INTRODUCTION

Sphingolipids have emerged as important components of signal transduction pathways involved in a variety of cellular processes. This is best illustrated in the case of the sphingomyelin (SM)¹ cycle (1), which has become a cornerstone in studying the role of sphingomyelin, ceramide, and other related products in cellular regulation. In this pathway, a number of extracellular agents such as tumor necrosis factor  (2), 1,25-dihydrovitamin D₃ (1), interleukin 1 (3, 4), nerve growth factor (5), chemotherapeutic agents (6, 7), and serum deprivation (8) results in activation of sphingomyelinases (SMases), which causes the hydrolysis of sphingomyelin, leading to the generation of ceramide. In turn, ceramide has been suggested to play important roles in differentiation, cell cycle arrest, apoptosis, inflammation, and the regulation of eukaryotic stress responses (9).

Although the ceramide generated in response to the action of extracellular agents appears to derive from the hydrolysis of sphingomyelin, not much is known about the mechanisms regulating the activity of the involved SMases. To date, at least five types of SMases have been identified. An acidic sphingomyelinase (A-SMase) was first discovered and found to have an optimum pH at 5.0 (10). Although the acidic sphingomyelinase has been primarily found to reside in the lysosomes, it has also been detected as a soluble form in cytosol and extracellular media (11). This enzyme has been cloned, and it is deficient in Neimann-Pick cells (12). It is also suggested to play a role in radiation-induced apoptosis (13) and has been described to be activated by Fas (14) and tumor necrosis factor- (15). A neutral pH-optimum and magnesium-dependent SMase (N-SMase) has also been described (16). This enzyme is associated with the plasma membrane. It is found to be activated in response to tumor necrosis factor-, Fas, Ara-C, and serum deprivation, and its activation appears to be closely related to growth suppression and apoptosis (6, 8, 17, 18). Recently, a cytosolic magnesium-independent neutral SMase was partially purified from HL-60 cells following treatment with vitamin D₃ (19). In addition, a zinc-dependent acidic SMase was detected in sera from a variety of species (20), and recent studies suggest that it is derived from the lysosomal A-SMase (21). Finally, an alkaline pH-optimum SMase has been described in intestinal cells (22).

In ongoing studies on characterizing N-SMase, we discovered that this enzyme was inhibited in vitro by physiologically relevant concentrations of glutathione (GSH) and it was activated in cells depleted of GSH. Glutathione, a tripeptide (L--glutamyl-L-cysteinyl-glycine), is the most abundant nonprotein thiol-containing small molecule in mammalian cells. It plays a critical role in cellular defense against oxidative stress by inactivating free radials, reactive oxygen species, and a variety of cytotoxic electrophiles including alkylating agents. GSH functions as an antioxidant by directly reacting with free radicals, conjugating with various agents to form S-substituted adducts nonenzymatically or through the action of GSH S-transferase, and serves as a substrate for GSH peroxidase (23). Under normal conditions, cells maintain a high level of intracellular GSH (1-20 mM). In this study, we report on the effects of GSH on N-SMase. The implications of inhibition of N-SMase by GSH will be discussed.

EXPERIMENTAL PROCEDURES

Materials

Molt-4 human leukemia cells were obtained from ATCC (Rockville, MD). RPMI 1640 and fetal calf serum were purchased from Life Technologies, Inc. Glutathione and related compounds were from Sigma. Leukotrienes were from Cayman Chemicals (Ann Arbor, MI). Stripped rat brains were from Pel-Freez Biologicals (Rogers, AR). DEAE-Sepharose was from Pharmacia Biotech Inc. [N-methyl-¹⁴C]Sphingomyelin was synthesized as described previously (19).

Methods

Molt-4 cells were grown in RPMI 1640 supplemented with 10% fetal calf serum at 37 °C in a humidified environment with 5% CO₂. Cells were seeded at a density of 2.5 × 10⁵/ml and grown to near confluence (1.5 × 10⁶/ml) before use.

