kfm212 44..53 by Ilhamjan Sabit

TOXICOLOGICAL SCIENCES 100(1), 44–53 (2007) doi:10.1093/toxsci/kfm212 Advance Access publication August 9, 2007

Glutathione-Dependent Reduction of Arsenate by Glycogen Phosphorylase—Responsiveness to Endogenous and Xenobiotic Inhibitors Zoltań Gregus1 and Bala´zs Ne´meti Department of Pharmacology and Pharmacotherapy, Toxicology Section, University of Pećs, Medical School, Szigeti u´t 12, H-7624 Pećs, Hungary Received June 13, 2007; accepted August 2, 2007

Rabbit muscle glycogen phosphorylase-a (GPa) reduces arsenate (As(V)) to the more toxic arsenite (As(III)) in a glutathione (GSH)-dependent fashion. To determine whether reduction of As(V) by GPa is countered by compounds known to inhibit GPcatalyzed glycogenolysis, the effects of thiol reagents, endogenous compounds (glucose, ATP, ADP) as well as nonspecific glycogen phosphorylase inhibitors (GPIs; caffeine, quercetin, flavopiridol [FP]), and specific GPIs (1,4-dideoxy-1,4-imino-D-arabinitol [DAB], BAY U6751, CP320626) were tested on reduction of As(V) by rabbit muscle GPa in the presence of glycogen (substrate), AMP (activator), and GSH, and the As(III) formed from As(V) was quantified by high-performance liquid chromatography-hydride generation-atomic fluorescence spectrometry. The As(V)-reducing activity of GPa was moderately sensitive to thiol reagents. Glucose above 5mM and ADP or ATP at physiological levels diminished GPa-catalyzed As(V) reduction. All GPIs inhibited As(V) reduction by GPa in a concentration-dependent fashion; however, their effects were differentially affected by glucose (10mM) or AMP (200mM instead of 25mM), known modulators of the action of some GPIs on the GP-catalyzed glycogenolysis. Inhibition of As(V) reduction by DAB and quercetin was not influenced by glucose or AMP. Glucose that potentiates the inhibitory effects of caffeine, BAY U6751, and CP320626 on the glycogenolytic activity of GPa also enhanced the inhibitory effects of these GPIs on GPa-catalyzed As(V) reduction. AMP at high concentration alleviated the inhibition by BAY U6751 and CP320626 (whose antagonistic effect on GP-catalyzed glycogen breakdown is also AMP sensitive), whereas the inhibition in As(V) reduction by FP or caffeine was little affected by AMP. Thus, GPIs inhibit both the glycogenolytic and As(V)-reducing activities of GP, supporting that the latter is coupled to glycogenolysis. It was also shown that a GPa-rich extract of rat liver contained GSHdependent As(V)-reducing activity that was inhibited by specific GPIs, suggesting that the liver-type GPa can also catalyze reduction of As(V). Key Words: arsenate; reduction; glycogen phosphorylase; glycogen phosphorylase inhibitors; glutathione.

To whom correspondence should be addressed. Fax: 36-72-536218. E-mail: zoltan.gregus@aok.pte.hu.

Exposure of arsenic-contaminated drinking water is a health hazard in several geographical areas (Yoshida et al., 2004). Under oxygenated conditions, well water contains predominantly arsenate (As(V)), a pentavalent inorganic arsenical (Cullen and Reimer, 1989). Reduction of As(V) to the much more toxic trivalent arsenite (As(III)) readily takes place in the body in a glutathione (GSH)-dependent manner (Csanaky and Gregus, 2005; Thomas et al., 2001). Without undergoing this biotransformation, As(V) would probably cause harm, by mimicking inorganic phosphate (Pi), only at extremely high levels of exposure. Two mammalian enzymes that can catalyze reduction of As(V) to As(III) have been reported on, namely purine nucleoside phosphorylase (PNP; Gregus and Ne´meti, 2002; Radabaugh et al., 2002) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; Gregus and Ne´meti, 2005); however, testing their contribution to reduction of As(V) in rats with the use of their specific inhibitors/inactivators administered individually have yielded negative or not entirely conclusive results (Ne´meti et al., 2003; Ne´meti et al., 2006). Nevertheless, the recognition that (1) PNP and GAPDH are phosphorolytic enzymes (i.e., they cleave substrate by Pi), which can also mediate arsenolysis when Pi is replaced with As(V), and that (2) the arsenolytic reaction catalyzed by either PNP or GAPDH is coupled to reduction of As(V) to As(III) in the presence of GSH has lead us to identification of yet another phosphorolytic/arsenolytic enzyme with GSH-dependent As(V)-reducing activity, namely glycogen phosphorylase-a (GPa; Ne´meti and Gregus, unpublished data). GP cleaves off a terminal glucose unit from glycogen by using Pi to produce glucose-1-phosphate; however, it forms glucose-1-arsenate when As(V) replaces Pi (Helmreich and Cori, 1964). Our experiments with rabbit muscle GPa supported the hypothesis that this enzyme catalyzes reduction of As(V) during arsenolysis of glycogen (Ne´meti and Gregus, 2007). However, we deemed it necessary to seek further support for this hypothesis and to determine whether the hepatic form of GPa can also mediate the reduction of As(V). To reach the first goal, we examined the responsiveness of As(V) reduction mediated by rabbit muscle GPa to a variety of

ARSENATE REDUCTION BY PHOSPHORYLASE, PART II

endogenous compounds, such as glucose, ADP, ATP, and glucose-6-phosphate (glucose-6P), and xenobiotics, such as thiol reagents, nonspecific GPIs, and specific GPIs (Fig. 3), which have been shown by others to inhibit the phosphorolytic/ arsenolytic breakdown of glycogen by GP. This approach with specific inhibitors of PNP and GAPDH (i.e., BCX-1777 and koningic acid, respectively) has been successfully applied in supporting the view that the phosphorolytic/arsenolytic activity of these enzymes is essential in their As(V)-reducing activities (Gregus and Ne´meti, 2002; Gregus and Ne´meti, 2005). Furthermore, as glucose and AMP are known to modulate inhibitory effects of GPIs on its glycogenolytic activity, it was also examined if these endogenous GP effectors similarly affected the As(V)-reducing activity of this enzyme. To reach the second objective, we tested whether the GPa-rich glycogen pellet prepared from rat liver could form As(III) from As(V) in a PNP- and GAPDH-independent manner and, if it could, whether this As(V) reduction was sensitive to specific GPIs. As(III) formed from As(V) in these assays was quantified by high-performance liquid chromatography—hydride generation— atomic fluorescence spectrometry (HPLC-HG-AFS).

