Apoptosis Induced by the Curcumin Analogue EF-24 Is Neither Mediated by Oxidative Stress-Related Mechanisms nor Affected by Expression of Main Drug Transporters ABCB1 and ABCG2 in Human Leukemia Cells

The synthetic curcumin analogue, 3,5-bis[(2-fluorophenyl)methylene]-4-piperidinone (EF-24), suppresses NF-κB activity and exhibits antiproliferative effects against a variety of cancer cells in vitro. Recently, it was reported that EF-24-induced apoptosis was mediated by a redox-dependent mechanism. Here, we studied the effects of N-acetylcysteine (NAC) on EF-24-induced cell death. We also addressed the question of whether the main drug transporters, ABCB1 and ABCG2, affect the cytotoxic of EF-24. We observed that EF-24 induced cell death with apoptotic hallmarks in human leukemia K562 cells. Importantly, the loss of cell viability was preceded by production of reactive oxygen species (ROS), and by a decrease of reduced glutathione (GSH). However, neither ROS production nor the decrease in GSH predominantly contributed to the EF-24-induced cell death. We found that EF-24 formed an adduct with GSH, which is likely the mechanism contributing to the decrease of GSH. Although NAC abrogated ROS production, decreased GSH and prevented cell death, its protective effect was mainly due to a rapid conversion of intra- and extra-cellular EF-24 into the EF-24-NAC adduct without cytotoxic effects. Furthermore, we found that neither overexpression of ABCB1 nor ABCG2 reduced the antiproliferative effects of EF-24. In conclusion, a redox-dependent-mediated mechanism only marginally contributes to the EF-24-induced apoptosis in K562 cells. The main mechanism of NAC protection against EF-24-induced apoptosis is conversion of cytotoxic EF-24 into the noncytotoxic EF-24-NAC adduct. Neither ABCB1 nor ABCG2 mediated resistance to EF-24.


Introduction
Many polyphenolic compounds extracted from plants have been demonstrated to have cancer-preventing activities in laboratory studies. Curcumin, a component of turmeric (Curcuma longa), is a typical example. This agent was reported to inhibit proliferation and survival of cancer cells in vitro, while retaining a pharmacologically safety profile in vivo. However, clinical studies show that curcumin has low efficacy due to rapid excretion in vivo [1,2]. This prompted the development of analogues, which are more potent inducers of apoptosis in in vitro assays and also more efficient in vivo.

EF-24 Did Not Activate Nuclear Factor-Erythroid 2 Related Factor 2 (Nrf2) in K562 Cells
To evaluate the significance of ROS production in the fate of the cell, we addressed the question of whether EF-24 is capable of activating the nuclear factor-erythroid 2 related factor 2 (Nrf2). It is well documented that cells respond to oxidative stress by activation of cellular defense mechanisms that comprise several antioxidant pathways regulated by Nrf2 [19,20]. However, our results clearly indicated that EF-24 did not activate the Nrf2 signaling pathway. Indeed, neither increased levels of Nrf2 nor elevated expression levels of stress-response proteins, including heme oxygenase-1 (HO-1) and NAD(P)H:quinon oxidoreductase 1 (NQO1) were observed in EF-24-treated K562 cells (Figure 4).
These results, together with those described in the previous paragraph, collectively suggest that increased production of ROS played only a marginal role in the induction of the death of EF-24-treated K562 cells.

