Glutathione levels and sensitivity to apoptosis are regulated by changes in transaldolase expression.

Transaldolase (TAL) is a key enzyme of the reversible nonoxidative branch of the pentose phosphate pathway (PPP) that is responsible for the generation of NADPH to maintain glutathione at a reduced state (GSH) and, thus, to protect cellular integrity from reactive oxygen intermediates (ROIs). Formation of ROIs have been implicated in certain types of apoptotic cell death. To evaluate the role of TAL in this process, Jurkat human T cells were permanently transfected with TAL expression vectors oriented in the sense or antisense direction. Overexpression of TAL resulted in a decrease in glucose 6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase activities and NADPH and GSH levels and rendered these cells highly susceptible to apoptosis induced by serum deprivation, hydrogen peroxide, nitric oxide, tumor necrosis factor-α, and anti-Fas monoclonal antibody. In addition, reduced levels of TAL resulted in increased glucose 6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase activities and increased GSH levels with inhibition of apoptosis in all five model systems. The effect of TAL expression on susceptibility to apoptosis through regulating the PPP and GSH production is consistent with an involvement of ROIs in each pathway tested. Production of ROIs in Fas-mediated cell death was further substantiated by measurement of intracellular ROI production with oxidation-sensitive fluorescent probes, by the protective effects of GSH precursor, N-acetyl cysteine, free radical spin traps 5,5-dimethyl-1-pyrroline-1-oxide and 3,3,5,5-tetramethyl-1-pyrroline-1-oxide, the antioxidants desferrioxamine, nordihydroguaiaretic acid, and Amytal, and by the enhancing effects of GSH depletion with buthionine sulfoximine. The results provide definitive evidence that TAL has a role in regulating the balance between the two branches of PPP and its overall output as measured by GSH production and thus influences sensitivity to cell death signals.

(ROIs) 1 have long been considered as toxic by-products of aerobic existence, but evidence is now accumulating that controlled levels of ROIs modulate cellular function and are necessary for signal-transduction pathways, including those mediating apoptosis (4 -7). Many of the chemical and physical stimuli that elicit programmed cell death generate ROIs such as H 2 O 2 and OH Ϫ (8). Low doses of H 2 O 2 induce apoptosis in a variety of cell types, thus establishing oxidative stress as a mediator of apoptosis (7). The ability of scavengers of ROIs such as N-acetylcysteine (NAC), a precursor of GSH (7,8) and of free radical spin traps such as 5,5-dimethyl-1-pyrroline-1oxide (DMPO) and 3,3,5,5-tetramethyl-1-pyrroline-1-oxide (TMPO) to inhibit apoptosis, supports this hypothesis (9). bcl-2, the prototype of a novel family of proto-oncogenes that inhibit apoptosis when it is induced by many diverse stimuli, was recently demonstrated to have antioxidant behavior (5,6). However, apoptosis and bcl-2 protection were demonstrated in very low oxygen pressure, suggesting that ROI may not be an absolute requirement for programmed cell death (10).
A normal reducing atmosphere, required for cellular integrity, is provided by reduced GSH, which protects cells from damage by ROIs (11). Synthesis of GSH from its oxidized form, GSSG, is completely dependent on NADPH produced by the pentose phosphate pathway (PPP) (11). In fact, a fundamental function of PPP is to maintain glutathione in a reduced state and thus provide protection of sulfhydryl groups and cellular integrity from emerging oxygen radicals. PPP comprises two separate branches: the oxidative and the nonoxidative. Reactions in the oxidative branch are irreversible, whereas all reactions of the nonoxidative branch are fully reversible. The two branches are functionally connected. The nonoxidative branch can convert ribose 5-phosphate into glucose 6-phosphate for the oxidative branch, and thus, indirectly, it can also contribute to generation of NADPH. The rate-limiting enzymes for the two branches are different. The oxidative phase is primarily dependent on G6PD (12). While the control of the nonoxidative branch is less well established, transaldolase (TAL) has been proposed as its rate-limiting enzyme (12,13).
