MST3 (mammalian sterile 20-like kinase 3) is a sterile 20 kinase reported to have a role in Fas-ligation- and staurosporine-induced cell death by unknown mechanism(s). We found that MST3-deficient cells are resistant to H2O2, which was reversed by reconstituting recombinant MST3. H2O2-induced JNK (c-Jun N-terminal kinase) activation was greatly enhanced in shMST3 cells (a cell line treated with short hairpin RNA against MST3). Suppression of JNK activity by the inhibitor SP600125 or by dominant-negative JNK2 re-sensitized cells to H2O2. Furthermore, c-Jun Ser-63 phosphorylation was augmented in shMST3 cells, whereas JunAA (dominant-negative c-Jun) reduced H2O2 resistance, implicating an AP-1 (activator protein 1) pathway in H2O2-induced survival signalling. Total cytoprotective HO-1 (haem oxygenase 1) expression, which was attenuated by JunAA, was induced up to 5-fold higher in shMST3 cells compared with controls. Zinc protoporphyrin IX, a potent inhibitor of HO reversed the H2O2-resistance of shMST3 cells. Our results reveal that H2O2-induced MST3-mediated cell death involves suppressing both a JNK survival pathway and up-regulation of HO-1.
Ste20 (sterile 20) kinase was originally described in the Saccharomyces cerevisiae pheromone signalling pathway that transmits extracellular signals through G-protein-coupled receptors to elicit the mating process . Homologous proteins in the mammalian lineage comprise approx. 30 members regulating diverse cellular events, including apoptosis, cellular proliferation and motility through cytoskeleton remodelling. MST3 (mammalian Ste20-like kinase 3), as well as MST4 and SOK1 (Ste20/oxidant-stress-response kinase), form the GCK (germinal centre kinase)-III subfamily due to common structural features [2–4].
A number of the mammalian Ste20 family proteins have been shown to be involved in apoptosis. For example, SOK1 was demonstrated, in an ischaemia model, to respond to ROS (reactive oxygen species), as well as chemical anoxia, mediating cell death by translocating from the Golgi to the nucleus . Silencing of SOK1 using shRNA (short hairpin RNA) protected cells from death following anoxic stress, although downstream effectors mediating this resistance are not currently known. Another member of the GCK family, MST1, was reported to mediate apoptosis in rat primary neurons through direct phosphorylation of FOXO3a (forkhead box O 3a) and FOXO3a-dependent up-regulation of the apoptotic Bim (Bcl-2-interacting mediator of cell death) protein . MST2, which shares the same GCK-II subfamily as MST1, is involved in the Hippo/Salvador/LATS (large tumour suppressor) pathway that culminates in transcription of the apoptotic protein, PUMA (p53 up-regulated modulator of apoptosis) [7–9]. Other Ste20-like proteins, for example, HPK (haemotopoietic progenitor kinase) and SLK (Ste20-like kinase) are cleaved by caspase-3 yielding activated kinases that potentiate apoptosis through JNK (c-Jun N-terminal kinase) activation [10–12]. Finally, PAK2 (p21-activated kinase 2) is also cleaved by caspases upon TNF-α (tumour necrosis factor α) or Fas stimulation and promotes apoptosis in Jurkat T-lymphocytes . MST3 itself has also been implicated in an apoptotic role on the basis of its ectopic expression . However, functions independent of apoptosis have also been described, namely in controlling cell migration .
