Polysulfide Na2S4 regulates the activation of PTEN/Akt/CREB signaling and cytotoxicity mediated by 1,4-naphthoquinone through formation of sulfur adducts

Electrophiles can activate redox signal transduction pathways, through actions of effector molecules (e.g., kinases and transcription factors) and sensor proteins with low pKa thiols that are covalently modified. In this study, we investigated whether 1,4-naphthoquinone (1,4-NQ) could affect the phosphatase and tensin homolog (PTEN)–Akt signaling pathway and persulfides/polysulfides could modulate this adaptive response. Simultaneous exposure of primary mouse hepatocytes to Na2S4 and 1,4-NQ markedly decreased 1,4-NQ-mediated cell death and S-arylation of cellular proteins. Modification of cellular PTEN during exposure to 1,4-NQ was also blocked in the presence of Na2S4. 1,4-NQ, at up to 10 µM, increased phosphorylation of Akt and cAMP response element binding protein (CREB). However, at higher concentrations, 1,4-NQ inhibited phosphorylation of both proteins. These bell-shaped dose curves for Akt and CREB activation were right-shifted in cells treated with both 1,4-NQ and Na2S4. Incubation of 1,4-NQ with Na2S4 resulted in formation of 1,4-NQ–S–1,4-NQ-OH. Unlike 1,4-NQ, authentic 1,4-NQ-S-1,4-NQ-OH adduct had no cytotoxicity, covalent binding capability nor ability to activate PTEN-Akt signaling in cells. Our results suggested that polysulfides, such as Na2S4, can increase the threshold of 1,4-NQ for activating PTEN–Akt signaling and cytotoxicity by capturing this electrophile to form its sulfur adducts.

. Cytotoxicity and chemical modification of cellular proteins during treatment of cells with 1,4-NQ, with or without Na 2 S 4 . (A,B) Primary mouse hepatocytes were exposed to 1,4-NQ with or without 10 (A) or 100 (B) µM Na 2 S 4 for 24 h. Cell viability was assessed with the MTT assay. Each value is the mean ± standard error for three independent experiments. *P < 0.05 and **P < 0.01, compared with the control. C, D: Primary mouse hepatocytes were exposed to 1,4-NQ, with or without 10 (C) or 100 (D) µM Na 2 S 4 for 1 h. Covalent modification of cellular proteins by 1,4-NQ was detected by western blotting using an anti-1,4-NQ antibody (upper), proteins were detected following SDS-PAGE by Coomassie Brilliant Blue staining (middle) and intensities of modified protein bands were calculated with Multi Gauge software (lower). Representative data are shown from two independent experiments. corresponding to Fig. 1C and D, Na 2 S 4 inhibited the covalent modification of PTEN by 1,4-NQ ( Fig. 2A). Recombinant human PTEN (0.73 µM) was also modified by 1,4-NQ in a concentration dependent manner (Fig. 2B). It is well recognized that 1,4-NQ undergoes S-arylation to proteins through an 1,4-addition reaction by nucleophiles, resulting in formation of 1,4-NQH 2 (MW = 158.02)-protein adduct 13 . We recently observed that this adduct readily underwent autooxidation, to yield a 1,4-NQ (MW = 156.02)-protein adduct 19 . Consistent with this, the trypsinized fragments detected by ultra-performance liquid chromatography (UPLC)-mass spectrometry (MS) in recombinant PTEN that had been modified by 1,4-NQ were 1,4-NQ-protein, not 1,4-NQH 2 -protein, adducts. We also found that the PTEN sites modified by 1,4-NQ were Cys71 and Cys83 ( Fig. 2C and Table 1). This suggested that 1,4-NQ activated Akt signaling through S-arylation to PTEN and that Na 2 S 4 suppressed this 1,4-NQ-mediated activation of the PTEN-Akt signaling pathway.