MATERIALS AND METHODS Chemicals. GPa from rabbit muscle, glycogen (from rabbit liver, type III), glycerol-2-phosphate, AMP, ADP, ATP, 1,4-dideoxy-1,4-imino-D-arabinitol (DAB), quercetin, and p-chloromercuribenzene sulfonic acid were purchased from Sigma-Aldrich. Reduced glutathione, glucose-6-phospate (glucose-6P), and disodium hydrogen arsenate (As(V)) were from Reanal Ltd. (Budapest, Hungary). Caffeine of Pharmacopoea Europaea quality was from the pharmacy, glucagon (GlucaGen) was the product of Novo Nordisk A/G, and calyculin A was purchased from LC Laboratories (Woburn, MA). The following enzyme inhibitors were generous gifts: BAY U6751 from Bayer HealthCare (Leverkusen, Germany), CP320626 from Pfizer (Groton, CT), flavopiridol (FP) hydrochloride from Sanofi-Aventis U.S. (Malvern, PA), BCX-1777 (also called Immucillin-H) from BioCryst Pharmaceuticals (Birmingham, AL), and koningic acid (also called heptelidic acid) from Professor Keiji Hasumi (Tokyo Noko University, Tokyo, Japan). The sources of chemicals used in arsenic speciation (Csanaky et al., 2003; Ne´meti and Gregus, 2002) and in assays for GP (Ne´meti and Gregus, 2007), PNP (Gregus and Ne´meti, 2002), and GAPDH (Ne´meti et al., 2006) have been given elsewhere. All other chemicals were of the highest purity commercially available. Testing the chemical responsiveness of As(V) reduction by rabbit muscle GPa. In order to test the effects of compounds on the As(V)-reducing activity of GPa, rabbit muscle GPa (0.5 U/ml) was typically preincubated in a buffer containing 25mM glycerol-2-phosphate, 1mM ethylenediamine tetraacetic acid (EDTA, pH 7.4) with glycogen (1%), GSH (10mM), and a test compound at 37 C for 5 min. Then AMP (200lM) and As(V) (50lM) were added to start the 60-min incubation. Variations from this procedure and other details of the incubation conditions are given under the figures. The incubation was terminated by addition to the 300-ll incubate of 100 ll 50mM CdSO4 solution followed by 100 ll 1.5M perchloric acid solution containing 50mM HgCl2. The incubates thus treated were stored at – 80 C until arsenic analysis. As(V) reductase activity was corrected for nonenzymatic activity measured when incubating As(V) under similar conditions (i.e., in the presence of GSH) but without GPa and expressed as the amount of As(III) formed per minute and unit GPa.

Testing the chemical responsiveness of As(V) reduction by hepatic GPa. In order to test the effect of compounds on hepatic GPa, a GPa-rich extract was prepared from rat liver, characterized with respect to the activity of enzymes known to be capable of reducing As(V) (i.e., GP, GAPDH, and PNP) and tested for As(V) reduction in the presence of known inhibitors of these enzymes. For this purpose, we used male Wistar rats from the SPF breeding house of the University of Pe´cs (Hungary) that were kept under standardized conditions and weighed 300–350 g. All procedures were carried out on animals according to the Hungarian Animals Act (Scientific Procedures, 1998), and the study was approved by the Ethics Committee on Animal Research of the University of Pe´cs. In order to prepare the glycogen pellet (with GPa attached to the glycogen particles), rats were anesthetized by ip injection of a mixture of fentanyl, midazolam, and droperidol (0.045, 4.5, and 5.5 mg/kg, respectively) and laparotomized. Subsequently, they were injected into a terminal branch of the portal vein with calyculin A (25 lg/kg) to inhibit dephosphorylation of GPa by protein phosphatases type-1 and -2A (Ishihara et al., 1989) and 7 min later with glucagon (100 lg/kg) to activate glycogen phosphorylase kinase (Fosgerau et al., 2000; Vandenheede et al., 1976). Ten min after calyculin A administration, the rats were exsanguinated, their liver removed, weighed, and homogenized in three volumes of ice-cold buffer (pH 7.4) containing 25mM glycerol-2-phosphate, 1mM EDTA, and 50nM calyculin A. The homogenate was centrifuged at 4 C, 10,000 3 g for 20 min to obtain the postmitochondrial supernatant, which was then centrifuged at 4 C, 100,000 3 g, for 75 min. Then the supernatant (cytosol) as well as the microsomal pellet was discarded. The surface of translucent glycogen pellet stuck to the bottom of the centrifuge tube was briefly rinsed, and then the pellet was removed and brought into solution by gentle homogenization in 0.25 ml/g liver of ice-cold calyculin A-containing homogenization buffer. The resultant liver extract was assayed for GP, PNP, and GAPDH activities and stored in aliquots at 80 C until assaying for As(V)-reducing activity within a week. The GP activity in the glycogen sediment was assayed in the glycogenolytic direction as described in detail by Ne´meti and Gregus (2007). This spectrophotometric assay measures the formation of NADPH during the GP-limited conversion of glycogen ultimately to 6-phosphogluconate in the presence of excess phosphoglucomutase, glucose-6P dehydrogenase, glucose-1,6-bisphosphate, and NADP. The GAPDH activity in the glycogen pellet was assayed spectrophotometrically based on the decrease of NADH concentration during the GAPDH-limited conversion of 3-phosphoglycerate ultimately to glyceraldehyde-3-phosphate in the presence of excess phosphoglycerate kinase and ATP, as described earlier (Gregus and Ne´meti, 2005). The PNP activity in the glycogen sediment was assayed as described in detail by Gregus and Ne´meti (2002). This spectrophotometric assay measures the formation of uric acid as the final product of the PNP-limited conversion of inosine (via hypoxanthine and xanthine) in the presence of excess Pi and xanthine oxidase. The As(V)-reducing activity of the glycogen sediment was assayed similarly to that of the rabbit muscle GPa (see above) except that the incubation mix was supplemented only with 0.5% glycogen and that it also contained 50nM calyculin A to inhibit possible dephosphorylation of the active GPa to inactive GPb. The volume of dissolved glycogen pellet used in the assay was selected so that the incubation mix should contain 0.5 U/ml pellet-derived GP. Other details of the assay condition are given under Figure 10. Arsenic analysis. The protein precipitant-treated incubates originating from the As(V) reductase assays were centrifuged at 10,000 3 g, 4 C for 10 min. As(III) and As(V) in the resultant supernatants were quantified by HPLC-HG-AFS, using Hamilton PRP X-100 strong anion exchange guard and analytical columns to separate and 60mM sodium phosphate buffer (pH 5.75) at a flow rate of 1.2 ml/min to elute the arsenicals. The details of this analysis have been given elsewhere (Gregus et al., 2000; Ne´meti et al., 2003). Statistics. Data were analyzed using one-way ANOVA followed by Duncan’s test or Student’s t-test with p < 0.05, as the level of significance.