H2O2 and other ROS are continuously produced as metabolic by-products of aerobic metabolism . Most ROS is efficiently detoxified; however, accumulation of ROS in certain pathological conditions can lead to DNA damage and the oxidative modification of proteins and lipids that can render them non-functional. The effect of H2O2 in cells is highly dose dependent, differing over a spectrum of concentrations, from cell proliferation to cell-cycle arrest to apoptosis and necrosis [16–19]. Interestingly, the MAPK (mitogen-activated protein kinase) signalling cascade, exemplified by the three MAPK families ERK (extracellular-signal-regulated kinase), JNK and p38, are all activated by H2O2 [20–23]. In particular, JNK proteins and their isoforms are important regulators of multiple cellular processes, including apoptosis, cell proliferation, survival, oncogenic transformation, migration and cell differentiation . Although there is much evidence for the role of JNKs in mediating apoptosis, their involvement in pro-survival signalling has also been reported. jnk1−/− and jnk2−/− mouse embryos, for example, display severe apoptosis in the forebrain [25,26]. Several B-lymphoma cells and primary murine cells tested exhibit high JNK expression that upon depletion resulted in elevated apoptosis and reduced proliferation . Furthermore, in MEFs (mouse embryonic fibroblasts) stimulated by TNF, JNK exerts a survival function through its target JunD . Interestingly, JNK phosphorylation of FOXO4 in jnk1−/− and jnk2−/− MEFs and in DLD1 cells reconstituted with jnk3 stimulated transcription of MnSOD (manganese superoxide dismutase) that protected against oxidative stress . Altogether, these observations highlight either apoptotic or pro-survival signalling by JNK depending on the environmental and internal stresses.
In summary, the current evidence reveals that a number of Ste20-like kinases in mammals have clearly evolved a role in cell death. Despite their similarities, these kinases exert their functions through different pathways to promote cell death. Although MST3 has been implicated in apoptosis, its mechanism of action is not known. We thus endeavoured to elucidate the role and mechanisms of MST3 in cell death. We subjected cells to a variety of stress-inducing agents and found that MST3 is a key protein in mediating cellular demise in response to oxidative stress, prompting us to investigate MAPK pathways and proteins that are known to be regulated by lethal concentrations of H2O2. Using this strategy, we were able to uncover a novel mechanism by which MST3 evokes cell death.
Materials and cell culture
Primary antibodies against c-Jun, phospho-c-Jun, JNK, phospho-JNK, p38 and phospho-p38 were from Cell Signaling Technology. Other primary antibodies used were targeted against β-actin (Sigma), HO-1 (haem oxygenase 1) (Assay Designs), MST3 (BD Transduction Laboratory) and phospho-MST3 (Epitomics). Anti-rabbit and anti-mouse secondary antibodies were purchased from Cell Signaling Technology and GE Healthcare respectively. H2O2 was purchased from Merck and the JNK inhibitor SP600125 was from Biomol. All other chemicals and reagents were obtained from Sigma, unless otherwise stated. The human colon carcinoma HCT116 cell line (A.T.C.C.) was maintained in McCoy's 5A medium supplemented with 10% (v/v) heat-inactivated fetal bovine serum (Hyclone), 100 units/ml penicillin and 100 μg/ml streptomycin in a humidified 5% CO2 incubator at 37°C.
RNA-interference-mediated knockdown of MST3
Stable MST3 knockdown clones were generated by transfecting HCT116 cells with plasmids (MISSION TRCN0000000641–TRCN0000000645; Sigma) encoding shRNA sequences aligning to the human MST3 mRNA transcript, using Lipofectamine™ 2000 (Invitrogen) according to the manufacturer's protocol. Following transfection, cells were selected using supplemented McCoy's 5A medium containing puromycin (1.5 μg/ml). Efficiency of knockdown was validated by immunoblotting and real-time PCR methods (see below). A similar protocol was applied to generate stable cell lines containing a non-targeting shRNA sequence (MISSION SHC002; Sigma).
Cells (2.5×106), seeded in 100-mm-diameter dishes, were harvested after 24 h by trypsinization, followed by centrifugation at 400 g for 5 min at 4°C. RNA was extracted from the resulting cell pellet using the RNeasy mini kit (Qiagen). Following first-strand synthesis using the Taqman Reverse Transcription kit (Applied Biosystems), targets were amplified from the resulting cDNA template and detected by real-time PCR (Rotorgene 6000; Corbett Life Science) using the Taqman Universal Master Mix (Applied Biosystems) and primers specific to MST3, β-actin or GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (Applied Biosystems). Relative MST3 expression was calculated by the comparative quantification method with β-actin or GAPDH used as the normalizing factor.
