Polyhydroxybenzoates inhibit ascorbic acid activation of mitochondrial glycerol-3-phosphate dehydrogenase: implications for glucose metabolism and insulin secretion.

Glycerol-3-phosphate dehydrogenase from pig brain mitochondria was stimulated 2.2-fold by the addition of 50 microm l-ascorbic acid. Enzyme activity, dependent upon the presence of l-ascorbic acid, was inhibited by lauryl gallate, propyl gallate, protocatechuic acid ethyl ester, and salicylhydroxamic acid. Homogeneous pig brain mitochondrial glycerol-3-phosphate dehydrogenase was activated by either 150 microm L-ascorbic acid (56%) or 300 microm iron (Fe(2+) or Fe(3+) (62%)) and 2.6-fold by the addition of both L-ascorbic acid and iron. The addition of L-ascorbic acid and iron resulted in a significant increase of k(cat) from 21.1 to 64.1 s(-1), without significantly increasing the K(m) of L-glycerol-3-phosphate (10.0-14.5 mm). The activation of pure glycerol-3-phosphate dehydrogenase by either L-ascorbic acid or iron or its combination could be totally inhibited by 200 microm propyl gallate. The metabolism of [5-(3)H]glucose and the glucose-stimulated insulin secretion from rat insulinoma cells, INS-1, were effectively inhibited by 500 microm or 1 mm propyl gallate and to a lesser extent by 5 mm aminooxyacetate, a potent malate-aspartate shuttle inhibitor. The combined data support the conclusion that l-ascorbic acid is a physiological activator of mitochondrial glycerol-3-phosphate dehydrogenase, that the enzyme is potently inhibited by agents that specifically inhibit certain classes of di-iron metalloenzymes, and that the enzyme is chiefly responsible for the proximal signal events in INS-1 cell glucose-stimulated insulin release.

Mitochondrial glycerol-3-phosphate dehydrogenase (mG-PDH) 1 (EC 1.1.99.5) plays a critical role in the shuttle of glycolytically generated reducing equivalents into mitochondrial electron transport and oxidative phosphorylation in numerous tissues (1)(2)(3)(4). In pancreatic islet ␤ cells, many studies support the significant participation of mGPDH and the L-␣-glycerol-3-phosphate shuttle in the proximal events that signal the release of insulin in response to increased glucose (5)(6)(7)(8)(9)(10)(11)(12). Recently, particular importance has been attributed to the role of NADH in the glucose-induced activation of mitochondrial metabolism and insulin secretion (13,14). These studies emphasized the essential roles played by both the glycerol-3-phosphate and the malate-aspartate shuttles in modulating the cytosolic NADH pool.
In confirmation of the original observations of Sigal and King that scorbutic guinea pigs demonstrated abnormal glucose tolerance (15), L-ascorbic acid was shown to be essential for the release of insulin from scorbutic guinea pig pancreatic islets (16,17). Further studies have demonstrated that L-ascorbic acid is an essential cofactor for the activation of mGPDH (oxidase) in mitochondria from guinea pig tissues and rat liver (18). In purified preparations of mGPDH from a variety of sources, both iron and acid extractable sulfur have been reported (2, 19 -21), suggesting that an iron/sulfur center is involved in the catalytic mechanism of this mitochondrial inner membrane bound enzyme. No further evidence, however, was obtained to support this suggestion, and the iron center of mGPDH has remained uncharacterized.
In addition to the need to clarify the iron/L-ascorbic acid relationship of mGPDH in intact mitochondria, it was essential to examine the effect of L-ascorbic acid, iron and specific di-iron metalloenzyme inhibitors on homogeneous mGPDH. In the present study, the effects of propyl gallate and other related polyhydroxybenzoate inhibitors on pig brain mGPDH in intact mitochondria and on preparations of pure mGPDH were examined. Because of the potential role of mGPDH for shuttling reducing equivalents into the mitochondria during glucoseinduced insulin release from pancreatic ␤ cells, we also investigated the effects of a di-iron metalloenzyme inhibitor, propyl gallate, on glucose usage and glucose-induced insulin release from the rat insulinoma cell line, INS-1.