GREGUS AND NE´METI

46 RESULTS

Effects of Thiol Reagents on the As(V)-Reducing Activity of GPa In order to determine whether the As(V)-reducing activity of GPa is dependent on the thiol groups of the enzyme, the effects of two chemicals that are covalently reactive with SH groups were tested on GPa-catalyzed As(V) reduction. Figure 1 demonstrates that when GPa was preincubated with N-ethyl maleimide (NEM) or p-chloromercuribenzene sulfonate the rate of As(III) formation was diminished by 33 and 49%, respectively, compared to the control. The NEM-induced inhibition in As(V) reduction was insignificantly alleviated, whereas the mercurialinduced inhibition was significantly diminished when these reagents were preincubated with GPa in the presence of AMP. Effects of Endogenous GP Inhibitors on the As(V)-Reducing Activity of GPa The effect of endogenous compounds known to impair the activity of GP, that is, glucose, ATP, ADP, and glucose-6P, was tested on the GPa-catalyzed As(V) reduction. Glucose, ATP, and ADP inhibited As(III) formation in a concentrationdependent fashion (Fig. 2). The inhibitory effect of glucose

FIG. 2. Effects of glucose, ADP, and ATP on the As(V)-reducing activity of GPa. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), and glucose or ADP or ATP (at the indicated concentrations) at 37 C for 5 min. Then AMP (200lM) and As(V) (50lM) were added to start the 60-min incubation. Symbols represent As(III) formation rates (mean ± SEM) in three incubations. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate from those observed in the absence of glucose, ADP, and ATP.

became significant at supraphysiological concentrations (10mM and above), whereas ADP and ATP brought about decreases in As(V) reduction at physiological levels (1–2mM and 4mM, respectively). Glucose-6P in the physiologic range (0.1– 1.0mM), however, failed to influence As(V) reduction (data not shown). Effects of Xenobiotic GP Inhibitors on the As(V)-Reducing Activity of GPa

FIG. 1. Effects of thiol reagents in the presence and absence of AMP on the As(V)-reducing activity of GPa. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%) plus NEM (1mM) or p-chloromercuribenzene sulfonate (0.5mM) in the absence or presence of AMP (1mM) at 37 C for 5 min. Then AMP (1mM) was added into preincubations lacking AMP, whereas GSH (10mM) and As(V) (50lM) were added into all preincubations to start the 60-min incubation. Bars represent As(III) formation rates (mean ± SEM) in three incubations. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rates from those observed in the absence of a thiol reagent. NS indicates no significant difference ( p > 0.05).

Several xenobiotics with diverse effects have been reported to inhibit GP, whereas others were developed specifically as GPIs for potential therapeutic application in diabetes. Three nonspecific GPIs and three specific GPIs (Fig. 3) have been tested on the As(V)-reducing activity of rabbit muscle GPa. Because the glycogenolysis-inhibitory action of some GPIs is synergized with glucose and/or counteracted by AMP, the effects of GPIs on the As(V)-reducing activity of GPa were tested under four conditions, that is, in the absence of glucose and presence of 25lM AMP or 200lM AMP and in the presence of 10mM glucose and 25lM AMP or 200lM AMP. The As(V)-reducing activities of GPa under these four conditions in the absence of a GPI were 113 ± 9, 111 ± 9, 76.6 ± 11, and 84.5 ± 8.3 pmol/min/U GP, respectively (n ¼ 19–21). Figures 4–9 demonstrate the As(V)-reducing activity of GPa in response to GPIs as percentage of these control values. Testing of the concentration-dependent effects of GPIs under these conditions permitted several statistical comparisons to be made. For clarity, however, only two types of comparisons are indicated in the figures, that is, (1) the difference caused by rising the AMP concentration from 25lM to 200lM in the absence of glucose (if significant, this difference is marked with an asterisk), and (2) the difference

