Oxidative species and S-glutathionyl conjugates in the apoptosis induction by allyl thiosulfate


  • Note
    Ridvan Nepravishta and Renato Sabelli contributed equally to this work

S. Melino, Dipartimento di Scienze e Tecnologie Chimiche, University of Rome ‘Tor Vergata’, via della Ricerca Scientifica, 00133-Rome, Italy
Fax: +39 067 259 4328
Tel: +39 067 259 4449
E-mail: melinos@uniroma2.it


Natural allyl sulfur compounds show antiproliferative effects on tumor cells, but the biochemical mechanisms underlying the antitumorigenic properties of the organ sulfur compounds are not yet fully understood. Sodium 2-propenyl-thiosulfate is a garlic water-soluble organo-sulfane sulfur compound able to promote apoptosis in cancer cells, affecting the ‘managing’ of the redox state in the cell. Our studies show that sodium 2-propenyl-thiosulfate reacts spontaneously with reduced glutathione at physiological pH, leading to the formation of S-allyl-mercapto-glutathione, radicals and peroxyl species, which are able to induce inhibition of enzymes with cysteine in the catalytic site, such as sulfurtransferases. S-Allyl-mercapto-glutathione was purified and characterized by NMR and MS, and its cytotoxic effect at 500 μm on HuT 78 cells was analyzed, showing activation of the p38–MAPK pathway. Many allyl sulfur compounds are also able to promote chemoprevention by induction of xenobiotic-metabolizing enzymes, inducing down-activation or detoxification of the carcinogens. Thus, the effects of the S-allyl-mercapto-glutathione on proteins involved in the cellular detoxification system, such as glutathione S-transferase, have been evaluated both in vitro and in HuT 78 cells. Although the antitumor properties of water-soluble sulfur compounds may arise from several mechanisms and it is likely that more cellular events occur simultaneously, a relevant role is played by the formation of both reduced glutathione conjugates and radical species that affect the activity of the thiol-proteins involved in fundamental cellular processes.




reduced glutathione




glutathione S-transferase


horseradish peroxidise


organ sulfur compound


tetramethyl benzidine


thiosulfate sulfurtransferase


The medicinal uses of Allium sativum have a long history [1] and its uses as a remedy for heart disease, tumors and headaches are documented in the Egyptian Codex Ebers, dating from 1550 bc. Recent studies have validated many of the medicinal properties attributed to garlic and its potential to lower the risk of disease. Thus, in recent years, several studies of either oil- or water-soluble garlic compounds have been carried out to verify their chemopreventive and anticarcinogenic effects, and to explain the mechanism of their action. Chemopreventive agents may exert antiproliferative effects via induction of cell-cycle arrest or apoptosis, induction of terminal differentiation and inhibition of oncogene activity or DNA synthesis [2].

Allicin is the main sulfur compound formed by alliin upon crushing of garlic cloves [3], it is an unstable compound and is rapidly transformed into secondary products in vivo, such as allylmercaptans and others that are also SH-modifying reagents [4–6]. Propenyl thiosulfinates are major species of organ sulfur compounds (OSCs) in fresh onion tissue macerates [7], thiosulfinates can freely permeate cell membranes and are most stable at the acidic pH of the stomach and also at duodenal pH; furthermore, the response to this compound has been shown to be also related to the number of sulfur atoms [3,7–9]. Recently, many of these OSCs have been shown to suppress the proliferation of cancer cells through the induction of programmed cell death. This induction leads to relevant questions about the metabolism of OSCs and their role in the activity of enzymes involved in detoxification of the cellular system. In recent years, it has been shown that increased garlic consumption diminishes the risk of stomach and colorectal cancer [10]. A variety of garlic allyl sulfur compounds have been reported to induce growth rate arrest in tumor cells, either in culture or in vivo [11–14]. Several proteins that regulate cellular proliferation can be damaged by S-allylsulfides and are considered to be the target proteins involved in the anticancer activity of garlic. Previous studies show that expression of kinase C, cyclin-dependent kinase 2, AP-1, p21waf1, p27Kip1, and especially NF-κB and p53, is changed most consistently by garlic treatment [13,15–17]. The biochemical mechanisms underlying the antitumorigenic and antiproliferative effects of garlic-derived OSCs are not yet fully understood, although it seems likely that the rate of clearance of allyl sulfur groups from cells is a determinant of the overall response. Recently, it has been proposed that exposure of tumor cells to allylsulfides and cysteinyl S-conjugates affects the intracellular redox status either by production of reactive oxygen species or reactive nitrogen species, or by altering the reduced glutathione (GSH) content [18]. GSH represents the main intracellular redox buffer, its concentration is organ dependent, varying between 2 and 10 mm [19], and its levels modulate the redox state of the cell. It is known that thiosulfinate, as well as some selenium compounds [20], can react instantaneously with GSH in the living cell to form intracellular GSH mixed-disulfide conjugates [6,21,22]. In this study, we used HPLC, NMR spectroscopy, MS and cyclic voltammetry to investigate the effects of the water–soluble proapoptotic garlic compound, 2-propenyl-thiosulfate (2-PTS) [23–25], on the antioxidant cellular system, by analyzing the products of the in vitro reaction with GSH. Moreover, the effects of the reaction products on enzymes, which are implied in the cellular detoxification system, were evaluated. Usually, targets for radical species in the proteins are highly conserved cysteine residues whose redox status is crucial for their activity. Sulfydril centers are reactive sites in redox-sensitive proteins that are essential in functional signal transduction and transcription events.