Conversion of Cytotoxic EF-24 Into the Non-Cytotoxic EF-24-NAC Adduct is the Main Mechanism of the NAC Protection Against the Cytotoxic Effects of EF-24
Excessive oxidation of GSH as a result of oxidative stress, or its conjugation with xenobiotics represent important mechanisms that may lead to the depletion of cellular GSH [22,23]. Given that no increased level of GSSG was found in cells ( Figure 2d) or culture medium (not shown), we studied the interaction of EF-24 and GSH in more detail. As expected, we found intracellular formation of the EF-24-GSH adduct in K562 cells (Figure 5a,b). The amount of EF-24-GSH adduct was proportional to the concentration of EF-24 (not shown). Formation of the EF-24-GSH adduct was significantly reduced by the addition of NAC (Figure 5b). While GSH depletion seemed to be related to the EF-24-GSH adduct formation, the restoration of GSH synthesis using NAC could not fully explain the mechanism of NAC protection against EF-24-induced apoptosis. Indeed, we found that adding NAC to the culture medium induced a rapid intra-and extra-cellular clearance of EF-24 (Figure 6a,b). This was accompanied by intra-and extra-cellular formation of the EF-24 adduct with NAC (Figure 6c-e). These results clearly indicated that conversion of cytotoxic EF-24 into the non-cytotoxic EF-24-NAC adduct was the main mechanism of NAC protection. It is important to point out that EF-24 formed monoand di-adducts with NAC and/or GSH, however, the mono-adducts were predominant (not shown). For this reason, the data presented here refer to the appropriate mono-adducts.

Conversion of Cytotoxic EF-24 Into the Non-Cytotoxic EF-24-NAC Adduct is the Main Mechanism of the NAC Protection Against the Cytotoxic Effects of EF-24
Excessive oxidation of GSH as a result of oxidative stress, or its conjugation with xenobiotics represent important mechanisms that may lead to the depletion of cellular GSH [22,23]. Given that no increased level of GSSG was found in cells ( Figure 2d) or culture medium (not shown), we studied the interaction of EF-24 and GSH in more detail. As expected, we found intracellular formation of the EF-24-GSH adduct in K562 cells (Figure 5a,b). The amount of EF-24-GSH adduct was proportional to the concentration of EF-24 (not shown). Formation of the EF-24-GSH adduct was significantly reduced by the addition of NAC (Figure 5b). While GSH depletion seemed to be related to the EF-24-GSH adduct formation, the restoration of GSH synthesis using NAC could not fully explain the mechanism of NAC protection against EF-24-induced apoptosis. Indeed, we found that adding NAC to the culture medium induced a rapid intra-and extra-cellular clearance of EF-24 (Figure 6a,b). This was accompanied by intra-and extra-cellular formation of the EF-24 adduct with NAC (Figure 6c-e). These results clearly indicated that conversion of cytotoxic EF-24 into the non-cytotoxic EF-24-NAC adduct was the main mechanism of NAC protection. It is important to point out that EF-24 formed mono-and di-adducts with NAC and/or GSH, however, the mono-adducts were predominant (not shown). For this reason, the data presented here refer to the appropriate mono-adducts.   To directly demonstrate that the EF-24-NAC adduct was not cytotoxic, we used a mixture of EF-24 and NAC in distilled water, where the reaction was allowed to proceed until 50% of EF-24 was converted into the EF-24-NAC adduct (not shown). K562 cells were then treated with the diluted reaction mixture. We observed that the proapoptotic effect of the reaction mixture, containing approximately one half of remnant free EF-24 and one half of EF-24-NAC adduct, was equivalent to the effect of remnant free EF-24 (Figure 7). For example, a diluted reaction mixture containing approximately 1 µM EF-24 and 1 µM EF-24-NAC adduct exhibited a proapoptotic effect, which was similar to that of 1 µM EF-24 itself rather than to the effect of 2 µM EF-24 (Figure 7).
To directly demonstrate that the EF-24-NAC adduct was not cytotoxic, we used a mixture of EF-24 and NAC in distilled water, where the reaction was allowed to proceed until 50% of EF-24 was converted into the EF-24-NAC adduct (not shown). K562 cells were then treated with the diluted reaction mixture. We observed that the proapoptotic effect of the reaction mixture, containing approximately one half of remnant free EF-24 and one half of EF-24-NAC adduct, was equivalent to the effect of remnant free EF-24 (Figure 7). For example, a diluted reaction mixture containing approximately 1 μM EF-24 and 1 μM EF-24-NAC adduct exhibited a proapoptotic effect, which was similar to that of 1 μM EF-24 itself rather than to the effect of 2 μM EF-24 (Figure 7). When 50% of EF-24 was converted into the EF-24-NAC adduct, reaction was stopped by dilution. Then K562 cells were treated with EF-24 alone or with diluted reaction mixture containing the same amount of EF-24, however, approximately 50% of it was converted into EF-24-NAC adduct. The experimental points represent mean values from three replicate experiments, with standard deviations; * denotes significant change in the number of cells with apoptotic nuclei (p < 0.05) between K562 cells treated with "free" EF-24 and cells treated with a diluted mixture containing the same amount of EF-24 (50% "free" EF-24 + 50% EF-24-NAC adduct).