ARSENATE REDUCTION BY PHOSPHORYLASE, PART II

FIG. 3. The chemical structures of xenobiotic GPI used.

caused by glucose in the presence of 200lM AMP (if significant, this difference is marked with a cross). The classical nonspecific GPI, caffeine, inhibited As(V) reduction by GPa in a concentration-dependent manner (Fig. 4) in the range of 0.25–2mM. In the presence of glucose and 200lM AMP, this methylxanthine compound was more inhibitory (IC50 ~0.25mM) than in the absence of glucose (IC50 ~1.3mM). Increasing AMP from 25 to 200lM slightly alleviated caffeine-induced inhibition in the absence of glucose but not in its presence. Quercetin antagonized GPa-catalyzed As(V) reduction in a concentration-dependent fashion between 5 and 100lM (Fig. 5). Glucose failed to significantly influence the inhibitory effect of this flavonol and so did AMP in general. Although the inhibitory effect of quercetin tended to decrease in the presence of AMP at the higher concentration, this typically did not reach the level of statistical significance. Between 2.5 and 50lM concentrations, FP gradually lowered the As(V)-reducing activity of GPa (Fig. 6). Glucose increased the sensitivity of GPa to this compound, as indicated by the change in approximate IC50 of FP from 13lM in the absence of glucose to 3lM in the presence of glucose. In the presence of 200lM AMP, FP tended to inhibit the GPa-catalyzed As(V) reduction less than in the presence 25lM AMP, with the differences reaching statistical significance only at certain drug concentrations. Each of the three specific GPIs tested impaired the As(V)reducing activity of rabbit muscle GPa in a concentrationdependent fashion in the range of 2.5–25lM. The inhibition produced by DAB was not influenced significantly either by the presence of glucose or by the increase in AMP level from 25 to 200lM (Fig. 7). In contrast, BAY U6751 became markedly more inhibitory when 10mM glucose was also present or when the concentration of AMP was lowered from 200 to 25lM (Fig. 8). For example, the approximate IC50 of BAY

U6751 in the absence of glucose and presence of 200lM AMP was 8.6lM. This value decreased to 2.4lM by adding glucose and to 1.5lM by decreasing the AMP level to 25lM. The inhibitory effect of CP320626 on As(V) reduction by GPa was also positively responsive to glucose and lowered AMP level (Fig. 9), as indicated by the shift in IC50 of this GPI compound from 5.9lM in the absence of glucose and the presence of 200lM AMP, to 1.8lM in the presence of glucose (10mM) with AMP (200lM), and to 1lM in the absence of glucose with only 25lM AMP. Effects of Specific GP Inhibitors on the As(V)-Reducing Activity of a GPa-Rich Extract from Rat Liver Because there are differences in structure and regulation between the muscle-type and liver-type GPs, it was of interest to test whether As(V) reduction is also catalyzed by liver-type GPa. For this reason, a GPa-rich extract was prepared from rat liver, characterized for activities of enzymes hitherto shown to be capable of reducing As(V) to As(III) (i.e., GAPDH, PNP, and GP), and tested for As(V) reduction as well as inhibition of As(V) reduction by specific inhibitors of these enzymes. The GPa-rich extract (i.e., the glycogen pellet from the liver of glucagon- and calyculin A–injected fed rats dissolved in a volume of 0.25 ml/g wet weight) exhibited 641 ± 14, 327 ± 1, and 1565 ± 39 U/ml GAPDH, PNP, and GP activities, respectively. When incubated with As(V) in the presence of added GSH, glycogen, and AMP, this extract formed As(III) from As(V) (Fig. 10). Omission of GSH abolished As(V)-reducing activity, whereas omission of exogenous glycogen or AMP diminished it by 20%. Addition of koningic acid, an inhibitor of GAPDH, to the complete incubation mix did not impair As(III) formation by the GParich liver extract, whereas addition of BCX-1777, an inhibitor of PNP, decreased it by 44%. GPIs at 50lM also significantly

GREGUS AND NE´METI

FIG. 4. Effect of caffeine on the As(V)-reducing activity of GPa as influenced by glucose and AMP. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), glucose (0 or 10mM), and caffeine (at the indicated concentrations) at 37 C for 5 min. Then AMP (25 or 200lM) and As(V) (50lM) were added to start the 60-min incubation. Symbols represent As(III) formation rates as percent of control (mean ± SEM) in four incubations. The control As(III) formation rates in the absence of glucose and presence of 25 and 200lM AMP were 113 ± 9 and 111 ± 9 pmol/min/U GP, respectively, whereas the control As(III) formation rates in the presence of glucose with 25 or 200lM AMP were 76.6 ± 11 and 84.5 ± 8.3 pmol/min/U GP, respectively. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate observed in the absence of glucose caused by rising the AMP concentration from 25 to 200lM. Crosses indicate significant difference ( p < 0.05) in As(III) formation rate observed in the presence of 200lM AMP caused by glucose.

lowered the As(V)-reducing activity of the glycogen pellet; DAB by 37%, CP320626 by 44%, BAY U6751 by 46%, and FP by 26%. In the presence of BCX-1777, however, these inhibitors, were comparatively more effective, as they impaired As(V)-reducing activity by 74, 73, 74, and 54%, respectively. DAB at 100lM caused an even larger inhibition (82%), whereas 200lM DAB virtually abolished the As(V)-reducing activity of the glycogen extract in the presence of the PNP inhibitor (data not shown).

FIG. 5. Effect of quercetin on the As(V)-reducing activity of GPa as influenced by glucose and AMP. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), glucose (0 or 10mM), and quercetin (at the indicated concentrations) at 37 C for 5 min. Then AMP (25 or 200lM) and As(V) (50lM) were added to start the 60-min incubation. Symbols represent As(III) formation rates as percent of control (mean ± SEM) in three incubations. The control As(III) formation rates are given under Figure 4. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate observed in the absence of glucose caused by raising the AMP concentration from 25 to 200lM.