TAL catalyzes the transfer of a 3-carbon fragment, corresponding to dihydroxyacetone, to D-glyceraldehyde 3-phosphate, D-erythrose 4-phosphate, and a variety of other acceptor aldehydes, including nonphosphorylated trioses and tetroses. Enzymatic activity of TAL is regulated in a tissue-specific (13, 14) and developmentally specific manner (15). In the brain, TAL is expressed selectively in oligodendrocytes at high levels (16). This is particularly interesting because myelin sheaths are formed by oligodendrocytes, and lesions in the most common demyelinating disease of the central nervous system, multiple sclerosis, are characterized by a progressive loss of oligodendrocytes and demyelination. Oligodendrocytes are exquisitely sensitive to damage by ROI, such as nitric oxide and tumor necrosis factor-␣ (TNF) released by activated macrophages and astrocytes (17). The present data provide evidence that levels of TAL expression can determine susceptibility to apoptosis signals. Overexpression of TAL in Jurkat human leukemic T cells was accompanied by a decrease in G6PD and 6PGD activities, a concomitant depletion in NADPH and GSH levels, and increased sensitivity to apoptosis provoked by serum deprivation, H 2 O 2 , nitric oxide (NO), TNF, and anti-Fas monoclonal antibody. In contrast, suppression of TAL activity increased G6PD and 6PGD activities, augmented GSH levels, and inhibited apoptosis. Thus, susceptibility to apoptosis could be regulated by TAL through control of the balance between the two branches of the PPP and its overall output as measured by NADPH and GSH production.
Construction and Transfection of Eukaryotic Expression Vectors-TAL-H cDNA clone 4/2-4/1 (18) was inserted into the HpaI site of the metallothionein promoter-driven pMAXRHneo-1 vector (19) following removal of the coding sequence of the p40/Tax protein of human T-cell lymphotrophic virus-I by cleavage with HpaI. Clones containing the TAL-H cDNA in the sense (pL26-3) and antisense orientation (pL18-3) were selected. 20 g of plasmid DNA linearized with KpnI were used to stably transfect Jurkat cells by electroporation as described (19). Transfected cells were grown in the presence of 750 g/ml G418 and cloned by limiting dilution. Levels of transaldolase expression were measured by enzyme assay and Western blot analyses in the absence and presence of 5 M CdCl 2 . TAL protein expression was assessed by Western blot analysis and compared with that of actin using an automated scanning densitometer (Bio-Rad model GS-670). Transfection of the pMAXRHneo-1 vector alone had no effect on TAL activity. Due to a leakiness of the promoter, cell lines producing increased and suppressed levels of TAL in the absence of CdCl 2 were obtained. While overexpression and suppression were doubled by incubating the cells with 5 M CdCl 2 , CdCl 2 was not utilized in the following experiments to preclude possible interference with apoptotic pathways (20).
Induction of Apoptosis-Cells were cultured in RPMI 1640 medium, supplemented with 10% fetal calf serum, 2 mM L-glutamine, 100 IU/ml penicillin, and 100 g/ml gentamicin at 37°C in a humidified atmosphere with 5% CO 2 . 24 h prior to assays, cells were fed with fresh medium and seeded at a density of 2 ϫ 10 5 cells/ml. CD95/Fas/Apo-1mediated cell death was induced with 50 or 100 ng/ml anti-Fas monoclonal antibody CH-11. TNF-mediated cell death was induced with 20 or 100 ng/ml human recombinant TNF-␣ as earlier described (21)(22)(23). Apoptotic cell death was also induced by withdrawal of fetal calf serum or treatment with 50 M or 100 M H 2 O 2 , or 2.5, 5, and 10 mM sodium nitroprusside, used as a source of exogenous NO (4). During treatment with sodium nitroprusside, cells were cultured in the dark (aluminum foil-covered plate). Apoptosis was monitored by observing cell shrinkage and nuclear fragmentation and was quantified by trypan blue exclusion (1). DNA fragmentation during apoptosis was monitored by agarose gel electrophoresis (24). Apoptosis was also measured by flow cytometry as described previously (25). Briefly, following induction of apoptosis, cells were washed in phosphate-buffered saline and resuspended in phosphate-buffered saline with 0.1% Triton X-100 and 50 g/ml propidium iodide. The DNA content was analyzed by using a Becton Dickinson FACStarplus flow cytometer. Hypodiploid cells containing a lower amount of DNA and a side scatter higher than that of G 0 /G 1 cells were considered to be apoptotic.
