The Glycolytic Enzyme Glyceraldehyde-3-Phosphate Dehydrogenase Works as an Arsenate Reductase in Human Red Blood Cells and Rat Liver Cytosol
http://www.100md.com
《毒物学科学杂志》
Department of Pharmacology and Pharmacotherapy, Toxicology Section, University of Pécs, Medical School, Pécs, Hungary
ABSTRACT
The mammalian enzymes responsible for reduction of the environmentally prevalent arsenate (AsV) to the much more toxic arsenite (AsIII) are unknown. In the previous paper (Németi and Gregus, 2005), we proposed that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and/or phosphoglycerate kinase (PGK) may catalyze reduction of AsV in human red blood cells (RBC), hemolysate, or rat liver cytosol. In testing this hypothesis, we show here that, if supplied with glutathione (GSH), NAD, and glycolytic substrate, the mixture of purified GAPDH and PGK indeed catalyzes the reduction of AsV. Further analysis revealed that GAPDH is endowed with AsV reductase activity, whereas PGK serves as an auxiliary enzyme, when 3-phosphoglycerate is the glycolytic substrate. The GAPDH-catalyzed AsV reduction required GSH, NAD, and glyceraldehyde-3-phosphate. ADP and ATP moderately, whereas NADH strongly inhibited the AsV reductase activity of the enzyme even in the presence of NAD. Koningic acid (KA), a specific and irreversible inhibitor of GAPDH, inhibited both the classical enzymatic and the AsV-reducing activities of the enzyme in a concentration-dependent fashion. To assess the contribution of GAPDH to the reduction of AsV carried out by hemolysate, rat liver cytosol, or intact erythrocytes, we determined the concentration-dependent effect of KA on AsV reduction by these cells and extracts. Inactivation of GAPDH by KA abolished AsV reduction in intact RBC as well as in the hemolysate and the liver cytosol, when GAPDH in the latter extracts was abundantly supplied with exogenous NAD and glycolytic substrate. However, despite complete inactivation of GAPDH by KA, the hepatic cytosol exhibited significant residual AsV-reducing activity in the absence of exogenous NAD and glycolytic substrate, suggesting that besides GAPDH, other cytosolic enzyme(s) may contribute to AsV reduction in the liver. In conclusion, the key glycolytic enzyme GAPDH can fortuitously catalyze the reduction of AsV to AsIII, if GSH, NAD, and glycolytic substrate are available. AsV reduction may take place during, or as a consequence of, the arsenolytic cleavage of the thioester bond formed between the enzyme's Cys149 and the 3-phosphoglyceroyl moiety of the substrate. Although GAPDH is exclusively responsible for reduction of AsV in human erythrocytes, its role in AsV reduction in vivo remains to be determined.
Key Words: arsenate; glyceraldehyde-3-phosphate dehydrogenase; koningic acid; glutathione; NAD; reduction.
INTRODUCTION
The environmentally prevalent arsenical, arsenate (AsV), is readily reduced to arsenite (AsIII) in living organisms (Thomas et al., 2001; Vahter, 1983). This process is considered as toxification, because AsIII is much more toxic, owing to its facile covalent reactivity with thiols, especially dithiols (Knowles and Benson, 1983). Despite its toxicological significance, the biochemical mechanism of AsV reduction has not been clarified in mammals. Rat liver mitochondria rapidly convert AsV to AsIII (Németi and Gregus, 2002); however, their contribution to AsV reduction in rats has not been proven. Purine nucleoside phosphorylase (PNP) also reduces AsV to AsIII, if its substrate (e.g., inosine) and an appropriate dithiol (e.g., dithiothreitol) are present (Gregus and Németi, 2002; Radabaugh et al., 2002), suggesting that this ubiquitous enzyme may play a role in AsV reduction in vivo. However, the testing of PNP for such a role in human erythrocytes and in rats failed to support the hypothesis that PNP is relevant in reduction of AsV in vivo (Németi et al., 2003).
It has recently been shown that human red blood cells (RBC) possess a PNP-independent AsV reductase activity (Németi and Gregus, 2004), and circumstantial evidence has been presented that this activity depends on the availabilities of glutathione (GSH) and NAD or NADP. Furthermore, it has been proposed that this activity is linked to one or more enzymes in the glycolytic pathway between glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and enolase. The previous work (Németi and Gregus, 2005) has further characterized this AsV reductase activity in human RBC lysate and rat liver cytosol, presented direct evidence for its NAD- and GSH-dependent nature, and confirmed the suggestion that it is associated with glycolysis. This work demonstrated that in addition to NAD, the glycolytic substrates fructose-1,6-bisphosphate (Fruc-1,6-BP), 3-phosphoglycerate (3-PGA), and 2- phosphoglycerate (2-PGA) supported reduction of AsV the greatest. The strong dependence of AsV reduction on NAD makes GAPDH a candidate enzyme with AsV reductase activity, because GAPDH is the only glycolytic enzyme, for which NAD is both a substrate and an activator (Tilton et al., 1991). However, the marked stimulation of the hemolysate- and cytosol-mediated AsV reduction by 3-PGA and 2-PGA (which can readily be converted to 3-PGA) makes phosphoglycerate kinase (PGK) also a candidate, because 3-PGA is a substrate for PGK, and because in the glycolytic pathway PGK is linked to GAPDH through their common substrate 1,3-bisphosphoglycerate (1,3-BPG) (see scheme in Németi and Gregus, 2005). Moreover, GAPDH can use AsV instead of inorganic phosphate (Pi) when converting glyceraldehyde-3-phosphate (Ga-3-P) into 1-arseno-3-phosphoglycerate, instead of 1,3-BPG.
The aim of the present study was to determine if GAPDH, PGK, or both enzymes can function as AsV reductase. For this purpose, we first tested whether the two purified enzymes supplemented with various combinations of their nucleotide and glycolytic substrates could carry out the reduction of AsV in the presence of GSH. These experiments have clarified that it is GAPDH and not PGK that catalyzes the reduction of AsV in the presence of GSH, NAD, and its glycolytic substrates. Therefore, we performed further studies in order to characterize the AsV reductase activity of this enzyme as well as to test, using the GAPDH inhibitor koningic acid (KA; Beisswenger et al., 2003; Kato et al., 1992; Sakai et al., 1988), whether GAPDH contributes to reduction of AsV in human RBC lysate, rat liver cytosol, and intact human erythrocytes.
MATERIALS AND METHODS
Chemicals.
BCX-1777 (also called Immucillin-H) was a generous gift from BioCryst Pharmaceuticals (Birmingham, AL). Koningic acid (also called heptelidic acid) was kindly provided by Professor Keiji Hasumi (Tokyo Noko University, Tokyo, Japan). 2,2'-Biquinoline-4,4'-dicarboxylic acid disodium salt (bicinchoninic acid) was from Fluka. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from rabbit muscle and human erythrocytes, D-gluconic acid sodium salt, N-(2-hydroxyethyl)-piperazine-N'-(2-ethanesulfonic acid) (HEPES), phosphoglyceric phosphokinase (PGK) from baker's yeast, fructose-1,6-bisphosphate (tetra)cyclohexylammonium salt, D,L-glyceraldehyde-3-phosphate diethyl acetal monobarium salt, and 3-phosphoglyceric acid disodium salt, were from Sigma. Reduced glutathione, sodium pyruvate, and potassium ferricyanide were from Reanal Ltd. (Budapest, Hungary). The sources of arsenic compounds and chemicals used in arsenic speciation have been given elsewhere (Csanaky et al., 2003; Németi and Gregus, 2002). All other chemicals were of the highest purity commercially available.
Assays of AsV reduction and GAPDH activity—general conditions.
If otherwise not specified, assays were carried out and substances were dissolved in sucrose buffer, containing 250 mM sucrose, 25 mM HEPES, 5 mM MgCl2, 2 mM EGTA, pH 7.4. AsV reductase assays were run in 300 μl final volume, in microcentrifuge tubes kept in a shaking water bath at 37°C. They were started by addition of AsV (50 μM) and terminated by sequential addition of 100 μl 25 mM CdSO4 solution followed by 100 μl 1.5 M perchloric acid solution containing 25 mM HgCl2. The rationale for this procedure is given in the adjoining paper (Németi and Gregus, 2005). The incubates thus treated were stored at –80°C until arsenic analysis. AsV reductase activity was expressed as the amount of AsIII formed per minute and quantity of enzyme source (e.g., unit GAPDH, ml packed RBC, mg cytosolic protein).
The GAPDH activity was assayed spectrophotometrically based on the decrease of NADH concentration (0.25 mM) during the GAPDH-limited conversion of 3-PGA (5 mM) to Ga-3-P in the presence of excess PGK (1 U) and ATP (5 mM). This procedure measures the rate of the reverse reaction rather than the forward reaction, which occurs during glycolysis. The assay was carried out in 1 ml final volume, in the spectrophotometer cell, at 25°C, and was started by addition of GAPDH or its source (e.g., diluted hemolysate, cytosol). The enzyme activity was calculated from the change of absorbance (A/min at 340 nm), taking 6220 M–1 as the molar extinction coefficient of NADH340nm. The GAPDH activity was expressed as unit per volume (U/ml, corresponding to μmol NADH utilized per min and ml) or U per quantity of enzyme source (e.g., ml packed RBC, mg cytosolic protein). For both assays, more details are given below.
Purified enzyme experiments.
Stock solutions from the commercially obtained GAPDH and PGK were prepared freshly and kept in ice until use within 2 h. The GAPDH activity in the stock solution was regularly quantified spectrophotometrically as described above.
When analyzing the role of GAPDH and PGK as well as their substrate supply in AsV reduction, the incubations contained these enzymes (both at 2 U/ml concentration): Ga-3-P (1 mM), NAD (1 mM), and ADP (1 mM) (when testing the forward reaction), or 3-PGA (1 mM), NAD (1 mM), and ATP (1 mM) (when testing the reverse reaction). The incubation was started by adding GSH (6 mM), GAPDH, PGK, and AsV (50 μM), and continued for 10 min. When characterizing the AsV reducing activity of purified GAPDH, incubations were performed in the presence of Ga-3-P and NAD (both 1 mM). The incubation was started by adding GSH (6 mM) and GAPDH (2 U/ml) followed by AsV (50 μM), and it lasted for 10 min. To determine the effect of koningic acid (dissolved in dimethyl sulfoxide (DMSO) on the AsV-reducing activity of purified GAPDH, the enzyme (2 U/ml) was preincubated for 10 min with koningic acid (2.5–50 μM) or DMSO (not exceeding 1% final concentration). Thereafter, GSH (6 mM), Ga-3-P (0.5 mM), NAD (1 mM) were added, followed by AsV (50 μM), and the incubation was continued for 10 min. To measure the effect of koningic acid on the classical enzymatic activity of GAPDH, the enzyme (2 U/ml) was incubated with koningic acid (or DMSO) for 10 min, and then its activity was measured immediately, using the spectrophotometric assay described above.
RBC experiments.
This research was approved by the Regional Scientific Research Ethics Committee of the University of Pécs, Center for Medical and Health Sciences. Blood (approximately 5 ml) was collected from healthy human volunteers after informed consent. For the experiments with hemolysate, the washed RBC were prepared with and incubated in sucrose buffer as described in the previous work (Németi and Gregus, 2005). For the experiments with intact RBC, however, the RBC suspension was prepared and incubated in a chloride-free gluconate buffer, containing 150 mM sodium gluconate, 10 mM HEPES, 5 mM glucose, pH 7.4. This was necessary, because chloride inhibits the uptake of AsV by the erythrocytes (Németi and Gregus, 2004).