DISCUSSION

Although the observations presented in this paper support a relationship between the glycogenolytic and As(V)-reducing activities of GP, their interpretation requires some insight into the structural and functional properties of this enzyme. GP is known as a highly regulated homodimeric enzyme. Its activity can be altered allosterically (i.e., by perturbation of its tertiary and quaternary structure) induced by covalent phosphorylation of Ser14 (whereby GPb is converted into more active GPa) and by noncovalent binding of endogenous metabolites (which may activate or inhibit this enzyme). A most relevant event of such structural changes is moving the so-called active site gate

ARSENATE REDUCTION BY PHOSPHORYLASE, PART II

FIG. 6. Effect of FP on the As(V)-reducing activity of GPa as influenced by glucose and AMP. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), glucose (0 or 10mM), and FP (at the indicated concentrations) at 37 C for 5 min. Then AMP (25 or 200lM) and As(V) (50lM) were added to start the 60-min incubation. Symbols represent As(III) formation rates as percent of control (mean Âą SEM) in four incubations. The control As(III) formation rates are given under Figure 4. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate observed in the absence of glucose caused by rising the AMP concentration from 25 to 200lM. Crosses indicate significant difference ( p < 0.05) in As(III) formation rate observed in the presence of 200lM AMP caused by glucose.

(or 280s loop) away from or to the active site, thereby making this site accessible to the substrates, that is, glycogen and Pi, (relaxed or R state) or inaccessible to them (tight or T state). Glucose, for example, while occupying the substrate-binding site in the catalytic center of GP, induces the T state of the enzyme and inhibits its glycogenolytic activity (Johnson, 1992). Conversely, phosphorylation of GP at Ser14 or its interaction with AMP at its intersubunit binding site induces the R state, thereby allowing glycogen and Pi to enter the catalytic site and activate the enzyme (Buchbinder et al., 2001). ATP, ADP, and glucose-6P inhibit GP by displacing the activator AMP from its allosteric activator site (Buchbinder

FIG. 7. Effect of DAB on the As(V)-reducing activity of GPa as influenced by glucose and AMP. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), glucose (0 or 10mM), and DAB (at the indicated concentrations) at 37 C for 5 min. Then AMP (25 or 200lM) and As(V) (50lM) were added to start the 60-min incubation. The control As(III) formation rates are given under Figure 4. Symbols represent As(III) formation rates as percent of control (mean Âą SEM) in three incubations.

et al., 2001). Our observations that glucose, ATP, and ADP diminishes the GSH- and glycogen-dependent reduction of As(V) by rabbit muscle GPa suggest that the glycogenolytic activity of GPa determines the rate of this reduction. Furthermore, the findings that inhibitory effects of ATP and ADP on GPa-supported As(V) reduction are moderate (Fig. 2) whereas glucose-6P lacks such an effect (data not shown) are in accordance with the assertion that while these metabolites are potent inhibitors of GPb their effects on GPa are far less pronounced (Oikonomakos et al., 1999). GP contains a number of cysteine residues; however, none of them contributes to the catalytic site (Newgard et al., 1989). This probably accounts for the slow inactivation of rabbit muscle GPa by iodoacetamide (Avramovic-Zikic et al., 1970) and for the comparatively low sensitivity of the As(V)

GREGUS AND NE´METI

FIG. 8. Effect of BAY U6751 on the As(V)-reducing activity of GPa as influenced by glucose and AMP. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), glucose (0 or 10mM), and BAY U6751 (at the indicated concentrations) at 37 C for 5 min. Then AMP (25 or 200lM) and As(V) (50lM) were added to start the 60-min incubation. Symbols represent As(III) formation rates as percent of control (mean ± SEM) in four incubations. The control As(III) formation rates are given under Figure 4. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate observed in the absence of glucose caused by rising the AMP concentration from 25 to 200lM. Crosses indicate significant difference ( p < 0.05) in As(III) formation rate observed in the presence of 200lM AMP caused by glucose.

FIG. 9. Effect of CP320626 on the As(V)-reducing activity of GPa as influenced by glucose and AMP. Rabbit muscle GPa (0.5 U/ml) was preincubated with glycogen (1%), GSH (10mM), glucose (0 or 10mM), and CP320626 (at the indicated concentrations) at 37 C for 5 min. Then AMP (25 or 200lM) and As(V) (50lM) were added to start the 60-min incubation. Symbols represent As(III) formation rates as percent of control (mean ± SEM) in three incubations. The control As(III) formation rates are given under Figure 4. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate observed in the absence of glucose caused by rising the AMP concentration from 25 to 200lM. Crosses indicate significant difference ( p < 0.05) in As(III) formation rate observed in the presence of 200lM AMP caused by glucose.

reduction supported by this enzyme to other thiol reagents (Fig. 1). This contrasts with the high susceptibility to mercurials of the As(V) reduction catalyzed by PNP (Gregus and Ne´meti, 2002) whose catalytic Pi-binding site possesses a cysteine. It is interesting to note that AMP can diminish thiol reagent–induced inactivation of GPa with respect to both its physiological enzymatic activity (Avramovic-Zikic et al., 1970) and its As(V)-reducing activity (Fig. 1), probably because the AMP-binding site of muscle GP contains a cysteine (Cys318) with reactive thiol group (Fletterick and Madsen, 1980).