For immunoblot analysis using anti-phosphoprotein antibodies, cells (5×105) were seeded in 6-well plates 24 h before treatment. Following addition of medium containing H2O2 (300 μM), cells were harvested by washing with ice-cold PBS at the appropriate time points (0–24 h) and lysed using Laemmli sample buffer (62.5 mM Tris/HCl, pH 6.8, 2% SDS, 10% glycerol, 50 mM dithiothreitol and 0.01% Bromophenol Blue) (Cell Signalling Technology), according to the manufacturer's instructions. Samples were sonicated for 15 s followed by boiling for 10 min and loading on to SDS polyacrylamide gels. After electrophoresis, proteins were transferred on to PVDF membranes (Millipore) by electrotransfer (Mini Transblot apparatus; Bio-Rad). PVDF membranes were blocked with blocking buffer [TBST (Tris-buffered saline containing 0.1% Tween 20) and 5% non-fat milk] for 1 h at room temperature with gentle shaking. Membranes were then incubated with primary antibodies diluted in TBST containing 5% BSA overnight at 4°C, with gentle mixing. Following three washes with TBST, membranes were incubated with secondary antibodies conjugated to horseradish peroxidase for 1 h at room temperature with shaking. Excess secondary antibodies were removed by further washing and the blot was developed using enhanced chemiluminescence (GE Healthcare). For immunoblotting of the HO-1 protein, cells (2.5×106) seeded in 100-mm-diameter dishes prior to treatment with H2O2 were harvested 14 h later by trypsinization and lysed in lysis buffer (1% Nonidet P40, 0.5% sodium deoxycholate and 0.1% SDS in PBS) containing protease inhibitors (Roche). Samples (100 μg) were loaded on to SDS polyacrylamide gels.
Cytotoxicity and viability assays
Cells (15×103) were seeded in 96-well microplates 24 h before treatment with H2O2 (0–10 mM). After 48 h of incubation, the medium was removed by careful aspiration, and the remaining adherent cells were stained with Crystal Violet solution (1% Crystal Violet in methanol) for 10 min. Following three washes with distilled water, the plates were air-dried. The stained cells were scored as a function of surviving cells by resolubilization of crystals in 1% SDS and detection with a microplate reader (Tecan) at 590 nm. Assigned values were normalized to those of untreated cells and expressed as the percentage of survival. Data from three independent experiments were fitted to multiple dose–response curves by non-linear regression using GraphPad Prism. Half-maximal effective concentration was derived from the resultant curve fit and expressed as the EC50 (means±S.E.M.).
Lactate dehydrogenase activity assay
Cells (15×103) seeded in 96-well microplates were treated 24 h later with H2O2 (0–600 μM). Following a 48 h incubation, the medium was collected and the remaining cells were lysed in McCoy's 5A medium containing 0.1% Triton X-100. Lactate dehydrogenase activity was measured and cytotoxicity was expressed as the percentage of activity in the supernatant/total activity (lysate+supernatant).
Following seeding (24 h), 4.5×105 cells were treated with H2O2 (0–400 μM) and incubated at 37°C for 32 h. The medium was collected and cells harvested by trypsinization were pooled together. Cell viability was measured using the ViCell XR counter (Beckman Coulter).
Transient overexpression of recombinant proteins
Following seeding (24 h; 1.5×106 cells) in T-25 flasks, cells were transfected with a plasmid encoding recombinant FLAG–MST3, pMST3 (Genecopoeia) or the empty vector pRECEIVER-M12 (Genecopoeia), using Fugene HD (Roche) in accordance with the manufacturer's instructions. After incubation with the transfection complex for 24 h, cells were collected by trypsinization and seeded for subsequent experiments. For other experiments employing transient expression of recombinant proteins, the same protocol was applied using the respective plasmids encoding dnJNK1 (dominant-negative JNK1) [a gift from Dr Roger J. Davis (Howard Hughes Medical Institute, Chevy Chase, MD, U.S.A.), dnJNK2 [a gift from Dr Shengcai Lin (Laboratories of Regulatory Biology, Institute of Molecular and Cell Biology, Singapore)] or JunAA (dominant-negative c-Jun) [a gift from Dr Dan Mercola (Department of Pathology and Laboratory Medicine, University of California, Irvine, CA, U.S.A.)] [29,30].