TAL Activities-The reverse reaction catalyzed by TAL was tested in the presence of 3.2 mM D-fructose 6-phosphate, 0.2 mM erythrose 4-phosphate, 0.1 mM NADH, and 10 g of ␣-glycerophosphate dehydrogenase/ triosephosphate isomerase at a 1:6 ratio in 40 mM triethanolamine, pH 7.6, 5 mM EDTA at room temperature by continuous absorbance reading at 340 nm for 8 min (18). The forward reaction catalyzed by TAL was measured at room temperature in the presence of 50 mM triethanolamine, pH 7.4, 5 mM MgCl, 3 mM ribose 5-phosphate, 0.9 mM xylulose 5-phosphate, 0.5 mM NADP, 0.2 mM thiamine pyrophosphate, 0.2 units/ml transketolase, 0.4 units/ml phosphoglucose isomerase, and 0.3 units/ml G6PD, following a 10-min lag period, by continuous absorbance reading at 340 nm for 8 min (26). The enzyme assays were conducted in the activity range of 0.001-0.01 units/ml. Unless indicated otherwise, TAL activity refers to enzyme activity measurements conducted in the reverse reaction.
Glutathione, NADPH, and NADH Levels-Total glutathione content was determined by the enzymatic recycling procedure essentially as described by Tietze (28). 10 6 cells were resuspended in 100 l of 4.5% 5-sulfosalicylic acid. The acid-precipitated protein was pelleted by centrifugation at 4°C for 10 min at 2000 ϫ g. The total protein content of each sample was determined using the Lowry assay (29). GSH content of the aliquot assayed was determined in comparison to reference curves generated with known amounts of GSH. For NADPH and NADH assays, 5 ϫ 10 6 cells were washed in phosphate-buffered saline and resuspended in 125 l of H 2 O, and pyridine nucleotides were extracted by adding 62.5 l of freshly prepared 1 M KOH in ethanol as earlier described (30).
Western Blot Analysis-Protein lysates were prepared from cell lines and quantified by the Lowry method. 40 g of protein lysate in 10 l/well was separated by SDS-polyacrylamide gel electrophoresis and electroblotted to nitrocellulose. Nitrocellulose strips were incubated in 100 mM Tris, pH 7.5, 0.9% NaCl, 0.1% Tween 20, and 5% skim milk with TAL-H-specific rabbit Ab 169 (18) or actin-specific murine monoclonal antibody C4 at a 1000-fold dilution at room temperature overnight. For detection of rabbit antibodies, after washing, the strips were incubated with horseradish peroxidase-conjugated goat anti-rabbit IgG (Boehringer Mannheim). For detection of murine antibodies, after washing, the strips were incubated with biotinylated goat anti-mouse serum, and subsequently with horseradish peroxidase-conjugated avidin (The Jackson Laboratory, West Grove, PA). Between the incubations, the strips were vigorously washed in 0.1% Tween 20, 100 mM Tris, pH 7.5, and 0.9% NaCl. The blots were developed with a substrate comprised of 1 mg/ml 4-chloronaphthol and 0.003% hydrogen peroxide.
Flow Cytometric Analysis of the Rates of Intracellular Oxidation-The production of ROIs was estimated fluorometrically using DCFH-DA and DHR as earlier described (31,32). Following apoptosis assay, cells were washed three times in 5 mM Hepes-buffered saline, pH 7.4, incubated in Hepes-buffered saline with 0.1 M DCFH-DA or DHR for 2 min, and samples were analyzed using a Becton Dickinson FAC-Star Plus flow cytometer equipped with an argon ion laser delivering 200 milliwatts of power at 488 nm. Fluorescence emission from 5,6carboxy-2Ј,7Ј-dichlorofluorescein (green) or DHR (green) was detected at a wavelength of 530 Ϯ 30 nm. Dead cells and debris were excluded from the analysis by the electronic gating of forward and side scatter measurements.
Statistics-Alterations in cell survival, NADPH, NADH, and GSH levels and the activity of PPP enzymes were analyzed by Student's t test. Changes were considered significant at p Ͻ 0.05.
