BPTES inhibition of hGA_124–551, a truncated form of human kidney-type glutaminase

Abstract

The initial transcript of the GLS1 gene undergoes alternative splicing to produce two glutaminase variants (KGA and GAC) that contain unique C-terminal sequences. A truncated form of human glutaminase (hGA_124–551) that lacks either C-terminal sequence was expressed in E.Coli and purified. This construct exhibits a hyperbolic glutamine saturation profile (K_m of 1.6 mM). BPTES, bis-2[5-phenylacetamido-1,2,4-thiadiazol-2-yl]ethylsulfide, functions as a potent uncompetitive inhibitor of this construct (K_i of 0.2 µM). The hGA_124–551 is inactive in the absence of phosphate, but exhibits a hyperbolic phosphate-dependent activation profile that is also inhibited by BPTES. Gel filtration studies indicate that hGA_124–551 forms a dimer in the absence or presence of 100 mM phosphate, whereas addition of BPTES causes the formation of an inactive tetramer. The combined data indicate that BPTES inhibits human glutaminase by a novel mechanism and that BPTES is a potential lead compound for development of an effective cancer chemotherapeutic agent.

Keywords::

• Recombinant glutaminase
• kinetic analysis
• uncompetitive inhibitor
• oligomerization
• cancer chemotherapy

Introduction

Most cancer cells take up large amounts of glucose that is consumed primarily by anaerobic glycolysis, even in the presence of oxygen^1,2. To maintain mitochondrial function, transformed cells also exhibit an increased catabolism of glutamine to lactate³. This pathway, termed as glutaminolysis, generates ATP, NADPH and the precursors necessary to support the synthesis of the nucleotides and lipids that are required for cell division. As a result, many transformed cells require glutamine as an essential nutrient. To support this addiction to glutamine, transformed cells frequently exhibit an increased expression of the kidney-type glutaminase (GA^4,5). Early studies used Ehrlich ascites tumor cells to demonstrate the role of GA in maintaining a transformed phenotype⁶. The tumor cells, which were stably transfected with a plasmid that encoded a GA antisense RNA, lost their transformed phenotype and failed to produce tumors when injected into mice. Interest in GA as a potential cancer chemotherapeutic target was kindled further by two recent studies. The initial study demonstrated that increased expression of the c-Myc oncogene in human P-493 B lymphoma cells resulted in increased expression of GA⁷. Further experiments demonstrated that c-Myc expression suppressed the synthesis of two microRNAs, miR-23a/b, that inhibit GA expression. The authors also demonstrated that siRNA knockdown of GA significantly decreased the rates of proliferation of P493-6 cells and human PC3 prostate cancer cells, two transformed cell lines that exhibit oncogenic levels of c-Myc. In the second study⁸, molecule 968 was identified as a potent inhibitor of the cellular transformation that was produced by expression of an oncogenic Dbl, a mutated form of a Rho-family guanine nucleotide exchange factor. Subsequent pull down and mass spectrometric analysis identified the mitochondrial GA as the target of 968. Additional studies established that siRNA knockdown of GA mimicked the effects of 968. Both treatments inhibited the ability of three constitutively activated Rac or Rho GTPases to stimulate growth in low serum or in soft agar. GA knockdown also inhibited proliferation of transformed NIH3T3 fibroblasts and breast cancer cells. Recent studies have also demonstrated that BPTES selectively inhibits the growth of glioma cells⁹.

Human GA is encoded by the GLS1 gene which spans 82 kb on chromosome 2 and contains 19 exons. The initial transcript of this gene undergoes alternative splicing to produce two isoforms, KGA and GAC^10,11. The KGA isoform is highly expressed in kidney, brain, intestine, and cells of the immune system, whereas GAC is expressed in heart, pancreas, placenta, lung and in many transformed cells¹⁰. The KGA and GAC isoforms share identical N-terminal and core amino acid sequences that are transcribed from exons 1–14, but have distinct C-terminal regions. The C-terminal amino acid sequence in GAC is derived from exon 15, whereas the C-terminal sequence of KGA is derived from exons 16 through 19. The N-terminal regions of various glutaminase proteins are also highly divergent, whereas the central core region is highly conserved from bacteria to humans¹². This homology suggests that the catalytic domain is located within the central region of the protein¹³. A unique catalytic property of GA is a potent activation by phosphate and other polyvalent anions¹⁴. The K_m for glutamine decreases in the presence of increasing phosphate concentration and phosphate activation correlates with the association of inactive dimers to form active tetramers or larger oligomers^15,16.