There are a number of xenobiotics that have been recognized by chance as GPI or developed to specifically inhibit GP for using them as blood glucose–lowering drugs in non-insulin– dependent diabetes (Henke and Sparks, 2006; Somsa´k et al., 2003; Oikonomakos, 2002; Fig. 3). Many of these compounds bind to specific allosteric sites on GP, such as the AMP allosteric effector site or AMP site (e.g., BAY U6751), the purine (or nucleoside) inhibitor site (e.g., caffeine, FP), and the new allosteric (or indole) site (e.g., CP320626), and thus promote formation of the T state of the enzyme. At variance with this mode of action, the iminosugar DAB (Fig. 3) in its protonated form acts as a transitional state analogue of the

ARSENATE REDUCTION BY PHOSPHORYLASE, PART II

FIG. 10. As(V) reduction by the GPa-rich glycogen pellet of rat liver—effects of omission of GSH, glycogen, or AMP, as well as of addition of compounds inhibiting GAPDH, PNP, or GP. For complete incubation (Compl), the dissolved glycogen pellet of rat liver (at a GP activity of 0.5 U/ml final incubation mix) was preincubated in a buffer containing calyculin A (50nM) with glycogen of rabbit liver origin (0.5%) and GSH (10mM) at 37 C for 5 min. Then, AMP (200lM) and As(V) (50lM) were added to start the 60-min incubation in a final volume of 0.3 ml. To test the effects of various enzyme inhibitors, the following compounds were also included in the preincubation mix when indicated in the figure: koningic acid (KA, 100lM), BCX-1777 (BCX, 20lM), DAB (50lM), CP320626 (CP, 50lM), BAY U6751 (BAY, 50lM), and FP (50lM). Bars represent As(III) formation rates (mean ± SEM) of three glycogen pellets. Asterisks indicate significant difference ( p < 0.05) in As(III) formation rate from the complete incubation performed in the absence of an enzyme inhibitor (the first gray bar in both panels). Crosses indicate significant difference ( p < 0.05) in As(III) formation rate from the complete incubation performed in the presence of BCX-1777 alone (the first black bar in the bottom panel).

terminal glucosyl unit of glycogen that forms an oxonium/ carbonium cation during its phosphorolytic cleavage from the glycogen chain. Therefore, DAB binds tightly to the catalytic site of GP, causing an uncompetitive inhibition of the enzyme (Oikonomakos et al., 2006). As discussed in more detail later,

GPIs may synergize with glucose and/or may be antagonized by AMP in inhibiting the GP-catalyzed glycogen breakdown; therefore, their effects on the GPa-supported As(V) reduction were examined in the presence and absence of glucose as well as in the presence of low and high concentrations of AMP. Testing the effect on GPa-catalyzed As(V) reduction of six GPIs in the presence of different concentrations of glucose and AMP revealed that, although each of these compounds inhibited As(V) reduction by GPa in a concentration-dependent fashion, these GPIs fall into three classes based on the influence of glucose and AMP on their inhibitory effects. One GPI group is represented by quercetin and DAB, whose inhibitory effects on the GPa-catalyzed As(V) reduction were not affected or were only slightly influenced by either glucose or change in AMP concentration (Figs. 5 and 7). In agreement with this observation, it has been shown that the strongly binding transition state substrate analogue DAB inhibits the enzymatic activity of GP independent of the presence of glucose (Andersen and Westergaard, 2002; Henke and Sparks, 2006) and possibly also irrespective of the concentration of AMP. However, neither the mechanism nor the responsiveness to glucose or AMP of the inhibitory effect of quercetin on the glycogenolytic activity of GP is known (Jakobs et al., 2006). A second group of GPIs includes caffeine and FP whose inhibitory effects on GPa-supported As(V) reduction were clearly synergized with glucose and slightly antagonized with AMP (Figs. 4 and 6). Similar findings have been reported on caffeine (Andersen and Westergaard, 2002; Kasvinsky et al., 1978) and FP (Kaiser et al., 2001; Oikonomakos et al., 2000) in terms of responsiveness of their inhibitory effects to these endogenous ligands on the physiological activity of GP. Alleviation by AMP of the caffeine- or FP-induced GP inhibition is explained by the fact that AMP is not only a highaffinity ligand at its AMP activator site but can also bind to the purine (or nucleoside) inhibitor site, and at high concentrations, it may thus displace these GPIs interacting with this latter GP domain (Kaiser et al., 2001; Newgard et al., 1989). Glucose is synergistic with these GPIs because it purportedly converts GP from the R state, which lacks the well-configured purine inhibitor site, into the T conformation, in which this site appears (Cuadri-Tome´ et al., 2006). BAY U6751 and CP320626 represent GPIs functionally different from the above-mentioned groups as they exhibited clear synergy with glucose and marked antagonism with AMP in their inhibitory effects on the As(V)-reducing activity of GPa (Figs. 8 and 9). CP320626, or its congener CP91149 (Andersen and Westergaard, 2002; Henke and Sparks, 2006; Oikonomakos et al., 1999), and BAY U6751, or its active enantiomer BAY W1807 (Bergans et al., 2000), are characterized by similar properties with respect to glucose- and AMP-responsiveness of their inhibitory effects on the GP-catalyzed glycogen breakdown. These GPIs act synergistically with glucose probably because they also stabilize the T conformation of the enzyme but act at different sites (Oikonomakos