To determine the effect of koningic acid on the PNP-independent AsV reductase activity of hemolysate, erythrocytes (50 μl packed cells) were preincubated at 37°C for 10 min in sucrose buffer with a PNP inhibitor (BCX-1777, 20 μM), detergent (Nonidet P-40, 0.067%), glucose oxidase (2 U), and koningic acid (2.5–100 μM) or DMSO. Thereafter, the incubation was started by adding GSH (6 mM), buffer or Fruc-1,6-BP (1 mM) plus NAD (1 mM) or 3-PGA (1 mM) plus NAD, followed by AsV (50 μM), and was continued for 2.5 min. To measure the effect of koningic acid on the GAPDH activity of hemolysate, RBC (50 μl packed cells) were incubated with koningic acid (or DMSO) for 10 min at 37°C in a final volume of 0.3 ml of sucrose buffer in the presence of detergent (Nonidet P-40, 0.067%). Thereafter, the incubate was diluted with ice-cold buffer approximately 17-fold, and its GAPDH activity was immediately assayed.
To determine the effects of koningic acid on AsV reductase activity of intact RBC, erythrocytes (15 μl packed cells) were preincubated with koningic acid (1–50 μM) or DMSO for 20 min in gluconate buffer. Thereafter, the incubation was started by adding buffer, pyruvate (0.25 mM), or ferricyanide (0.25 mM) followed by AsV (50 μM), and was continued for 30 min. The incubations were stopped as described above except that the CdSO4 solution contained 1% Triton X-100. To measure the effect of koningic acid on the GAPDH activity of intact erythrocytes, RBC were incubated in gluconate buffer with koningic acid for 20 or 50 min (corresponding to time points when incubations with AsV would start and end). Therefater, RBC were pelleted by brief centrifugation (10,000 x g, 30 sec), and the supernatant was carefully removed. The cells were resuspended in 33 volumes of ice-cold saline and centrifuged again. After careful supernatant removal, RBC were lysed by adding 99 volumes of water and brief sonication. The GAPDH activity of RBC lysate thus prepared was immediately assayed spectrophotometrically.
Cytosol experiments.
To prepare rat liver cytosol, male Wistar rats weighing 250–270 g and maintained under standardized conditions were used. All procedures were carried out according to the Hungarian Animals Act (Scientific Procedures, 1998), and the study was in agreement with the rules of the Ethics Committee on Animal Research of the University of Pécs. The cytosolic fraction was prepared by differential centrifugation of the liver homogenate made in sucrose buffer as described in the preceding paper (Németi and Gregus, 2005). Cytosolic protein concentration was quantified by the bicinchoninic acid method according to Brown et al. (1989).
To determine the effects of koningic acid on the PNP-independent AsV reductase activity of cytosol, rat liver cytosol (5 mg protein/ml) was preincubated for 10 min in sucrose buffer with a PNP inhibitor (BCX-1777, 20 μM), glucose oxidase (2 U), and koningic acid (2.5–100 μM) or DMSO. Thereafter, the incubation was started by adding GSH (10 mM), buffer or Fruc-1,6-BP (1 mM) plus NAD (1 mM), or 3-PGA (1 mM) plus NAD, followed by AsV (50 μM), and was continued for 2.5 min. To measure the effects of koningic acid on the GAPDH activity of cytosol, rat liver cytosol (5 mg protein/ml) was incubated with koningic acid (2.5–100 μM) or DMSO for 10 min. Thereafter, the GAPDH activity of the cytosol was immediately assayed, using the spectrophotometric procedure.
Arsenic analysis.
The incubates having been subjected to protein precipitation were centrifuged at 10,000 x g, 4°C for 10 min. AsIII and AsV in the resultant supernatants were separated and quantified by HPLC-hydride generation-atomic fluorescence spectrometry (HPLC-HG-AFS), the details of which have been given elsewhere (Gregus et al., 2000; Németi et al., 2003).
Statistics.
Data were analyzed using one-way ANOVA followed by Duncan's test or Students' t-test with p < 0.05, as the level of significance.
RESULTS
Reduction of AsV to AsIII by a Mixture of Purified GAPDH and PGK
Two enzymatically interconvertible phosphoglycerates (i.e., 3-PGA and 2-PGA) and NAD were most effective in supporting the reduction of AsV either by hemolysate or by cytosol (Németi and Gregus, 2005). NAD is substrate for GAPDH, whereas 3-PGA is for PGK, hence we investigated if AsV reduction was also carried out in the presence of purified GAPDH and PGK, neighboring enzymes of the glycolytic pathway. In these assays, the enzymes were supported with 3-PGA plus ATP plus NAD or Ga-3-P plus ADP plus NAD. These experiments were carried out with GAPDH purified from either rabbit muscle or human erythrocytes and PGK prepared from yeast.
Figure 1 demonstrates that, when 3-PGA, ATP, and NAD were the added substrates, the mixture of GAPDH (rabbit) and PGK in the presence of GSH reduced AsV at a rate of 31.5 ± 5.0 pmol/minute. Omitting any one of the components from the complete incubations practically abolished AsIII formation, although the lack of exogenous NAD resulted in strong but incomplete inhibition of AsV reduction. Replacement of NAD with NADH or ATP with ADP in the complete incubation mixture also markedly diminished AsV-reducing activity.
When Ga-3-P, NAD, and ADP were the substrates added to the mix of rabbit GAPDH and PGK, the reduction rate was 13.4 ± 2.2 pmol/min (Fig. 2). Upon omission of GAPDH, Ga-3-P, or NAD, the AsIII formation ceased, and so did upon replacement of NAD with NADH. In contrast, omission of PGK did not affect the rate of AsV reduction, while omission of ADP significantly increased it. The presence of ATP instead of ADP tended to increase AsIII formation.
When performing these experiments with GAPDH originating from human erythrocytes instead of rabbit GAPDH, we obtained qualitatively similar findings, though the reduction rate was a little higher in the presence of the human enzyme under every condition tested (Figs. 1 and 2).
Reduction of AsV to AsIII by Purified GAPDH
Since the observations presented above strongly suggested that AsV reduction was carried out by GAPDH, we tested if this enzyme alone supported by its two substrates (i.e., Ga-3-P and NAD) was capable of reducing AsV in the presence of GSH. As demonstrated in Figure 3, the reduction rate increased proportionately to the concentration of rabbit muscle GAPDH up to 2 units/ml. Above this concentration, the reduction rate increased less than proportionately. The human GAPDH seemed to have somewhat higher AsV-reducing activity than the rabbit muscle enzyme.
We characterized the AsV reductase activity of GAPDH with respect to its dependence on concentrations of the substrates (i.e., Ga-3-P and NAD) and GSH, and to its responsiveness to selected nucleotides. The rate of AsV reduction by purified GAPDH changed with the concentration of the two substrates differently. The increase in Ga-3-P concentration below 0.5 mM resulted in steep increase in AsIII formation (Fig. 4, left). However, above 0.5 mM, the rate of AsV reduction declined below the rate at 0.5 mM Ga-3-P. With increase in NAD concentration, the GAPDH-catalyzed AsIII formation rose to approach its maximum at 1 mM NAD (Fig. 4, right). With further increase in NAD supply, the AsIII formation rate remained unchanged. It is noteworthy that AsV reduction catalyzed by the human erythrocyte-derived GAPDH responded to changes in substrate availability similarly, but the human enzyme exhibited a slightly higher activity.
Both the rabbit and the human GAPDH failed to catalyze reduction of AsV in the absence of GSH, even if Ga-3-P and NAD were provided (Fig. 5). Increasing the concentration of GSH enhanced the formation of AsIII from AsV in a concentration-dependent exponential fashion. As previously observed, the AsV reduction by GAPDH enzymes from rabbit muscle and human RBC responded similarly, and the latter exhibited a slightly higher activity.
The effects of nucleotides on the AsV reductase activity of GAPDH were tested in the presence of GAPDH substrates (i.e., Ga-3-P and NAD) and GSH (Fig. 6). While AMP and NADPH did not influence AsV reduction, ADP, ATP, and NADP moderately, whereas NADH strongly inhibited AsIII formation by GAPDH of both sources.
The Effects of Koningic Acid on the Classical Enzymatic and AsV Reductase Activities of Purified GAPDH
Because KA is a potent specific inhibitor of GAPDH (Sakai et al., 1988), we tested its effects on the classical enzymatic activity of GAPDH (i.e., NAD-dependent Ga-3-P oxidation and phosphorylation) as well as the AsV reductase activity of GAPDH purified from rabbit muscle. In the absence of KA, GAPDH and AsV reductase activities were 2.0 ± 0.05 U/ml and 67.06 ± 2.42 pmol AsIII formed per minute and U GAPDH, respectively. Preincubation of GAPDH for 10 min with this compound (with only the enzyme and KA being present in the preincubation solution) decreased the classical enzymatic activity in a concentration-dependent manner, abolishing it at 25 μM (Fig. 7, top). In a similar fashion, preincubation of GAPDH with KA under the same conditions diminished the AsV reductase activity of the enzyme as well (Fig. 7, bottom). Inclusion of GSH or NAD in the preincubation mix did not influence the inhibitory effect of KA on AsV reduction (Table 1). In contrast, inclusion of Ga-3-P either alone or in combination with NAD partially protected the AsV reductase activity of the enzyme from the inhibitory effect of KA.
Effects of Koningic Acid on the GAPDH and AsV Reductase Activities of RBC Lysate and Rat Liver Cytosol
Because KA effectively inhibited both the classical and the AsV reductase activities of purified rabbit muscle GAPDH, we investigated if KA exhibited similar effects on the GAPDH and AsV-reducing activities of hemolysate and rat liver cytosol. To exclude the contribution of PNP to AsV reduction, PNP was inactivated by BCX-1777 included in the preincubation mix.
In the absence of KA, the GAPDH activity of the hemolysate was 19.6 ± 1.86 U/ml packed cell. Preincubation with KA diminished GAPDH activity of the RBC lysate in a concentration-dependent fashion (Fig. 8, top), causing an approximately 20% inhibition at a concentration as low as 2.5 μM, and near complete inhibition at 50–100 μM. AsIII formation rates in the hemolysate without substrates, with Fruc-1,6-BP plus NAD, and with 3-PGA plus NAD were 0.46 ± 0.09, 6.15 ± 0.87, and 12.3 ± 0.71 nmol/min/ml packed cell, respectively. Without added substrate, preincubation with KA decreased the AsV-reducing activity significantly at 10 μM concentration and above (Fig. 8, bottom); however, 17–20% activity remained in the presence of even 50–100 μM KA. The Fruc-1,6-BP plus NAD and 3-PGA plus NAD-stimulated AsV reduction exhibited much higher sensitivity toward KA. The diminution in AsIII formation caused by preincubation with KA was significant at 10 μM concentration and above; AsV reduction was barely detectable at 25 μM KA and was abolished by KA at 50 μM and above.
In rat liver cytosol, the control GAPDH activity was 0.47 ± 0.02 U/mg protein. Preincubation with KA decreased this activity in a concentration-dependent manner, causing significant inhibition at 10 μM KA and above and complete inactivation at 100 μM (Fig. 9, top). AsIII formation rates in the cytosol without substrates, with Fruc-1,6-BP plus NAD, and with 3-PGA plus NAD were 17.7 ± 1.9, 413 ± 41, and 302 ± 39 pmol/min/mg protein, respectively. KA also inhibited the cytosolic AsV reduction (Fig. 9, bottom). The stimulated AsIII formation (i.e., in the presence of Fruc-1,6-BP plus NAD or 3-PGA plus NAD) decreased to low rates at 50 μM KA and almost ceased at 100 μM KA. However, the basal AsV reductase activity (i.e., observed in the absence of exogenous substrates) was only moderately affected, as KA even at 100 μM concentration inhibited AsIII formation by approximately 45%.