Most GA inhibitors are affinity labeling reagents or substrate analogs that have a broad specificity and a mM affinity. By contrast, the structure of BPTES differs greatly from that of glutamine¹⁷. In the current study, we characterized the kinetics of BPTES inhibition of a highly purified, truncated form of human kidney-type glutaminase (hGA_124–551) that lacks the C-terminal sequence from either the KGA or the GAC isoform. This protein retained the basic kinetic properties characteristic of the native enzyme, including a phosphate-dependent activation profile. Kinetic analysis indicated that BPTES functions as a potent uncompetitive inhibitor with a K_i of 0.2 µM. Gel filtration experiments indicated that this form of glutaminase retains a dimeric structure even when activated by phosphate. However, the binding of BPTES still promotes the formation of an inactive tetramer.

Materials and methods

Materials

Bovine liver glutamate dehydrogenase and all other biochemical reagents were obtained from Sigma, St. Louis, MO. BPTES (bis-2 [5-phenylacetamido-1, 2, 4-thiadiazol-2-yl] ethyl sulfide) was obtained from MGI-Pharma, Baltimore, MD.

Cloning of hGA_124–551

A cDNA segment encoding amino acids 124 through 551 of the hGA protein was PCR amplified from the hGA-4 plasmid¹⁸. The forward and reverse primers incorporated NdeI and BamHI restriction sites at the 5′ and 3′ ends, respectively. The PCR product was initially cloned into the pGEM-T easy vector, amplified, and then ligated into the pET-15b expression vector using the Quick Stick Ligation kit (Fermentas, Glen Burnie, ND). The resulting pET-hGA_124–551 plasmid was sequenced across the entire insert to confirm the identity of the plasmid.

Protein expression and purification

Expression of the recombinant hGA_124–551 protein was performed using an E. coli BL-21 (DE3) strain containing the Rosetta pRARE plasmid that was obtained from the Structural Genomics Consortium (Karolinska Institutet, Solna, Sweden). Cells were grown in 6 L of 2xYT medium at 37°C to an O.D.₆₀₀ of 1.0–1.5 and then cooled to 18°C before being induced with 0.5 mM IPTG. Cultures were harvested the following morning by centrifugation at 4000 rpm at 4°C for 30 min. The cell pellets were resuspended in 50 mL of lysis buffer containing 300 mM KCl, 10% glycerol (v/v), 10 mM potassium phosphate, 10 mM imidazole, 0.5 mM Tris(2-carboxyethyl)phosphine hydrochloride (TCEP), 10 mM Tris-Cl, pH 8.0, and one tablet of Complete Protease Inhibitor Cocktail (Roche, Indianapolis, IN). Resuspended cell pellets were then stored in a −80°C freezer. The cell pellets were thawed in an ice/water bath and disrupted by sonication in the presence of 2000–4000 units of Benzonase. The lysate was centrifuged at 35,000 rpm at 4°C for 25 min. The supernatant was pooled and filtered through a 0.22-μ syringe filter. The filtered supernatant was manually loaded onto a 5-mL HiTrap nickel-chelating column (Amersham Biosciences, Piscataway, NJ) using a peristaltic pump with a flow rate of ∼1.0 mL/min. The column was subsequently washed using 100 mL of lysis buffer containing 75 mM imidazole and the N-terminal His₆-tagged hGA_124–551 was eluted with lysis buffer containing 400 mM imidazole. The hGA_124–551 was recovered in a single 3-mL fraction. This fraction was loaded onto a S200 Superdex™ 16/60 gel filtration column that was pre-equilibrated with 300 mM KCl, 10% glycerol (v/v), 0.5 mM TCEP, and 10 mM Tris-Cl, pH 8.0. The column was run at a flow rate of 1.0 mL/min and the eluant was collected in 3.0-mL fractions. The purified recombinant protein was stored at 4°C. Aliquots of samples collected at various steps in the purification were assayed for protein¹⁹ and glutaminase activity²⁰.