At least three independent experiments were performed. Results are expressed as the means±S.E.M. Differences between two or three groups were determined using Student's t test or ANOVA respectively. Bonferroni correction was applied for post-hoc multiple comparisons. The Wilcoxon signed-rank test was applied to determine differences between two groups in non-parametric testing. P<0.05 was statistically significant. All statistics were calculated using GraphPad Prism.
Oxidative stress induces phosphorylation of MST3
In the first approach to elucidate the role of MST3 in cell death, we subjected HCT116 colon carcinoma cells to selected challenges and monitored the subsequent activation of MST3 kinase using anti-phospho-MST3 antibodies. This antibody targets phosphorylated Thr-178 in the activation loop of the MST3 kinase subdomain VIII, which is required for its activation . Our screen revealed that H2O2 at a lethal concentration (300 μM) induces rapid phosphorylation of MST3 as early as 15 min following treatment, suggesting that MST3 is activated (Figure 1). Interestingly, MST3 phosphorylation consistently occurred in a biphasic manner, in that a second phosphorylation event occurred 6 h post-treatment (Figure 1), and was sustained for at least 24 h.
Oxidative stress induces phosphorylation of MST3
Since anti-phospho-MST3 antibodies recognize residues surrounding the phosphorylated Thr-178 that are shared among all GCK-III family members, namely MST4 and SOK1, it follows that MST4 and SOK1 are also likely to be detected in the immunoblots (Figure 1) . The phosphorylated form of MST3 can be distinguished from phosphorylated MST4 and SOK1 by comparison with MST3 knockdown cells (see below) subjected to the same treatment. The absence of the upper band in MST3 knockdown lysates allows the unequivocal identification of phosphorylated MST3 in the parental HCT116 cell samples.
MST3 depletion increases resistance to oxidative stress
Using three independent shRNA sequences that target different regions along the MST3 mRNA, we successfully produced multiple clonal lines derived from the HCT116 colon carcinoma background with substantially reduced MST3 expression as validated by protein and mRNA transcript quantification (Figure 2). Clones D and E expressed virtually no detectable MST3 protein and were thus selected for all further experiments (Figure 2B).
Stable MST3 knockdown clones (A–E) generated in the HCT116 cell line using RNA interference
Previous work by Huang et al.  revealed that ectopic expression of native, but not the kinase-dead mutant of MST3 caused apoptosis. Our finding that MST3 protein becomes phosphorylated upon exposure to H2O2 led us to hypothesize that MST3 might be a key player in oxidative-stress-induced cell death. Indeed, using several complementary methods to evaluate cytotoxicity, MST3 knockdown clones showed elevated resistance to H2O2 challenge compared with the parental HCT116 and negative control (shNON-TARGETING) cell lines (Figure 3). Crystal Violet staining of surviving cells after H2O2 exposure revealed approx. 2–3.6-fold higher EC50 values in MST3 knockdown cells than either the parental or negative control cells (Figure 3A). Consistently, cytotoxicity assays, measuring lactate dehydrogenase activity as an indicator of cell death (Figure 3B), and the Trypan Blue dye exclusion method for viable cells (Figure 3C) both corroborated findings that cells lacking MST3 are relatively refractory to oxidative-stress-induced death.