Effects of Koningic Acid on the GAPDH and AsV Reductase Activities of Intact RBC
To assess the contribution of GAPDH to the PNP-independent AsV reduction in intact RBC, the effect of KA on the AsV-reducing activity of GAPDH of these cells was tested in the presence of BCX-1777. Since these incubations lasted for 30 min, GAPDH activity was determined both at the beginning and at the end of the incubation (i.e., at 20 and at 20 + 30 min after starting the preincubation with KA). In the absence of KA, the GAPDH activity of the intact RBC at 20 and 50 min were 17.9 ± 1.1 and 18.3 ± 1.2 U/ml packed cell, respectively. As demonstrated in Figure 10 (top), the inhibition of GAPDH activity brought about by KA was not only concentration-dependent but also time-dependent. Inhibition of GAPDH activity became significant at 2.5 μM KA after 20-min preincubation, but at 1 μM KA after 50 min. Above 10 μM, KA caused near complete and complete GAPDH inhibition after 20 and 50 min, respectively. The AsIII formation rates without an NADH oxidant added, with pyruvate, and with ferricyanide added were 2.17 ± 0.25, 5.76 ± 0.41, and 5.06 ± 0.61 nmol/min/ml packed cell, respectively. KA markedly diminished these rates in a concentration-dependent fashion, and above 10 μM, it inhibited AsV reduction almost completely or completely under all three experimental conditions (Fig. 10, bottom).
DISCUSSION
In a previous work we presented circumstantial evidence that reduction of AsV to AsIII in intact RBC is linked to the part of the glycolytic pathway between GAPDH and enolase (Németi and Gregus, 2004). Characterization of this GSH-dependent AsV reductase activity in hemolysate and rat liver cytosol (Németi and Gregus, 2005) has led us to narrow down the candidate glycolytic enzymes to the neighboring pair of GAPDH and PGK. The present work demonstrates that the mixture of GAPDH and PGK can indeed reduce AsV in the presence of GSH and their substrates. As discussed below, further experimentation has revealed that, of the two enzymes, it is GAPDH, and not PGK, that is endowed with AsV reductase activity.
The mix of GAPDH and PGK formed AsIII from AsV when supplemented with Ga-3-P, NAD, and ADP (Fig. 2). Under this condition, GAPDH converts Ga-3-P to 1,3-BPG (in the presence of Pi) or 1-arseno-3-phosphoglycerate (in the presence of AsV), which in turn may be converted by PGK to 3-PGA (see insert in Fig. 2). Two findings indicate that PGK in this system does not contribute to AsV reduction. First, omission of PGK alone did not influence the AsV reductase activity (Fig. 2). Second, omission of ADP alone, which would prevent PGK from converting the GAPDH-produced 1,3-BPG (or 1-arseno-3-phosphoglycerate) to 3-PGA, even enhanced the reduction of AsV (Fig. 2). However, omission of PGK did abolish the AsV reductase activity of the mix of GAPDH and PGK, when this enzyme pair was supplemented with 3-PGA, ATP, and NAD (Fig. 1). Under this condition, PGK converts 3-PGA to 1,3-BPG for GAPDH, which in turn can 1-dephosphorylate 1,3-BPG and bind the resultant 3-phosphoglyceroyl group, but cannot reduce it to Ga-3-P in the absence of NADH. In the light of the findings discussed so far, the role of PGK under the conditions shown in Figure 1 can be interpreted as that of an auxiliary enzyme merely providing substrate for GAPDH, the actual NAD- and GSH-dependent AsV reductase. This interpretation is confirmed by the observation that GAPDH alone is capable of catalyzing reduction of AsV to AsIII in the presence of three indispensable ingredients, namely its glycolytic substrate Ga-3-P (Fig. 4, left), NAD (Fig. 4, right), and GSH (Fig. 5).
AsV reductase activity of purified GAPDH exhibited distinct concentration dependence on GSH, NAD, and Ga-3-P. The rate of the GAPDH-catalyzed AsV reduction kept increasing with an increase in GSH concentration (Fig. 5) similarly to that seen with the rate of AsV reduction catalyzed by hemolysate or cytosol (Fig. 1 in Németi and Gregus, 2005). The hyperbolic curve representing the concentration-dependent effect of NAD on GAPDH-catalyzed AsV reduction reflects most likely the saturation of the enzyme with this coenzyme at 1 mM, above which NAD causes no additional increase in AsV reduction (Fig. 4, right). To rationalize the peculiar concentration-dependent effect of Ga-3-P on the AsV reductase activity of GAPDH (Fig. 4, left), it should be considered that Ga-3-P is not only a substrate of GAPDH but also its allosteric inhibitor (Tomschy et al., 1993). Furthermore, it should also be realized that out of necessity we used racemic Ga-3-P (i.e., a mixture of the enzymatically active D-isomer and the inactive L-isomer). It has been reported that at higher concentrations, the racemic Ga-3-P produced stronger inhibition of GAPDH activity than the pure D-isomer, indicating that the L-isomer is more potent allosteric inhibitor of the enzyme than the D-isomer (Tomschy et al., 1993). Thus, the observed decline in the GAPDH-catalyzed AsIII formation from AsV above 0.5 mM Ga-3-P (Fig. 4, left) can most likely be ascribed to the inhibitory effect of Ga-3-P, especially of the L-isomer present in the commercially available racemic mixture. The inhibitory effect of Ga-3-P on AsV reduction by the hemolysate in the absence of exogenous NAD described in the previous paper (Németi and Gregus, 2005) may, in part, originate from the presence of L-Ga-3-P also. In addition, when Ga-3-P is the added substrate, the enzyme produces NADH, which strongly inhibits the GAPDH-catalyzed AsV reduction even in the presence of 1 mM NAD (Fig. 6). Adenine nucleotides (i.e., AMP, ADP, and ATP) are known to weakly inhibit GAPDH by competing with NAD for binding (Nakamura et al., 1982). This may account for the moderate inhibition of AsV reduction catalyzed by purified GAPDH (Fig. 6), hemolysate, or cytosol (Németi and Gregus, 2005) brought about by these nucleotides. Furthermore, 2,3-bisphosphoglycerate (2,3-BPG) also inhibits GAPDH (Srivastava and Beutler, 1972); this may explain why this compound decreased AsV reduction when added to hemolysate or rat liver cytosol (Németi and Gregus, 2005).
The experiments carried out with KA also corroborate the role for GAPDH in the reduction of AsV to AsIII. The epoxide group-containing KA is a potent and specific inhibitor of GAPDH; it irreversibly inhibits the enzyme by forming a thioether bond with the active site cysteine at position 149, especially when the enzyme is complexed with NAD (Beisswenger et al., 2003; Kato et al., 1992; Sakai et al., 1991). As demonstrated in Figure 7, KA inhibited not only the classical enzymatic activity of GAPDH, but also its AsV reductase activity in a concentration-dependent manner. Nevertheless, Ga-3-P markedly decreased the inhibitory effect of KA on the GAPDH-catalyzed AsV reduction, when the enzyme was preincubated in the combined presence of KA and Ga-3-P (Table 1). This protective effect of Ga-3-P against KA results from the fact that Ga-3-P competes with KA for binding to GAPDH and may account, at least in part, for the incomplete inhibition by KA of both the GAPDH and the AsV-reducing activities of hemolysate and cytosol (Figs. 8 and 9, respectively). Importantly, GSH did not protect GAPDH from inactivation by KA (Table 1), indicating that the thiol reactivity of KA is not universal. Furthermore, KA-induced inactivation of GAPDH depends not only on KA concentration, but also on the incubation time (Sakai et al., 1988). Accordingly, the inhibition of GAPDH in intact RBC became more pronounced after longer incubations with KA (Fig. 10).
Demonstration that KA inhibits the AsV reductase activity of purified GAPDH permitted us to use KA as an experimental tool to assess the contribution of GAPDH to the PNP-independent reduction of AsV by the hemolysate, the liver cytosol, and the intact erythrocytes. The experiments with KA strongly suggest that the PNP-independent AsV reductase activity of the human lysed RBC and rat liver cytosol can almost solely be attributed to GAPDH under conditions that provide abundant NAD and glycolytic substrate supply for this enzyme (i.e., in the presence of NAD plus Fruc-1,6-BP or 3-PGA), because the GAPDH and AsV reductase activities decreased in parallel in both the hemolysate and the cytosol in response to increased concentration of KA, and when KA abolished the activity of GAPDH, formation of AsIII from AsV ceased (Figs. 8 and 9, respectively). However, when GAPDH was not provided with exogenous NAD and substrates, even 100 μM KA failed to inhibit AsV reduction by the hemolysate and the liver cytosol completely (Figs. 8 and 9, bottom panels), even though the GAPDH activities in these cell extracts were practically abolished. While the residual AsV-reducing activity in the presence of 100 μM KA was only 18% of control in the hemolysate (Fig. 8, bottom), in the cytosol it was as high as 60% (Fig. 9, bottom). Since both PNP and GAPDH activities in the rat liver cytosol were blocked (by BCX-1777 and KA, respectively), this finding supports the hypothesis put forward in the preceding paper (Németi and Gregus, 2005) that, besides GAPDH, there is another hitherto unidentified cytosolic enzyme, which is able to reduce AsV and is supported by GSH and NAD(P).
In intact human RBC, irrespective of whether the glycolysis was or was not stimulated with NADH oxidants (pyruvate or ferricyanide), KA was much more effective in inhibiting AsV reductase activity than in the hemolysate or the cytosol. For example, KA concentrations as low as 10 μM caused a near complete inhibition of AsIII formation in the intact erythrocytes (Fig. 10). As to the increased sensitivity of erythrocytes to KA, one may raise the possibility that these cells may concentrate the inhibitor, or GAPDH may be in a conformation in situ that is more susceptible to inhibition by KA. In addition, in the light of the discussion above, it may be speculated that GAPDH in intact erythrocytes is well supplied with NAD and glycolytic substrate, conditions which would give GAPDH a proportionally larger role in AsV reduction in these cells than in the hemolysate without optimal NAD and substrate supply. Nevertheless, it is to be emphasized that our observation on AsV reduction in RBC should not give the impression that erythrocytes play a significant role in AsV reduction in the body. Such a role is severely limited by Pi and chloride ions present in the plasma (but not in our incubation buffer), which inhibit erythrocytic uptake of AsV (Németi and Gregus, 2004).
The mechanism of the GAPDH-catalyzed AsV reduction is unknown. This work, however, provides some clues to put forward a hypothesis, which we present after briefly describing the catalytic cycle of this enzyme. GAPDH is composed of four identical subunits with one molecule of NAD bound strongly to each. During glycolysis, this enzyme catalyzes both the oxidation of Ga-3-P and the incorporation of Pi into Ga-3-P to produce 1,3-BPG. This process involves four main steps. In step 1, Ga-3-P covalently binds to Cys149 of the enzyme forming a thiohemiacetal (Harris and Waters, 1976; Nagradova, 2001; Nagradova and Schmalhausen, 1998). In step 2, the bound Ga-3-P is oxidized into 3-phosphoglyceric acid still bound to Cys149, now with a thioester bond, while NAD is reduced to NADH. In step 3, the NADH formed in step 2 is replaced by NAD. This is necessary because NAD binding induces a conformational change, which will dislocate the phosphate group of the 3-phosphoglyceroyl moiety from the binding site for Pi making it accessible for Pi (Nagradova, 2001) or, in our case, for AsV. In step 4, the thioester bond is cleaved by Pi (termed phosphorolytic cleavage), releasing the 3-phosphoglyceroyl moiety as 1,3-BPG from Cys149 of GAPDH. Instead of Pi, AsV can also cleave this bond (by arsenolytic cleavage) to produce the purportedly unstable 1-arseno-3-phopshoglycerate, instead of 1,3-BPG.