Kinetic assays

hGA_124–551 activity was determined using a two-step assay that was performed in a 96-well plate and a final volume of 227 µL¹⁷. For the standard assay, each well contained 20 µL of 20 mM glutamine, 150 mM potassium phosphate, 0.2 mM EDTA, 50 mM Tris-acetate, pH 8.6. The assay was started by adding 5 µL of an appropriate dilution of the recombinant hGA_124–551. The reaction was incubated for 10 min in a 37°C incubator and then stopped by the addition of 2 µL of 3N HCl, which denatures the glutaminase. Next, 200 µL of an assay mix containing 0.4 mg/mL glutamate dehydrogenase, 0.2 mM NAD⁺, 200 mM hydrazine, 2.5 mM ADP, and 80 mM Tris-acetate, pH 9.4 was added and incubated for 35 min at room temperature. In this reaction, glutamate dehydrogenase converts the glutamate, produced in the initial reaction, to α-ketoglutarate and generates an equivalent amount of NADH. The added hydrazine reacts with α-ketoglutarate forming a hydrazide to ensure that the reaction proceeds to completion. The absorbance of the samples at 340 nm was measured using a Wallac Victor³V plate reader (PerkinElmer, Waltham, MA) versus a blank in which the 3N HCl was added immediately before the addition of hGA_124–551.

Glutamine saturation profiles were conducted using 0.5 mM–80 mM glutamine. Inhibition studies were performed by repeating the glutamine saturation profiles in the presence of 0.1 µM, 0.3 µM, 0.6 µM and 1.0 µM BPTES. The BPTES was added to the initial reaction mixture in a 1 µL volume. Because BPTES is dissolved in DMSO, a control was performed with addition of 1 µL of DMSO. Data analysis to determine the K_m and V_max were performed using KaleidaGraph software (Synergy Software, Reading, PA) to fit the Michaelis–Menten equation. To determine the K_i for BPTES, the apparent V_max and apparent K_m were plotted against the concentration of BPTES using KaleidaGraph software to fit the equations:

where [I] is equal to the concentration of BPTES.

Phosphate activation profiles were performed using 5 mM–150 mM phosphate. Inhibition studies were performed using 0.5 µM, 1.0 µM, 3.0 µM and 10 µM BPTES along with a DMSO control. The Hill coefficient and K_0.5 were determined using KaleidaGraph software to perform a non-linear least squares fit to the Hill equation:

where v_o/V_max is the fractional saturation of hGA_124–551 with phosphate, [P] is the concentration of phosphate, h is the Hill coefficient, and K_a is the dissociation constant. All kinetic assays were performed with 5 replicates and the data are reported as the mean +/- SE.

Size exclusion gel filtration chromatography

Gel filtration chromatography was performed at 4°C using a 15-mL Bio-Silect SEC 250 column attached to an AKTA FPLC system¹⁷. The SEC column was pre-equilibrated with 300 mM KCl, 10% glycerol (v/v), 0.5 mM TCEP and 10 mM Tris-Cl, pH 8.0 either with or without 100 mM potassium phosphate. The hGA_124–551 was concentrated by centrifuging at 1200g at 4°C in a 20-mL Vivaspin concentrator with a 3.0-kDa cut-off. The concentrated samples were dialyzed for 4 h at 4°C against SEC buffer containing 0 or 100 mM potassium phosphate. An 80-µL aliquot containing ∼1.0 mg/mL of the dialyzed hGA_124–551 (∼20 μM) was loaded onto the Bio-Silect 250 column and the elution profile was determined by measuring the absorbance at 230 nm. Bio-Rad gel filtration standards were used to calibrate the column and to determine the apparent molecular weight.

Results

Protein expression and purification

The pET-hGA_124–551 plasmid, when transformed into the BL-21 (DE3) strain of E. coli, produced very little recombinant hGA_124–551 protein in the absence of IPTG. However, expression of hGA_124–551 was strongly induced by maintaining the cells overnight at 18°C in the presence of 0.5 mM IPTG. Comparison of the cell lysate and the supernatant fraction indicated that ∼80 % of the recombinant protein was solubilized by the sonication procedure (Table 1). Subsequent chromatography on the nickel column resulted in a 60-fold enrichment in the specific activity of the hGA_124–551. Further chromatography on a Superdex S200 column was necessary to remove a few minor contaminants that were evident by SDS-PAGE. However, this step did not yield a further increase in specific activity. The final specific activity of nearly 600 µmol/min•mg was 2-fold greater than the specific activity of native GA isolated from rat kidney¹⁴.

Table 1.  Purification of recombinant hGA_124–551.