Overexpression of ABCB1 and ABCG2 Did Not Compromise the Antiproliferative Effects of EF-24 in K562 Cells
Decreased intracellular drug levels, which prevent effective interaction between the drug and its cellular target, is a generally accepted mechanism of resistance mediated by ABC transporters [24,25]. Therefore, we analyzed extracts from cells expressing ABCB1 and ABCG2, the main drug transporters (Figures 8 and 9). We found that neither the overexpression of ABCB1 nor ABCG2 significantly reduced the intracellular levels of EF-24 ( Figure 9). Accordingly, neither ABCB1 nor ABCG2 expression mediated resistance to EF-24 (Table 2). Importantly, EF-24 exhibited a proapoptotic effect in ABCB1-and ABCG2-expressing cells similar to that found in parental K562 cells (not shown). Our results clearly indicate that EF-24 would exhibit antiproliferative effects, irrespective of ABCB1 or ABCG2 expression. Figure 7. Proapoptotic effects of the EF-24-NAC adduct. EF-24 (50 µM) was mixed with 2 mM NAC in distilled water at ambient temperature and reaction was monitored using LC/MS/MS analysis. When 50% of EF-24 was converted into the EF-24-NAC adduct, reaction was stopped by dilution. Then K562 cells were treated with EF-24 alone or with diluted reaction mixture containing the same amount of EF-24, however, approximately 50% of it was converted into EF-24-NAC adduct. The experimental points represent mean values from three replicate experiments, with standard deviations; * denotes significant change in the number of cells with apoptotic nuclei (p < 0.05) between K562 cells treated with "free" EF-24 and cells treated with a diluted mixture containing the same amount of EF-24 (50% "free" EF-24 + 50% EF-24-NAC adduct).

Overexpression of ABCB1 and ABCG2 Did Not Compromise the Antiproliferative Effects of EF-24 in K562 Cells
Decreased intracellular drug levels, which prevent effective interaction between the drug and its cellular target, is a generally accepted mechanism of resistance mediated by ABC transporters [24,25]. Therefore, we analyzed extracts from cells expressing ABCB1 and ABCG2, the main drug transporters (Figures 8 and 9). We found that neither the overexpression of ABCB1 nor ABCG2 significantly reduced the intracellular levels of EF-24 ( Figure 9). Accordingly, neither ABCB1 nor ABCG2 expression mediated resistance to EF-24 (Table 2). Importantly, EF-24 exhibited a proapoptotic effect in ABCB1and ABCG2-expressing cells similar to that found in parental K562 cells (not shown). Our results clearly indicate that EF-24 would exhibit antiproliferative effects, irrespective of ABCB1 or ABCG2 expression.