GREGUS AND NE´METI

et al., 2000). While the mechanism whereby AMP relieves the inhibition of GP by CP320626 remains to be explained, the antagonism between AMP and the active enantiomer of BAY U6751 is attributed to the fact that the latter GPI competes with AMP for the allosteric AMP activator site (Zographos et al., 1997). Although the muscle and liver GPs catalyze the same reaction and possess identical amino acids at the catalytic center, they exhibit differences, for example, in structure, in responsiveness to the AMP and other allosteric effectors, and in specific activity, which is threefold higher for the muscle enzyme than for the hepatic GPa (Johnson, 1992). Therefore, it was of interest to determine whether or not As(V) reduction is supported not only by the muscle GPa but also by the liver enzyme. Our approach was indirect; it was tested whether the glycogen pellet of rat liver homogenate rich in GPa and supplemented with GSH and AMP exhibited As(V)-reducing activity that could be inhibited by GPIs. DAB (Henke and Sparks, 2006), CP320626 (Hoover et al., 1998), and BAY U6751 (Bergans et al., 2000) are known to inhibit the glycogenolytic activity of both hepatic and muscle GPa. Our studies indicated that the liver extract did contain GSHdependent As(V)-reducing activity that was lowered by GPIs. However, part of this activity resulted from the presence of PNP in the glycogen pellet because BCX-1777, a specific inhibitor of PNP-catalyzed As(V) reduction (Gregus and Ne´meti, 2002), also diminished reduction of As(V) by the pellet (Fig. 10). In the presence of BCX-1777, however, addition of GPIs further diminished As(III) formation by the glycogen pellet, indicating that it does contain GPa-catalyzed As(V)-reducing activity. Interestingly, the glycogen preparation from rat liver also contained GAPDH activity; however, the GAPDH inhibitor koningic acid that blocks the GAPDH-catalyzed As(V) reduction (Gregus and Ne´meti, 2005) did not lower the As(V)-reducing activity of the glycogen pellet (Fig. 10), indicating that under the assay conditions GAPDH did not function as an As(V) reductase, probably because of the absence of a glycolytic substrate. In conclusion, this study demonstrates that endogenous compounds (e.g., glucose, ATP, ADP) and xenobiotics (e.g., thiol-reactive chemicals, nonspecific and specific GP inhibitor drugs) that are known to decrease the glycogenolytic activity of GP also diminish the As(V)-reducing activity of rabbit muscle GPa and, at least the tested GPIs, the activity of the GPa-rich glycogen pellet of rat liver homogenate. Moreover, modulation by AMP or glucose of the inhibitory effects of GPIs on the GPa-mediated As(V) reduction mimicked the reported alterations by these ligands in the inhibitory effects of GPIs on the GP-catalyzed glycogenolysis. Therefore, in addition to the findings of the companion paper (Ne´meti and Gregus, 2007), the observations presented in this article further support the contention that As(V) is reduced by GPa in the course or as a consequence of arsenolytic glycogen breakdown. Theoretically, the capacity of specific GPIs to antagonize the

As(V)-reducing activity of GPa may be exploited experimentally for assessing whether or not this enzyme contributes to reduction of As(V) in vivo. However, the use of GPIs for this purpose may not be definitely expedient. This cautionary note is justified because GPa is so far the third of the known mammalian enzymes (besides PNP and GAPDH) which can catalyze the phosphorolytic/arsenolytic cleavage of their substrates with simultaneous reduction of As(V) and which can potentially contribute to reduction of As(V) to the much more harmful As(III) in the body.

FUNDING

Hungarian National Research Fund (OTKA) and the Hungarian Ministry of Health.

ACKNOWLEDGMENTS The authors wish to thank Gyo¨ngyi Sz}oke and Istva´n Schweibert for their excellent assistance in the experimental work. Conflicts of Interest: None declared.

REFERENCES Andersen, B., and Westergaard, N. (2002). The effect of glucose on the potency of two distinct glycogen phosphorylase inhibitors. Biochem. J. 367, 443–450. Avramovic-Zikic, O., Smillie, L. B., and Madsen, N. B. (1970). The sulfhydryl groups of muscle phosphorylase. IV. Reactivities as related to changes in protein structure. J. Biol. Chem. 245, 1558–1565. Bergans, N., Stalmans, W., Goldmann, S., and Vanstapel, F. (2000). Molecular mode of inhibition of glycogenolysis in rat liver by the dihydropyridine derivative, BAY R3401: Inhibition and inactivation of glycogen phosphorylase by an activated metabolite. Diabetes 49, 1419–1426. Buchbinder, J. L., Rath, V. L., and Fletterick, R. J. (2001). Structural relationships among regulated and unregulated phosphorylases. Annu. Rev. Biophys. Biomol. Struct. 30, 191–209. Csanaky, I., and Gregus, Z. (2005). Role of glutathione in reduction of arsenate and of gamma-glutamyltranspeptidase in disposition of arsenite in rats. Toxicology 207, 91–104. Csanaky, I., Ne´meti, 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. Cuadri-Tome, C., Baron, C., Jara-Perez, V., Parody-Morreale, A., Martinez, J. C., and Camara-Artigas, A. (2006). Kinetic analysis and modelling of the allosteric behaviour of liver and muscle glycogen phosphorylases. J. Mol. Recognit. 19, 451–457. Cullen, W. R., and Reimer, K. J. (1989). Arsenic speciation in the environment. Chem. Rev. 89, 713–764. Fletterick, R. J., and Madsen, N. B. (1980). The structures and related functions of phosphorylase a. Annu. Rev. Biochem. 49, 31–61. Fosgerau, K., Westergaard, N., Quistorff, B., Grunnet, N., Kristiansen, M., and Lundgren, K. (2000). Kinetic and functional characterization of 1,4-dideoxy-1, 4-imino-d-arabinitol: A potent inhibitor of glycogen phosphorylase with antihyperglycaemic effect in ob/ob mice. Arch. Biochem. Biophys. 380, 274–284.

ARSENATE REDUCTION BY PHOSPHORYLASE, PART II ´ ., and Csanaky, I. (2000). Biliary and urinary excretion Gregus, Z., Gyurasics, A of inorganic arsenic: Monomethylarsonous acid as a major biliary metabolite in rats. Toxicol. Sci. 56, 18–25. Gregus, Z., and Ne´meti, B. (2002). Purine nucleoside phosphorylase as a cytosolic arsenate reductase. Toxicol. Sci. 70, 13–19.