Subsequently, the effect of TAL expression on activities of other key PPP enzymes and intracellular levels of NADPH, NADH, and GSH was evaluated. Overexpression and increased activity of TAL had no significant effect on TK activities in L26-3/4 and L26-3/2D1 cells. Suppression of TAL was associated with a decrease of TK activity in L18-3/1 cells, while TK activity remained unchanged in L18-3/1D9 cells. Since the effect of TAL suppression was deeper in L18-3/1D9 than that in L18-3/1 cells, reduction of TK in L18-3/1 cells may not be related to diminished TAL activity. Activities of enzymes of the oxidative branch of the PPP, G6PD and 6PGD, which are directly responsible for NADPH production, were inversely correlated with TAL expression (Table I). Accordingly, NADPH levels were decreased in L26-3/4 and L26-3/2D1 cells in comparison to control Jurkat cells. GSH levels were depleted in these cell lines. NADH levels were also diminished in L26-3/4 and L26-3/2D1 cells. Increased TAL expression occurred with coordinated changes that included (a) down-regulation of G6PD and 6PGD activities and (b) a decrease of NADPH and GSH synthesis in both L26-3/4 and L26-3/2D1 cells. Along the same line, decreased TAL expression occurred with up-regulation of G6PD and 6PGD activities and an increase of GSH levels in L18-3/4 and L18-3/1D9 cells. However, NADPH and NADH levels were unchanged in these cell lines.

. Western blot analysis of TAL protein levels in cell lines stably transfected with TAL expression vectors oriented in the sense (L26-3/4) and antisense directions (L18-3/1), in comparison to untransfected Jurkat cells (control)
. 40 g of protein lysate was loaded in each lane. The TAL protein (38 kDa) was detected with rabbit antibody 169 while actin (42 kDa) was visualized with mouse monoclonal antibody C4. Level of TAL expression relative to actin content was determined by scanning densitometry. TAL/actin ratio in control Jurkat cells was considered as 100%. TAL expression was reduced by 25% in L18-3/1 cells and increased 2.6-fold in L26-3/4 cells.

Sensitivity to Apoptotic Signals Is Influenced by Levels of TAL Expression-Stably transfected cell lines with increased
and depressed TAL activities have been maintained in culture for over 2 years showing viability similar to that of control Jurkat cells (Ͼ99%). However, levels of TAL expression had a dramatic influence on susceptibility to apoptotic cell death induced by H 2 O 2 , NO, TNF-␣, anti-Fas monoclonal antibody, or withdrawal of fetal calf serum from the culture medium (Fig.  2). Apoptosis was monitored by DNA ladder formation (Fig. 3) and flow cytometry as shown for Fas simulation (Fig. 4). Cell death was particularly accelerated in TAL-overproducing cells in comparison to control Jurkat cells. Suppressed TAL expression inhibited cell death produced by all stimuli tested. Cell survival inversely correlated with TAL expression levels (Table I).

Fas-mediated Apoptosis Is Associated with Production of ROI-
The effect of TAL expression on susceptibility to apoptosis through regulating the PPP and GSH production suggested the involvement of ROIs in each pathway tested. While the production of ROIs has been associated with apoptosis induced by H 2 O 2 , NO, TNF, and serum deprivation, involvement of ROIs in Fas-dependent signaling has been controversial. To assess changes in intracellular ROIs we used oxidation-sensitive fluorescent probes DCFH-DA and DHR (32,33). DCFH-DA is nonfluorescent, readily accumulates within cells, and following deacetylation to DCFH is oxidized to the fluorescent compound dichlorofluorescein. Similarly, DHR is nonfluorescent, uncharged, and readily taken up by cells, whereas R123, the product of DHR oxidation, is fluorescent, positively charged, and trapped within cells. We evaluated the rates of increase in fluorescence of cells treated with 100 M H 2 O 2 and 50 ng/ml anti-Fas monoclonal antibody. As shown in Fig. 5A, relative to H 2 O 2 a smaller but consistent increase in ROI was detected in Fas-stimulated cells as compared with untreated cells. In agreement with earlier data (33), DHR was a significantly more sensitive detector of increases in ROI levels than DCFH. Production of ROI correlated with the rate of cell death (p Ͻ 0.01; Fig. 5B).