Step	Total volume (mL)	Activity (U/mL)	Total activity (U)	Protein (mg/mL)	Total protein (mg)	Specific activity (U/mg)
Lysate	31.6	263	8311	31.2	985	8.4
Supernatant	26.8	261	6995	22.9	614	10
Nickel affinity chromatography	3.0	539	1617	0.9	2.7	591

The activity of hGA_124–551 was measured at various steps in the purification using the standard glutamine assay ¹⁴, which has a final assay volume of 1.11 ml. The protein concentration was determined using the Bradford assay ¹⁹. The hGA_124–551 is reported in units(U) of μmol of product/min.

Kinetics

To determine the kinetic mechanism of BPTES inhibition, glutamine saturation profiles were performed with 30 mM phosphate in the absence or presence of 0.1, 0.3, 0.6 and 1.0 μM BPTES (Figure 1). All of the data fit to hyperbolic saturation profiles. In the absence of BPTES, the K_m for glutamine was 1.6 ± 0.13 mM. Increasing concentrations of BPTES produced approximately a 5-fold decrease in both the observed V_max and the K_m, indicative of uncompetitive or a mixed non-competitive inhibition. However, a double reciprocal plot of the data (Figure 2) produced parallel lines, indicative of uncompetitive inhibition. Uncompetitive inhibition occurs when the inhibitor binds to the enzyme-substrate complex and alters the K_m and catalytic activity to a similar extent. The K_i value for BPTES was determined by plotting the apparent V_max or the apparent K_m against the concentration of BPTES (Figure 3). This analysis yielded K_i values of 0.24 ± 0.03 μM and 0.16 ± 0.02 μM from the apparent V_max and apparent K_m values, respectively. The slight difference in the calculated K_i values supports the conclusion that BPTES functions as an uncompetitive inhibitor with respect to glutamine.

Figure 1.  BPTES inhibition of hGA_124–551 with respect to glutamine. Glutamine saturation profiles were produced in the presence of 30 mM phosphate and increasing concentrations of BPTES. The measured activity was plotted against the concentration of glutamine. The data are plotted using KaleidaGraph software and fit to the Michaelis–Menten equation using a non-linear regression. Data are the mean ± standard error of five replicates.

Figure 2.  Double-reciprocal plots of BPTES inhibition of hGA_124–551 activity. The glutamine saturation data collected in the presence of 30 mM phosphate were plotted using the Lineweaver-Burke double-reciprocal plot transformation of the Michaelis–Menten equation. The concentrations of BPTES ranged from 0.1 to 0.6 µM. The lines are the least squares fit of the data to the straight lines obtained by plotting the reciprocal of hGA_124–551 activity versus the reciprocal of the glutamine concentration.

Figure 3.  Determination of the inhibition constant for BPTES. The K_i for BPTES was determined by using KaleidaGraph software to fit the plots of the apparent V_max (open circles) and the apparent K_m (open squares) versus the concentration of BPTES to the equations listed in the Materials and Methods section. The derived K_i values for BPTES were 0.24 µM ± 0.03 and 0.16 µM ± 0.02 calculated from the apparent V_max and the apparent K_m, respectively.

Phosphate activation profiles were performed using 20 mM glutamine in the absence or presence of BPTES (Figure 4). The hGA_124–551 was essentially inactive in the absence of phosphate, but the activity increased with increasing concentrations of phosphate. When the data were fit to the Hill equation using KaleidaGraph software, the activation profile appeared to be hyperbolic with only a very slight degree of cooperativity (Hill coefficient of 1.3). Increasing concentrations of BPTES caused a pronounced decrease in the maximal activity of hGA_124–551. However, the K_0.5 values for the phosphate activation increased from 50 mM to nearly 3 M, suggesting that phosphate and BPTES compete for binding to hGA_124–551. The calculated Hill coefficients also increased to 2.2 as the concentration of BPTES was increased to 10 µM. This is evident in the activation profile with 10 µM BPTES, which is clearly sigmoidal (Figure 4).

Figure 4.  BPTES inhibition of hGA_124–551 with respect to phosphate. Phosphate activation profiles where produced using 20 mM glutamine in the absence and presence of 0.5 to 10 µM BPTES. The control profile contained an equivalent amount of DMSO. The data are the mean ± standard error of five measurements. The lines are a non-linear least squares fit to the Hill equation that was performed using KaleidaGraph software.