The cyanide:thiosulfate sulfurtransferase (TST; EC enzyme, which is characterized by the presence of a catalytic cysteine residue and by high structural homology with Cdc25 phosphatases [26], was used as a model to investigate the effects of the products of the 2-PTS–GSH reaction on the activity of the homologous classes of enzymes with sulfydril active centers. S-allyl-mercapto-glutathione (GSSA) was also generated with high yield by a 2-PTS–GSH reaction in vitro, and this glutathione derivative was characterized and tested for its anticarcinogenic properties, suggesting a possible role for GSH in the antiproliferating effect of 2-PTS. It has been observed that thiosulfinates and polysulfides can modify the drug detoxification system by upregulation of phase II enzymes [27–29] and furthermore, changes in the activity of detoxification enzymes may be due to reaction with thiol groups [18,25,30]. Thus the effects of GSSA on the cell detoxification system were investigated in vitro and in treated HuT 78 cells.


2-PTS and the cell redox system: GSSA synthesis

Previously, the effects of 2-PTS on the redox system were investigated and oxidation of thioredoxin was observed [25]. Thioredoxin and GSH represent the two principal molecules that regulate the cell redox state. In this study, we analyzed the interaction of 2-PTS with GSH in vitro.

The reaction between GSH and 2-PTS was performed in 50 mm Tris/HCl buffer, pH 7.4, using a 4 : 1 2-PTS/GSH molar ratio and then analyzed by RP-HPLC. Figure 1A shows the RP-HPLC profile of the 2-PTS/GSH mixture and it is possible to see, in addition to the a and b peaks which correspond to the mix of 2-PTS, GSH, their oxidated forms and Tris, the presence of a stable compound eluted at a retention time of 21.81 min (peak c), which represents the third principal component of the mixture. Peak c was collected and then analyzed by ESI MS yielding a mass of 380.07 ± 0.07 m/z (Fig. 1B). The molecular mass of peak c was compatible with a glutathione-propenyl-sulfide derivative, also named GSSA [6]. Structural characterization of peak c was also performed by 1H NMR spectroscopy (Fig. 1C) showing the characteristics resonances of the S-allyl-mercapto-glutathione (Table 1). GSSA formation obtained in vitro also suggests a possible interaction of 2-PTS with GSH in vivo, leading to GSSA production. A 2-PTS/GSH mixture obtained using a 2 : 1 molar ratio was also analyzed by RP-HPLC, resulting in a GSSA amount lower than that obtained using a molar ratio of 4 : 1. As previously observed in the case of GSSA formation by the allicin/GSH reaction [6], also in this case, a dependence of the reaction on high pH values was observed, indicating dependence on the concentration of GS.

Figure 1.

 Characterization of the 2-PTS/GSH mixture. (A) RP-HPLC of the PTS/GSH mixture (25 mm GSH and 100 mm 2-PTS) after 1 h of the reaction. The elution was performed with a linear gradient of 80% CH3CN, 0.1% TFA as described in Materials and Methods (B) ESI mass spectrum of the c peak, 380.07 ± 0.07 m/z. (C) 1H NMR spectrum of the c peak in D2O solution.

Table 1. 1H NMR chemical shifts for glutathione, sodium-2-propenyl-thiosulfate (2-PTS) and their disulfide conjugates in D2O. s, singlet; d, doublet; t, triplet; m, other multiplets.
CompoundsGroupsH chemical shift (ppm)aMultiplicityb
  1. a Proton chemical shift are reported with reference to 3-(trimethylsilyl)-propionic-2,2,3,3-d4 acid sodium salt at 0 ppm. b The multiplicity given here was observed in conventional 1D spectra recorded at 400 MHz.