Our observations presented here are compatible with the following tentative suggestions: (1) the enzyme is ready to catalyze the reduction of AsV when it forms the 3-phosphoglyceroyl-enzyme : NAD complex (i.e., it carries bound NAD and the 3-phosphoglyceroyl group bound via a thioester bond to Cys149), and (2) the GAPDH-mediated AsV reduction takes place during, or as a consequence of, the arsenolytic cleavage of the thioester bond formed between the enzyme's Cys149 and the 3-phosphoglyceroyl group of the substrate. The first suggestion is supported by the observation that GAPDH catalyzed the reduction of AsV in the presence of PGK, ATP, 3-PGA, and NAD (Fig. 1). Under this condition, PGK uses ATP to phosphorylate 3-PGA into 1,3-BPG (see insert in Fig. 1), which in turn will acylate GAPDH at Cys149 while releasing Pi. Thus, the critical 3-phosphoglyceroyl-enzyme : NAD complex is formed, which cannot undergo reduction to Ga-3-P because NADH is absent, but can undergo phosphorolysis or arsenolysis. As to the second suggestion, a prerequisite for the phosphorolysis (or arsenolysis) is that the NADH formed as a result of substrate oxidation (step 2) be replaced by NAD. The arsenolytic (or phosphorolytic) cleavage is strongly inhibited by NADH (Nagradova and Schmalhausen, 1998), and so is the GAPDH-catalyzed AsV reduction (Fig. 6), supporting the importance of the arsenolytic cleavage in the reduction of AsV.
When proposing a mechanism for AsV reduction by GAPDH, it is important to note that three thiols cooperate in the process catalyzed by microbial AsV reductases. The E. coli (Martin et al., 2001) and yeast (Mukhopadhyay et al., 2000) AsV reductases are monothiol enzymes, for which GSH and glutaredoxin provide the two additional thiol groups, whereas in the Staphylococcus aureus enzyme all three thiols belong to the enzyme (Messens et al., 2002). In this latter enzyme, two thiols bind and reduce AsV to AsIII, while they form a disulfide bond, which will be reduced by the third thiol. The disulfide bond then formed is reduced by thioredoxin, completing a process termed thiol-disulfide cascade. In the microbial reductases, the catalytically important thiol-group is activated by a nearby positive charge. The S. aureus enzyme may functionally model GAPDH, because three thiols are also present and located in GAPDH near the site of the phosphorolytic/arsenolytic cleavage. One of these is GSH, which is associated with GAPDH very strongly (Krimsky and Racker, 1952) and is claimed to protect the 3-phosphoglyceroyl-enzyme : NAD complex against cleavage of the thioester bond by water (i.e., hydrolysis) (Kuzminskaya et al., 1993). Another thiol is the enzyme's Cys153, which is located at the active site of the GAPDH close to the catalytically important Cys149 in the same -helix, and the third is Cys149, which becomes free upon the arsenolytic cleavage of the thioester bond. Like the microbial enzymes, the active site of GAPDH also contains a thiol-activating positive charge on the imidazole ring of His176 (Nagradova and Schmalhausen, 1998). Thus, the three thiols might contribute to the reduction of AsV when it is in 1-arseno-3-phosphoglycerate, or after it is released from the latter unstable compound. The formed AsIII may then be released and complexed by GSH present in large excess in the ambient solution. GSH may also serve to reduce the disulfide bonds formed during AsV reduction. Theoretically, it is also possible that the contribution of GAPDH to the reduction of AsV is indirect, and GSH directly reduces AsV while being in mixed anhydride with 3-PGA as 1-arseno-3-phosphoglycerate. Even then, GAPDH is required, because formation of 1-arseno-3-phosphoglycerate must precede the reduction of AsV. Whether the role of GAPDH in AsV reduction is direct or indirect, the inhibitory effect of KA on the GAPDH-mediated AsV reduction is compatible with our hypotheses, as covalent binding of KA to Cys149 (Sakai et al., 1991) prevents formation of the 3-phosphoglyceroyl-enzyme and 1-arseno-3-phosphoglycerate. Nevertheless, more research is needed on the mechanism of GAPDH-mediated AsV reduction.
In summary, this work demonstrates that GAPDH can fortuitously function as an AsV reductase, provided NAD, glycolytic substrate, and GSH are available. We hypothesize that GAPDH-catalyzed AsV reduction takes place during, or as a result of, the arsenolytic cleavage of its substrate from the thioester bond between the active site cysteine and the substrate. As studies with the specific GAPDH inhibitor KA indicate, GAPDH is exclusively responsible for the PNP-independent AsV reductase activity in human erythrocytes and significantly contributes to the AsV reduction in rat liver cytosol. Further research should clarify whether or not GAPDH contributes to the reduction of AsV to AsIII in vivo.
NOTES
The authors certify that all research involving human subjects was done under full compliance with all government policies and the Helsinki Declaration.
ACKNOWLEDGMENTS
This publication is based on a work supported by the Hungarian National Scientific Research Fund (OTKA) and the Hungarian Ministry of Health. The authors wish to thank Mónika Agyaki and István Schweibert for their excellent assistance in the experimental work.
REFERENCES
Beisswenger, P. J., Howell, S. K., Smith, K., and Szwergold, B. S. (2003). Glyceraldehyde-3-phosphate dehydrogenase activity as an independent modifier of methylglyoxal levels in diabetes. Biochim. Biophys. Acta 1637, 98–106.
Brown, R. E., Jarvis, K. L., and Hyland, K. J. (1989). Protein measurement using bicinchoninic acid: Elimination of interfering substances. Anal. Biochem. 180, 136–139.
Csanaky, I., Németi, B., and Gregus, Z. (2003). Dose-dependent biotransformation of arsenite in rats – not S-adenosylmethionine depletion impairs arsenic methylation at high dose. Toxicology 183, 77–91.
Gregus, Z., Gyurasics, á., and Csanaky, I. (2000). Biliary and urinary excretion of inorganic arsenic: Monomethylarsonous acid as a major biliary metabolite in rats. Toxicol. Sci. 56, 18–25.
Gregus, Z., and Németi, B. (2002). Purine nucleoside phosphorylase as a cytosolic arsenate reductase. Toxicol. Sci. 70, 13–19.
Harris, J. I., and Waters, M. (1976). Glyceraldehyde-3-phosphate dehydrogenase. In The Enzymes, 3rd ed., (P. D. Boyer, Ed.), Vol. 13, pp. 1–49. Academic Press, New York.
Kato, M., Sakai, K., and Endo, A. (1992). Koningic acid (heptelidic acid) inhibition of glyceraldehyde-3-phosphate dehydrogenases from various sources. Biochim. Biophys. Acta 1120, 113–116.
Knowles, F. C., and Benson, A. A. (1983). The biochemistry of arsenic. Trends Biochem. Sci. 8, 178–180.
Krimsky, I., and Racker, E. (1952). Glutathione, a prosthetic group of glyceraldehyde-3-phosphate dehydrogenase. J. Biol. Chem. 198, 721–729.
Kuzminskaya, E. V., Asryants, R. A., and Nagradova, N. K. (1993). SH-containing compounds as allosteric effectors of glyceraldehyde-3-phosphate dehydrogenase. FEBS Lett. 336, 208–210.
Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P., and Edwards, B. F. P. (2001). Insights into the structure, solvation, and mechanism of ArsC arsenate reductase, a novel arsenic detoxification enzyme. Structure 9, 1071–1081.
Messens, J., Martins, J. C., Van Belle, K., Brosens, E., Desmyter, A., De Gieter, M., Wieruszeski, J. M., Willem, R., Wyns, L., and Zegers, I. (2002). All intermediates of the arsenate reductase mechanism, including an intramolecular dynamic disulfide cascade. Proc. Natl. Acad. Sci. U.S.A. 99, 8506–8511.
Mukhopadhyay, R., Shi, J., and Rosen, B. P. (2000). Purification and characterization of Acr2p, the Saccharomyces cerevisiae arsenate reductase. J. Biol. Chem. 275, 21149–21157.
Nagradova, N. K. (2001). Study of the properties of phosphorylating D-glyceraldehyde-3-phosphate dehydrogenase. Biochemistry (Moscow) 66, 1067–1076.
Nagradova, N. K., and Schmalhausen, E. V. (1998). D-glyceraldehyde-3-phosphate dehydrogenase: Structural basis and functional role of the acyl transfer reactions. Biochemistry (Moscow) 63, 504–515.
Nakamura, M., Fujiwara, A., Yasumasu, I., Okinaga, S., and Arai, K. (1982). Regulation of glucose metabolism by adenine nucleotides in round spermatids from rat testes. J. Biol. Chem. 257, 13945–13950.
Németi, B., Csanaky, I., and Gregus, Z. (2003). Arsenate reduction in human erythrocytes and rats – Testing the role of purine nucleoside phosphorylase. Toxicol. Sci. 74, 22–31.
Németi, B., and Gregus, Z. (2002). Mitochondria work as reactors in reducing arsenate to arsenite. Toxicol. Appl. Pharmacol. 182, 208–218.
Németi, B., and Gregus, Z. (2004). Glutathione-dependent reduction of arsenate in human erythrocytes – a process independent of purine nucleoside phosphorylase. Toxicol. Sci. 82, 419–428.
Németi, B., and Gregus, Z. (2005). Reduction of arsenate to arsenite by human erythrocyte lysate and rat liver cytosol – Characterization of a glutathione- and NAD-dependent arsenate reduction linked to glycolysis. Toxicol. Sci. 85, 000–000.
Radabaugh, T. R., Sampayo-Reyes, A., Zakharyan, R. A., and Aposhian, H. V. (2002). Arsenate reductase II. Purine nucleoside phosphorylase in the presence of dihydrolipoic acid is a route for reduction of arsenate to arsenite in mammalian systems. Chem. Res. Toxicol. 15, 692–698.
Sakai, K., Hasumi, K., and Endo, A. (1988). Inactivation of rabbit muscle glyceraldehyde-3-phosphate dehydrogenase by koningic acid. Biochim. Biophys. Acta 952, 297–303.
Sakai, K., Hasumi, K., and Endo, A. (1991). Identification of koningic acid (heptelidic acid)-modified site in rabbit muscle glyceraldehyde-3-phosphate dehydrogenase. Biochim. Biophys. Acta 1077, 192–196.
Srivastava, S. K., and Beutler, E. (1972). The effect of normal red cell constituents on the activities of red cell enzymes. Arch. Biochem. Biophys. 148, 249–255.
Thomas, D. J., Styblo, M., and Lin S. (2001). The cellular metabolism and systemic toxicity of arsenic. Toxicol. Appl. Pharmacol. 176, 127–144.
Tilton, W. M., Seaman, C., Carriero, D., and Piomelli, S. (1991). Regulation of glycolysis in the erythrocyte: Role of the lactate/pyruvate and NAD/NADH ratios. J. Lab. Clin. Med. 118, 146–152.
Tomschy, A., Glockshuber, R., and Jaenicke, R. (1993). Functional expression of D-glyceraldehyde-3-phosphate dehydrogenase from the hyperthermophilic eubacterium Thermotoga maritima in Escherichia coli. Eur. J. Biochem. 214, 43–50.