To address whether 1,4-NQ would activate Akt signaling and Na 2 S 4 would modulate this effect, we exposed primary mouse hepatocytes to 1,4-NQ, with or without Na 2 S 4 , and then measured phosphorylation of Akt and its downstream protein CREB by western blotting. As shown in Fig. 3, 1,4-NQ, at up to 10 µM, increased phosphorylation of Akt and CREB in a concentration dependent manner. However, phosphorylation, hence activation, of both proteins was inhibited by 1,4-NQ at higher concentrations. The bell-shaped 1,4-NQ dose responses for effects on this redox signal transduction pathway agreed with our findings that MeHg, at lower concentrations, activated PTEN-Akt-CREB signaling through S-modification of PTEN and, at higher concentrations, disrupted the cascade through non-specific S-modification of CREB 16 . At 10 µM, Na 2 S 4 significantly decreased 1,4-NQ induced phosphorylation of Akt and CREB (Fig. 3A). Treatment with 1,4-NQ and 100 µM Na 2 S 4 together led to a markedly right-shifted bell-shaped response curves for Akt and CREB phosphorylation (Fig. 3B), relative to those obtained with 1,4-NQ alone.
Determination of 1,4-NQ-sulfur adducts formed by reaction of 1,4-NQ with Na 2 S 4 . We recently found that incubation of 1,4-NQ with Na 2 S 4 consumed 1,4-NQ, suggesting that 1,4-NQ can react with Na 2 S 4 to form sulfur adducts. Such reaction products might then become less reactive with primary mouse hepatocytes because of their decreased electrophilicity. To test this hypothesis, the sulfur adducts formed in a reaction mixture of 1,4-NQ with Na 2 S 4 were separated on a preparative ODS-column, eluted with 20% acetonitrile and monitored spectrophotometrically at 250 nm. Fractions ranging from 14 to 18 min (Fraction I) and from 30 to 33 min (Fraction II) were collected (Fig. 4A) and reaction products were analyzed in each. The molecular masses of the reaction products in Fractions I and II were mainly m/z 345 and 361, respectively, by UPLC-MS analysis. This  Table 1.

Discussion
In our study, the atmospheric electrophile 1,4-NQ activated PTEN-Akt signaling at lower concentrations but disrupted it at higher concentrations. In addition, 1,4-NQ-mediated redox signaling was negatively regulated by a model polysulfide, Na 2 S 4 , through formation of 1,4-NQ sulfur adducts (Fig. 6). Under basal conditions, PTEN can negatively regulate the Akt cascade by dephosphorylating the substrate of phosphoinositide 3-kinase, which phosphorylates Akt 21 . Reactive oxygen species, nitric oxide and endogenous electrophiles, such as Δ12-prostaglandin J 2 and 4-hydroxynonenal, can activate the PTEN-Akt signaling pathway though modification of cysteine residues in PTEN, which has 10 cysteine residues (both in mouse, NP_032986, and in human, NP_000305) [22][23][24][25] . For example, hydrogen peroxide can oxidize PTEN to form a disulfide bond between Cys71 and Cys124, which are located close to one other 22,26 . Numajiri et al. found that S-nitrosylation through Cys83 in PTEN regulated Akt signalling in vivo 27 . Although the pKa value of cysteine is 8-9, the pKa value of the cysteine thiol proximal to basic amino acids, such as histidine, lysine and arginine, was decreased 2 . Of interest, Cys71, Cys83 and Cys124 are located near basic amino acids, including arginine and lysine, indicating that 1,4-NQ could potentially modify these cysteine residues. Consistent with this, we identified Cys71 and Cys83 as modification sites for 1,4-NQ (Fig. 2C), but did not detect modification of Cys124 under these conditions. Shearn et al. reported that 4-hydroxynonenal modified An additional singlet proton signal in the high field at δ 6.07 (s, 1 H) must be attributable to H-3. An aromatic OH group must be located at the C-3′ position, although this OH signal was not detected. The COSY NMR spectrum showed that two triplets at δ 7.74 and δ 7.63 ppm were correlated to one another and to two doublets at δ 7.97 and δ 7.9 ppm, respectively. These signals must be attributable to H-5′, H-6′, H-7′ and H-8′. The other two triplets at δ 7.86 and δ 7.83 ppm were correlated to one another and to two doublets at δ7.93 and δ 8.05 ppm, respectively. These signals must be attributable to H-5, H-6, H-7 and H-8. D: MS spectrum of the sulfur adduct (m/z 361) of 1,4-NQ formed during incubation with Na 2 S 4 . The purified sulfur adduct was analyzed by UPLC-MS. Representative data are shown from three independent experiments.