MST3 knockdown cells are more resistant to H2O2 challenge compared with wild-type (WT) or shNON-TARGETING cells as determined by alternative assays for cytotoxicity
Reconstitution of MST3 into knockdown cells restores sensitivity to H2O2
To ensure that the observed alteration in response to H2O2 was specific to MST3, we performed a reconstitution experiment. The knockdown phenotype of clone A is a result of silencing by shRNA targeting the 3′ untranslated region of the MST3 mRNA (Figure 2A). Therefore, a plasmid encoding only the full-length open reading frame of MST3 (recombinant MST3) can be used to replenish the protein in this clone. The original knockdown clone A exhibits >80% reduction in MST3 protein levels, as measured by densitometry (Figures 2A and 2B), and shows resistance towards H2O2 comparable with clones D and E (Figure 3A). Following transient expression of recombinant MST3, the reconstituted knockdown clone A was challenged with oxidative stress under the same experimental conditions. The data show that ectopic MST3 expression restored sensitivity of the knockdown cells to H2O2 (Figure 3D). This result indicates that MST3 is important for the cellular oxidative stress response in cell death.
H2O2-induced activation of JNK is enhanced in MST3 knockdown cells
Because H2O2 has been reported to activate MAPK signalling cascades, including the ERK, JNK and p38 pathways, we investigated the response of these canonical pathways to ROS signalling in the context of MST3. As expected, all three MAPKs were phosphorylated upon H2O2 treatment (Figure 4; results not shown for ERK) [20–23]. Interestingly, phosphorylation of JNK was augmented in both MST3 knockdown clones, with a concomitant increase in phosphorylation of its target, c-Jun (Figure 4; results not shown for clone D). p38 phosphorylation in MST3 knockdown cells was also enhanced compared with parental or negative control cells. Both JNK and p38 activation occurred rapidly (Figure 4A) and was sustained over several hours (Figure 4B). These findings raise the possibility that JNK(s) and p38 may be involved in boosting cellular resistance to oxidative stress.
MST3 modulates JNK, c-Jun and p38 phosphorylation during oxidative stress
In order to probe further, we asked if inhibiting JNK activity would reverse the protection from oxidative stress seen in MST3 knockdown clones. Cells pre-incubated with the JNK inhibitor SP600125 regained sensitivity to H2O2 to an extent similar to that of the parental and negative control cells (Figures 5A and 5B), supporting our earlier proposition that in the absence of MST3, JNK signalling contributes towards survival during oxidative stress. Next, we employed MST3 knockdown cells transiently expressing dnJNK1 or dnJNK2. Our results reveal that in cells expressing dnJNK2 (Figure 5C), sensitivity to H2O2 was partially restored, but not restored in dnJNK1-expressing cells (results not shown), implying that JNK2 might be the JNK isoform mediating survival signals induced by oxidative stress.
Inhibition of JNK activity restores H2O2 sensitivity
Because c-Jun phosphorylation was robustly augmented in MST3 knockdown cells exposed to H2O2 (Figure 4), we next tested if interfering with c-Jun activity itself might alter the response of MST3 knockdown cells. Our experiments show that transient expression of a plasmid encoding JunAA (pJunAA), indeed, attenuated the resistance of MST3 knockdown cells to H2O2, suggesting that c-Jun mediates protection from oxidative stress (Figure 5D).
HO-1 induction is highly elevated in MST3 knockdown cells
In order to determine why shMST3 cells (a cell line treated with shRNA against MST3) display enhanced survival in the face of oxidative stress, we sought to identify downstream effectors of H2O2 signalling. We therefore probed for changes in the abundance of candidate proteins that have previously been shown to produce antioxidant effects. Of the proteins tested, only HO-1 expression levels were significantly elevated in MST3 knockdown cells compared with control cells (Figure 6). On the other hand, levels of another obvious candidate, MnSOD, remained unchanged between shMST3 and control cells (results not shown), suggesting that HO-1 might mediate resistance to H2O2 in the MST3 knockdown cells.
HO-1 is abundantly expressed in MST3 knockdown cells
To further examine the cytoprotective role of HO-1 in our model, we asked if the increased survival rate seen in MST3 knockdown cells was dependent on HO-1 activity. We pre-treated cells with ZnPP (zinc protoporphyrin IX), a potent inhibitor of HO , followed by H2O2 challenge. Our results show that ZnPP re-sensitized MST3 knockdown cells to oxidative stress, supporting the suggestion that HO-1 imparts a survival advantage to these cells (Figure 7A). Negative control cells exposed to ZnPP displayed even higher sensitivity to H2O2 than ZnPP-treated MST3 knockdown cells, presumably due to their overall lower basal and induced HO-1 levels.