Vahter, M. (1983). Metabolism of arsenic. In Biological and Environmental Effects of Arsenic (B. A. Fowler, Ed.), pp. 170–197. Elsevier Science, New York.(Zoltán Gregus and Balázs )
ABSTRACT
The mammalian enzymes responsible for reduction of the environmentally prevalent arsenate (AsV) to the much more toxic arsenite (AsIII) are unknown. In the previous paper (Németi and Gregus, 2005), we proposed that glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and/or phosphoglycerate kinase (PGK) may catalyze reduction of AsV in human red blood cells (RBC), hemolysate, or rat liver cytosol. In testing this hypothesis, we show here that, if supplied with glutathione (GSH), NAD, and glycolytic substrate, the mixture of purified GAPDH and PGK indeed catalyzes the reduction of AsV. Further analysis revealed that GAPDH is endowed with AsV reductase activity, whereas PGK serves as an auxiliary enzyme, when 3-phosphoglycerate is the glycolytic substrate. The GAPDH-catalyzed AsV reduction required GSH, NAD, and glyceraldehyde-3-phosphate. ADP and ATP moderately, whereas NADH strongly inhibited the AsV reductase activity of the enzyme even in the presence of NAD. Koningic acid (KA), a specific and irreversible inhibitor of GAPDH, inhibited both the classical enzymatic and the AsV-reducing activities of the enzyme in a concentration-dependent fashion. To assess the contribution of GAPDH to the reduction of AsV carried out by hemolysate, rat liver cytosol, or intact erythrocytes, we determined the concentration-dependent effect of KA on AsV reduction by these cells and extracts. Inactivation of GAPDH by KA abolished AsV reduction in intact RBC as well as in the hemolysate and the liver cytosol, when GAPDH in the latter extracts was abundantly supplied with exogenous NAD and glycolytic substrate. However, despite complete inactivation of GAPDH by KA, the hepatic cytosol exhibited significant residual AsV-reducing activity in the absence of exogenous NAD and glycolytic substrate, suggesting that besides GAPDH, other cytosolic enzyme(s) may contribute to AsV reduction in the liver. In conclusion, the key glycolytic enzyme GAPDH can fortuitously catalyze the reduction of AsV to AsIII, if GSH, NAD, and glycolytic substrate are available. AsV reduction may take place during, or as a consequence of, the arsenolytic cleavage of the thioester bond formed between the enzyme's Cys149 and the 3-phosphoglyceroyl moiety of the substrate. Although GAPDH is exclusively responsible for reduction of AsV in human erythrocytes, its role in AsV reduction in vivo remains to be determined.
Key Words: arsenate; glyceraldehyde-3-phosphate dehydrogenase; koningic acid; glutathione; NAD; reduction.
INTRODUCTION
The environmentally prevalent arsenical, arsenate (AsV), is readily reduced to arsenite (AsIII) in living organisms (Thomas et al., 2001; Vahter, 1983). This process is considered as toxification, because AsIII is much more toxic, owing to its facile covalent reactivity with thiols, especially dithiols (Knowles and Benson, 1983). Despite its toxicological significance, the biochemical mechanism of AsV reduction has not been clarified in mammals. Rat liver mitochondria rapidly convert AsV to AsIII (Németi and Gregus, 2002); however, their contribution to AsV reduction in rats has not been proven. Purine nucleoside phosphorylase (PNP) also reduces AsV to AsIII, if its substrate (e.g., inosine) and an appropriate dithiol (e.g., dithiothreitol) are present (Gregus and Németi, 2002; Radabaugh et al., 2002), suggesting that this ubiquitous enzyme may play a role in AsV reduction in vivo. However, the testing of PNP for such a role in human erythrocytes and in rats failed to support the hypothesis that PNP is relevant in reduction of AsV in vivo (Németi et al., 2003).
It has recently been shown that human red blood cells (RBC) possess a PNP-independent AsV reductase activity (Németi and Gregus, 2004), and circumstantial evidence has been presented that this activity depends on the availabilities of glutathione (GSH) and NAD or NADP. Furthermore, it has been proposed that this activity is linked to one or more enzymes in the glycolytic pathway between glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and enolase. The previous work (Németi and Gregus, 2005) has further characterized this AsV reductase activity in human RBC lysate and rat liver cytosol, presented direct evidence for its NAD- and GSH-dependent nature, and confirmed the suggestion that it is associated with glycolysis. This work demonstrated that in addition to NAD, the glycolytic substrates fructose-1,6-bisphosphate (Fruc-1,6-BP), 3-phosphoglycerate (3-PGA), and 2- phosphoglycerate (2-PGA) supported reduction of AsV the greatest. The strong dependence of AsV reduction on NAD makes GAPDH a candidate enzyme with AsV reductase activity, because GAPDH is the only glycolytic enzyme, for which NAD is both a substrate and an activator (Tilton et al., 1991). However, the marked stimulation of the hemolysate- and cytosol-mediated AsV reduction by 3-PGA and 2-PGA (which can readily be converted to 3-PGA) makes phosphoglycerate kinase (PGK) also a candidate, because 3-PGA is a substrate for PGK, and because in the glycolytic pathway PGK is linked to GAPDH through their common substrate 1,3-bisphosphoglycerate (1,3-BPG) (see scheme in Németi and Gregus, 2005). Moreover, GAPDH can use AsV instead of inorganic phosphate (Pi) when converting glyceraldehyde-3-phosphate (Ga-3-P) into 1-arseno-3-phosphoglycerate, instead of 1,3-BPG.
The aim of the present study was to determine if GAPDH, PGK, or both enzymes can function as AsV reductase. For this purpose, we first tested whether the two purified enzymes supplemented with various combinations of their nucleotide and glycolytic substrates could carry out the reduction of AsV in the presence of GSH. These experiments have clarified that it is GAPDH and not PGK that catalyzes the reduction of AsV in the presence of GSH, NAD, and its glycolytic substrates. Therefore, we performed further studies in order to characterize the AsV reductase activity of this enzyme as well as to test, using the GAPDH inhibitor koningic acid (KA; Beisswenger et al., 2003; Kato et al., 1992; Sakai et al., 1988), whether GAPDH contributes to reduction of AsV in human RBC lysate, rat liver cytosol, and intact human erythrocytes.
MATERIALS AND METHODS
Chemicals.
BCX-1777 (also called Immucillin-H) was a generous gift from BioCryst Pharmaceuticals (Birmingham, AL). Koningic acid (also called heptelidic acid) was kindly provided by Professor Keiji Hasumi (Tokyo Noko University, Tokyo, Japan). 2,2'-Biquinoline-4,4'-dicarboxylic acid disodium salt (bicinchoninic acid) was from Fluka. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from rabbit muscle and human erythrocytes, D-gluconic acid sodium salt, N-(2-hydroxyethyl)-piperazine-N'-(2-ethanesulfonic acid) (HEPES), phosphoglyceric phosphokinase (PGK) from baker's yeast, fructose-1,6-bisphosphate (tetra)cyclohexylammonium salt, D,L-glyceraldehyde-3-phosphate diethyl acetal monobarium salt, and 3-phosphoglyceric acid disodium salt, were from Sigma. Reduced glutathione, sodium pyruvate, and potassium ferricyanide were from Reanal Ltd. (Budapest, Hungary). The sources of arsenic compounds and chemicals used in arsenic speciation have been given elsewhere (Csanaky et al., 2003; Németi and Gregus, 2002). All other chemicals were of the highest purity commercially available.
Assays of AsV reduction and GAPDH activity—general conditions.
If otherwise not specified, assays were carried out and substances were dissolved in sucrose buffer, containing 250 mM sucrose, 25 mM HEPES, 5 mM MgCl2, 2 mM EGTA, pH 7.4. AsV reductase assays were run in 300 μl final volume, in microcentrifuge tubes kept in a shaking water bath at 37°C. They were started by addition of AsV (50 μM) and terminated by sequential addition of 100 μl 25 mM CdSO4 solution followed by 100 μl 1.5 M perchloric acid solution containing 25 mM HgCl2. The rationale for this procedure is given in the adjoining paper (Németi and Gregus, 2005). The incubates thus treated were stored at –80°C until arsenic analysis. AsV reductase activity was expressed as the amount of AsIII formed per minute and quantity of enzyme source (e.g., unit GAPDH, ml packed RBC, mg cytosolic protein).
The GAPDH activity was assayed spectrophotometrically based on the decrease of NADH concentration (0.25 mM) during the GAPDH-limited conversion of 3-PGA (5 mM) to Ga-3-P in the presence of excess PGK (1 U) and ATP (5 mM). This procedure measures the rate of the reverse reaction rather than the forward reaction, which occurs during glycolysis. The assay was carried out in 1 ml final volume, in the spectrophotometer cell, at 25°C, and was started by addition of GAPDH or its source (e.g., diluted hemolysate, cytosol). The enzyme activity was calculated from the change of absorbance (A/min at 340 nm), taking 6220 M–1 as the molar extinction coefficient of NADH340nm. The GAPDH activity was expressed as unit per volume (U/ml, corresponding to μmol NADH utilized per min and ml) or U per quantity of enzyme source (e.g., ml packed RBC, mg cytosolic protein). For both assays, more details are given below.
Purified enzyme experiments.
Stock solutions from the commercially obtained GAPDH and PGK were prepared freshly and kept in ice until use within 2 h. The GAPDH activity in the stock solution was regularly quantified spectrophotometrically as described above.
When analyzing the role of GAPDH and PGK as well as their substrate supply in AsV reduction, the incubations contained these enzymes (both at 2 U/ml concentration): Ga-3-P (1 mM), NAD (1 mM), and ADP (1 mM) (when testing the forward reaction), or 3-PGA (1 mM), NAD (1 mM), and ATP (1 mM) (when testing the reverse reaction). The incubation was started by adding GSH (6 mM), GAPDH, PGK, and AsV (50 μM), and continued for 10 min. When characterizing the AsV reducing activity of purified GAPDH, incubations were performed in the presence of Ga-3-P and NAD (both 1 mM). The incubation was started by adding GSH (6 mM) and GAPDH (2 U/ml) followed by AsV (50 μM), and it lasted for 10 min. To determine the effect of koningic acid (dissolved in dimethyl sulfoxide (DMSO) on the AsV-reducing activity of purified GAPDH, the enzyme (2 U/ml) was preincubated for 10 min with koningic acid (2.5–50 μM) or DMSO (not exceeding 1% final concentration). Thereafter, GSH (6 mM), Ga-3-P (0.5 mM), NAD (1 mM) were added, followed by AsV (50 μM), and the incubation was continued for 10 min. To measure the effect of koningic acid on the classical enzymatic activity of GAPDH, the enzyme (2 U/ml) was incubated with koningic acid (or DMSO) for 10 min, and then its activity was measured immediately, using the spectrophotometric assay described above.
RBC experiments.
This research was approved by the Regional Scientific Research Ethics Committee of the University of Pécs, Center for Medical and Health Sciences. Blood (approximately 5 ml) was collected from healthy human volunteers after informed consent. For the experiments with hemolysate, the washed RBC were prepared with and incubated in sucrose buffer as described in the previous work (Németi and Gregus, 2005). For the experiments with intact RBC, however, the RBC suspension was prepared and incubated in a chloride-free gluconate buffer, containing 150 mM sodium gluconate, 10 mM HEPES, 5 mM glucose, pH 7.4. This was necessary, because chloride inhibits the uptake of AsV by the erythrocytes (Németi and Gregus, 2004).