Effect of Antioxidants on TAL and G6PD Activities, GSH Levels, and Fas-induced Cell Death-The involvement of ROI in Fas-dependent signaling was suggested by our observation of (i) an increased sensitivity to Fas-induced death of cells with increased TAL expression and decreased GSH content and (ii) the production of ROI during Fas-mediated apoptosis. We therefore examined whether changes in GSH levels can influence Fas-induced programmed cell death. Under the conditions utilized, none of the agents had a significant effect on the binding of anti-Fas antibody to its receptor based on flow cytometry (data not shown). Intracellular GSH levels were raised by as much as 2-fold using NAC, a precursor of glutathione (34), or suppressed to less than 15% of baseline by BSO, an inhibitor of ␥-glutamyl-cysteine synthetase (34) (Fig. 6A). Neither NAC nor BSO influenced TAL activity. In contrast, G6PD activity was increased by both NAC (p Ͻ 0.01) and BSO (p Ͻ 0.05) after 24 h incubation. In accordance with earlier observations (35), cell viability was not affected up to 4 days in culture by NAC (up to 3 mM), BSO (up to 1 mM), or any of the other antioxidants tested (data not shown). However, Fas-mediated cell death was markedly influenced by NAC and BSO (Fig. 6B). Prior to stimulation with anti-Fas antibody, Jurkat cells were pretreated with NAC or BSO for 24 h. 1 mM BSO substantially accelerated apoptosis, while 3 mM NAC inhibited Fas-mediated cell death (p Ͻ 0.01; Fig. 6B). The antioxidants amytal, desferrioxamine, and nordihydroguaiaretic acid also inhibited Fas-induced apoptosis (Fig. 6B). DMPO and TMPO are nitrones that react with ROIs to form more stable nitroxide radical products (36) and have protected thymocytes against apoptosis (9). While DMPO or TMPO had no significant effect on GSH levels (not shown), they also inhibited Fas-induced apoptosis (p Ͻ 0.01; Fig. 6B).

DISCUSSION
The PPP fulfills two essential functions consisting of the formation of pentose phosphates for synthesis of nucleotides, RNA, and DNA and the generation of NADPH for biosynthetic reactions and to maintain glutathione at a reduced state, thus protecting sulfhydryl groups and cellular integrity from oxygen radicals. A number of different approaches have been applied to delineate the mechanism by which PPP is controlled: 1) identification of the rate-limiting enzymes, 2) comparison of both the mass action ratios with the equilibrium constant and steady-state concentrations with the actual flux of intermediates for each enzyme, and 3) identification of enzymes under hormonal or environmental control (12). Finding a unifying approach has been complicated by the fact that PPP is comprised of two separate branches, oxidative and nonoxidative. Reactions in the oxidative branch are irreversible, whereas all reactions of the nonoxidative phase are fully reversible. The rate-limiting enzymes for the two branches are different. The oxidative branch is primarily dependent on G6PD (12). The control of the nonoxidative branch, between transaldolase and transketolase, is less well established based on enzyme kinetic studies. Tissue-specific variations in enzymatic activities suggested that TAL may be a rate-limiting enzyme of the nonoxidative branch of the PPP (13, 14). This study provides evidence that TAL may have a pivotal role in regulating the entire pathway. Overexpression of TAL in Jurkat human T cells resulted in down-regulation of G6PD and 6PGD activities and a decrease of NADPH and GSH levels. NADH levels were also reduced in TAL-overproducing cells, which was consistent with the tendency to maintain NADPH at the expense of NADH by transhydrogenases (37). Alternatively, decreased TAL expression led to up-regulation of G6PD and 6PGD activities and increased GSH levels.