Molt-4 cells (2.5 × 10⁹) were washed three times with phosphate-buffered saline and then homogenized with a Teflon pestle glass homogenizer in lysis buffer (25 mM Tris-HCl, pH 7.4, 2 mM EDTA, 5 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 20 µg/ml each of chymostatin, leupeptin, antipain, and pepstatin A). The post-nuclear homogenate was centrifuged for 1 h at 100,000 × g, and the pellet was dissolved in 1% Triton X-100 in lysis buffer. The Triton X-100-soluble fraction was loaded onto a DEAE-Sepharose column (1 × 10 cm) equilibrated in buffer A (20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 1 mM EGTA, and 1 mM phenylmethylsulfonyl fluoride). After washing off the unbound proteins, the enzymes were eluted by stepping to 1 M NaCl in buffer A, followed by a gradient of Triton X-100 from 0-1% in buffer A. Fractions (4 ml) were collected and assayed for activities of N-SMase and A-SMase. The recovery of the N-SMase over the DEAE-Sepharose column was typically greater than 85%. For isolating SMases from rat brain, 4 rat brains were used and the procedure was the same as that for Molt-4 cells as described above. All procedures were performed at 4 °C.

The activity of both N-SMase and A-SMase were determined using a mixed micelle assay system (19). For determining N-SMase activity, enzyme preparation (5-10 µl, 0.5-1 µg of protein) diluted to 50 µl with 20 mM Tris-HCl, pH 7.4, was mixed with 10 nmol of [¹⁴C]sphingomyelin (100,000 dpm) dissolved in 0.05% Triton X-100 in 50 µl of 100 mM Tris-HCl, pH 7.4, containing 5 mM magnesium chloride. The reaction proceeded for 30 min at 37 °C and was then terminated by the addition of 1.5 ml of chloroform:methanol (2:1, v/v) followed by 0.2 ml of H₂O. After phase separation, a portion of the upper phase was removed and the radioactivity determined by liquid scintillation counting. A-SMase activity was determined using [¹⁴C]sphingomyelin resuspended in 50 µl of 100 mM sodium acetate, pH 5.0. The identity of the product in the upper phase was confirmed by thin layer chromatography as described (19).

Molt-4 cells were seeded at 2.5 × 10⁵/ml in complete medium, rested for 24 h, and then treated with or without 250 µM L-buthionine-(SR)-sulfoximine (BSO) for various time intervals. Afterward, cells were pelleted and total cellular lipids extracted by the method of Bligh and Dyer (49). Ceramide content was determined using a modified diacylglycerol kinase assay as described previously (7, 8, 17). SM level was measured as described previously (8). Briefly, cellular lipids were separated by TLC using a solvent system of chloroform/methanol/acetic acid/water (50:30:8:5). The sphingomyelin spots were visualized by iodine staining, scraped, and measured for phosphate. Sphingomyelin was quantitated using an external standard and was normalized to total phosphate.

Intracellular GSH content was determined by the Griffith (50) modification of Tietze's enzymatic procedure (51). Briefly, cells (2 × 10⁷) were resuspended in 400 µl of 10% 5-sulfosalicylic acid. The acid-precipitated proteins were pelleted by centrifugation at 4 °C for 10 min at 2000 × g. To determine the GSH content, aliquots of the acid-soluble supernatant were mixed with 125 mM sodium phosphate, pH 7.5, 6.3 mM EDTA, 0.21 mM NADPH, and 0.6 mM 5,5-dithiobis(2-nitrobenzoic acid) in a total volume of 1 ml. Upon addition of glutathione reductase, the increase in absorption at 412 nm was monitored and used to determine the amount of GSH in the sample by comparison to a reference curve generated with known amounts of GSH standard. Protein content was determined by the Bradford dye binding assay (52) using bovine serum albumin as standard.

The formation of GSH fragments, -glutamyl-cysteine and cysteinyl-glycine, was examined by incubating N-SMase preparation with GSH for a given time, and an aliquot of the mixture was separated on a silica thin layer chromatographic plate developed with a solvent system of 1-butanol, pyridine, and water (1:1:1, v/v) as reported by Tate and Meister (48). The compounds were identified by spraying with ninhydrin and their migration compared with authentic standards.