Glutathione reduced form (GSH)
 Glycine moietyCH23.95s
 Cysteine moietyCH4.56dd
 Glutamate moietyCH3.82t
Glutathione oxidated form (GSSG)
 Glycine moietyCH23.91s
 Cysteine moietyCH3.28dd
 Glutamate moietyCH3.78t
S-allyl-mercapto-glutathione (GSSA)
 Glycine moietyCH23.95s
 Cysteine moietyCH3.24dd
 Glutamate moietyCH3.79t
 2-PTS moiety–CH23.38d

Cysteine–enzyme inhibition and radical species formation by 2-PTS–GSH reaction

In a previous study, we showed the ability of 2-PTS to induce inhibition of TST. TST from Azotobacter vinelandii was used as model to assay the effects of both a 2-PTS/GSH mixture and GSSA on thiol-containing proteins, for high 3D structural homology with Cdc25 phosphatase and for the characteristic presence of only one cysteine residue, which is also the catalytic residue. The effects of the 2-PTS/GSH mixture on TST activity were analyzed in vitro (Fig. 2), showing that the mixture is more able to inhibit TST activity in the E form of the enzyme. Previously, we have showed that 2-PTS is able to bind the sulfur-free form (E) of TST, inhibiting its activity by thiolation of the catalytic cysteine [25], but inhibition of the 2-PTS/GSH mixture was higher than 2-PTS alone at the same concentrations. By contrast, purified GSSA is not able to significantly inhibit TST activity in vitro (Fig. 2B), although a quenching of the intrinsic fluorescence of both forms (ES and E) of TST changed with the increase in GSSA concentration in solution (Fig. 3A,C). Quenching of the intrinsic fluorescence of both forms of the enzyme, which are characterized by two different fluorescence spectra [31,32], and the absence of inhibition may indicate unspecific interaction of the GSSA with the surface of the enzyme that does not involve the active site of the enzyme or induce relevant conformation changes in the protein.

Figure 2.

 2-PTS/GSH mixture and GSSA affect the TST activity. (A) Inhibitory effect of 2-PTS/GSH mixture on TST activity. Time-dependence decreasing of TST activity of 5 μm TST in 50 mm Tris/HCl buffer, pH 8.0, 0.3 m NaCl in the presence of a threefold molar excess of KCN (15 μm) and of 200-fold molar excess of 2-PTS (1 mm) (□) or 2-PTS/GSH mixture (1 mm and 250 μm respectively) (Δ) at 37 °C. (B) Time dependence decreasing of TST activity (8.75 μm TST) in the presence of a threefold molar excess of KCN (□) (26.25 μm) and of 26.25 μm KCN and a 400-fold molar excess of GSSA (3.5 mm) (inline image) at 37 °C. All values are expressed in percent of TST activity value of ES form.

Figure 3.

 Effects of GSSA on the TST intrinsic fluorescence. Intrinsic fluorescence of 3 μm TST (E) at different molar ratios of GSSA (A) and GSH (B), and of 3 μm TST (ES) at different molar ratios of GSSA (C). 1/0 (____), 1/1 (inline image), 1/10 (_ _ _ _), 1/100 (- - - -) and 1/150 (inline image) TST/GSH or GSSA c/c.

TST activity was also unaffected by preincubation of the E form in the presence of GSH, at the same concentrations and conditions used for the mixture (data not shown) and no changes in the intrinsic fluorescence of the E form were observed after addition of GSH (Fig. 3B).

Taken together, these results suggest that TST inhibition in the presence of the the 2-PTS–GSH mixture may be due to the formation of oxidative species during the 2-PTS–GSH reaction. The reductive ability of thiols, particularly GSH, makes them an effective antioxidant in vivo and in vitro.

Thus, we investigated the species produced during the reaction that led to the formation of GSSA. The formation of radicals and peroxyl species during the 2-PTS–GSH reaction was analyzed by cyclic voltammetry after 30 min of the reaction. A horseradish peroxidase (HRP; EC reaction was used in the presence of tetramethyl benzidine (TMB) as an electron donor [33,34]. Cyclic voltammetry was carried out using screen-printed electrodes (2 × 10−4 m) in 0.1 m citrate–phosphate buffer, pH 5.0. The 2-PTS–GSH mixture was analyzed after 1 min incubation with HRP and TMB and 70 μL of the reaction mixture was put directly onto the screen-printed electrode for electrochemical measurement of the enzyme product [TMB(Ox)].

Figure 4A shows the current values of the oxidation peak of TMB as a function of the concentration of 2-PTS alone compared with the reaction mixture 2-PTS–GSH. As expected, in the case of the 2-PTS mixture, the current increases with increasing reagent concentrations, indicating the formation of a larger amount of product. The current value of the peak is an indirect measurement of the 2-PTS–GSH reaction products. In fact, TMB(Ox) formation, which is due to a reaction catalyzed by HRP, indicates the presence of peroxide/peroxyl species in the solution. Oxidation of thiols by peroxidase has been shown to yield thiyl radicals [33,35], thus we have performed an analysis of both 2-PTS (Fig. 4A) and GSH (Fig. S1) separately. By contrast to the mixture, single compounds do not show a concentration-dependent increase in the positive band in the presence of HRP, indicating the absence of peroxyl species in the solution. Figure 4B shows the voltammograms for the mixture at different times of the 2-PTS–GSH reaction at room temperature after reaction with HRP and it is possible to observe the time-dependent formation of the radical species.

Figure 4.