Materials
DL-␣-Glycerol phosphate, mannitol, MOPS, NaN 3 , Triton X-100, SHAM, PCAEE, DEAE-Sepharose (fast flow), and Sephacryl S-300 (fast flow) were purchased from Sigma. Sucrose, HEPES, Tris-base, glutathione, and NADH were from Roche Molecular Biochemicals. L-Ascorbic acid and glycerol were from J. T. Baker, Inc. L-ascorbic acid-2-phosphate (Mg 2ϩ ) was purchased from Wako Pure Chemical Industries, Ltd. Menadione was a product of Nutritional Biochemicals Corp. Lauryl gallate, propyl gallate, and 2-(4-iodophenyl)-3-(4-nitrophenyl)-5-phenyltetrazolium chloride were purchased from Aldrich. Hydroxylapatite, 40% acrylamide/bis solution, and low molecular weight protein standards were obtained from Bio-Rad. Succinic acid was a product of Mallinckrodt Chemical Works.  H]D-Glucose (10 -20 Ci/mmol) was purchased from PerkinElmer Life Sciences. Rat insulin antibodies and rat insulin standards were purchased from Linco (St. Charles, MO). All other chemicals used were of A.C.S. reagent grade. Pig brains were kindly provided by Thomas Fortan from Michigan State University Meats laboratory. Cerebra were removed shortly after slaughter, chilled on ice, washed in 0.25 M sucrose to remove excess blood, and frozen (Ϫ70°C) for storage.

Methods
Isolation of Mitochondria-Frozen pig cerebrum was homogenized in an all glass Dounce homogenizer with a 10ϫ volume of homogenizing fluid described by Greenawalt (22) consisting of 200 mM mannitol, 80 mM sucrose, and 10 mM potassium HEPES, pH 7.4. The mitochondrial fraction was isolated by differential centrifugation following the sedimentation protocol reported by Lai and Clark (23). Briefly, homogenization was performed on ice with eight thrusts of the Dounce plunger. All subsequent steps were conducted on ice or in a refrigerated centrifuge (Sorval-RC2-B) at 4°C. The homogenate was first centrifuged at 1,000 ϫ g for 10 min, and the supernatant was removed. The supernatant fraction was then centrifuged at 15,000 ϫ g for 10 min. The upper layer was decanted off, and the pellet was gently rinsed with small amounts of homogenizing fluid to remove broken mitochondrial remnants. The pellet was gently resuspended in 5 volumes of homogenizing fluid using a Dounce homogenizer and centrifuged at 12,000 ϫ g for 10 min. The supernatant was discarded, and the pellet was gently washed with homogenizing fluid to remove light weight fluffy mitochondrial fragments. The pellet was resuspended in 2.5 volumes of homogenizing fluid in a Dounce homogenizer, and the suspension was centrifuged at 12,000 ϫ g for 10 min. The supernatant was decanted off, and the pellet was resuspended in ϳ1 ml of 250 mM sucrose, 10 mM Tris-HCl, pH 7.5, per g original tissue weight.
Electron Transport Particles-Mitochondria were isolated as described above and placed in a freezer at Ϫ20°C until needed. ETPs were prepared according the method of Green and Ziegler (24). The thawed mitochondrial preparations were centrifuged at 12,000 ϫ g for 10 min, and the pellet was resuspended to a concentration of ϳ20 mg protein/ml in a solution of 250 mM sucrose, 10 mM Tris-HCl, pH 7.5, 15 mM MgCl 2 , and 1 mM ATP at 4°C. The suspension was sonicated by a Branson Sonifier 450 with 3 ϫ 30 s bursts of energy (40 watts). The suspension was centrifuged at 1000 ϫ g for 10 min, and the supernatant fraction was further centrifuged at 100,000 ϫ g for 40 min in a Spinco Ultracentrifuge 40K rotor. The pellet was resuspended in 250 mM sucrose, 10 mM Tris-HCl, pH 7.5, containing 10 mM MgCl 2 and 1 mM ATP. The suspended ETPs at ϳ5-6 mg protein/ml were stored at Ϫ20°C until assayed. Protein concentrations of mitochondria or ETP preparations were determined by the BCA protein assay protocol according to the manufacturer's direction (Pierce) using bovine serum albumin as standard.
Mitochondrial Enzyme Assays-mGPDH activity in intact mitochondria was measured by oxygen uptake using a Clark oxygen electrode. mGPDH from freshly prepared mitochondrial suspensions of 5-7 g protein/ml in 250 mM sucrose, Tris-HCl, pH 7.5, from frozen pig brain was assayed as described previously (18). The Clark oxygen electrode chamber of 3.0 ml (model 53 oxygen monitor, Yellow Springs Instrument Co. Yellow Springs, Ohio) contained 250 mM sucrose, 10 mM Tris-HCl, pH 7.5, 1 mM GSH, 0.5-0.8 mg of mitochondrial protein, with or without 50 M L-ascorbic acid. Typically, oxygen uptake was measured for 3-6 min followed by the addition of 50 mM DL-glycerol-3phosphate, pH 7.5, at 37°C. The rate of oxygen uptake was measured with a Linseis recorder, and the rate, in nmol oxygen uptake/min, was corrected for the basal oxygen uptake rate before addition of the substrate. Values were expressed as nmol oxygen uptake/min/mg protein. Stock solutions of the inhibitor compounds were made up in absolute ethanol. When added to the assay mixture, ethanol aliquots of 30 l or less were preincubated with the mitochondria for 3 min followed by the sequential addition of 50 M L-ascorbic acid (3 min of incubation) and 50 mM DL-glycerol-3-phosphate. Controls of up to 30 l of ethanol in the 3.0-ml reaction volume had no effect on the oxygen uptake rates (data not shown).