Gel filtration chromatography

Gel filtration chromatography was performed to determine the effects of phosphate and of BPTES on the apparent molecular weight of hGA_124–551 (Figure 5). In the absence of BPTES, the addition of 100 mM phosphate had little effect on the elution volume of the hGA_124–551. However, the addition of 2 µM BPTES caused the hGA_124–551 to elute from the column with a significantly shorter retention time. The apparent molecular weight of the hGA_124–551 was determined by calibrating the Bio-Silect 250 column with Bio-Rad gel filtration standards. In the absence of phosphate and BPTES the apparent molecular weight was 103 kDa, whereas in the presence of 100 mM phosphate the apparent molecular weight of the hGA_124–551 was 124 kDa. However, with the addition of 2 µM BPTES, the apparent molecular weight in the absence of phosphate was 226 kDa and in the presence of 100 mM phosphate was 245 kDa. The calculated molecular weight of the hGA_124–551 monomer is 47-kDa. These data suggest that in the absence of BPTES, hGA_124–551 exists in a dimer. Furthermore, the phosphate activation of this form of glutaminase may occur without formation of higher oligomers. However, the addition of BPTES in the presence or absence of phosphate still results in the formation of an inactive tetramer.

Figure 5.  Effects of phosphate and BPTES on the apparent molecular weight of hGA_124–551. Gel filtration chromatography was conducted by loading 80 µL of concentrated hGA_124–551 onto a 14-mL Bio-Silect 250 column equilibrated with 0 or 100 mM phosphate in the absence (Upper Panel) or presence of 2 µM BPTES (Lower Panel). Samples were run at 0.7 mL/min and the elution of hGA_124–551 was tracked by measuring the absorbance of the eluant at 230 nm. The apparent molecular weight of the hGA1_24-551 was calculated by comparing its peak elution volume to that of a series of standard proteins: Thy, thyroglobulin (670 kDa); γ-G, γ-globulin (158 kDa); Oval, ovalbumin (44 kDa); and Myo, myoglobin (17 kDa).

Discussion

The core region of glutaminase is highly conserved throughout evolution from bacteria to humans. This segment of the human glutaminase, which contains amino acids 221 through 532, was recently crystallized in the presence of glutamate by the Structural Genomics Consortium (PDB ID: 3CZD). This segment forms a compact globular structure that retains the same secondary and tertiary structure determined for the glutaminases from E. coli and B. subtilis¹². In addition, all of the active site residues identified in the bacterial glutaminases are conserved within this segment of the human enzyme. Thus, we initially planned to characterize the kinetic properties and BPTES inhibition of an hGA_221–532 construct. However, expression and purification of recombinant hGA_221–532 proved to be problematic. Most of the expressed protein was insoluble. The small proportion that was soluble rapidly lost activity and precipitated out of solution. By contrast, the recombinant hGA_124–551 protein, which contains 97 additional N-terminal and 19 additional C-terminal amino acids, was more soluble and its activity was stable when stored at 4°C. Thus, this construct was used for the current analysis.

The expression and purification protocols developed by the Structural Genomics Consortium were adopted to express and purify the hGA_124–551 protein. The E. coli cells were grown to a high density and then cooled to 18°C before adding a low level of IPTG. This protocol slowed the rate of synthesis and apparently facilitated the correct folding of the recombinant protein. As a result, a greater percent of the expressed hGA_124–551 was soluble and fully active compared to the previous protocol used to express the rat GA_Δ1 protein¹⁷. The Ni⁺⁺-affinity purification step was very effective and removed nearly all of the contaminating proteins. The step elution protocol produced a very concentrated, but slightly turbid solution of hGA_124–551. The gel filtration chromatography removed the remaining contaminants and the imidazole used to elute the protein from the affinity column. It also diluted the hGA_124–551 concentration to <1.0 mg/mL. In this form, the hGA_124–551 remained soluble and fully active for 2–3 weeks at 4°C. Concentrating the hGA_124–551 to >1.0 mg/mL caused the enzyme to precipitate from solution.

The truncated hGA_124–551 enzyme exhibits kinetic profiles similar to the native glutaminase²¹ and the recombinant rGA_Δ1construct¹⁷. Thus, the C-terminal segments that are unique to the KGA and GAC variants are unlikely to affect the binding of glutamine or impart the phosphate-dependent activation that is characteristic of the kidney-type glutaminase. By contrast, the bacterial glutaminases, which are homologous to the core region, are fully active in the absence of phosphate¹². Thus, the additional 97 N-terminal amino acids found in hGA_124–551 protein may contain a binding site for multivalent anions that imparts the phosphate dependence. It is unknown if this segment folds as a separate domain or if it integrates with the globular structure of the core domain. The addition of DMSO caused a 20–30% decrease in the activity of the hGA_124–551 without affecting the K_m for glutamine (data not shown).