To determine the effect of koningic acid on the PNP-independent AsV reductase activity of hemolysate, erythrocytes (50 μl packed cells) were preincubated at 37°C for 10 min in sucrose buffer with a PNP inhibitor (BCX-1777, 20 μM), detergent (Nonidet P-40, 0.067%), glucose oxidase (2 U), and koningic acid (2.5–100 μM) or DMSO. Thereafter, the incubation was started by adding GSH (6 mM), buffer or Fruc-1,6-BP (1 mM) plus NAD (1 mM) or 3-PGA (1 mM) plus NAD, followed by AsV (50 μM), and was continued for 2.5 min. To measure the effect of koningic acid on the GAPDH activity of hemolysate, RBC (50 μl packed cells) were incubated with koningic acid (or DMSO) for 10 min at 37°C in a final volume of 0.3 ml of sucrose buffer in the presence of detergent (Nonidet P-40, 0.067%). Thereafter, the incubate was diluted with ice-cold buffer approximately 17-fold, and its GAPDH activity was immediately assayed.
To determine the effects of koningic acid on AsV reductase activity of intact RBC, erythrocytes (15 μl packed cells) were preincubated with koningic acid (1–50 μM) or DMSO for 20 min in gluconate buffer. Thereafter, the incubation was started by adding buffer, pyruvate (0.25 mM), or ferricyanide (0.25 mM) followed by AsV (50 μM), and was continued for 30 min. The incubations were stopped as described above except that the CdSO4 solution contained 1% Triton X-100. To measure the effect of koningic acid on the GAPDH activity of intact erythrocytes, RBC were incubated in gluconate buffer with koningic acid for 20 or 50 min (corresponding to time points when incubations with AsV would start and end). Therefater, RBC were pelleted by brief centrifugation (10,000 x g, 30 sec), and the supernatant was carefully removed. The cells were resuspended in 33 volumes of ice-cold saline and centrifuged again. After careful supernatant removal, RBC were lysed by adding 99 volumes of water and brief sonication. The GAPDH activity of RBC lysate thus prepared was immediately assayed spectrophotometrically.
Cytosol experiments.
To prepare rat liver cytosol, male Wistar rats weighing 250–270 g and maintained under standardized conditions were used. All procedures were carried out according to the Hungarian Animals Act (Scientific Procedures, 1998), and the study was in agreement with the rules of the Ethics Committee on Animal Research of the University of Pécs. The cytosolic fraction was prepared by differential centrifugation of the liver homogenate made in sucrose buffer as described in the preceding paper (Németi and Gregus, 2005). Cytosolic protein concentration was quantified by the bicinchoninic acid method according to Brown et al. (1989).
To determine the effects of koningic acid on the PNP-independent AsV reductase activity of cytosol, rat liver cytosol (5 mg protein/ml) was preincubated for 10 min in sucrose buffer with a PNP inhibitor (BCX-1777, 20 μM), glucose oxidase (2 U), and koningic acid (2.5–100 μM) or DMSO. Thereafter, the incubation was started by adding GSH (10 mM), buffer or Fruc-1,6-BP (1 mM) plus NAD (1 mM), or 3-PGA (1 mM) plus NAD, followed by AsV (50 μM), and was continued for 2.5 min. To measure the effects of koningic acid on the GAPDH activity of cytosol, rat liver cytosol (5 mg protein/ml) was incubated with koningic acid (2.5–100 μM) or DMSO for 10 min. Thereafter, the GAPDH activity of the cytosol was immediately assayed, using the spectrophotometric procedure.
Arsenic analysis.
The incubates having been subjected to protein precipitation were centrifuged at 10,000 x g, 4°C for 10 min. AsIII and AsV in the resultant supernatants were separated and quantified by HPLC-hydride generation-atomic fluorescence spectrometry (HPLC-HG-AFS), the details of which have been given elsewhere (Gregus et al., 2000; Németi et al., 2003).
Statistics.
Data were analyzed using one-way ANOVA followed by Duncan's test or Students' t-test with p < 0.05, as the level of significance.
RESULTS
Reduction of AsV to AsIII by a Mixture of Purified GAPDH and PGK
Two enzymatically interconvertible phosphoglycerates (i.e., 3-PGA and 2-PGA) and NAD were most effective in supporting the reduction of AsV either by hemolysate or by cytosol (Németi and Gregus, 2005). NAD is substrate for GAPDH, whereas 3-PGA is for PGK, hence we investigated if AsV reduction was also carried out in the presence of purified GAPDH and PGK, neighboring enzymes of the glycolytic pathway. In these assays, the enzymes were supported with 3-PGA plus ATP plus NAD or Ga-3-P plus ADP plus NAD. These experiments were carried out with GAPDH purified from either rabbit muscle or human erythrocytes and PGK prepared from yeast.
Figure 1 demonstrates that, when 3-PGA, ATP, and NAD were the added substrates, the mixture of GAPDH (rabbit) and PGK in the presence of GSH reduced AsV at a rate of 31.5 ± 5.0 pmol/minute. Omitting any one of the components from the complete incubations practically abolished AsIII formation, although the lack of exogenous NAD resulted in strong but incomplete inhibition of AsV reduction. Replacement of NAD with NADH or ATP with ADP in the complete incubation mixture also markedly diminished AsV-reducing activity.
When Ga-3-P, NAD, and ADP were the substrates added to the mix of rabbit GAPDH and PGK, the reduction rate was 13.4 ± 2.2 pmol/min (Fig. 2). Upon omission of GAPDH, Ga-3-P, or NAD, the AsIII formation ceased, and so did upon replacement of NAD with NADH. In contrast, omission of PGK did not affect the rate of AsV reduction, while omission of ADP significantly increased it. The presence of ATP instead of ADP tended to increase AsIII formation.
When performing these experiments with GAPDH originating from human erythrocytes instead of rabbit GAPDH, we obtained qualitatively similar findings, though the reduction rate was a little higher in the presence of the human enzyme under every condition tested (Figs. 1 and 2).
Reduction of AsV to AsIII by Purified GAPDH
Since the observations presented above strongly suggested that AsV reduction was carried out by GAPDH, we tested if this enzyme alone supported by its two substrates (i.e., Ga-3-P and NAD) was capable of reducing AsV in the presence of GSH. As demonstrated in Figure 3, the reduction rate increased proportionately to the concentration of rabbit muscle GAPDH up to 2 units/ml. Above this concentration, the reduction rate increased less than proportionately. The human GAPDH seemed to have somewhat higher AsV-reducing activity than the rabbit muscle enzyme.
We characterized the AsV reductase activity of GAPDH with respect to its dependence on concentrations of the substrates (i.e., Ga-3-P and NAD) and GSH, and to its responsiveness to selected nucleotides. The rate of AsV reduction by purified GAPDH changed with the concentration of the two substrates differently. The increase in Ga-3-P concentration below 0.5 mM resulted in steep increase in AsIII formation (Fig. 4, left). However, above 0.5 mM, the rate of AsV reduction declined below the rate at 0.5 mM Ga-3-P. With increase in NAD concentration, the GAPDH-catalyzed AsIII formation rose to approach its maximum at 1 mM NAD (Fig. 4, right). With further increase in NAD supply, the AsIII formation rate remained unchanged. It is noteworthy that AsV reduction catalyzed by the human erythrocyte-derived GAPDH responded to changes in substrate availability similarly, but the human enzyme exhibited a slightly higher activity.
Both the rabbit and the human GAPDH failed to catalyze reduction of AsV in the absence of GSH, even if Ga-3-P and NAD were provided (Fig. 5). Increasing the concentration of GSH enhanced the formation of AsIII from AsV in a concentration-dependent exponential fashion. As previously observed, the AsV reduction by GAPDH enzymes from rabbit muscle and human RBC responded similarly, and the latter exhibited a slightly higher activity.
The effects of nucleotides on the AsV reductase activity of GAPDH were tested in the presence of GAPDH substrates (i.e., Ga-3-P and NAD) and GSH (Fig. 6). While AMP and NADPH did not influence AsV reduction, ADP, ATP, and NADP moderately, whereas NADH strongly inhibited AsIII formation by GAPDH of both sources.
The Effects of Koningic Acid on the Classical Enzymatic and AsV Reductase Activities of Purified GAPDH
Because KA is a potent specific inhibitor of GAPDH (Sakai et al., 1988), we tested its effects on the classical enzymatic activity of GAPDH (i.e., NAD-dependent Ga-3-P oxidation and phosphorylation) as well as the AsV reductase activity of GAPDH purified from rabbit muscle. In the absence of KA, GAPDH and AsV reductase activities were 2.0 ± 0.05 U/ml and 67.06 ± 2.42 pmol AsIII formed per minute and U GAPDH, respectively. Preincubation of GAPDH for 10 min with this compound (with only the enzyme and KA being present in the preincubation solution) decreased the classical enzymatic activity in a concentration-dependent manner, abolishing it at 25 μM (Fig. 7, top). In a similar fashion, preincubation of GAPDH with KA under the same conditions diminished the AsV reductase activity of the enzyme as well (Fig. 7, bottom). Inclusion of GSH or NAD in the preincubation mix did not influence the inhibitory effect of KA on AsV reduction (Table 1). In contrast, inclusion of Ga-3-P either alone or in combination with NAD partially protected the AsV reductase activity of the enzyme from the inhibitory effect of KA.
Effects of Koningic Acid on the GAPDH and AsV Reductase Activities of RBC Lysate and Rat Liver Cytosol
Because KA effectively inhibited both the classical and the AsV reductase activities of purified rabbit muscle GAPDH, we investigated if KA exhibited similar effects on the GAPDH and AsV-reducing activities of hemolysate and rat liver cytosol. To exclude the contribution of PNP to AsV reduction, PNP was inactivated by BCX-1777 included in the preincubation mix.
In the absence of KA, the GAPDH activity of the hemolysate was 19.6 ± 1.86 U/ml packed cell. Preincubation with KA diminished GAPDH activity of the RBC lysate in a concentration-dependent fashion (Fig. 8, top), causing an approximately 20% inhibition at a concentration as low as 2.5 μM, and near complete inhibition at 50–100 μM. AsIII formation rates in the hemolysate without substrates, with Fruc-1,6-BP plus NAD, and with 3-PGA plus NAD were 0.46 ± 0.09, 6.15 ± 0.87, and 12.3 ± 0.71 nmol/min/ml packed cell, respectively. Without added substrate, preincubation with KA decreased the AsV-reducing activity significantly at 10 μM concentration and above (Fig. 8, bottom); however, 17–20% activity remained in the presence of even 50–100 μM KA. The Fruc-1,6-BP plus NAD and 3-PGA plus NAD-stimulated AsV reduction exhibited much higher sensitivity toward KA. The diminution in AsIII formation caused by preincubation with KA was significant at 10 μM concentration and above; AsV reduction was barely detectable at 25 μM KA and was abolished by KA at 50 μM and above.
In rat liver cytosol, the control GAPDH activity was 0.47 ± 0.02 U/mg protein. Preincubation with KA decreased this activity in a concentration-dependent manner, causing significant inhibition at 10 μM KA and above and complete inactivation at 100 μM (Fig. 9, top). AsIII formation rates in the cytosol without substrates, with Fruc-1,6-BP plus NAD, and with 3-PGA plus NAD were 17.7 ± 1.9, 413 ± 41, and 302 ± 39 pmol/min/mg protein, respectively. KA also inhibited the cytosolic AsV reduction (Fig. 9, bottom). The stimulated AsIII formation (i.e., in the presence of Fruc-1,6-BP plus NAD or 3-PGA plus NAD) decreased to low rates at 50 μM KA and almost ceased at 100 μM KA. However, the basal AsV reductase activity (i.e., observed in the absence of exogenous substrates) was only moderately affected, as KA even at 100 μM concentration inhibited AsIII formation by approximately 45%.