Scientific RepoRts | 7: 4814 | DOI:10.1038/s41598-017-04590-z Cys71, but not Cys124, in PTEN 28 . This suggested that the organic electrophile had a preference for Cys71, rather than Cys124, in formation of the electrophile adducts detected by MS/MS. Band intensities were normalized to those for total Akt and CREB, respectively (lower). Intensities are expressed as fold induced, relative to results with 0 µM 1,4-NQ exposure. Each value is the mean ± the standard error for three independent experiments. *P < 0.05 and *P < 0.01, compared with the 0 µM 1,4-NQ exposure. Whereas 1,4-NQ exposure caused cytotoxicity and covalent modification of cellular proteins in primary mouse hepatocytes, the additional presence of Na 2 S 4 blocked this protein modification, thereby protecting the cells from cytotoxicity (Fig. 1). In another study, we demonstrated that reactive persulfide and polysulfide species suppressed 1,4-NQ-and Cd-mediated activation, through electrophilic HSP90 modification, of HSP90-HSF1 signaling 18,19 . In this study, we found that S-arylation of PTEN by 1,4-NQ was also suppressed by cotreatment with Na 2 S 4 (Fig. 6). Because PTEN can negatively regulate the Akt cascade, we postulated that 1,4-NQ would activate the PTEN-Akt signaling pathway and that this activation could be modulated by polysulfides such as Na 2 S 4 . Consistent with these predictions, 1,4-NQ activated Akt and its downstream protein CREB in a concentration dependent manner, up to 10 µM, an effect suppressed by addition of Na 2 S 4 (Fig. 3). At higher 1,4-NQ concentrations, however, Akt and CREB activation did not occur, indicating a bell-shaped dose response (Fig. 3). We obtained similar results in our previous study, in which MeHg, at lower concentrations, activated PTEN-Akt-CREB signaling and, at higher concentrations, disrupted the signaling because of nonspecific binding of MeHg to Akt and CREB 16 . Consistent with this, while nitric oxide, at lower concentrations, is an endogenous activator of Akt signaling through S-nitrosylation to PTEN, at higher concentrations nitric oxide can inactivate PTEN-Akt signaling, disrupting Akt function through S-nitrosylation 27 . Disruption of redox signaling through S-modification may protect cells by preventing overactivation of these pathways. Of interest, Na 2 S 4 elevated the threshold of 1,4-NQ-mediated Akt and CREB phosphorylation in primary mouse hepatocytes, at least in part by suppression of 1,4-NQ-dependent S-modification of proteins, including PTEN, by the polysulfide (Figs 1C,D, 2A and 3). These observations suggested that the polysulfide Na 2 S 4 modulated adaptive responses, such as activation of redox signaling caused by electrophilic modifications.
In summary, previous studies showed that 1,2-NQ activated the Keap1-Nrf2 and PTP1B-EGFR signaling pathways through covalent binding to Cys151, Cys273, Cys288, Cys257 and Cys488 in Keap1 34 and Cys121 in PTP1B 9 . Its isomer, 1,4-NQ, activated the HSP90-HSF1 pathway through covalent modification to Cys412 and Cys564 in HSP90 19 . The present study showed, in addition, that 1,4-NQ activated the PTEN-Akt signaling pathway through modification to Cys71 and Cys83 on PTEN. From these observations, it seems likely that environmental electrophiles at lower concentrations can activate redox signaling pathways by electrophilic modification of thiol groups in sensor proteins. This would result in adaptive responses useful for cell survival, cell proliferation, detoxification and excretion of electrophiles and quality control of cellular proteins. Reactive polysulfides can negatively regulate the quinone-mediated activation of redox signaling, such as the HSP90-HSF1 and PTEN-Akt pathways, by capturing environmental electrophiles to form inert sulfur adducts. Environmental electrophiles can react with not only Na 2 S 4 , the agent used in this study, but also endogenous per/polysulfides such as GSSH, GSSSG and CysSSH 19,29,35 . Endogenous H 2 S and persulfides/polysulfides are produced by enzymatic reaction of CSE, CBS, and 3-mercaptopyruvate sulfurtransferase [36][37][38] , which may play an important role on protection against electrophiles mediated toxicity. We showed that: 1) NAPQIH 2 -SSSCys and NAPQIH 2 -SSG adducts were detected in biological samples from mice given acetaminophen; 2) (MeHg) 2 S was produced from the reaction of MeHg with GSSH and/or GSSSG; and 3) 1,4-NQ reacted with Cys persulfide, and/or its polysulfide generated enzymatically by CSE, to yield 1,4-NQ-SCys, 1,4-NQ-SH, and 1,4-NQ-S-1,4-NQ adducts 19 . Taken together, these findings suggested that reactive per/polysulfide species have the potential to modulate the adaptive responses caused by environmental electrophile exposures. Thus, supplementation or other simultaneous intake of per/polysulfide species might decrease the health risks of environmental electrophile exposures.