NADH dehydrogenase (oxidase) from pig brain ETPs was assayed spectrophotometrically by a modification of the method of Singer (25). The reaction mixture contained 100 mM potassium phosphate, pH 7.4, 1 mM GSH, 0.26 mM NADH, 0.1-0.12 mg ETP protein, with or without 3 min of preincubation with various levels of inhibitors or 50 M Lascorbic acid at 30°C in a total of 500 l. The reaction was initiated by the addition of NADH, and the decrease in absorbance at 340 nm was recorded in a Gilford Response II recording spectrophotometer. NADH dehydrogenase activity (nmol/min/mg protein) were calculated from the extinction coefficient of NADH of 6.22 ϫ 10 3 M Ϫ1 cm Ϫ1 . Because 5 l of ethanol were without effect on enzyme activity, all additions of inhibitors were made with less than 5 l of ethanol.
Succinoxidase activity in ETP preparations was assayed by a modification of the method of Green and Ziegler (24). The rate of succinate oxidation at 37°C was measured by oxygen uptake analysis using a Clark oxygen electrode. The 3.0-ml reaction chamber contained 100 mM potassium phosphate, pH 7.4, 1 mM GSH, 0.3 mg of ETP protein, 50 mM potassium succinate, and with or without various concentrations of inhibitors or 50 M L-ascorbic acid. When inhibitors or L-ascorbic acid were included, ETPs were preincubated for 3 min. Ethanol, up to 30 l/3.0 ml reaction volume, did not affect the succinoxidase activity.
mGPDH Purification-All steps in the purification of mGPDH were monitored for activity spectrophotometrically by the method of Garrib and McMurray (26). The assay cuvettes contained in a final volume of 0.5 ml, 90 mM potassium phosphate, pH 7.5, 1 mM KCN, 3 mM INT, 0.6 mM menadione, and various enzyme fractions. The reaction was initiated by the addition of 30 mM DL-glycerol-3-phosphate and was monitored at 500 nm. The extinction coefficient of reduced INT was taken as 11.5 ϫ 10 3 M Ϫ1 cm Ϫ1 at 500 nm (26). Rates were expressed as nmol/min, and protein was determined by the bicinchoninic acid protocol described above.
The purification of mGPDH was a modification of the method of Garrib and McMurray (27). In a typical preparation, 65-100 g of frozen pig brain were thawed and homogenized at 4°C in 650 -1000 ml of homogenizing buffer containing 200 mM mannitol, 80 mM sucrose, 0.2 mM L-ascorbic acid, 1 mM DTT, and 10 mM HEPES, pH 7.4, in a Waring blender for 1 min at low speed. The suspension was centrifuged at 1000 ϫ g for 10 min in a Sorval RC2-B refrigerated centrifuge at 4°C. The supernatant was stored on ice or kept at 4°C for all subsequent steps, and the pellet was resuspended in 300 ml of homogenizing buffer and homogenized with an all glass Dounce homogenizer using five strokes of the pestle. The homogenate was diluted with 300 ml of homogenizing buffer, stirred for 5 min, and then centrifuged at 1000 ϫ g for 10 min. The supernatant fractions of the previous steps were combined and centrifuged at 15,000 ϫ g for 10 min. The supernatant was discarded, and the pellets were resuspended in 300 ml of 0.15 M KCl containing 1 mM DTT and centrifuged at 15,000 ϫ g for 10 min. The pellets were resuspended in 40 ml of 10 mM Tris-HCl, pH 7.5, containing 250 mM sucrose followed by dilution 10-fold with 1 mM DTT. The slurry was stirred for 60 min on ice, and the diluted suspension was centrifuged at 15,000 ϫ g for 15 min. The pellet was washed with 300 ml of 50 mM potassium phosphate, pH 7.5, 0.1 mM DTT, and 0.02% NaN 3 followed by centrifugation at 15,000 ϫ g for 15 min. The pellets were resuspended in 50 mM potassium phosphate, 0.1 mM DTT, and 0.02% NaN 3 , and the protein content was determined by the BCA method described previously. The suspension was stored at Ϫ70°C until required.