Discussion
A large number of studies have demonstrated aberrant NF-κB signaling in solid cancers, as well as in various types of hematologic malignancies. NF-κB is involved in regulating the expression of a number of genes that affect proliferation, cell survival, tumor metastasis, angiogenesis and inflammation. Therefore, targeting aberrant NF-κB activation together with its upstream and downstream interacting regulatory molecules using low molecular weight inhibitors, may be useful in clinical settings for the treatment of solid cancers and hematological malignancies [18].
Our results indicate the strong proapoptotic potential of EF-24 in K562 cells (Figure 1, Table 1). Its proapoptotic effect is much higher than that demonstrated in colorectal or gastric cancer cells [16,17]. This is not surprising, since chronic myelogenous leukemia cells rely on aberrant activation of NF-κB [18], and EF-24 efficiently suppresses NF-κB signaling through direct inhibition of IKK [5,12]. However, several laboratories have shown that the proapoptotic effects of EF-24 are also mediated by a redox-dependent mechanism [14][15][16][17]. A recent publication concluded that EF-24 induces apoptosis via ROS-dependent mitochondrial dysfunction in human colorectal cancer cells [17]. This is supported by the findings that the EF-24-induced cell death was preceded by production of ROS, GSH depletion and mainly by the observation that NAC, a well-known radical scavenger and a precursor of GSH synthesis, prevented cell death [23]. Our results suggest that the EF-24induced cell death with apoptotic features (Figure 1) was preceded by transient production of ROS (not shown) with a peak at 3 h after the EF-24 addition, and by a decrease in GSH levels in human leukemia K562 cells (Figure 2). However, in contrast to the conclusions published by He and coworkers [17], we do not think that ROS production itself is the direct cause of cell death. This