The microtiter plate assay was rapidly initiated and terminated by using an automated multi-channel pipettor (Rainin, Oakland, CA). As a result, it was possible to perform a complete saturation profile with five replicates of each concentration in a single assay. This modification greatly improved the accuracy and precision of the activity assay. Previous studies with the rGA_Δ1 enzyme indicated that BPTES was a potent inhibitor that bound with a K_i of 3 µM¹⁷. However, the kinetic data was not sufficiently accurate to determine if BPTES functioned as a mixed non-competitive or an uncompetitive inhibitor. By contrast, the data reported in the current study firmly establishes that BPTES is an uncompetitive inhibitor. When plotted as a double-reciprocal plot, the data form parallel lines that are characteristic of uncompetitive inhibition. An uncompetitive inhibitor binds only to enzyme-substrate complex and prevents formation of product. To further investigate the mode of inhibition, the K_i for BPTES was calculated using both the apparent K_m and the apparent V_max values derived from the glutamine saturation data. For an uncompetitive inhibitor, the K_i values calculated from the two terms should be identical. Thus, the finding that both calculations yield a K_i of 0.2 µM validates the conclusion that BPTES is a potent uncompetitive inhibitor of glutaminase. The observed inhibition of the core segment suggests that BPTES will effectively inhibit both the KGA and GAC isoforms of human glutaminase.

Phosphate is a known positive allosteric regulator of the mammalian GA²². Both the native enzyme¹⁴ and the rGA_Δ1 construct¹⁷ exhibit sigmoidal phosphate activation profiles. Similarly, hGA_124–551 is essentially inactive in the absence of phosphate, but is activated with increasing concentrations of phosphate. However, the resulting activation profile is nearly hyperbolic. Thus, the C-terminal portion of the GA protein, which is deleted in the hGA_124–551 construct, may contribute to the cooperativity of the phosphate activation profile of the native glutaminase. The addition of BPTES reduced the phosphate activation by decreasing the apparent V_max and greatly increasing the K_0.5 for phosphate. Increasing concentrations of BPTES also increased the cooperativity of the residual phosphate activation profile. This is evidenced by the finding that the Hill coefficient was increased from 1.3 ± 0.1 to 2.2 ± 0.2 in the absence and presence of 10 µM BPTES, respectively.

Gel filtration chromatography was used to estimate the effects of BPTES and of phosphate on the oligomerization of hGA_124–551. The phosphate-dependent activation of the native GA correlates with the conversion of inactive dimers to active tetramers and higher oligomers¹⁶. Previous studies demonstrated that the rGA_Δ1 undergoes a similar phosphate-dependent oligomerization, while BPTES causes the formation of a stable, but inactive tetramer¹⁷. The gel filtration data suggest hGA_124–551 also exists as a dimer in the absence of phosphate. However, the phosphate-dependent activation of hGA_124–551 occurs without formation of higher order structures. This observation is consistent with the observed lack of cooperativity in the phosphate activation profile. By contrast, the addition of BPTES promotes the formation of an apparent tetramer which may contribute to the restoration of the sigmoidal phosphate activation profile of the residual glutaminase activity.

Conclusions

The combined kinetic profiles predict a complex model for the activation and inhibition of hGA_124–551. In the absence of inhibitor, increasing concentrations of phosphate drive the dimeric enzyme into an active conformation without causing the formation of tetramers. The resulting enzyme-phosphate-substrate complex undergoes catalysis to accomplish the hydrolysis of glutamine. BPTES acts as an uncompetitive inhibitor and thus it preferentially binds to the phosphate-activated enzyme-substrate complex and decreases both the V_max and K_m values. The binding of BPTES may cause this complex to form a stable, but inactive tetramer. In addition, the gel filtration data suggest that BPTES also promotes the formation of a tetramer in the absence of glutamine and phosphate. Further insight into the unique mechanism of BPTES inhibition will require the crystallization of the BPTES/hGA_124–551 complex and the determination of its 3-dimensional structure.

Declaration of interest

This research was supported in part by National Institutes of Diabetes and Digestive and Kidney Diseases Grants DK-37124 awarded to N. P. Curthoys.