RESULTS

As part of the biochemical characterization of N-SMase, we examined the effect of reducing agents on the in vitro activity of the enzyme. To this end, we partially purified the N-SMase from the membranes of Molt-4 cells which have very little cellular acidic SMase activity. On DEAE-Sepharose column, three peaks of SMase activity were resolved (see Fig. 4A). Peak I (flow-through) contained predominantly A-SMase and peak II (salt elution) had a small quantity of A-SMase. Peak III (detergent elution) had the highest activity of N-SMase (4 times that of peak II) and no A-SMase activity. The N-SMase in peak III was typically 50-100-fold purified over post-nuclear homogenate, and the N-SMase in peak II was 10-20-fold purified. Therefore, the N-SMase in peak III was used for this study. The partially purified N-SMase was preincubated with the reducing agents of interest, and this was followed by incubation with 10 mol% of radiolabeled sphingomyelin substrate in a Triton X-100 mixed micelle assay in the presence of 5 mM magnesium. The radiolabeled product formed after a 30-min incubation with substrate was separated by organic solvent extraction and its quantity determined by liquid scintillation counting. The reaction was linear within the range of the amount of enzyme protein, concentration of substrate, and incubation time used in this study. As shown in Fig. 1, preincubation of N-SMase for 15 min at 37 °C with concentrations of GSH ranging from 1 to 20 mM resulted in an inhibition of the N-SMase within a very narrow GSH concentration range. Little inhibition was detected at concentrations of GSH below 2 mM, whereas GSH above 4 mM gave rise typically to a 90-100% inhibition of N-SMase (Fig. 1A). Dithiothreitol or -mercaptoethanol at concentrations ranging from 1 to 20 mM were completely ineffective in inhibiting the N-SMase (Fig. 1A). Taken together, these results demonstrate that physiological concentrations of GSH are inhibitory to the neutral magnesium-dependent sphingomyelinase and that this inhibitory property of GSH does not appear to be a general property of thiol-containing reducing agents.

-ME

The inhibition by GSH of the N-SMase was enhanced by preincubation of GSH with the enzyme. When no preincubation time was allowed, GSH at 5 mM inhibited the enzyme activity by 30% (Fig. 1B). When the enzyme was preincubated with GSH for time periods ranging from 5 to 15 min, the degree of inhibition progressively increased such that preincubation with 5 mM GSH for 5, 10, and 15 min resulted in a 70, 90, and 100% inhibition, respectively (Fig. 1B). Preincubation of the enzyme with GSH for time periods longer than 15 min did not result in further inhibition (data not shown). Interestingly, for preincubation time intervals longer than 5 min, similar degree of inhibition was observed for GSH concentrations between 3 and 5 mM (Fig. 1B). To rule out effects of preincubation on stability of the enzyme or GSH, the following experiments were performed. First, preincubation of N-SMase, in the absence of GSH, for up to 20 min at 37 °C prior to the initiation of activity assay (30 min) did not affect the enzyme activity (data not shown). Second, thin layer chromatography analysis of GSH (5 mM) following incubation with N-SMase preparation for up to 60 min at 37 °C under the conditions identical to SMase assay did not indicate any metabolism of GSH to either -glutamyl-cysteine or cysteinyl-glycine (data not shown). To study the reversibility of GSH inhibition of N-SMase, enzyme was preincubated with 3 mM GSH for 15 min at 37 °C and then assayed for SMase activity or diluted 5 times to achieve a final concentration of GSH of 0.6 mM before activity assay. N-SMase activity (standardized based on equal amount of enzyme) was completely recovered after diluting out GSH. These results show that inhibition is reversible, and that the enhancement of inhibition by preincubation is not a result of metabolism of GSH or an irreversible change in the enzyme.