Inhibition of HO-1 activity and expression
JunAA expression inhibits HO-1 induction
Alam  reported the presence of AP-1 (activator protein 1)-binding sites in the HO-1 promoter. Because we also found that JunAA and ZnPP could independently sensitize MST3 knockdown cells to H2O2, we next examined HO-1 expression levels in the context of altered c-Jun activity. Our results show that transient expression of JunAA inhibited H2O2-induced HO-1 expression in both MST3 knockdown and negative control cells, suggesting that c-Jun is involved in the induction of HO-1 expression during oxidative stress (Figures 7B and 7C).
MST3 has previously been implicated to induce apoptosis when overexpressed ectopically; however, the mechanism has, so far, remained unknown. Our study has uncovered some probable novel mechanisms by which MST3 promotes cell death. We report for the first time that MST3 is phosphorylated upon induction of oxidative stress. Phosphorylation occurs at Thr-178, which has previously been shown to be critical for MST3 kinase activity . The activation of MST3, as indicated by the phosphorylation of Thr-178, together with the observed alteration in behaviour of the MST3 knockdown cells thus suggest that MST3 activation plays a role in the response of cells to oxidative stress.
Our study not only reveals that MST3 responds to H2O2 treatment by its activation, but that cells devoid of this protein display remarkable resistance to oxidative stress. Recently, Wu et al. , using trophoblast cells derived from the placenta, showed that MST3 expression is up-regulated by H2O2 at lethal concentrations. That MST3 was found to be involved in cellular response to H2O2 in both the present study and Wu et al.  independently provides mutually supportive evidence that MST3 does, indeed, play a role in H2O2-induced apoptosis. Unlike Wu et al. , however, we did not detect any changes in MST3 expression in our model system when cells were subjected to any of the multitude of stress stimuli tested: DNA damage agents, DNA synthesis inhibitors and oxidative stress (results not shown). It is likely that this disparity is a reflection of the different sources of tissue from which the cell lines in each study were derived.
Much previous work mutually corroborates the finding that oxidative stress can trigger activation of MAPKs, notably ERK1/2, JNKs and p38 [20–23]. We systematically investigated these pathways and their dependence on MST3. Our studies imply that MST3 depresses JNK activation during oxidative stress. On the other hand, cells lacking MST3 display highly enhanced JNK activation in parallel with a more resistant phenotype. The JNK pathways are known to transmit signals in response to stress, differentiation, development, inflammation and apoptosis [24,34]. However, more recent studies have begun to illuminate an opposing role for JNKs in pro-survival signalling in the event of certain stress stimuli [25,26,28]. Our experiments further strengthen the perspective of this new pro-survival role for JNKs, particularly in the context of oxidative stress. We were able to show that, in the presence of lethal concentrations of H2O2, MST3 indirectly retards both JNK activation and activity in order to drive the cell towards death. It is not yet known how MST3 modulates JNK activation; one possibility is that MST3 increases the activity of a phosphatase that maintains JNK in a dephosphorylated state. Since JNK is primarily phosphorylated by MKK4 (MAPK kinase 4) and MKK7 , and given that the phosphorylation of JNK is highly augmented in the MST3 knockdown cells, it is likely that the activity of JNK kinases or upstream proteins is altered, rather than JNK itself. Another possibility is that MST3 could alter the phosphorylation state of a JNK scaffolding protein, which in turn affects the status of JNK activity during oxidative stress. We were able to further confirm that JNK plays an important role in MST3-dependent cell death, as the specific JNK inhibitor, SP600125, reversed the resistance of MST3 knockdown cells to oxidative challenge. In addition, expression of dnJNK2 partially re-sensitized the knockdown cells to H2O2, further supporting this conclusion.