DEAE-Sepharose Chromatography-Washed and stored mitochondria were thawed at 37°C, chilled, and kept at 4°C throughout each subsequent purification step. The mGPDH was solubilized by addition of appropriate amounts of 10% Triton X-100 (Triton X-100/protein ϭ 0.25 mg/mg protein) and was stirred for 60 min. The suspension was centrifuged at 60,000 ϫ g for 30 min, and the supernatant fraction was saved for further purification. A 2.5 ϫ 15 cm column of DEAE-Sepharose (fast flow) was equilibrated with buffer A consisting of 50 mM potassium phosphate, pH 7.5, 0.1% Triton X-100, 0.02% NaN 3 , 0.1 mM DTT, and 10% glycerol. After loading the sample, the column was washed with buffer A, and the enzyme was eluted with a linear NaCl gradient consisting of 150 ml of buffer A and 150 ml of buffer A containing 0.6 M NaCl at a flow rate of 1 ml/min. The fractions with enzyme activity were pooled and diluted with triple distilled water to 10 mM potassium phosphate, pH 7.5, 0.1 mM DTT, 0.1% Triton X-100, 2% glycerol, and 0.02% NaN 3 .
Hydroxylapatite Chromatography-A column of hydroxylapatite (2.5 ϫ 6.0 or 12.0 cm) was equilibrated with buffer B consisting of 10 mM potassium phosphate, pH 7.5, 0.1 mM DTT, 0.1% Triton X-100, 100 mM NaCl, 5% glycerol, and 0.02% NaN 3 . The diluted enzyme sample from the DEAE column was applied to the hydroxylapatite and washed with buffer B. The enzyme was eluted by a linear sodium phosphate gradient consisting of 50 ml of buffer B and 50 ml of buffer B with 150 mM sodium phosphate, pH 7.5. The enzyme containing fractions were subjected to LDS-PAGE to monitor purity. The fractions with highest activity and purity were pooled and concentrated to 1.5 ml by an Amicon Centiprep YM-10 apparatus.
Sephacryl S-300-HR Chromatography-The concentrated enzyme fraction was applied to a Sephacryl S-300 (fast flow) column (2.5 ϫ 92 cm) equilibrated with buffer C (50 mM potassium phosphate, pH 7.5, 0.05% Triton X-100, 0.02% NaN 3 , 0.1 mM DTT, 20% glycerol, and 0.2 M NaCl at a flow rate of 0.25 ml/min. The fractions containing mGPDH activity were pooled, concentrated as described above, and subjected to LDS-PAGE analysis to assess purity. LDS-PAGE-LDS-PAGE was performed on samples of purified mG-PDH by a modification (28) of the basic method of Laemmli (29). The stacking and separating gels (0.75 mm) had polyacrylamide concentrations of 5 and 8%, respectively. The gels were run at a constant 100V in a Bio-Rad minigel apparatus, and the gel was stained with Coomassie Brilliant Blue R-250.
Homogeneous mGPDH Assay-For measurement of the mGPDH specific activity and kinetic properties, preparations of the pure enzyme were assayed by oxygen uptake using the Clark electrode following a modification of the assay described by Beleznai et al. (30). Homogeneous mGPDH (2-5 g), stabilized in a solution of 50 mM HEPES, pH 7.5, 0.1 mM DTT, 0.05% Triton X-100, 0.02% NaN 3 , and 20% glycerol, was placed in a 3.0-ml chamber at 37°C containing 100 mM potassium phosphate, pH 7.6, 0.6 mM GSH, 0.05% Triton X-100, and 0.15 mM menadione (freshly prepared as a 10 mM stock solution in absolute ethanol) as electron acceptor. When L-ascorbic acid was added, stock solutions of 10 mM crystalline L-ascorbic acid were freshly prepared in Chelex treated triple distilled water and stored on ice. For the addition of Fe 2ϩ or Fe 3ϩ , freshly prepared stock solutions of 10 mM FeSO 4 or FeCl 3 were added prior to the preincubation period. After a 2-min preincubation, followed by a period (3-5 min) of linear recording for blank oxygen consumption, rates for mGPDH were initiated by the addition of 75 mM DL-glycerol-3-phosphate (stock solution of 1.5 M substrate dissolved in 30 mM potassium phosphate, pH 7.6). In the absence of enzyme, the addition of substrate caused no change in background oxygen uptake rate. For inhibition analysis, stock solutions of 60 mM propyl gallate in absolute ethanol were stored at Ϫ20°C. When propyl gallate was included, it was added to the incubation mixture at 37°C for 2 min prior to measurement of control oxygen uptake rates and those following addition of DL-glycerol-3-phosphate.