Isolation and culture of primary mouse hepatocytes. All animal protocols were approved by the University of Tsukuba Animal Care and Use Committee and were performed strict adherence to the committee's guidelines for alleviation of suffering. Primary mouse hepatocytes were isolated from 6-11-wk-old C57BL/6 J female mice as described previously 39 . Briefly, the hepatocytes (8 × 10 4 cells/cm 2 ) were seeded in William's Scientific RepoRts | 7: 4814 | DOI:10.1038/s41598-017-04590-z medium E containing 10% fetal bovine serum, 2 mM glutaMAX-I (Thermo Fisher Scientific, Waltham, MA, USA) and antibiotics (100 units/mL penicillin and 100 µg/mL streptomycin) on culture plates coated with fetal bovine type I collagen (Corning Inc., Corning, NY, USA) and were maintained at 37 °C in a humidified atmosphere containing 95% air and 5% CO 2 . The cells were cultured for 2 d after isolation and then starved overnight by incubation in serum-free medium before exposure to 1,4-NQ.
Lysate preparation. After exposure to 1,4-NQ, with or without Na 2 S 4 , primary mouse hepatocytes were washed twice with ice-cold phosphate-buffered saline. A cell lysate was then prepared by sonicating the cells in radioimmunoprecipitation assay (RIPA) buffer [25 mM Tris-HCl (pH 7.5), 150 mM sodium chloride, 1% NP40 and 0.5% sodium deoxycholic acid] containing 1% protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA). The cells lysed in RIPA buffer were centrifuged for 10 min at 14,000 g. Protein concentrations were determined using the bicinchoninic acid assay (Thermo Fisher Scientific).
Western blot analysis. Samples, adjusted for equal protein contents, were each mixed with a half volume of SDS-PAGE loading buffer [62.5 mM Tris-HCl (pH 6.8), 8% glycerol (v/v), 2% SDS (w/v) and 0.005% bromophenol blue (w/v)] containing either 15 mM 2-mercaptoethanol or 50 mM tris(2-carboxyethyl)phosphine. Each mixture was then heated to 95 °C for 5 min and applied to a SDS-polyacrylamide gel. The proteins were separated by SDS-PAGE and electro-transferred onto polyvinylidene difluoride membranes (Bio-Rad Laboratories, Hercules, CA, USA) at 2 mA/cm 2 for 1 h. The membranes were blocked in 5% skim milk at 25 °C for 1 h, incubated with primary antibodies at 4 °C overnight and then incubated with secondary antibodies coupled to HRP at room temperature for 2 h. Stained protein bands were detected using an enhanced chemiluminescence system (Nacalai Tesque, Kyoto, Japan) using a LAS 3000 imager (Fujifilm, Tokyo, Japan).
Preparation of recombinant PTEN. The entire coding sequence of human wildtype PTEN was amplified from a human cDNA library by the polymerase chain reaction (PCR). The cDNA encoding PTEN was subcloned into the pGEX-6P-1 vector. The recombinant human PTEN was expressed as an N-terminal glutathione S-transferase (GST)-tagged fusion protein in BL21 (DE3) cells transformed with the pGEX-PTEN vector. Protein production was induced by 0.5 mM isopropyl β-D-thiogalactopyranoside (Nacalai Tesque) at 25 °C for 12 h. GST-PTEN was affinity purified on Glutathione 4B Sepharose, eluted with 10 mM reduced glutathione in 50 mM Tris-HCl (pH 7.5), 150 mM sodium chloride and 1 mM dithiothreitol (DTT). Thiol groups oxidized during purification were reduced by incubation with 20 mM DTT for 1 h. Free glutathione and DTT were removed by buffer exchange to 50 mM potassium phosphate buffer (pH 7.0) using an ultrafiltrator (Piece Concentrators 9 K; Thermo Fisher Scientific). Samples were stored in at −80 °C before use.