Discussion
A large number of studies have demonstrated aberrant NF-κB signaling in solid cancers, as well as in various types of hematologic malignancies. NF-κB is involved in regulating the expression of a number of genes that affect proliferation, cell survival, tumor metastasis, angiogenesis and inflammation. Therefore, targeting aberrant NF-κB activation together with its upstream and downstream interacting regulatory molecules using low molecular weight inhibitors, may be useful in clinical settings for the treatment of solid cancers and hematological malignancies [18].
Our results indicate the strong proapoptotic potential of EF-24 in K562 cells (Figure 1, Table 1). Its proapoptotic effect is much higher than that demonstrated in colorectal or gastric cancer cells [16,17]. This is not surprising, since chronic myelogenous leukemia cells rely on aberrant activation of NF-κB [18], and EF-24 efficiently suppresses NF-κB signaling through direct inhibition of IKK [5,12]. However, several laboratories have shown that the proapoptotic effects of EF-24 are also mediated by a redox-dependent mechanism [14][15][16][17]. A recent publication concluded that EF-24 induces apoptosis via ROS-dependent mitochondrial dysfunction in human colorectal cancer cells [17]. This is supported by the findings that the EF-24-induced cell death was preceded by production of ROS, GSH depletion and mainly by the observation that NAC, a well-known radical scavenger and a precursor of GSH synthesis, prevented cell death [23]. Our results suggest that the EF-24-induced cell death with apoptotic features (Figure 1) was preceded by transient production of ROS (not shown) with a peak at 3 h after the EF-24 addition, and by a decrease in GSH levels in human leukemia K562 cells (Figure 2). However, in contrast to the conclusions published by He and co-workers [17], we do not think that ROS production itself is the direct cause of cell death. This conclusion stems from following findings. First, although CAT prevented ROS production (Figure 3a), it failed to reduce the antiproliferative effects of EF-24 (Table 1, Figure 3c). Second, even though ROS production and reduced intracellular level of GSH are typical signs of oxidative stress, these were not accompanied by elevated levels of GSSG (Figure 2d). In contrast to expectations, we observed that, both GSH and GSSG decreased, however, the decrease in GSSG was somewhat lower than that of GSH for 2 µM EF-24 (Figure 2d). There was no evidence for any other mechanisms of GSH depletion, such as leakage of GSH and/or GSSG into the growth medium (not shown). Third, we also failed to find any signs of Nrf2 activation. Nrf2 is a master regulator of cell response to oxidative stress [19,20]. Under "normal" conditions it is maintained at low levels and resides predominantly in the cytoplasm. Upon oxidative stress, Nrf2 levels increase due to diminished proteasomal degradation, and it translocates to the nucleus to trigger transcription of a plethora of genes to cope with the oxidative stress [19,26]. However, no elevated Nrf2 level (Figure 4a,b) or increased expression of antioxidant genes, including HO-1 (Figure 4c,d) or NQO1 (Figure 4e,f) were found in the present study.
Similarly, the decrease in GSH level itself is probably not a direct cause of cell death. Indeed, EF-24, up to 1 µM concentration, had no effect on inducing a significant decrease in GSH, while it induced significant cell death (Figures 1 and 2). In addition, NAC at sub-millimolar concentrations, affected cell survival only partially, despite significantly reduced the drop in GSH level (not shown).
Instead, EF-24 may serve as a Michael acceptor and form adducts with thiols [15]. Similar to others [15], we found that EF-24 forms mono-adducts ( Figure 5) and di-adducts (not shown) with GSH. Since the di-adducts represented only a minor part of the adducts, we analyzed mono-adducts. We hypothesize that the formation of EF-24-GSH adducts may contribute to the decrease in GSH levels (Figures 2c and 3b), since we found no evidence for any other mechanism of GSH depletion, such as leakage of GSH and/or GSSG into the growth medium (not shown).
NAC prevented the adverse effects of EF-24, including cell death (Figures 1 and 2). However, we do not agree with the interpretation that NAC prevented EF-24-induced cell death due to the abrogation of ROS production and restoration of GSH synthesis [14,15,17]. Our results strongly suggest that addition of NAC at millimolar concentrations into the culture medium induces a rapid conversion of EF-24 into EF-24-NAC adduct ( Figure 6) which is non-cytotoxic (Figure 7). Based on our results, we believe that this is the main mechanism of NAC protection against EF-24 cytotoxicity. A summary of our proposed mechanism is shown in Figure 10. It is necessary to note that EF-24 forms mono-adducts (Figure 6c,d) and di-adducts (not shown) with NAC. The di-adducts represented only a minor percentage of adducts. The results presented refer to the mono-adducts for this reason. conclusion stems from following findings. First, although CAT prevented ROS production ( Figure  3a), it failed to reduce the antiproliferative effects of EF-24 (Table 1, Figure 3c). Second, even though ROS production and reduced intracellular level of GSH are typical signs of oxidative stress, these were not accompanied by elevated levels of