Enzyme Kinetics-Oxygen uptake measurements were carried out as described above using varying concentrations of L-glycerol-3-phosphate (as the DL mixture) with 4.4 g of protein in a 3.0-ml oxygen electrode chamber. GSH (0.6 mM), which had no effect on the oxygen uptake reaction rates, was added to each sample to maintain the L-ascorbic acid, when added, at the level designated, 150 M. When iron was added to the reaction mixture, it was provided as 150 -300 M FeSO 4 or 300 M FeCl 3 . Reaction rates after substrate addition were corrected by rates measured in the absence of substrate. Other controls were run in the absence of enzyme when appropriate. Values were expressed as nmol/min and kinetic constants, K m(app) and V max(app) , were determined using the advanced kinetics software provided by Gilford for the Re-sponse II spectrophotometer. The values for k cat were calculated by dividing V max(app) by the molar concentration of pig brain mGPDH with the molecular weight taken as 75,000 (21). Insulin Secretion Studies-For static secretion studies, INS-1 cells were plated onto 12-well plates at a density of 1.5 ϫ 10 6 cells/well in RPMI 1640 medium plus supplements and grown for 24 h. The growth medium was then changed to RPMI 1640 containing 4 mM glucose plus the supplements described above, and cells were cultured for an additional 30 h. Cells were then incubated for 60 min at 37°C in Krebs Ringer bicarbonate buffer (KRB buffer) (118.5 mM NaCl, 2.54 mM CaCl 2 , 1.19 mM KH 2 PO 4 , 4.74 mM KCl, 25 mM NaHCO 3 , 1.19 mM MgCl 2 , 10 mM HEPES, pH 7.4, 0.1% bovine serum albumin) containing 4.0 mM glucose. Cells were then incubated for 20 min at 37°C in KRB buffer containing 100 M 3-isobutyl-1-methylxanthine and either 4 or 16.7 mM glucose in the absence or presence of metabolic inhibitors as indicated in the figure legends. Propyl gallate was dissolved in absolute ethanol, and therefore all controls contained equal amounts of ethanol. Concentrations of insulin released into the medium were determined by insulin enzyme-linked immunosorbent assay using a modification of the procedure described by Kekow et al. (32). Insulin released into the medium was normalized to cellular protein concentrations determined according to Lowry et al. (33).

INS-1 Cell
Glucose Utilization Studies-Glucose usage was measured using a modification of the method of Zawalich and Matschinsky (34,35). INS-1 cells were plated onto 12-well plates at a density of 1.5 ϫ 10 6 cells/well and grown for 24 h. The growth medium was then changed to RPMI 1640 medium containing 4 mM glucose and supplements described above and incubated for an additional 30 h. Cells were then incubated for 60 min at 37°C in KRB buffer containing 4.0 mM glucose. Glucose and the scintillation vials were sealed tightly and incubated at 50°C for 18 h. After cooling, the Eppendorf tubes were removed, 10 ml of Safety-Solve scintillation mixture were added to the vials, and the samples were counted in a Beckman scintillation counter. Glucose utilization was then determined from the following formula and expressed as picomoles of glucose metabolized per min per mg protein. The equilibration coefficient (EQC) was determined with 3 [H]H 2 O following the procedure outlined above. Glucose usage ϭ (dpm Ϫ blank)/(specific activity ϫ EQC ϫ min).

RESULTS
Pig Brain mGPDH-Pig brain mGPDH activity in isolated intact mitochondria was 28.6 Ϯ 6.6 nmol/min/mg protein (n ϭ 11). Addition of 50 M L-ascorbic acid increase mGPDH activity to 62.9 Ϯ 10.7 nmol/min/mg protein (n ϭ 11). These activities were completely inhibited by 10 mM KCN (data not shown), indicating that a functional cytochrome c oxidase was required to complete the reaction with oxygen whether L-ascorbic acid was present or not. SHAM, PCAEE, propyl gallate, and lauryl gallate were potent inhibitors of pig brain mGPDH activity in intact mitochondria stimulated by L-ascorbic acid (Fig. 1). In contrast, these agents were without effect on the basal activity, i.e. activity in the absence of L-ascorbic acid (data not shown). The concentration of each compound calculated to cause 50% inhibition of the L-ascorbic acid stimulated activities were: SHAM, 27.7 Ϯ 6.9 M; PCAEE, 585 Ϯ 203 nM; propyl gallate, 305 Ϯ 113 nM; and lauryl gallate, 111 Ϯ 42 nM.