 Cyclic voltammetric investigation of the 2-PTS/GSH mixture in 0.1 m citrate–phosphate buffer, pH 5.0, with 0.1 m KCl. (A) Currents with respect to TMB(Ox) at different concentrations of a 2-PTS/GSH mixture (40 and 10 mm respectively) and of 2-PTS solution (40 mm) alone, in 50 mm Tris/HCl, pH 8.0, buffer after 30 min incubation at 25 °C. Before analysis, the solutions were incubated with 20 U·mL−1 HRP for 1 min at room temperature and 0.2 mm TMB was added to the solutions. (B) Cyclic voltammetry of a 2-PTS/GSH mixture after 30 min (b) and 60 min (c) of the reaction. Before the measurements, the solutions were incubated with 20 U·mL−1 HRP for 5 min at room temperature and 0.2 mm TMB was added to the solutions. (a) Voltammograms of TMB in buffer.

Cyclic voltammetry of the mixture and of the individual components in the absence of HRP was also performed and a significant shift in the oxidation current over time was observable only in the case of the 2-PTS–GSH mixture, indicating formation of the radical species in the reaction (Fig. S2). Moreover, analysis of the mixture at different times of the 2-PTS–GSH reaction was also performed using a Prussian Blue screen-printed electrode, indicating the formation of the radical species, in agreement with previous results (data not shown).

Taking into account that Allium derivatives S-allyl-l-cysteine and diallyl trisulfide are hydrogen sulfide (H2S)-releasing, formation of H2S during the 2-PTS–GSH reaction was also evaluated by methylene blue formation, but no significant H2S formation was detected (data not shown).

Inhibition of cell-cycle progression of the HuT 78 cells by GSSA

The effects of purified GSSA on cell growth in the human T lymphoblastoid cell line, HuT 78, were checked and a typical time- and dose-dependent inhibition of cell growth in these cells was observed. A statistically significant decrease in the number of viable cells was evidenced at different concentrations of GSSA (from 0 to 0.5 mm) and at different times, compared with the control (Fig. 5A). Approximately 90% and 80% of the HuT 78 cells were viable following exposure to 0.5 mm GSSA for 24 and 48 h, respectively (Fig. 5B). The growth inhibitory effects of GSSA were assessed by Trypan Blue dye exclusion assay. Flow cytometric analysis of HuT 78 cells after 24 h treatment with 0.5 mm GSSA resulted in a statistically significant increase in the fraction of subG1 over the control and a blockage in the G2/M phase (Fig. 5).

Figure 5.

 Effects of GSSA on cell growth of HuT 78 cells. (A) HuT 78 cell viability after addition in culture of different concentrations of GSSA (0, 30, 60, 125, 250 and 500 μm). Trypan Blue staining was used to differentiate viable cells. (B) Cell viability dependence at 24 and 48 h as a function of the GSSA concentration. Data are expressed as mean ± SD. *P < 0.005, **P < 0.001 (n = 4). (C) Percentage cell-cycle distribution of HuT 78 cells after 24 and 48 h of treatment with 0.5 mm GSSA detected using FACSCalibur flow cytometer. Data are expressed as mean ± SD. *P < 0.001, **P < 0.05 (n = 3). (D) Western immunoblotting showing P-p38 expression in HuT 78 cell lysates after GSSA treatment, 30 μg of proteins from the lysates of untreated cells (CTRL) and 0.5 mm of GSSA-treated cells after 4, 9 and 24 h incubation. A monoclonal anti-actin IgG was used as a control.

We also investigated regulatory proteins that might be involved in cell-cycle progression. p38 MAP kinase is a member of a pathway that responds to cellular stress and is linked to the cell cycle through senescence and the differentiation pathway [36,37]. Western blot analysis of HuT 78 lysates after GSSA treatment (Fig. 5C) shows that expression of the phosphorylated form of p38 (P-p38) was markedly increased, suggesting that the antiproliferative effect of GSSA is related to activation of the p38 pathway in time-dependent manner in HuT 78 cells.

Effects of GSSA on glutathione S-transferase expression and activity in vitro and in HuT 78 cells

Interestingly, it has recently been demonstrated that GSSA is a potent quinone reductase inductor at micromolar concentrations [38]. In this study, the effects of GSSA on the xenobiotic export system have been investigated, evaluating its influence on glutathione S-transferase (GST; EC activity either in vitro or in a cell system. GST activity in the presence of different concentrations of GSSA was assayed, as described in Materials and Methods. GSSA was able to inhibit GST activity in a concentration-dependent manner (Fig. 6A). IC50 and Ki were measured at ∼ 0.6 and 0.1 mm, respectively, and were comparable with inhibition of the S-methylglutathione inhibitor, suggesting a similar mechanism of inhibition.

Figure 6.