Detection of cellular PTEN modified by 1,4-NQ.
Primary mouse hepatocytes were exposed to 1,4-NQ (10 µM) for 30 min, with or without Na 2 S 4 , then lysates were prepared using RIPA buffer, as described above under lysate preparation. Anti-rabbit IgG conjugated magnetic beads (100 µL, Dynabeads M-280-sheep anti-rabbit IgG) were washed three times with Tris-buffered saline and Tween 20, then incubated with anti-PTEN antibodies (5 µL, Cell Signaling Technology, #9552) at 4 °C for 3 h. The unbound antibodies were removed and the beads resuspended in 500 µL cell lysate (1 µg/µL). The mixture was then incubated, with rotation, at 4 °C overnight. The beads were then washed four times with RIPA buffer and protein complexes eluted by adding 40 µL RIPA buffer and a half volume of SDS-PAGE loading buffer containing 50 mM tris(2-carboxyethyl)phosphine. The eluted proteins were incubated at 95 °C for 5 min, then analyzed by western blotting.

Recombinant PTEN modification and LC-MS E analysis.
Recombinant GST-PTEN (1 μg) was incubated with 1,4-NQ at 25 °C for 1 h. The reaction mixture was then analyzed by western blotting with an anti-1,4-NQ antibody. For LC-MS analysis, 1.7 μg protein was incubated with 1,4-NQ (10 μM) at 25 °C for 30 min in a total volume of 10 μL 50 mM Tris-HCl (pH 7.5). Samples of native and 1,4-NQ-modified GST-PTEN were incubated with 2 mM tris(2-carboxyethyl)phosphine at 25 °C for 10 min in a total volume of 20 μL 50 mM ammonium bicarbonate solution. Each mixture was then alkylated by adding 2.5 μL 30 mM 2-iodoacetamide in 50 mM ammonium bicarbonate solution and incubating the mixture at 25 °C for 20 min in the dark. The GST-PTEN was digested by adding 2.5 μL MS-grade modified trypsin (100 ng) and incubating the mixture at 37 °C overnight. Nano UPLC-tandem MS (MS E ) analysis was performed using a nanoAcquity UPLC system (Waters, Milford, MA, USA), equipped with a BEH130 nanoAcquity C 18 column (100 mm long, 75 μm i.d., 1.7 μm particle size; Waters), maintained at 35 °C. The analysis was performed in direct injection mode. Mobile phases A (0.1% formic acid) and B (acetonitrile with 0.1% formic acid) were mixed using a gradient system, at a flow rate of 0.3 μL/min. The mobile phase program started at 3% B for 1 min, then linearly increased over 74 min to 40% B, which was maintained for 4 min, then linearly increased over 1 min to 95% B, which was maintained for 5 min, then linearly decreased over 1 min to 3% B. The total run time (including conditioning the column at the initial conditions) was 100 min. The eluted peptides were transferred to the nano-electrospray source of a quadrupole time-of-flight mass spectrometer (a Synapt High Definition Mass Spectrometry system; Waters) through a Teflon capillary union and a precut PicoTip (Waters). The initial Synapt mass spectrometer parameters were capillary voltage of 2.8 kV, sampling cone voltage of 35 V and source temperature of 100 °C. A low (6 eV) or elevated (stepped from 15 to 30 eV) collision energy was used to generate either intact peptide precursor ions (low energy) or peptide product ions (elevated energy). The detector was operated in positive ion mode. The mass spectrometer performed survey scans from m/z 50 to 1990. All analyses were performed using an independent reference, glu-1-fibrinopeptide B (m/z 785.8426), which was infused through the NanoLockSpray ion source and sampled every 10 s and used as an external mass calibrant. Data were collected using MassLynx version 4.1 software (Waters). Biopharmlynx version 1.2 software (Waters) was used to perform baseline subtraction, smoothing, de-isotoping, de novo peptide sequence identification and database searches.