NADH dehydrogenase and succinoxidase activities from pig brain ETPs were compared in the presence and absence of the four hydroxybenzoic acid derivatives. Minimal concentrations of each agent, previously found to completely inhibit the Lascorbic acid-stimulated mGPDH activity, had no effect on these two well established iron/sulfur enzymes (data not shown). In addition, L-ascorbic acid had no stimulatory effect on the activity of either NADH dehydrogenase or succinoxidase (data not shown).
Purification of mGPDH-The purification of mGPDH from pig brain was accomplished by a series of steps including Triton X-100 extraction of washed mitochondria, DEAE-Sepharose chromatography, Bio-Gel HT hydroxylapatite chromatography, and Sephacryl S-300 gel chromatography (Table I). mGPDH was judged to be homogeneous based on LDS polyacrylamide gel electrophoresis (Fig. 2) with a molecular weight of 75,000. This value is in good agreement with that reported by Cottingham and Ragan (21).
mGPDH Activity-The basal activity of purified mGPDH was 14.0 Ϯ 2.2 mol/min/mg protein (Table II). This activity could be stimulated 56% to 21.9 Ϯ 3.3 mol/min/mg (p ϭ Ͻ0.001) by the addition of 150 M L-ascorbic acid. The addition of 300 M Fe 2ϩ (optimal level based on dose-response curve; data not shown) also activated the purified enzyme by 62% to a value of 22.7 Ϯ 4.2 mol/min/mg (p ϭ Ͻ0.005), suggesting that there was a partial loss of iron from the enzyme binding centers during the purification steps. The addition of both L-ascorbic acid and Fe 2ϩ resulted in a 2.6-fold increase in activity over that of the control (36.7 Ϯ 3.0 mol/min/mg, p ϭ Ͻ0.001). Preincubation of the enzyme with 200 M propyl gallate caused a decrease in activity from 14.0 to 9.1 Ϯ 1.6 mol/min/mg (p ϭ Ͻ0.005), which we attribute to the complete blockage of the iron pathway of electron transport between substrate and the acceptor, menadione. From these data, it can be estimated that 35% of the basal activity utilized the iron-mediated pathway, whereas 65% followed the FAD/FADH 2 linked pathway to menadione. In contrast, in the presence of 150 M L-ascorbic acid, 52% of the substrate oxidation passed along the iron-dependent pathway, whereas 48% followed the FAD/FADH 2 -dependent pathway. The difference between control samples with inhibitor and L-ascorbic acid supplemented samples with inhibitor (9.1 versus 11.4 mol/min/mg) were not significantly different by statistical analysis. Propyl gallate also completely inhibited  2. LDS-PAGE analysis of pig brain mitochondrial glycerol-3-phosphate dehydrogenase. Lane 1, Bio-Rad low molecular weight standards; lane 2, 5.6 g of protein taken from the reactive fraction isolated via the Sephacryl S-300 gel chromatography purification step. The band indicated by an arrow migrated with an estimated molecular weight of 75,000. The gel was stained with Coomassie Brilliant Blue R-250.  Table III. In these studies, the values for basal control are compared with those obtained for enzyme supplemented with 150 M L-ascorbic acid and 150 M Fe 2ϩ . The K m(app) for basal enzyme was 10.0 Ϯ 1.2 mM L-glycerol-3phosphate (in good agreement with reference number 3), whereas that for the stimulated enzyme was 14.5 Ϯ 4.9 mM (not statistically different). L-Ascorbic acid and Fe 2ϩ supplementation increased the V max(app) from 56.9 Ϯ 13.2 nmol/min to 161.9 Ϯ 24,7 nm/min (p ϭ Ͻ0.0001). Expressed as k cat values, the controls were 21.1 Ϯ 9.

Effect of Propyl Gallate on Glucose Utilization and Insulin
Secretion from INS-1 Cells-In pancreatic ␤ cells, it is well established that there is a relationship between glucose metabolism through mGPDH and insulin release (5-9, 11, 12, 36). Therefore, agents that effectively inhibit L-ascorbic acid-induced mGPDH activity may have profound effects on glucose utilization and insulin release from pancreatic ␤ cells. To examine this possibility, the effects of propyl gallate on glucose utilization and insulin secretion from INS-1 cells were determined.