 Effects of GSSA on GST activity and expression. (A) In vitro GST activity in the presence of different concentrations of GSSA. The inhibition of the GST activity by GSSA showed an IC50 value of ∼ 0.6 ± 0.07 mm and apparent ki of 0.10 ± 0.01 mm. (Inner) Plot of the GST activity. (B) GST activity of HuT 78 cells after 4, 9 and 24 h treatment with 0.5 mm GSSA. (C) Western immunoblotting showing expression in HuT 78 cell lysates after GSSA treatment, 30 μg of proteins from the lysates of untreated cells after 24 h (CTRL) and treated cells with 0.5 mm GSSA after 4, 9 and 24 h incubation. A monoclonal anti-actin IgG was used as control of the concentrations.

The effects of GSSA on GSTP1-1 enzyme expression and total cellular GST activity in HuT 78 cells were analyzed. Figure 5C shows the western blot of the HuT 78 lysates after 24 and 48 h of treatment with 0.5 mm GSSA. Densitometry measurements of western blots, corrected for actin expression, show that no significant variation in GST expression with respect to controls was induced by treatment with GSSA. Total GST activity was also assayed after GSSA treatment, as shown in Fig. 6B, and no variation in total GST activity was observed, although the sulfur compound was able to inhibit GST activity in vitro.


Allyl sulfur compounds are the dominant species in both garlic (50–94%) and fresh onion tissue macerates [7], and are widely believed to be responsible for the health benefits of these plants [3,7]. The study of the biological effects of their mixed-disulfide conjugate derivatives may highlight on their health benefits. The bioavailability of thiosulfinates approaches 100%, and they are metabolized rapidly in human blood in vitro, with a half-life of < 1 min [39]. Their cellular membrane permeability and high reactivity with thiols [6,21,40] may account for the short half-life of thiosulfinates in physiological systems. In recent years, several studies have focused on elucidation of the mechanism of biological activity in garlic and its allyl sulfide derivatives, such as allicin, diallyl disulfide, diallyl trisulfide and, more recently, 2-PTS [25,41–46]. Not all cells are equally susceptible to the deleterious effects of the garlic sulfur compounds and, in particular, non-neoplastic cells tend to be less susceptible, suggesting that the uncontrolled proliferation of the neoplastic cell might be also related to a sulfane sulfur deficiency due to an incorrect functionality of the enzymes involved in their catabolism [47]. Previously, it has been suggested that erythrocyte GSH accelerates the oxidative damage produced by 2-PTS, indicating that this compound may react with GSH, which might convert to GSSG, resulting in the activation of n-propylthiosulfate as an oxidant [48]. In this study, we have shown that 2-PTS, at physiological conditions, is able to react spontaneously with reduced GSH, forming the S-allyl-mercaptoglutathione. Mixed-disulfide conjugates of thiosulfinates with biological/physiological thiols represent major derivatives that form in the gut or through metabolic processes in vivo. GSH mixed-disulfide conjugates would most likely be formed intracellularly after consumption of thiosulfinates [6] or through metabolic activity [49,50]. Moreover, GSH mixed-disulfide conjugate species might remain in the extracellular space until metabolized, thus it is likely that GSH mixed-disulfide conjugates and cysteine-conjugates would be accumulated and act in different extracellular compartments, affecting redox modulation. The biological effects on thiol-containing proteins of both a 2-PTS–GSH mixture and purified GSSA have been tested in vitro using TST from A. vinelandii, showing the major ability of the mixture to inhibit TST activity with respect to 2-PTS alone. By contrast, purified GSSA was able to interact with the enzyme, but showed very low ability to inhibit TST activity at high concentrations. The evidence for production during the reaction of oxidative species, such as peroxyl species and thiyl and/or oxygen radicals, was obtained by cyclic voltammetry analysis. In fact, the oxidizing agent is reduced and the thiols are rapidly oxidized, generating thiyl and thiyl-derived radicals. Thiyl radicals are oxidizing radicals with a potential redox E° (RS·, H+/RSH) = 1.35 ± 0.04 V [51]. Once formed, thiyl radicals can react with the thiolate or a parent thiol to form the disulfide radical anion (RSSR·). Both these reactive species can react with molecular oxygen, and the first can form the thiyl peroxyl radical [52–54], which may be a damaging species involved in protein deactivation [55]. The presence of oxygen can also lead to sulfonylperoxyl radicals (RSOO·, RSO·, RSO2OO·) [55], which also exhibit oxidizing properties. The disulfide radical anion can also react with oxygen, generating the superoxide radical anion: GSSG· + O2 GSSG + O2·.

Thus, TST inhibition may be related to the presence of these species in solution, which was shown to be able to inhibit TST activity in vitro [56].