GSSG (Figure 2d). In contrast to expectations, we observed that, both GSH and GSSG decreased, however, the decrease in GSSG was somewhat lower than that of GSH for 2 μM EF-24 (Figure 2d). There was no evidence for any other mechanisms of GSH depletion, such as leakage of GSH and/or GSSG into the growth medium (not shown). Third, we also failed to find any signs of Nrf2 activation. Nrf2 is a master regulator of cell response to oxidative stress [19,20]. Under "normal" conditions it is maintained at low levels and resides predominantly in the cytoplasm. Upon oxidative stress, Nrf2 levels increase due to diminished proteasomal degradation, and it translocates to the nucleus to trigger transcription of a plethora of genes to cope with the oxidative stress [19,26]. However, no elevated Nrf2 level (Figure 4a Similarly, the decrease in GSH level itself is probably not a direct cause of cell death. Indeed, EF-24, up to 1 μM concentration, had no effect on inducing a significant decrease in GSH, while it induced significant cell death (Figures 1 and 2). In addition, NAC at sub-millimolar concentrations, affected cell survival only partially, despite significantly reduced the drop in GSH level (not shown).
Instead, EF-24 may serve as a Michael acceptor and form adducts with thiols [15]. Similar to others [15], we found that EF-24 forms mono-adducts ( Figure 5) and di-adducts (not shown) with GSH. Since the di-adducts represented only a minor part of the adducts, we analyzed mono-adducts. We hypothesize that the formation of EF-24-GSH adducts may contribute to the decrease in GSH levels (Figures 2c and 3b), since we found no evidence for any other mechanism of GSH depletion, such as leakage of GSH and/or GSSG into the growth medium (not shown).
NAC prevented the adverse effects of EF-24, including cell death (Figures 1 and 2). However, we do not agree with the interpretation that NAC prevented EF-24-induced cell death due to the abrogation of ROS production and restoration of GSH synthesis [14,15,17]. Our results strongly suggest that addition of NAC at millimolar concentrations into the culture medium induces a rapid conversion of EF-24 into EF-24-NAC adduct ( Figure 6) which is non-cytotoxic ( Figure 7). Based on our results, we believe that this is the main mechanism of NAC protection against EF-24 cytotoxicity. A summary of our proposed mechanism is shown in Figure 10. It is necessary to note that EF-24 forms mono-adducts (Figure 6c,d) and di-adducts (not shown) with NAC. The di-adducts represented only a minor percentage of adducts. The results presented refer to the mono-adducts for this reason.  We believe that this mechanism of NAC protection plays a crucial role in experimental systems with other cytotoxic agents. For example, cell death induced by geldanamycin (GDN) or by carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone (FCCP) at high micromolar concentrations is preceded by ROS production and GSH depletion, and can be prevented by NAC addition to the culture medium. In these experimental systems, NAC prevents GDN and FCCP cytotoxicity by conversion of these cytotoxic drugs into corresponding non-cytotoxic adducts with NAC [27,28].
EF-24 may have promise as an anticancer compound applicable in cases of drug resistance, specifically for those which express main drug transporters, ABCB1 and ABCG2 [29]. Indeed, neither ABCB1 nor ABCG2 overexpression significantly reduced the intracellular level of EF-24 in K562 cells ( Figure 9). Correspondingly, cells overexpressing ABCB1 or ABCG2 were sensitive to EF-24 in a way that is similar to the parent K562 cells ( Table 2). Since drug resistance strongly depends on transporter expression levels [30][31][32][33], we used cells with various expression levels of ABCB1 or ABCG2 in this study. Importantly, cells with high expression levels of drug transporters, which are often used in laboratory experiments, might exhibit a distinct resistance to the studied drug. In contrast, cells expressing lower levels of drug transporters, which may occur in clinical samples [34][35][36], might exhibit a much lower degree of resistance to the particular drug, or the resistance can be completely lost. However, neither moderate nor high expression levels of either transporter failed to mediate resistance to EF-24 in our study (Figures 8 and 9, Table 2). These results indicate that the antiproliferative potential of EF-24 cannot easily be decreased by overexpression of the main drug transporters, ABCB1 and ABCG2, neither in clinical settings nor in the laboratory. In this context, the properties of EF-24 are very similar to those of curcumin, which is not transported by ABCB1 or ABCG2 [37,38].
In conclusion, the results strongly suggest only marginal contribution of a redox-dependent mechanism to EF-24 induced apoptosis in human leukemia K562 cells. The main mechanism of NAC protection against EF-24-induced apoptosis is the conversion of cytotoxic EF-24 into the non-cytotoxic EF-24-NAC adduct. Drug transporters, ABCB1 and ABCG2, do not reduce the antiproliferative effects of EF-24 in K562 cells.
K562/ABCG2CL10 and K562/ABCG2CL1 with high and low expression level of ABCG2, respectively, were used in this study. They were established by a single cell cloning by limiting dilution of K562/ABCG2 cells [31].
Resistant cells were cultured under the same conditions as maternal K562 cells.