We next examined the specificity for GSH inhibition of the N-SMase by testing the effects on the enzyme activity of GSH related compounds. As shown in Fig. 2, glutamic acid, glutamine, and glycine at concentrations up to 10 mM were totally ineffective in inhibiting the enzyme (Fig. 2A). Interestingly, cysteine exhibited a triphasic mode of action for its effect on the N-SMase. An initial inhibitory phase was observed with concentrations of cysteine from 0.05 to 1 mM, and this was followed by a stimulatory phase for concentrations between 1 and 2.5 mM. At concentrations starting at 2.5 mM and up to 10 mM, cysteine was inhibitory, almost completely coinciding with that of GSH (Fig. 2B).

-Glu-Cys

When the enzyme was preincubated with the GSH fragment cysteinyl-glycine, a biphasic mode of action was observed. At concentrations ranging from 0.01 to 1 mM, cysteinyl-glycine was partially inhibitory of the enzyme, which was followed by reversal of its inhibition at concentrations from 1 to 10 mM, closely resembling the initial two phases for cysteine (Fig. 2C). In contrast to cysteine, at concentrations beyond 2 mM, cysteinyl-glycine did not show further inhibition toward N-SMase. The effect of another fragment of GSH, -glutamyl-cysteine, on the enzyme activity was very similar to that of GSH except that -glutamyl-cysteine was slightly more potent such that the range for total inhibition was from 1-3 mM for this fragment (Fig. 2C; 2-4 mM for GSH). These results showed that the minimal configuration of GSH required for inhibition resided in the -glutamyl-cysteine segment of the molecule, which faithfully reproduced the effects of GSH on N-SMase.

Since DTT and -mercaptoethanol did not inhibit the enzyme, we next addressed the requirement for the free sulfhydryl group on GSH. For these experiments, we used oxidized GSH, GSSG, and the thiol-modified GSH molecules, S-methyl GSH and S-ethyl GSH. Both S-methyl GSH and S-ethyl GSH were equally potent as GSH in inhibiting the enzyme (Fig. 2D). Surprisingly, oxidized glutathione GSSG was more effective than GSH in inhibiting the enzyme, with an apparent EC₅₀ of between 0.5 and 1 mM compared with 2.5 mM for GSH (Fig. 2D). The greater potency of GSSG did not seem to arise from spontaneous reduction of GSSG to two molecules of GSH. First, GSSG was more than twice as potent as GSH. Second, the inhibition of N-SMase activity by GSSG was not altered in an oxidizing environment created by co-incubation with H₂O₂ (100 µM). In this case, the EC₅₀ remained at 0.5 mM.

Because leukotrienes are GSH-modified bioactive lipid molecules important in cellular regulations, we also examined their effects on the activity of the N-SMase in vitro. As shown in Fig. 2F, amounts of leukotrienes C₄ and D₄ well above physiologically relevant concentrations had no effect on the activity of the enzyme. Taken together, these results suggested that the minimal requirement for inhibiting the N-SMase was -glutamyl-cysteine and that the free sulfhydryl group on the GSH molecule was not essential for the inhibitory effect.

Next, we examined the interaction of GSH with the substrate sphingomyelin and the activator magnesium. To study the interaction of GSH with sphingomyelin, enzyme was preincubated with 0-2.5 mM GSH followed by incubation with 1-20 mol% of sphingomyelin. As shown in Fig. 3A, concentrations of GSH ranging from 1.5 to 2.5 mM gave rise to similar degrees of inhibition over a broad range of substrate concentrations. Because of the steep nature of the inhibitory effect of GSH on the N-SMase, we could not construct reliable double-reciprocal plots for these studies. However, since increasing concentrations of substrate did not overcome inhibition by GSH, the mode of inhibition did not appear to be of the competitive type. Similarly, when the enzyme was preincubated with 0-2.5 mM GSH followed by incubation with 10 mol% of sphingomyelin in the presence of 0-10 mM magnesium, a sharp inhibition was observed between 1.5 and 2 mM GSH over a broad range of magnesium concentrations. Again, the inhibition did not appear to be of a competitive nature (Fig. 3B).