Reactive oxygen species formation in tumor cells after 2-PTS treatment was also observed in our previous study [25] and was suggested as the earliest event in the cascade of apoptosis induction by 2-PTS treatment of HuT 78 cells. Recently, it has been also shown that 2-PTS increased 4β-phorbol 12-myristate 13-acetate-induced O2 generation, although it is also able to accelerate the reaction catalyzed by NADPH oxidase in canine polymorphonuclear leukocytes [57,58]. This evidence is also in agreement with recent studies on diallyl trisulfide compound, which suggest that diallyl trisulfide can mediate the decline in Cdc25C protein levels in cancer cells because of an oxidation component [44]; this is consistent with redox regulation of Cdc25C stability by oxidation of two specific cysteine residues [59]. Therefore, oxidative stress, as suggested for other OSCs [60], might represent the triggering factor in induced apoptosis by 2-PTS. Reactive oxygen species formation may be also related to an oxidative unbalance due to both the consumption of GSH and to the TST–thioredoxin–thioredoxin reductase system [25].

The effects of 2-PTS glutathionyl derivative GSSA on HuT 78 cells were analyzed, showing a typical time- and dose-dependent inhibition in cell growth. Flow cytometric analysis of HuT 78 cells after GSSA treatment resulted in an antiproliferative effect with a blockage in the G2/M phase.

These results are in agreement with the antiproliferative mechanism of most garlic compounds, which involves cell arrest in the G2/M phase of the cell cycle in cancer cells [44,61]. An increase in the expression of the activated form of p38MAPK after GSSA treatment was also observed, suggesting that the antiproliferative effect of GSSA may be due to activation of the p38 pathway. p38MAPK is a key member of the MAPK family and it is known that p38 is rapidly activated in response to proinflammatory cytokines and stress and cellular damage [62–67]. Most studies are consistent with the interpretation that p38 plays a critical role in progression through G2 and not mitosis. It has been shown that activated P-p38 is distributed throughout the cytoplasm of mitotic cells [68], and also that it is localized in the centrosomes during spindle assembly [69,70]. Recently, it has been demonstrated that diallyl disulfide- and diallyl trisulfide-induced apoptosis in some tumor cell lines is related to activation of the p38MAPK signaling pathway [71–73]. Thus p38MAPK activation might be a common mechanism of action of the allyl sulfur compounds due to the formation of their glutathione conjugate.

The induction of programmed cell death by sulfane sulfur compounds [alk(en)yl thiosulfate, selenodiglutathione, allyl disulfide, etc.] poses significant questions concerning their catabolism in the cell and on the role in cancerogenesis of proteins involved in the detoxification process.

Interestingly, glutathione mixed-disulfide conjugates were found to be able to inhibit nitric oxide production in macrophage cells, showing their potential ability to prevent cancer and the inflammatory process, and to induce quinone reductase in murine hepatoma cells (Hepa 1c1c7) at micromolar concentrations [39]. It is noteworthy, in fact, that anticancer effects can be mediated through the antioxidant response element to afford upregulation of phase II enzymes [27].

It has been shown that the antiproliferative effects might be attributable to a multifactorial mechanism in which the GSTs may play a role, and the detoxification system of tumor cells might be affected by the glutathione derivative of OSCs. Several studies using allyl sulfides have demonstrated a significant decrease in GST activity in hepatocytes after treatment with a high concentration of them [74]. Our results show that, although GSSA is able to inhibit GST activity in vitro, similar to the GST inhibitor S-methylglutathione, both GSTP1-1 expression and cellular GST activity do not change after GSSA treatment of HuT 78 cells. In general, GSH-conjugates are exported from cells by energy-dependent GS-xenobiotic pumps; also known as multidrug resistance proteins, these transport mechanisms, designated as phase III of the detoxification process [75], are therefore an essential component of cellular defense mechanisms against toxic chemicals [76]. The phase II reaction products must eventually be transported to complete detoxification, because accumulation of these products can not only cause toxicity, but also inhibit the phase II reactions. It has been suggested that the chemopreventive agents present in garlic might have a significant effect on chemotherapeutic treatments using GS-xenobiotic pump substrates [77]. It has been observed that diallyldisulfide (DAS) is a selective and highly potent modulator of P-glycoprotein-mediated multidrug resistance in human K562 leukemic cells and in rodent liver [78]. Thus, further studies are required to evaluate the effects of GSSA on the cell detoxification system and the possibility that this glutathione mixed-disulfide conjugate may act on the P-glycoprotein protein cannot be ruled out. Further studies on the relationship between sulfur metabolism and the p38MAPK signal transduction pathway in cellular apoptosis will lead to better clarification of the mechanism of apoptosis and will have important significance for the investigation into the antitumor mechanisms of OSCs and the design of new drugs.

Materials and methods

2-PTS synthesis

2-PTS was synthesized according to a method described by Chapelet et al. [79]. The product was dried in vacuum and extracted with methanol. The extract was then purified by silica gel chromatography (methanol/chloroform 45 : 55 v/v). The purity and structure of the compound were evaluated by RP-HPLC, LC-MS and 1H NMR [25].