Sensitivity to apoptosis was effectively controlled by regulating the activity of TAL, a pivotal enzyme of the PPP. Overexpression of TAL increased sensitivity, while suppression of TAL decreased sensitivity to five different apoptotic signals indicating that TAL expression levels can profoundly influence susceptibility to programmed cell death. The mechanism of this regulatory function may be explained by a considerable difference in forward TAL catalytic activity that favors production of glucose 6-phosphate, compared with reverse TAL activity that depletes glucose 6-phosphate (11). Maximal velocity of TAL in the forward direction was only about one-third of that of the reverse direction in the yeast (38). Reversibility of the TAL reaction has been proposed as a possible control mechanism for the entire pathway in the yeast (26). In L26-3/4 and L26/2D1 cells overexpressing TAL, the nonoxidative branch was pushed in the reverse direction, depleting glucose 6-phosphate (not shown). This effect may be directly responsible for diminished G6PD activities and GSH levels and increased sensitivity to apoptotic signals.
The involvement of ROIs in each of the apoptosis signaling pathways examined here was suggested by the finding that TAL expression, via regulation of the PPP and of GSH production, modulated susceptibility to apoptosis. Apoptosis triggered by serum withdrawal (39) and NO has been associated with the production of ROIs (4). Because GSH is the most abundant intracellular thiol compound that neutralizes ROI, a direct correlation of GSH levels with resistance to apoptosis induced by either H 2 O 2 , NO, or serum withdrawal was predictable. In contrast, an involvement of ROIs has not been clearly defined in Fas-and TNF-induced apoptotic pathways that are signaled through specific cell surface receptors (2). The Fas/Apo-1 antigen and the TNF-␣ receptor are members of the TNF/nerve growth factor receptor superfamily. All these receptors contain canonical cysteine-rich extracellular domains; Fas and the type I TNF receptor, mainly responsible for mediating cytolytic activity of TNF, also have a common 70-amino acid intracellular sequence, which may play a role in triggering common cytoplasmic death signals (2). Indeed, TNF-mediated apoptosis is known to involve oxidative stress based on (a) the induction of ROI in response to TNF (40,41) and (b) the inhibition of TNF-induced killing by free radical scavengers (40 -42). Schulze-Osthoff et al. (42) found that TNF-mediated but not Fas-mediated cell death can be inhibited by ROI-scavenging compounds. No requirement of ROIs in either TNF-or Fasmediated apoptosis was suggested by Hug et al. (43). In contrast, the effect of the deletion of the cytoplasmic "death domains" shared by Fas and the type I TNF receptor (2) and the involvement of common signal transducer suggest that both receptors may mediate cell death via a similar mechanism (44). Indeed, our data indicated that the rate of both TNF-and Fas-mediated cell death programs correlated with intracellular GSH levels as regulated by TAL expression. Increased produc-tion of ROI by Fas-mediated signaling was demonstrated using oxidation-sensitive fluorescent probes DCFH-DA and DHR. In agreement with other studies, we found that DHR is a much more sensitive probe of ROI than DCFH-DA (Fig. 5), since the product of DHR oxidation, R123, is effectively trapped within the cell (33). In correlation with the rate of Fas-induced cell death, production of ROIs was accelerated in L26-3/4 and L26-3/2D1 cells, while intracellular ROI levels were not significantly increased in L18-3/1 and L18-1/1D9 cells (not shown). Moreover, several different pretreatment regimens including NAC (which increases GSH levels), DMPO and TMPO (two free radical spin traps that form relatively stable nitroxide radical products with ROIs) (40), and various antioxidants (desferrioxamine, nordihydroguaiaretic acid, and Amytal) all protected against cell death. Our observation of greatly accelerated Fasinduced killing following depletion of GSH by BSO provides additional evidence for the involvement of ROI in Fas-dependent signaling. Thus, cells with higher GSH levels were less sensitive, whereas cells with diminished GSH levels were more sensitive to Fas-mediated apoptosis. We conclude that the Fas signaling pathway is also associated with the formation of ROIs.
The present results support the notion that the PPP represents an important biochemical mechanism regulating sensitivity to cell death programs dependent on formation of ROIs. This is consistent with an essential role of the pathway in the generation of NADPH for the synthesis of GSH which, in turn, protects cellular integrity from oxygen radicals. Levels of TAL expression may have a dominant role in regulating the balance between the two branches of PPP and the ultimate output of GSH. We propose that TAL, and possibly other PPP enzymes, serves as a critical determinant of tissue-and cell type-specific sensitivity to apoptotic signals.