparatus. The predicted absorbance change for a single turnover agrees closely with the observed burst. This observation is consistent with the rapid formation of an acyl-enzyme intermediate, followed by its slow hydrolysis.   Fig. 1. B, the extent of processing was determined using densitometric measurements as described under "Experimental Procedures" and plotted as a function of time. The rate of processing exhibited t1 ⁄2 ϭ 5.75 Ϯ 0.82 days. Inset, the catalytic activity of T380S plotted as a function of processing indicates a 1:1 correlation.
in the wild-type and HpGT-Duet proteins (Fig. 6, compare  lanes 1 and 2). Only the 40-kDa ␣-subunit had a hexahistidine tag for purification, suggesting that a properly folded ␣ 2 ␤ 2 -heterotetramer was formed. Analysis by electrospray mass spectrometry indicated that the initiator methionine had been removed from the 20-kDa subunit during expression in E. coli. The 20-kDa subunit generated via autoprocessing had an observed mass of 20,390.0 Da compared with 20,389.8 Da for that generated via the coexpression system (predicted molecular mass of 20,391.25 Da). HpGT-Duet had kinetic constants comparable with those of the wild-type enzyme ( Table 2) and thus successfully uncoupled maturation from enzymatic activity.
Using the newly developed coexpression construct, Thr 380 was again replaced with an alanine residue. Whereas the T380A mutant produced in the original construct persisted as an unprocessed 60-kDa protein (Fig. 6, lane 3), purified T380A-Duet contained both the ␣and the ␤-subunits (Fig. 6, lane 4), suggesting that the protein was correctly assembled. Removal of the initiator methionine from the ␤-subunit of the mutant protein was confirmed by mass spectrometry. Like the wildtype and HpGT-Duet proteins, the T380A-Duet protein formed an ␣ 2 ␤ 2 -heterotetramer in solution as judged by dynamic light scattering experiments. The T380A-Duet protein did not exhibit any catalytic activity with the substrate analog GNA ( Table 2), indicating that Thr 380 is required for enzymatic activity.

DISCUSSION
HpGT has been implicated in gut colonization and persistence by H. pylori as well as in the progression of diseases resulting from infection by these bacteria (3,4). To gain insights into the mechanisms by which HpGT confers a growth advantage, we have biochemically examined the enzymatic and processing activities of the enzyme. We were able to isolate an unprocessed precursor protein and to characterize its maturation in detail. The precursor protein lacks hydrolase activity, indicating that autoprocessing is required for enzymatic activation. In complementary studies, we successfully uncoupled processing from enzymatic activity using a coexpression system. These recombinant protein expression systems allowed us to assess the contributions of a conserved threonine residue (Thr 380 ) to autoprocessing independent of glutathione hydrolase activity.
In many organisms, ␥GT catalyzes the first step in the reclamation of extracellular glutathione and its conjugates by cleaving the ␥-glutamyl peptide bond. HpGT and human ␥GT share Ͼ30% sequence identity and exhibit similar hydrolase activities. However, a search of the H. pylori genome did not reveal any proteins with significant similarity to known glutathione biosynthetic enzymes, suggesting that H. pylori does not produce glutathione to modulate cellular redox levels (18,19). The bacteria are equipped with numerous peroxidases, catalase, and a functional thioredoxin system that may protect against oxidative damage (26). Perhaps, as suggested by in vivo studies (3,4), HpGT provides the bacteria with a growth advantage within the gut by salvaging extracellular glutathione to obtain cysteine for subsequent protein synthesis. Both the autoprocessing and hydrolase activities of HpGT are maximal at neutral pH and suggest a functional role for the enzyme following colonization. Studies in other model systems are consistent with ␥GT providing necessary precursors for protein and glutathione biosynthesis. Cysteine is often limiting in these processes and is obtained either via the transsulfuration pathway, which converts homocysteine to cysteine, or via the salvage of excreted glutathione and its conjugates. ␥GT-deficient mice appear normal at birth, exhibit slower growth, do not sexually mature, and have shorter life spans (27)(28)(29)(30)(31)(32)(33)(34). They are also considerably more susceptible to cataract formation (29) and damage by reactive oxygen species (32). Supplementation with N-acetylcysteine largely restores the native phenotype, suggesting that the primary function of ␥GT in normal cells is the recovery of cysteine for use in protein and glutathione synthesis, which is important for adequate growth and oxidative protection (28). In agreement with these findings, ␥GT is found up-regulated in several cancer types and has been shown to accelerate tumor growth and to increase tumor resistance to damage induced by chemotherapy and radiation treatment (35)(36)(37). Thus, it is reasonable to suppose that expression of a homologous enzyme would also facilitate growth of the bacteria.