Effects of Koningic Acid on the GAPDH and AsV Reductase Activities of Intact RBC
To assess the contribution of GAPDH to the PNP-independent AsV reduction in intact RBC, the effect of KA on the AsV-reducing activity of GAPDH of these cells was tested in the presence of BCX-1777. Since these incubations lasted for 30 min, GAPDH activity was determined both at the beginning and at the end of the incubation (i.e., at 20 and at 20 + 30 min after starting the preincubation with KA). In the absence of KA, the GAPDH activity of the intact RBC at 20 and 50 min were 17.9 ± 1.1 and 18.3 ± 1.2 U/ml packed cell, respectively. As demonstrated in Figure 10 (top), the inhibition of GAPDH activity brought about by KA was not only concentration-dependent but also time-dependent. Inhibition of GAPDH activity became significant at 2.5 μM KA after 20-min preincubation, but at 1 μM KA after 50 min. Above 10 μM, KA caused near complete and complete GAPDH inhibition after 20 and 50 min, respectively. The AsIII formation rates without an NADH oxidant added, with pyruvate, and with ferricyanide added were 2.17 ± 0.25, 5.76 ± 0.41, and 5.06 ± 0.61 nmol/min/ml packed cell, respectively. KA markedly diminished these rates in a concentration-dependent fashion, and above 10 μM, it inhibited AsV reduction almost completely or completely under all three experimental conditions (Fig. 10, bottom).
DISCUSSION
In a previous work we presented circumstantial evidence that reduction of AsV to AsIII in intact RBC is linked to the part of the glycolytic pathway between GAPDH and enolase (Németi and Gregus, 2004). Characterization of this GSH-dependent AsV reductase activity in hemolysate and rat liver cytosol (Németi and Gregus, 2005) has led us to narrow down the candidate glycolytic enzymes to the neighboring pair of GAPDH and PGK. The present work demonstrates that the mixture of GAPDH and PGK can indeed reduce AsV in the presence of GSH and their substrates. As discussed below, further experimentation has revealed that, of the two enzymes, it is GAPDH, and not PGK, that is endowed with AsV reductase activity.
The mix of GAPDH and PGK formed AsIII from AsV when supplemented with Ga-3-P, NAD, and ADP (Fig. 2). Under this condition, GAPDH converts Ga-3-P to 1,3-BPG (in the presence of Pi) or 1-arseno-3-phosphoglycerate (in the presence of AsV), which in turn may be converted by PGK to 3-PGA (see insert in Fig. 2). Two findings indicate that PGK in this system does not contribute to AsV reduction. First, omission of PGK alone did not influence the AsV reductase activity (Fig. 2). Second, omission of ADP alone, which would prevent PGK from converting the GAPDH-produced 1,3-BPG (or 1-arseno-3-phosphoglycerate) to 3-PGA, even enhanced the reduction of AsV (Fig. 2). However, omission of PGK did abolish the AsV reductase activity of the mix of GAPDH and PGK, when this enzyme pair was supplemented with 3-PGA, ATP, and NAD (Fig. 1). Under this condition, PGK converts 3-PGA to 1,3-BPG for GAPDH, which in turn can 1-dephosphorylate 1,3-BPG and bind the resultant 3-phosphoglyceroyl group, but cannot reduce it to Ga-3-P in the absence of NADH. In the light of the findings discussed so far, the role of PGK under the conditions shown in Figure 1 can be interpreted as that of an auxiliary enzyme merely providing substrate for GAPDH, the actual NAD- and GSH-dependent AsV reductase. This interpretation is confirmed by the observation that GAPDH alone is capable of catalyzing reduction of AsV to AsIII in the presence of three indispensable ingredients, namely its glycolytic substrate Ga-3-P (Fig. 4, left), NAD (Fig. 4, right), and GSH (Fig. 5).
AsV reductase activity of purified GAPDH exhibited distinct concentration dependence on GSH, NAD, and Ga-3-P. The rate of the GAPDH-catalyzed AsV reduction kept increasing with an increase in GSH concentration (Fig. 5) similarly to that seen with the rate of AsV reduction catalyzed by hemolysate or cytosol (Fig. 1 in Németi and Gregus, 2005). The hyperbolic curve representing the concentration-dependent effect of NAD on GAPDH-catalyzed AsV reduction reflects most likely the saturation of the enzyme with this coenzyme at 1 mM, above which NAD causes no additional increase in AsV reduction (Fig. 4, right). To rationalize the peculiar concentration-dependent effect of Ga-3-P on the AsV reductase activity of GAPDH (Fig. 4, left), it should be considered that Ga-3-P is not only a substrate of GAPDH but also its allosteric inhibitor (Tomschy et al., 1993). Furthermore, it should also be realized that out of necessity we used racemic Ga-3-P (i.e., a mixture of the enzymatically active D-isomer and the inactive L-isomer). It has been reported that at higher concentrations, the racemic Ga-3-P produced stronger inhibition of GAPDH activity than the pure D-isomer, indicating that the L-isomer is more potent allosteric inhibitor of the enzyme than the D-isomer (Tomschy et al., 1993). Thus, the observed decline in the GAPDH-catalyzed AsIII formation from AsV above 0.5 mM Ga-3-P (Fig. 4, left) can most likely be ascribed to the inhibitory effect of Ga-3-P, especially of the L-isomer present in the commercially available racemic mixture. The inhibitory effect of Ga-3-P on AsV reduction by the hemolysate in the absence of exogenous NAD described in the previous paper (Németi and Gregus, 2005) may, in part, originate from the presence of L-Ga-3-P also. In addition, when Ga-3-P is the added substrate, the enzyme produces NADH, which strongly inhibits the GAPDH-catalyzed AsV reduction even in the presence of 1 mM NAD (Fig. 6). Adenine nucleotides (i.e., AMP, ADP, and ATP) are known to weakly inhibit GAPDH by competing with NAD for binding (Nakamura et al., 1982). This may account for the moderate inhibition of AsV reduction catalyzed by purified GAPDH (Fig. 6), hemolysate, or cytosol (Németi and Gregus, 2005) brought about by these nucleotides. Furthermore, 2,3-bisphosphoglycerate (2,3-BPG) also inhibits GAPDH (Srivastava and Beutler, 1972); this may explain why this compound decreased AsV reduction when added to hemolysate or rat liver cytosol (Németi and Gregus, 2005).
The experiments carried out with KA also corroborate the role for GAPDH in the reduction of AsV to AsIII. The epoxide group-containing KA is a potent and specific inhibitor of GAPDH; it irreversibly inhibits the enzyme by forming a thioether bond with the active site cysteine at position 149, especially when the enzyme is complexed with NAD (Beisswenger et al., 2003; Kato et al., 1992; Sakai et al., 1991). As demonstrated in Figure 7, KA inhibited not only the classical enzymatic activity of GAPDH, but also its AsV reductase activity in a concentration-dependent manner. Nevertheless, Ga-3-P markedly decreased the inhibitory effect of KA on the GAPDH-catalyzed AsV reduction, when the enzyme was preincubated in the combined presence of KA and Ga-3-P (Table 1). This protective effect of Ga-3-P against KA results from the fact that Ga-3-P competes with KA for binding to GAPDH and may account, at least in part, for the incomplete inhibition by KA of both the GAPDH and the AsV-reducing activities of hemolysate and cytosol (Figs. 8 and 9, respectively). Importantly, GSH did not protect GAPDH from inactivation by KA (Table 1), indicating that the thiol reactivity of KA is not universal. Furthermore, KA-induced inactivation of GAPDH depends not only on KA concentration, but also on the incubation time (Sakai et al., 1988). Accordingly, the inhibition of GAPDH in intact RBC became more pronounced after longer incubations with KA (Fig. 10).
Demonstration that KA inhibits the AsV reductase activity of purified GAPDH permitted us to use KA as an experimental tool to assess the contribution of GAPDH to the PNP-independent reduction of AsV by the hemolysate, the liver cytosol, and the intact erythrocytes. The experiments with KA strongly suggest that the PNP-independent AsV reductase activity of the human lysed RBC and rat liver cytosol can almost solely be attributed to GAPDH under conditions that provide abundant NAD and glycolytic substrate supply for this enzyme (i.e., in the presence of NAD plus Fruc-1,6-BP or 3-PGA), because the GAPDH and AsV reductase activities decreased in parallel in both the hemolysate and the cytosol in response to increased concentration of KA, and when KA abolished the activity of GAPDH, formation of AsIII from AsV ceased (Figs. 8 and 9, respectively). However, when GAPDH was not provided with exogenous NAD and substrates, even 100 μM KA failed to inhibit AsV reduction by the hemolysate and the liver cytosol completely (Figs. 8 and 9, bottom panels), even though the GAPDH activities in these cell extracts were practically abolished. While the residual AsV-reducing activity in the presence of 100 μM KA was only 18% of control in the hemolysate (Fig. 8, bottom), in the cytosol it was as high as 60% (Fig. 9, bottom). Since both PNP and GAPDH activities in the rat liver cytosol were blocked (by BCX-1777 and KA, respectively), this finding supports the hypothesis put forward in the preceding paper (Németi and Gregus, 2005) that, besides GAPDH, there is another hitherto unidentified cytosolic enzyme, which is able to reduce AsV and is supported by GSH and NAD(P).
In intact human RBC, irrespective of whether the glycolysis was or was not stimulated with NADH oxidants (pyruvate or ferricyanide), KA was much more effective in inhibiting AsV reductase activity than in the hemolysate or the cytosol. For example, KA concentrations as low as 10 μM caused a near complete inhibition of AsIII formation in the intact erythrocytes (Fig. 10). As to the increased sensitivity of erythrocytes to KA, one may raise the possibility that these cells may concentrate the inhibitor, or GAPDH may be in a conformation in situ that is more susceptible to inhibition by KA. In addition, in the light of the discussion above, it may be speculated that GAPDH in intact erythrocytes is well supplied with NAD and glycolytic substrate, conditions which would give GAPDH a proportionally larger role in AsV reduction in these cells than in the hemolysate without optimal NAD and substrate supply. Nevertheless, it is to be emphasized that our observation on AsV reduction in RBC should not give the impression that erythrocytes play a significant role in AsV reduction in the body. Such a role is severely limited by Pi and chloride ions present in the plasma (but not in our incubation buffer), which inhibit erythrocytic uptake of AsV (Németi and Gregus, 2004).
The mechanism of the GAPDH-catalyzed AsV reduction is unknown. This work, however, provides some clues to put forward a hypothesis, which we present after briefly describing the catalytic cycle of this enzyme. GAPDH is composed of four identical subunits with one molecule of NAD bound strongly to each. During glycolysis, this enzyme catalyzes both the oxidation of Ga-3-P and the incorporation of Pi into Ga-3-P to produce 1,3-BPG. This process involves four main steps. In step 1, Ga-3-P covalently binds to Cys149 of the enzyme forming a thiohemiacetal (Harris and Waters, 1976; Nagradova, 2001; Nagradova and Schmalhausen, 1998). In step 2, the bound Ga-3-P is oxidized into 3-phosphoglyceric acid still bound to Cys149, now with a thioester bond, while NAD is reduced to NADH. In step 3, the NADH formed in step 2 is replaced by NAD. This is necessary because NAD binding induces a conformational change, which will dislocate the phosphate group of the 3-phosphoglyceroyl moiety from the binding site for Pi making it accessible for Pi (Nagradova, 2001) or, in our case, for AsV. In step 4, the thioester bond is cleaved by Pi (termed phosphorolytic cleavage), releasing the 3-phosphoglyceroyl moiety as 1,3-BPG from Cys149 of GAPDH. Instead of Pi, AsV can also cleave this bond (by arsenolytic cleavage) to produce the purportedly unstable 1-arseno-3-phopshoglycerate, instead of 1,3-BPG.