Preparation of RhdA and GST proteins

The plasmid pQER1 containing the gene coding for N-terminal His-tag A. vinelandii TST was expressed and purified as previously described [80]. The protein concentration was determined using inline image = 1.3 [81] and the molecular mass of the E form was estimated at 31 063.8 kDa by ESI MS. TST(E) form was prepared by adding a 10-fold molar excess of cyanide to ES rhodanese or using 4 mm dithiothreitol, in 50 mm Tris/HCl buffer pH 7.4 and 0.3 m NaCl followed by 30-min incubation at room temperature. Excess cyanide, thiocyanate or dithiothreitol were removed using gel-filtration chromatography (Nap™ 25 column; GE Amersham, Milan, Italy) or by 48 h dialysis at 4 °C. ES → E form conversion was monitored by increasing the fluorescence quantum yield, which is accompanied by removal of the persulfide sulfur [31,32]. GST expression and purification were performed in Escherichia coli, as descried by Battistoni et al. [82].

Kinetic analysis

TST activity was measured using the discontinuous colorimetric assay, as described by Sörbo [83], in which the production of thiocyanate from thiosulfate and cyanide was followed at 460 nm using a Perkin–Elmer spectrometer. GST activity was measured using the GST assay kit (CS0410; Sigma-Aldrich, Milan, Italy). The reaction is measured by observing the conjugation of 1-chloro-2,4-dinitrobenzene with GSH, which induces an increase in absorbance at 340 nm. One unit of enzyme will conjugate 10.0 nmol of 1-chloro-2,4-dinitrobenzene with reduced glutathione per min at 25 °C [84–86]. The Ki value was calculated with the Cheng–Prusoff equation [87]:


where the IC50 is the concentration of the competitive inhibitor producing a 50% inhibition, S is the substrate concentration under which the study is performed and Km is the Michaelis constant of the substrate for the enzyme and assuming a 0.19 mminline image and a competitive behavior against GSH. The activity measurements were performed at least three times using cellular lysates from two different cellular treatments.

Fluorescence measurements

All fluorescence measurements were made using a LS50 Perkin–Elmer spectrofluorimeter equipped with a thermostated stirrer cell holder. The temperature was always maintained at 23 °C. The excitation and emission bandwidths were 5 and 3 nm, respectively. The excitation wavelength was set at 286 nm and the spectra were recorded from 300 to 400 nm. The fluorescence measurements were performed in the presence of different concentrations of allyl compound with 3 or 5 μm enzyme in 50 mm Tris/HCl buffer, pH 7.2. E form of the enzyme for the fluorescence spectra was obtained by treatment with dithiothreitol in a molar ratio of 1 : 100 TST/dithiothreitol and after treatment a gel-filtration chromatography of the sample was performed to remove the dithiothreitol using a Nap™ 25 column (GE Amersham).

S-Allyl-mercapto-glutathione analysis

S-Allyl-mercapto-glutathione synthesis was performed using 25 mm GSH and 100 mm 2-PTS (1 : 4 v/v), in Tris/HCl buffer pH 7.4. GSSA formation was evaluated using RP-HPLC (mod. LC-10AVP; Shimadzu, Milan, Italy) with an solvent B gradient (0–5 min, 0%; 5–25 min, 15%; 25–27 min, 15% and 27–30 min 90%), using 0.1% trifluoric acid as solvent A and 80% CH3CN, 0.1% trifluoric acid as solvent B, and a Brouwnlee C18 column (OD300, 250 × 4.6 mm, 7 μm). The elute was monitored at 220 nm by UV detector (Shimadzu, Milan, Italy). Samples were then analyzed by ESI MS (CEINGE Advanced Biotechnologies, Naples, Italy). High-resolution 1H NMR experiments (25 °C) were performed at 400 MHz (Bruker AVANCE spectrometer, Milan, Italy). The compounds were resuspended and 500 μL of D2O (Sigma-Aldrich, Milan, Italy) containing 0.1 mm 3-(trimethylsilyl)-propionic-2,2,3,3-d4 acid sodium salt was added as the internal standard (Merck, Montreal, Canada). A quantitative determination of the GSSA solution was performed by 1H NMR. 1H NMR spectra of samples were obtained using RF pulses for excitation, water signal presaturation, data processing and data analysis as described previously [25,88]. The solution was used to form a calibration curve by RP-HPLC using the same column and conditions described above.

Cyclic voltammetry

Electrochemical measurements were performed at room temperature, using a computer-controlled system, AUTO-LAB model GPSTAT-12 with gpes software (Ecochemie, Utrecht, The Netherlands). Screen-printed electrodes were home-produced with a 245 DEK (Weymouth, UK) screen-printing machine. Graphite-based ink (Elettrodag 421) from Acheson (Milan, Italy) was used to print the working and counter electrode. The substrate was a flexible polyester film (Autostat HT5) obtained from Autotype Italia (Milan, Italy).