Incubation of INS-1 cells in 16.7 mM glucose led to a 4.83 Ϯ 0.11-fold (n ϭ 4) increase in the rate of conversion of 5-[ 3 H]glucose to [ 3 H]H 2 O compared with cells incubated in 4.0 mM glucose (Fig. 3). The addition of 500 M propyl gallate led to a 23.64 Ϯ 0.57% (n ϭ 4, p Ͻ 0.006) and 13.61 Ϯ 3.42% (n ϭ 4, p Ͻ 0.008) reduction in glucose usage in cells incubated in 4 or 16.7 mM glucose, respectively. The addition of 1 mM propyl gallate led to a 75.20 Ϯ 2.32% (n ϭ 4, p Ͻ 0.0001) and 80.75 Ϯ 1.23% (n ϭ 4, p Ͻ 0.0001) reduction in cells incubated in 4 or 16.7 mM glucose, respectively. Treatment of cells with propyl gallate concentrations lower than 250 M had no significant effects on glucose usage in cells incubated in 4.0 or 16.7 mM glucose. Next the ability of aminooxyacetate (AOA), an inhibitor of aspartate aminotransferases in the malate-aspartate shuttle (37), to potentiate the propyl gallate-mediated inhibition of glucose usage was determined. Incubation of cells in 5 mM AOA led to a 27.08 Ϯ 1.47% (n ϭ 4, p Ͻ 0.004) and 23.77 Ϯ 3.10% (n ϭ 4, p Ͻ 0.005) reduction in glucose utilization in cells cultured in 4.0 mM or 16.7 mM glucose, respectively. Combined treatment of cells with 5 mM AOA and 500 M propyl gallate led to a further reduction in glucose usage in cells incubated in 16.7 mM glucose. Nevertheless, combined treatment of 5 mM AOA and 1 mM propyl gallate were not able to reduce glucose usage below that observed with 1 mM propyl gallate alone.
Treatment of cells with 250 M propyl gallate had no significant effects on the ability of 16.7 mM to induce insulin release. Next the ability of AOA to potentiate the propyl gallate-mediated inhibition of insulin release was determined. Incubation of cells in 5 mM AOA led to a 36.62 Ϯ 3.35% (n ϭ 4, p Ͻ 0.03) and 58.75 Ϯ 3.78% (n ϭ 4, p Ͻ 0.001) reduction in insulin release from cells cultured in 4.0 or 16.7 mM glucose, respectively. Combined treatment of cells with 5 mM AOA and 500 M or 1 mM propyl gallate were not able to further reduce insulin release from cells incubated in either 4.0 or 16.7 mM glucose when compared with cells incubated with 500 M or 1 mM propyl gallate alone. DISCUSSION mGPDH of mammalian origin has been purified to homogeneity by other laboratories with results, suggesting that the inner mitochondrial membrane bound enzyme contains iron and acid releasable sulfur (2, 19 -21). However, the amount of iron found after purification ranged between 1 mol of iron/ 100,000 -350,000 ϫ g of enzyme protein. Because pig brain mGPDH is a monomer of 75 kDa, these iron level estimations suggested that significant loss of iron occurred during the pu- rifications previously reported. In agreement with this suggestion, the activity of our pure enzyme preparations was stimulated by the addition of iron. It is difficult to account for the origin of the acid releasable sulfur from homogeneous mGPDH reported by earlier workers (2, 3, 19 -21). Little further evidence has been published to support the existence of a typical iron/sulfur center (38,39). Moreover, NADH dehydrogenase and succinoxidase, key enzymes in mitochondrial electron transport and well established iron/sulfur metalloenzymes (4), were not activated by L-ascorbic acid nor inhibited by propyl gallate or its homologues in the present study.
The finding of potent inhibition of the L-ascorbic acid-stimulated activity of mGPDH by propyl gallate and related compounds in intact mitochondria or that inhibition by propyl gallate of combined iron and L-ascorbic acid activation of homogeneous mGPDH is consistent with results expected for a di-iron metalloenzyme (40). For example, Schonbaum et al. (41) reported that substituted benzohydroxamic acids specifically inhibited the cyanide-insensitive "alternative oxidase" electron transport pathway in isolated plant mitochondria. Siedow et al. (42) identified conserved amino acids, including two copies of the iron-binding motif, EXXH, in the C-terminal domain of the alternative oxidase that suggested the presence of a hydroxobridged binuclear iron center. Recently, mGPDH was cloned and sequenced from rat (9), human (43), and Drosophila melanogaster (44) sources. Inspection of the sequences of the mammalian enzymes revealed the presence of conserved sequences at 139EAL142H and another sequence at 360DVY363H (D instead of E) that could function as a di-iron binding site (9,43). Future studies are needed to elucidate the possible site or sites of iron binding in mGPDH and the action of L-ascorbic acid in maximizing the flow of electrons into the ubiquinone pool.