Determination of Cell Survival and Proliferation
Cell viability and proliferation was determined using the MTT assay as described previously [42].

Analysis of Cell Cycle and Apoptotic Cells
Flow cytometric measurements of DNA content were used to analyze the cell cycle as well as for identification of apoptotic cells (fraction of cells in the sub G1 phase), as described previously [27,43]. Apoptotic cells are expressed as a percentage of cells in sub G1 phase.

Morphological Analysis of Apoptosis
Fixed cells were stained with Hoechst 33342 Sigma-Aldrich (St. Louis, MO, USA) and morphology of cell nuclei was examined using an Olympus BX60 (Olympus, Hamburg, Germany) fluorescence microscope as described previously [44].

Measurement of Caspase-3 Enzymatic Activity
Caspase-3 enzymatic (DEVDase) activity was measured in cytoplasmic extracts using the fluorescent substrate Ac-DEVD-AMC [45]. It is necessary to note that Ac-DEVD-AMC substrate is efficiently cleaved also by caspase-7. Therefore, we used the term "DEVDase activity" when refering to the cleavage of Ac-DEVD-AMC substrate in the text.

Measurement of Caspase-3 Processing
Western blot analysis was used to directly demonstrate capase-3 processing. Briefly, protein extracts and sample preparation were done as described previously [46]. Caspase-3 processing was determined using Western blot analysis with a polyclonal anti-caspase-3 antibody (1:1000; Cell Signaling Technology, Danvers, MA, USA) recognizing both pro-and active protease forms and polyclonal anti-HSP90 antibody (1:2000; Cell Signaling Technology, Danvers, MA, USA) for detection of reference protein. A horseradish peroxidase-conjugated secondary anti-rabbit antibody (1:2000; Dako, Glostrup, Denmark) in combination with an enhanced chemiluminiscence (ECL; Amersham, Little Chalfont, UK) was used for signal detection.

Western Blot Analysis of Activation of Nrf2 and Stress-Response Pathway
Western blot analysis was used to demonstrate the activation of Nrf2 and its downstream regulated genes, including HO-1 and NQO1 [20]. Protein extracts were done as described previously [46]. Mouse

Preparation of Cell Extracts
An optimized acidic extraction of cells after their separation from the culture medium by centrifugation through a layer of silicone oil with a slight modification was used [27,28,47]. Briefly, cells at a density of 5 × 10 5 /mL were incubated in the culture medium (with or without EF-24) for appropriate time periods at 37 • C. Afterwards, cells were centrifuged through silicone oil and cell pellets were extracted as follows: ice cold 5% (v/v) formic acid was used for GSH and EF-24 analysis; or ice cold 4% (v/v) formic acid in 40% (v/v) methanol in water was used for EF-24-GSH and EF-24-NAC adduct analysis. Clarified cell extracts (centrifugation: 40,000× g 10 min at 4 • C) were diluted with distilled water and analyzed by liquid chromatography coupled with a low-energy collision tandem mass spectrometer (LC/MS/MS). Alternatively, clarified cell extracts were stored at −80 • C.

High-Performance Liquid Chromatography (HPLC) Analysis of Glutathione (GSH) and Oxidized Glutathione (GSSG)
Quantitative analysis of GSH and GSSG was done using LC/MS/MS during one run with the specific parameters. The chromatographic separations were performed using the high-performance liquid chromatography (HPLC) tower system UltiMate 3000 (Dionex, Germering, Germany), a Polaris C18-A, 5 µm, 250 × 2.0 mm HPLC column (Varian Inc., Lake Forest, CA, USA), and a guard C18, 4.0 × 2.0 mm precolumn (Phenomenex, Torrance, CA, USA). The chromatographic parameters were as follows: the binary gradient of mobile phase A (95% methanol in 0.25% formic acid, v/v) and B (0.25% formic acid in water, v/v) from 0-3 min (5 → 23% of solvent A), from 3-4 min (23 → 95% of solvent A), from 4-6 min (95 → 5% of solvent A) and from 6-10 min (5% of solvent A); the flow rate at 0.3 mL/min; the sample injection volume at 5 µL. The API 3200 triple quadrupole mass spectrometer (MDS SCIEX, Concord, ON, Canada) with the TurboIonSpray interface in the positive ion mode was applied for quantification of analytes. The Product Ion Scan mode (GSH: Q1 quadrupole 308.1 amu, Q3 quadrupole at scale 178.95-179.05 amu and GSSG: Q1 quadrupole 613.1 amu, Q3 quadrupole at scale 230.6-231.4 amu) was used. The mass-dependent parametres were optimized: the collision energy and the declustering potential for GSH standard were 17 V and 26 V, and for GSSG standard were 45 V and 51 V, respectively. Ion spray probe parameters were set for GSH and GSSG standards: needle voltage 5500 V and temperature 450 • C. Data were acquired using Analyst ® software, ver. 1.5.1 (MDS SCIEX, Concord, ON, Canada).

Statistical Analysis
Data are reported as means ± S.D. All statistical analyses were performed using SigmaPlot 11.0 software package (Systat Software Inc., San Jose, CA, USA). Statistical significance of differences was determined by Student's t-tests and one-way ANOVA. p values equal to or less than 0.05 were considered significant.