We next investigated whether the inhibitory effect of GSH was specific to the N-SMase. To this end, we resolved, on a DEAE-Sepharose column the A-SMase and N-SMase prepared from detergent-solubilized rat brain homogenate which contains substantial amounts of both enzymes. As shown in Fig. 4A, the majority of the A-SMase was found in peak I (flow-through) with a small amount in peak II, while peaks II and III contained approximately 35% and 65% of the N-SMase activity, respectively. The N-SMase in peak III was more pure than that in peak II and was essentially free of A-SMase activity (Fig. 4A). Therefore, we compared the effect of GSH on the activity of enzyme preparations from peak I (A-SMase) and peak III (N-SMase). When preincubated with 1-10 mM GSH, N-SMase from rat brain was inhibited by GSH with a complete inhibition at 2 mM GSH, while DTT was totally ineffective (Fig. 4B). In contrast, GSH up to 10 mM had little effect on the activity of A-SMase from rat brain, while DTT significantly abolished the activity of the acid SMase (Fig. 4C) as reported previously (24). Similar results for GSH and DTT were obtained for purified human placenta A-SMase of commercial source (data not shown).

Finally, we analyzed the effect of the levels of GSH in intact cells on the hydrolysis of SM and the generation of ceramide. Cultured Molt-4 cells were treated with 250 µM BSO for 12-48 h, and the total intracellular GSH content was determined. As shown in Fig. 5A, treatment of cells with BSO resulted in a time-dependent decrease in the intracellular level of GSH. The GSH level of the BSO-treated Molt-4 cells dropped to 18 and 9% of that of untreated cells at 24 and 48 h post-treatment, respectively. When ceramide content in the BSO-treated cells was examined, an increase in the cellular ceramide level was observed as early as 12 h post-treatment of cells with BSO (Fig. 5B). The ceramide level at 24 and 48 h after BSO treatment was significantly higher than that of untreated cells (68% and 102% over that of corresponding control, respectively). A 32% decrease in SM content was observed in cells treated with 250 µM BSO for 48 h (Fig. 5C), indicating that the ceramide generated was due to the activation of a sphingomyelinase.

DISCUSSION

In this study we demonstrate that glutathione at physiologically relevant concentrations inhibits the neutral magnesium-dependent sphingomyelinase but not the acidic sphingomyelinase and that depletion of cellular GSH results in the hydrolysis of SM and generation of ceramide. The inhibition of the neutral magnesium-dependent sphingomyelinase does not appear to be a general characteristic of thiol-containing reducing agents since dithiothreitol and -mercaptoethanol were ineffective. Moreover, the free sulfhydryl of GSH is not required since inhibition was still observed with the thiol-modified GSHs, S-ethyl and S-methyl GSH, as well as with oxidized glutathione, GSSG. Studies with analogs and fragments of GSH suggest that the structural requirements for inhibition reside in the -glutamyl-cysteine moiety of GSH.

This structural specificity in inhibiting N-SMase suggests that GSH may function as a specific allosteric regulator of N-SMase. However, since partially purified sphingomyelinases were used in this study, we cannot rule out the possibility of an indirect mechanism of action for GSH inhibition of the N-SMase. The precise mechanism responsible for GSH inhibition of the N-SMase remains to be elucidated.

Our results raise a number of physiologic implications. Current estimation of intracellular concentrations of free GSH in most cell types suggests a range between 1 and 10 mM (25, 26). In hepatic cells, the lower limit of intracellular GSH level is believed to be around 3 mM and the higher limit 20 mM. This leads us to speculate that under resting and stable conditions, the N-SMase in the cells exists as an inactive enzyme in the presence of GSH. Depletion of GSH from the cell may then relieve the inhibition by GSH, resulting in activation of N-SMase.