The electrodes were produced in foils of 48 strips. Each sensor consists of three screen-printed elements: two carbon electrodes acting as working and counter, respectively, and a silver electrode acting as pseudoreference. The diameter of the working electrode was 0.3 cm. After the printing step, the foils were stored dry, at room temperature, in the dark. The screen-printed electrodes were modified with Prussian Blue [potassium iron (III)hexacyanoferrate(II)] prior to enzyme immobilization. The working electrode surface-modification protocol was based on the procedure optimized by Ricci et al. [89]. In general, 5 μL of each Prussian Blue precursor solution, consisting of 0.1 m of K3Fe(CN)6 in 10 mm HCl and 20 μL of 0.1 m FeCl6 in 10 mm HCl, were applied to the working electrode. The electrodes were incubated for 10 min in the dark and then rinsed with 10 mm HCl. The electrodes were then left for 1 h in the oven at 100 °C to obtain a dry and stable active layer of Prussian Blue. The Prussian Blue-modified electrodes were stored dry at room temperature in the dark until use. All measurements were performed at least three times showing good reproducibility.

H2S assay

To estimate H2S production, a methylene blue formation assay was performed following the method of Schmidt [90] with minor modifications. Briefly, the reaction (200 μL) was incubated at 37 °C for 10 min, and terminated by the addition of 20 μL of solution I (20 mmN′,N’-dimethyl-p-phenylenediamine dihydrochloride in 7.2 m HCl) and 20 μL of solution II (30 mm FeCl3 in 1.2 m HCl). After incubation for 30 min at room temperature, methylene blue formation was examined spectrophotometrically at 670 nm.

Cell proliferation assay

HuT 78 human T lymphoblastoid cells were purchased from the Istituto Superiore di Sanità (Italy). HuT 78 (0.2 × 106) cells were preincubated for 24 h in RPMI 1640 (GIBCO, Milan, Italy) in the presence of 1% glutamine, 10% heat-inactivated fetal bovive serum and antibiotics (1% penicillin and streptomycin sulfate) at 37 °C in air supplemented with 5% CO2. HuT 78 cells were treated with different concentrations of GSSA and then monitored for 8, 24 and 48 h. The cells were then collected and counted after Trypan Blue staining (0.4% Tripan Blue solution; Sigma-Aldrich, Milan, Italy) by optical microscopy using a Thoma chamber. The rates of growth inhibition were calculated with respect to the control culture, which was taken as 100% growth.

Cell-cycle analysis

The cell-cycle distribution of HuT 78 cells was measured by flow cytometry. Harvested cells (∼ 0.5 × 106 cells) were stained with 50 μg·mL−1 propidium iodide (Sigma-Aldrich) in NaCl/Pi buffer with 0.1% Triton X-100 and 1 mg·mL−1 sodium citrate. They were immediately analyzed using a flow cytometer FACSCalibur (Becton-Dickinson, San Josè, CA, USA) and the percentage of cells in each phase of cell cycle was evaluated according to Nicoletti et al. [91].

Protein extraction and western blot analysis

Proteins were extracted from HuT 78 cells in 200 μL of 50 mm Tris/HCl, pH 7.09, containing a protease inhibitor cocktail (Sigma-Aldrich, Milan, Italy) and pervanadate as the phosphatase inhibitor, and sonicated using four steps of 5 s with 1 min standby in ice. Samples were centrifuged for 10 min at 12 000 g at 4 °C. Protein contents was determined by bicinconinic acid protein assay (Sigma-Aldrich), cell extracts (30 μg of protein) were electrophoresed on 15% polyacrylamide gel, electroblotted on a poly(vinylidene difluoride) membrane (Applied Biosystems, Monza, Italy), expression levels of GST were analyzed using anti-(human GSTP1-1) monoclonal rabbit Ig (Calbiochem, Milan, Italy) and phosphorylated p38 using anti-(P-p38) monoclonal IgG (Sigma-Aldrich). An immunoblot with anti-actin monoclonal IgG (Santa-Cruz D.B.A. Italia, Milan, Italy) was also probed for controlling the protein loading. The protein complex formed upon incubation with specific secondary antibodies (dilution 1 : 10 000) (Sigma-Aldrich) was identified using a Fluorchem Imaging system (Alpha Innotech Corp., Analitica De Mori, Milan, Italy) after incubation with ChemiGlow chemiluminescence substrate. Densitometry analysis of western blots was performed using quantity one software (Bio-Rad, Milan, Italy).

Statistical analysis

All experiments were carried out at least three time (= 3), unless otherwise indicated. Data were expressed as mean ± SD. Comparisons between control and treated cells were made using Student’s t-test. Statistical significance was defined as P < 0.05.


We thank Prof. P. Tagliatesta for his help in the organic synthesis of the 2-PST compound; Dr G. Viticchiè for her help in some experiments. We are grateful to Prof. S. Pagani for kindly giving us the plasmid coding for the RhdA protein.