HpGT exhibits similar rates of hydrolysis with glutathione, glutamine, and the substrate analog GNA (Table 1), and these values are comparable with the rates observed for other ␥GTs isolated from E. coli (38), Bacillus subtilis (39), rat (40), and human (24). Pre-steady-state kinetic studies with HpGT (Fig. 4) indicated that hydrolysis of the ␥-glutamyl-enzyme intermediate is rate-limiting, as has been demonstrated in the rat homolog (23). In the presence of 20 mM glycylglycine, HpGT exhibits a modest 1.2-fold rate increase relative to hydrolysis and does not exhibit saturation kinetics with respect to this model acceptor peptide ( Table 1). Each of the 20 common L-amino acids, as well as reduced and oxidized glutathione, was tested as a potential acceptor substrate. Minimal HpGT transpeptidase activity was observed for each acceptor substrate examined, and none performed better than glycylglycine (data not shown). The inability of HpGT to efficiently catalyze the transfer of glutamate to the acceptor amino acids and peptides tested may be a general feature of bacterial ␥GT (38, 39) and differentiates HpGT from eukaryotic homologs (24,40).
In studies with rat ␥GT, Keillor et al. demonstrated that nucleophilic attack upon the acyl-enzyme intermediate is ratelimiting (23) and that activation of the amino acid acceptor is catalyzed by a general base (41). Furthermore, they confirmed that, like the human enzyme, rat ␥GT shows a strict requirement for L-amino acids as the acceptor. Therefore, for mammalian ␥GT, transpeptidation is significantly faster than hydrolysis because the acceptor peptide is a more effective nucleophile than water. For HpGT, the modest rate increases observed in the presence of high concentrations of acceptor substrates are probably the result of nonspecific nucleophilic attack by the ␣-amino group. Alternatively, the diminished transpeptidase activity of HpGT may be because the tested compounds are not physiological acceptor substrates, and additional studies to investigate this possibility are ongoing.
In previous reports describing the expression and purification of ␥GT, the purified enzyme had been processed into an ϳ40and an ϳ20-kDa heterodimer (3,6,38,42). In one exceptional case, a fusion protein between maltose-binding protein and E. coli ␥GT was reported purified in the uncleaved form; this protein was expressed in inclusion bodies in E. coli and had to be refolded to regain limited processing and catalytic activities (15). Thus, kinetic parameters for the processing and hydrolase activities were not reported. Despite these limitations, the authors were able to demonstrate that a conserved threonine residue (Thr 391 ) is involved in the intramolecular autoprocessing reaction. In the present study, recombinant HpGT was isolated as a kinetically competent precursor that quantitatively underwent autoprocessing to generate a mature, fully active enzyme (Fig. 1). The processing reaction was most proficient at neutral pH and was considerably impeded at extreme pH values. The strong correlation between the extent of processing and enzymatic activity with both wild-type HpGT and the slow-processing T380S mutant strongly suggests that processing is required for enzymatic activity.
Attempts to coexpress the two subunits of mature Ntn hydrolases have been successful in a number of cases (43,44). For E. coli ␥GT, however, coexpression of the 20-and 40-kDa subunits does not result in significant quantities of active enzyme (45). In contrast, coexpression of the ␣and ␤-subunits of HpGT results in an enzyme with kinetic constants comparable with those of the mature wild-type enzyme ( Fig. 6 and Table  2). Furthermore, expression levels (15-25 mg/liter of bacterial culture) are comparable with those of the wild-type enzyme. The ability to uncouple enzymatic activity from maturation, as well as to monitor the autoprocessing reaction, makes HpGT an excellent model system to examine structure-function relationships in ␥GT.
The conserved threonine residue at the newly formed N terminus of mature ␥GT has been suggested to be the relevant nucleophile in both the autoprocessing and hydrolase reaction mechanisms (14,15,46,47). Chemical modification studies with E. coli ␥GT identified Thr 391 as a highly reactive group within the enzyme active site (47). However, its direct involvement in catalysis could not be verified; substitution with an alanine residue results in a protein incapable of maturation. Although this observation clearly implicates the conserved threonine in the autoprocessing reaction, it does not demonstrate that the threonine residue is the catalytic nucleophile in the hydrolase reaction. Comparable mutagenesis studies with rat ␥GT have implicated the conserved threonine in the processing reaction (14), but again, its role in the hydrolase mechanism could not be tested. With HpGT as a model system, the contributions of Thr 380 to both the autoprocessing and enzymatic activities of the enzyme were accessible to quantification.
As observed in other systems, substitution of Thr 380 with alanine (T380A) in HpGT results in a protein incapable of maturation. A serine residue can substitute for Thr 380 , but results in an ϳ80-fold reduction in the maturation rate ( Fig. 5 and Table  1), suggesting that the ␥-methyl group of Thr 380 helps to position its hydroxyl group for optimal processing. Similarly, mature T380S has reduced hydrolase activity (Table 1), underscoring the importance of the ␥-methyl group in orienting the nucleophilic hydroxyl group for the enzymatic reaction. To determine whether the hydroxyl group of Thr 380 is indeed critical for hydrolase activity, a coexpression system was used to produce a T380A mutant in a mature form (T380A-Duet). Although the T380A-Duet mutant formed an ␣ 2 ␤ 2 -heterotetramer, it completely lacked enzymatic activity when using the substrate analog GNA (Table 2). These observations strongly support the proposed function of Thr 380 as the required nucleophile in both the processing and enzymatic reactions.