Depletion of GSH has been described for several agents such as oxidative and alkylating agents, tumor necrosis factor-, and Fas in various cell types (27). These include lung epithelial cells exposed to high doses of O₂ (28), the leukemic cell lines HL-60, U937, and K562 exposed to the alkylating agent melphalan (29), and U937 cells treated with hydrogen peroxide and protein synthesis inhibitors (30). In addition to depletion, GSH levels may be reduced by extrusion from the cell as has been observed with Fas, which causes a rapid efflux of GSH from cells. Cells with reduced GSH levels either undergo apoptosis or become more sensitive to various death-inducing agents (29, 31, 32). For example, Kain and associates reported that depletion of GSH in a neural cell line committed it to apoptosis (33). Restoring and/or maintaining the intracellular GSH content through increased synthesis or addition of exogenous GSH or GSH methyl esters appears to protect cells from damage/death otherwise inflicted by the cytotoxic agents. In fact, Chiba et al. reported that a Fas-resistant human T cell line reverted to Fas sensitivity upon inhibition of its GSH synthesis capacity (34).

Since activation of N-SMase leads to the generation of ceramide, changes in GSH levels may influence ceramide levels through N-SMase. In turn, this direct biochemical coupling may tie-in the various regulators and modulators of the redox state of the cell (and of GSH levels) with ceramide-mediated cellular responses. These include cell differentiation (5), apoptosis (35), cell cycle arrest (8), Rb dephosphorylation (36), activation of a protein phosphatase (37) and a protein kinase (38), inhibition of cellular protein kinase C- (39), phospholipase D (40), and cellular senescence (41). Generation of ceramide has been attributed to the activation of N-SMase in studies with serum starvation (8), Fas-induced apoptosis (17), and cellular senescence studies (41). The fact that GSH depletion is frequently observed in most of these situations suggests that the levels of cellular GSH regulates, at least in part, ceramide generation. We show that, in Molt-4 cells, the GSH synthesis inhibitor BSO not only, as predicted, depletes the GSH content of these cells, but also causes the hydrolysis of SM and generation of ceramide. These data demonstrate that GSH may indeed play a role in regulating SM hydrolysis and ceramide generation in cells.

As a corollary to this hypothesis, N-SMase in the cells may remain inactive until GSH is depleted. This would argue against the possibility that under physiological conditions N-SMase can be activated without the depletion of GSH. Alternatively, a mechanism may exist for activating N-SMase by overriding the inhibition by GSH.

While the intracellular GSH levels are between 1 and 20 mM, the level of free intracellular cysteine is believed to be in the high micromolar range. In studying the uptake of radiolabeled L-[³⁵S]cysteine by human fibroblasts, Bannai and Kitamura (42) estimated that the intracellular concentration of cysteine was approximately 0.1 mM. In another study by Lu et al. (43), the ratio of total hepatic GSH to cysteine was found to be around 42 to 1. Combined with the fact that the intracellular GSH in hepatocytes is 3-20 mM and that intracellular cysteine transported from outside of the cell is rapidly metabolized (26), the actual level of intracellular cysteine is probably very low and in the micromolar range (75-500 µM). Therefore, the inhibition by cysteine of N-SMase observed in vitro may not be very relevant physiologically, but cannot be discounted at this point. Interestingly, depletion of both GSH and cysteine has been reported in several studies (44-46).

Similarly, the in vitro inhibition of N-SMase by the oxidized glutathione, GSSG, observed in this study may not have much physiological relevance. GSSG is formed as a result of the reaction of GSH with reactive oxygen species. Compared with GSH, GSSG exists in cells at a much lower concentration (20-40 µM) due to the activity of glutathione reductase, which favors the formation of GSH (23, 47). In this study, GSSG was found to be approximately 3 times more effective than GSH in inhibiting N-SMase, whereas in cells the ratio of GSH/GSSG is generally considered to be in the range of 100-500. Therefore, in cells under normal conditions N-SMase is more likely to be inhibited by GSH, but not GSSG.

In conclusion, inhibition of N-SMase by GSH as shown in this study serves as a direct link between two important areas involved in the regulation of mammalian stress responses, the oxidative stress, and the sphingomyelin/ceramide cycle. A diverse array of agents and factors initiate the hydrolysis of sphingomyelin and the generation of ceramide leading to cell cycle arrest and apoptosis. In addition, many of these same agents and factors cause the depletion of GSH in the course of inducing apoptosis. Further investigation is required to understand the relationship between GSH and sphingomyelinases at the cellular level.