Our observations presented here are compatible with the following tentative suggestions: (1) the enzyme is ready to catalyze the reduction of AsV when it forms the 3-phosphoglyceroyl-enzyme : NAD complex (i.e., it carries bound NAD and the 3-phosphoglyceroyl group bound via a thioester bond to Cys149), and (2) the GAPDH-mediated AsV reduction takes place during, or as a consequence of, the arsenolytic cleavage of the thioester bond formed between the enzyme's Cys149 and the 3-phosphoglyceroyl group of the substrate. The first suggestion is supported by the observation that GAPDH catalyzed the reduction of AsV in the presence of PGK, ATP, 3-PGA, and NAD (Fig. 1). Under this condition, PGK uses ATP to phosphorylate 3-PGA into 1,3-BPG (see insert in Fig. 1), which in turn will acylate GAPDH at Cys149 while releasing Pi. Thus, the critical 3-phosphoglyceroyl-enzyme : NAD complex is formed, which cannot undergo reduction to Ga-3-P because NADH is absent, but can undergo phosphorolysis or arsenolysis. As to the second suggestion, a prerequisite for the phosphorolysis (or arsenolysis) is that the NADH formed as a result of substrate oxidation (step 2) be replaced by NAD. The arsenolytic (or phosphorolytic) cleavage is strongly inhibited by NADH (Nagradova and Schmalhausen, 1998), and so is the GAPDH-catalyzed AsV reduction (Fig. 6), supporting the importance of the arsenolytic cleavage in the reduction of AsV.
When proposing a mechanism for AsV reduction by GAPDH, it is important to note that three thiols cooperate in the process catalyzed by microbial AsV reductases. The E. coli (Martin et al., 2001) and yeast (Mukhopadhyay et al., 2000) AsV reductases are monothiol enzymes, for which GSH and glutaredoxin provide the two additional thiol groups, whereas in the Staphylococcus aureus enzyme all three thiols belong to the enzyme (Messens et al., 2002). In this latter enzyme, two thiols bind and reduce AsV to AsIII, while they form a disulfide bond, which will be reduced by the third thiol. The disulfide bond then formed is reduced by thioredoxin, completing a process termed thiol-disulfide cascade. In the microbial reductases, the catalytically important thiol-group is activated by a nearby positive charge. The S. aureus enzyme may functionally model GAPDH, because three thiols are also present and located in GAPDH near the site of the phosphorolytic/arsenolytic cleavage. One of these is GSH, which is associated with GAPDH very strongly (Krimsky and Racker, 1952) and is claimed to protect the 3-phosphoglyceroyl-enzyme : NAD complex against cleavage of the thioester bond by water (i.e., hydrolysis) (Kuzminskaya et al., 1993). Another thiol is the enzyme's Cys153, which is located at the active site of the GAPDH close to the catalytically important Cys149 in the same -helix, and the third is Cys149, which becomes free upon the arsenolytic cleavage of the thioester bond. Like the microbial enzymes, the active site of GAPDH also contains a thiol-activating positive charge on the imidazole ring of His176 (Nagradova and Schmalhausen, 1998). Thus, the three thiols might contribute to the reduction of AsV when it is in 1-arseno-3-phosphoglycerate, or after it is released from the latter unstable compound. The formed AsIII may then be released and complexed by GSH present in large excess in the ambient solution. GSH may also serve to reduce the disulfide bonds formed during AsV reduction. Theoretically, it is also possible that the contribution of GAPDH to the reduction of AsV is indirect, and GSH directly reduces AsV while being in mixed anhydride with 3-PGA as 1-arseno-3-phosphoglycerate. Even then, GAPDH is required, because formation of 1-arseno-3-phosphoglycerate must precede the reduction of AsV. Whether the role of GAPDH in AsV reduction is direct or indirect, the inhibitory effect of KA on the GAPDH-mediated AsV reduction is compatible with our hypotheses, as covalent binding of KA to Cys149 (Sakai et al., 1991) prevents formation of the 3-phosphoglyceroyl-enzyme and 1-arseno-3-phosphoglycerate. Nevertheless, more research is needed on the mechanism of GAPDH-mediated AsV reduction.
In summary, this work demonstrates that GAPDH can fortuitously function as an AsV reductase, provided NAD, glycolytic substrate, and GSH are available. We hypothesize that GAPDH-catalyzed AsV reduction takes place during, or as a result of, the arsenolytic cleavage of its substrate from the thioester bond between the active site cysteine and the substrate. As studies with the specific GAPDH inhibitor KA indicate, GAPDH is exclusively responsible for the PNP-independent AsV reductase activity in human erythrocytes and significantly contributes to the AsV reduction in rat liver cytosol. Further research should clarify whether or not GAPDH contributes to the reduction of AsV to AsIII in vivo.
NOTES
The authors certify that all research involving human subjects was done under full compliance with all government policies and the Helsinki Declaration.
ACKNOWLEDGMENTS
This publication is based on a work supported by the Hungarian National Scientific Research Fund (OTKA) and the Hungarian Ministry of Health. The authors wish to thank Mónika Agyaki and István Schweibert for their excellent assistance in the experimental work.
REFERENCES
Beisswenger, P. J., Howell, S. K., Smith, K., and Szwergold, B. S. (2003). Glyceraldehyde-3-phosphate dehydrogenase activity as an independent modifier of methylglyoxal levels in diabetes. Biochim. Biophys. Acta 1637, 98–106.
Brown, R. E., Jarvis, K. L., and Hyland, K. J. (1989). Protein measurement using bicinchoninic acid: Elimination of interfering substances. Anal. Biochem. 180, 136–139.
Csanaky, I., Németi, B., and Gregus, Z. (2003). Dose-dependent biotransformation of arsenite in rats – not S-adenosylmethionine depletion impairs arsenic methylation at high dose. Toxicology 183, 77–91.
Gregus, Z., Gyurasics, á., and Csanaky, I. (2000). Biliary and urinary excretion of inorganic arsenic: Monomethylarsonous acid as a major biliary metabolite in rats. Toxicol. Sci. 56, 18–25.
Gregus, Z., and Németi, B. (2002). Purine nucleoside phosphorylase as a cytosolic arsenate reductase. Toxicol. Sci. 70, 13–19.
Harris, J. I., and Waters, M. (1976). Glyceraldehyde-3-phosphate dehydrogenase. In The Enzymes, 3rd ed., (P. D. Boyer, Ed.), Vol. 13, pp. 1–49. Academic Press, New York.
Kato, M., Sakai, K., and Endo, A. (1992). Koningic acid (heptelidic acid) inhibition of glyceraldehyde-3-phosphate dehydrogenases from various sources. Biochim. Biophys. Acta 1120, 113–116.
Knowles, F. C., and Benson, A. A. (1983). The biochemistry of arsenic. Trends Biochem. Sci. 8, 178–180.
Krimsky, I., and Racker, E. (1952). Glutathione, a prosthetic group of glyceraldehyde-3-phosphate dehydrogenase. J. Biol. Chem. 198, 721–729.
Kuzminskaya, E. V., Asryants, R. A., and Nagradova, N. K. (1993). SH-containing compounds as allosteric effectors of glyceraldehyde-3-phosphate dehydrogenase. FEBS Lett. 336, 208–210.
Martin, P., DeMel, S., Shi, J., Gladysheva, T., Gatti, D. L., Rosen, B. P., and Edwards, B. F. P. (2001). Insights into the structure, solvation, and mechanism of ArsC arsenate reductase, a novel arsenic detoxification enzyme. Structure 9, 1071–1081.
Messens, J., Martins, J. C., Van Belle, K., Brosens, E., Desmyter, A., De Gieter, M., Wieruszeski, J. M., Willem, R., Wyns, L., and Zegers, I. (2002). All intermediates of the arsenate reductase mechanism, including an intramolecular dynamic disulfide cascade. Proc. Natl. Acad. Sci. U.S.A. 99, 8506–8511.
Mukhopadhyay, R., Shi, J., and Rosen, B. P. (2000). Purification and characterization of Acr2p, the Saccharomyces cerevisiae arsenate reductase. J. Biol. Chem. 275, 21149–21157.
Nagradova, N. K. (2001). Study of the properties of phosphorylating D-glyceraldehyde-3-phosphate dehydrogenase. Biochemistry (Moscow) 66, 1067–1076.
Nagradova, N. K., and Schmalhausen, E. V. (1998). D-glyceraldehyde-3-phosphate dehydrogenase: Structural basis and functional role of the acyl transfer reactions. Biochemistry (Moscow) 63, 504–515.
Nakamura, M., Fujiwara, A., Yasumasu, I., Okinaga, S., and Arai, K. (1982). Regulation of glucose metabolism by adenine nucleotides in round spermatids from rat testes. J. Biol. Chem. 257, 13945–13950.
Németi, B., Csanaky, I., and Gregus, Z. (2003). Arsenate reduction in human erythrocytes and rats – Testing the role of purine nucleoside phosphorylase. Toxicol. Sci. 74, 22–31.
Németi, B., and Gregus, Z. (2002). Mitochondria work as reactors in reducing arsenate to arsenite. Toxicol. Appl. Pharmacol. 182, 208–218.
Németi, B., and Gregus, Z. (2004). Glutathione-dependent reduction of arsenate in human erythrocytes – a process independent of purine nucleoside phosphorylase. Toxicol. Sci. 82, 419–428.
Németi, B., and Gregus, Z. (2005). Reduction of arsenate to arsenite by human erythrocyte lysate and rat liver cytosol – Characterization of a glutathione- and NAD-dependent arsenate reduction linked to glycolysis. Toxicol. Sci. 85, 000–000.
Radabaugh, T. R., Sampayo-Reyes, A., Zakharyan, R. A., and Aposhian, H. V. (2002). Arsenate reductase II. Purine nucleoside phosphorylase in the presence of dihydrolipoic acid is a route for reduction of arsenate to arsenite in mammalian systems. Chem. Res. Toxicol. 15, 692–698.
Sakai, K., Hasumi, K., and Endo, A. (1988). Inactivation of rabbit muscle glyceraldehyde-3-phosphate dehydrogenase by koningic acid. Biochim. Biophys. Acta 952, 297–303.
Sakai, K., Hasumi, K., and Endo, A. (1991). Identification of koningic acid (heptelidic acid)-modified site in rabbit muscle glyceraldehyde-3-phosphate dehydrogenase. Biochim. Biophys. Acta 1077, 192–196.
Srivastava, S. K., and Beutler, E. (1972). The effect of normal red cell constituents on the activities of red cell enzymes. Arch. Biochem. Biophys. 148, 249–255.
Thomas, D. J., Styblo, M., and Lin S. (2001). The cellular metabolism and systemic toxicity of arsenic. Toxicol. Appl. Pharmacol. 176, 127–144.
Tilton, W. M., Seaman, C., Carriero, D., and Piomelli, S. (1991). Regulation of glycolysis in the erythrocyte: Role of the lactate/pyruvate and NAD/NADH ratios. J. Lab. Clin. Med. 118, 146–152.
Tomschy, A., Glockshuber, R., and Jaenicke, R. (1993). Functional expression of D-glyceraldehyde-3-phosphate dehydrogenase from the hyperthermophilic eubacterium Thermotoga maritima in Escherichia coli. Eur. J. Biochem. 214, 43–50.
Vahter, M. (1983). Metabolism of arsenic. In Biological and Environmental Effects of Arsenic (B. A. Fowler, Ed.), pp. 170–197. Elsevier Science, New York.(Zoltán Gregus and Balázs )