Ne´meti, B., and Gregus, Z. (2007). Glutathione-dependent reduction of arsenate by glycogen phosphorylase - A reaction coupled to glycogenolysis. Toxicol. Sci. Advance Access published August 9, 2007, 10.1093/toxsci/kfm211. Newgard, C. B., Hwang, P. K., and Fletterick, R. J. (1989). The family of glycogen phosphorylases: Structure and function. Crit. Rev. Biochem. Mol. Biol. 24, 69–99.

Gregus, Z., and Ne´meti, B. (2005). The glycolytic enzyme glyceraldehyde-3phosphate dehydrogenase works as an arsenate reductase in human red blood cells and rat liver cytosol. Toxicol. Sci. 85, 859–869.

Oikonomakos, N. G. (2002). Glycogen phosphorylase as a molecular target for type 2 diabetes therapy. Curr. Protein Pept. Sci. 3, 561–586.

Helmreich, E., and Cori, C. F. (1964). The effects of pH and temperature on the kinetics of phosphorylase reaction. Proc. Natl. Acad. Sci. USA 52, 647–654.

Oikonomakos, N. G., Schnier, J. B., Zographos, S. E., Skamnaki, V. T., Tsitsanou, K. E., and Johnson, L. N. (2000). Flavopiridol inhibits glycogen phosphorylase by binding at the inhibitor site. J. Biol. Chem. 275, 34566–34573.

Henke, B. R., and Sparks, S. M. (2006). Glycogen phosphorylase inhibitors. Mini Rev. Med. Chem. 6, 845–857.

Oikonomakos, N. G., Tiraidis, C., Leonidas, D. D., Zographos, S. E., Kristiansen, M., Jessen, C. U., Norskov-Lauritsen, L., and Agius, L. (2006). Iminosugars as potential inhibitors of glycogenolysis: Structural insights into the molecular basis of glycogen phosphorylase inhibition. J. Med. Chem. 49, 5687–5701. Oikonomakos, N. G., Tsitsanou, K. E., Zographos, S. E., Skamnaki, V. T., Goldmann, S., and Bischoff, H. (1999). Allosteric inhibition of glycogen phosphorylase a by the potential antidiabetic drug 3-isopropyl 4-(2chlorophenyl)-1,4-dihydro-1-ethyl-2-methyl-pyridine-3,5,6-tricarbo xylate. Protein Sci. 8, 1930–1945.

Hoover, D. J., Lefkowitz-Snow, S., Burgess-Henry, J. L., Martin, W. H., Armento, S. J., Stock, I. A., McPherson, R. K., Genereux, P. E., Gibbs, E. M., and Treadway, J. L. (1998). Indole-2-carboxamide inhibitors of human liver glycogen phosphorylase. J. Med. Chem. 41, 2934–2938. Ishihara, H., Martin, B. L., Brautigan, D. L., Karaki, H., Ozaki, H., Kato, Y., Fusetani, N., Watabe, S., Hashimoto, K., Uemura, D., et al. (1989). Calyculin A and okadaic acid: Inhibitors of protein phosphatase activity. Biochem. Biophys. Res. Commun. 159, 871–877. Jakobs, S., Fridrich, D., Hofem, S., Pahlke, G., and Eisenbrand, G. (2006). Natural flavonoids are potent inhibitors of glycogen phosphorylase. Mol. Nutr. Food Res. 50, 52–57. Johnson, L. N. (1992). Glycogen phosphorylase: Control by phosphorylation and allosteric effectors. FASEB J. 6, 2274–2282. Kaiser, A., Nishi, K., Gorin, F. A., Walsh, D. A., Bradbury, E. M., and Schnier, J. B. (2001). The cyclin-dependent kinase (CDK) inhibitor flavopiridol inhibits glycogen phosphorylase. Arch. Biochem. Biophys. 386, 179–187.

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. Somsak, L., Nagy, V., Hadady, Z., Docsa, T., and Gergely, P. (2003). Glucose analog inhibitors of glycogen phosphorylases as potential antidiabetic agents: Recent developments. Curr. Pharm. Des. 9, 1177–1189. Thomas, D. J., Styblo, M., and Lin, S. (2001). The cellular metabolism and systemic toxicity of arsenic. Toxicol. Appl. Pharmacol. 176, 127–144.

Kasvinsky, P. J., Shechosky, S., and Fletterick, R. J. (1978). Synergistic regulation of phosphorylase a by glucose and caffeine. J. Biol. Chem. 253, 9102–9106.

Vandenheede, J. R., Keppens, S., and De Wulf, H. (1976). The activation of liver phosphorylase b kinase by glucagon. FEBS Lett. 61, 213–217.

Ne´meti, 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.

Yoshida, T., Yamauchi, H., and Sun, G. F. (2004). Chronic health effects in people exposed to arsenic via the drinking water: Dose-response relationships in review. Toxicol. Appl. Pharmacol. 198, 243–252.

Ne´meti, B., Csanaky, I., and Gregus, Z. (2006). Effect of an inactivator of glyceraldehyde-3-phosphate dehydrogenase, a fortuitous arsenate reductase, on disposition of arsenate in rats. Toxicol. Sci. 90, 49–60.

Zographos, S. E., Oikonomakos, N. G., Tsitsanou, K. E., Leonidas, D. D., Chrysina, E. D., Skamnaki, V. T., Bischoff, H., Goldmann, S., Watson, K. A., and Johnson, L. N. (1997). The structure of glycogen phosphorylase b with an alkyldihydropyridine-dicarboxylic acid compound, a novel and potent inhibitor. Structure 5, 1413–1425.

Ne´meti, B., and Gregus, Z. (2002). Mitochondria work as reactors in reducing arsenate to arsenite. Toxicol. Appl. Pharmacol. 182, 208–218.