Trypanosomes, isolated from the blood of hosts, exhibit a SHAM-sensitive, cyanide-insensitive terminal oxidase that utilizes glycerol-3-phosphate as substrate by participating in a NAD ϩ regenerating cycle (45). This glycerol phosphate oxidase system has been the subject of studies aimed at designing drugs, such as homologues of SHAM and the inhibitors selected in the present work, that specifically kill trypanosomes in vivo without producing severe toxic side effects in the host (46 -49). The metal chelation properties of these drugs have been implicated in their inhibitory action (46). Alternatively, the effect of hydroxamic acids on the glycerol phosphate oxidase system of trypanosomes was proposed to be due to their ability to competitively displace ubiquinol, the putative electron carrier, from the dehydrogenase to the terminal oxidase of the glycerol phosphate oxidase complex (50). Whether either of these two unique systems, plant alternative oxidase and trypanosome terminal oxidase, are activated by L-ascorbic acid remains to be explored.
If mGPDH is indeed an essential enzyme in the proximal events linked to glucose-stimulated glucose metabolism and insulin release as previously reported (5,8,11,51), specific binuclear iron inhibitors such as propyl gallate should significantly inhibit the process. This hypothesis was tested in the INS-1 cell line because they have been previously reported to have high levels of mGPDH activity similar to that observed in rat pancreatic ␤ cells (12,53). Incubation of INS-1 cells with propyl gallate at concentrations of 500 M and 1 mM effectively reduced [5-3 H]D-glucose metabolism and glucose-induced insulin release. The ability of propyl gallate to reduce glucose utilization is consistent with the hypothesis that mGPDH inhibition would reduce the reoxidation of cytosolic NADH and thereby inhibit glycolysis at the level of triose phosphates. Blockage of mGPDH would also cause a reduction in shuttling of cytosolic NADH generated from glycolysis and pyruvate into the mitochondria, thus leading to an overall reduction in ATP generation and thereby markedly reducing insulin release. Under similar culture conditions, 5 mM AOA was only partially effective in inhibiting glucose utilization and led to a 50% reduction in glucose-induced insulin release. The ability of AOA to inhibit glucose-induced insulin release in INS-1 cells is consistent with previous reports showing that millimolar concentrations of AOA inhibit insulin secretion from rat islets by 50% (37,54). Our results suggest that in INS-1 cells the glycerol phosphate shuttle is more active than the malate-aspartate shuttle in the regeneration of NAD ϩ consumed during glycolysis because propyl gallate is more effective than AOA at suppressing both glucose utilization and glucose-induced insulin release. This conclusion, however, directly contradicts Ishihara et al. (52) results that suggest that the malate-aspartate shuttle is more active than the glycerol phosphate shuttle in INS-1 cells. Our observation also contradict results from Eto et al. (14) demonstrating that in mouse islets glucose-induced insulin release is only markedly suppressed when activities of both the glycerol phosphate and malate/aspartate shuttles are impaired. Nevertheless, our results are in general agreement with those of Sekine et al. (12), Dukes et al. (13), and Eto et al. (14) regarding the importance of cytosolic NADH and its subsequent oxidation through the glycerol-3-phosphate and malate-aspartate shuttles.
Our previous studies have shown L-ascorbic acid is essential for insulin release from scorbutic guinea pig islets (16,17) and that L-ascorbic acid serves as an essential cofactor for mGPDH (oxidase) activation in mitochondria isolated from guinea pig tissues and rat liver (18). However, the requirement for ascorbic acid in glucose-induced insulin release from INS-1 cells has been difficult to directly access. In unpublished studies, we have established that INS-1 cells are capable of dephosphorylating exogenously added ascorbic acid 2-phosphate, thus releasing ample amounts of vitamin C. Addition of 1 mM ascorbic acid 2-phosphate, however, did not enhance glucose utilization or glucose-induced insulin release from INS-1 cells. 2 The in- ability of exogenously added ascorbic acid 2-phosphate to stimulate both glucose utilization and insulin release is most likely due to the difficulty of establishing scorbutic INS-1 cells because these cells are capable of scavenging trace amounts of ascorbic acid present in our commercial source of fetal bovine serum. 2 Nevertheless, the ability of propyl gallate to block ascorbic acid-activated mGPDH activity in vitro and to reduce both glucose utilization and insulin release from INS-1 cells suggests that ascorbic acid plays an essential role in glucoseinduced insulin release.
Overall we conclude that mGPDH is activated by L-ascorbic acid via a potential di-iron reactive center and is effectively inhibited by propyl gallate and other polyhydroxybenzoic acid derivates. Furthermore, the glycerol phosphate shuttle, in comparison with the malate-aspartate shuttle, is crucial to the release of insulin from INS-1 cells in response to elevated glucose levels. Use of propyl gallate and other related polyhydroxybenzoate inhibitors may serve as effective tools for studying the role of mGPDH in glucose-induced insulin secretion.