Consistent with elevated JNK activation in MST3 knockdown cells, our data also reveals that c-Jun was robustly phosphorylated at the Ser-63 and Ser-73 positions, which are known target sites of JNK (results not shown for c-Jun Ser-73 phosphorylation) . In addition to c-Jun, we also tested phosphorylation levels of other AP-1 components targeted by JNK, i.e. JunD and ATF2 (activating transcription factor 2). Enhanced phosphorylation was observed in both of these transcription factors. However, in contrast with c-Jun, the differences compared with control cells were not great (results not shown), implying that mainly c-Jun mediates survival signals in the absence of MST3. The effect of JunAA expression on H2O2-induced cell death in our experiments further supports this suggestion.
We examined a number of probable effector proteins that could impart antioxidant effects. We found that the expression of HO-1 in MST3 knockdown cells was highly up-regulated, but not that of MnSOD. HO-1 is highly inducible by an immense variety of stimuli, including diverse chemicals, UV, heat shock, ischaemia, hypoxia and hyperoxia [37,38]. HO-1 catalyses the conversion of haem into carbon monoxide, iron and the compound biliverdin, which is a precursor for a potent antioxidant, bilirubin . Bilirubin forms a part of the arsenal of non-enzymatic scavengers for H2O2-induced ROS, together with vitamins A, C and E, and urate, and therefore could be an effector molecule in detoxification of the MST3 knockdown cells . Indeed, we show that ZnPP, an inhibitor of HO-1 activity, reduced cellular resistance to oxidative stress. Thus the higher basal and total inducible HO-1 levels in MST3 knockdown cells imply that these cells are poised for protection against oxidative stress.
Our experiments demonstrate that when JunAA was overexpressed, levels of H2O2-induced HO-1 was significantly diminished, thereby evoking c-Jun activity in HO-1 induction in our model. Supporting evidence for this JNK–c-Jun–HO-1 axis can be derived from work by Aggeli et al. , in which it was demonstrated in the cardiomyoblast model that c-Jun and ATF2 phosphorylation mediate up-regulation of HO-1 under hyperoxic conditions . These authors found that phosphorylation of c-Jun and ATF2 is dependent on JNK and p38 activity respectively, which could be suppressed by SP600125 and the p38 inhibitor SB203580. Moreover, the observation that the promoter of HO-1 contains AP-1-binding sites [32,41], and that the DNA-binding activity of AP-1 can be inhibited by SP600125, further support the suggestion that HO-1 aids survival of cells during oxidative stress and the lesser known paradigm of JNK-mediated pro-survival signalling in these conditions.
We show that activation of MST3 by oxidative stress consistently occurs in a temporally biphasic manner. Although the function of this biphasic activation is not yet known, one possible explanation is that MST3 might promote distinct cellular responses depending on the duration of its signal, whether in a transient or sustained context. Indeed, the rapid phosphorylation response in MST3 when stimulated is consistent with the role of many Ste20-like kinases as MAP4K (MAPK kinase kinase kinase) proteins, which may function as sentinels of specific environmental stimuli . Further investigation will be needed in order to determine the physiological importance of this biphasic phenomenon.
Comparative survival rates between MST3 knockdown cells and control cells subjected to diverse challenges also revealed significant survival differences for the drug mimosine, a cell-cycle inhibitor (results not shown). Mimosine is used routinely as a cell-cycle synchronizing agent because of its ability to reversibly arrest cells in G1 phase [43,44]. Mimosine can also induce apoptosis through the mitochondrial permeability transition and production of H2O2 . MST3 has been reported to directly phosphorylate NDR (nuclear Dbf2-related) kinases, whose orthologues in S. cerevisiae are essential for cell-cycle progression and for the mitotic exit network [46–48]. Thus the effect of survival during mimosine challenge in MST3-depleted cells may occur through events within the cell-cycle machinery or through secondary ROS generation. Uncovering the mechanisms of mimosine resistance in the MST3-depleted model will undoubtedly require further investigation.
In conclusion, our studies reveal the outline of a novel mechanism of cell-death signalling by MST3. These findings reveal further insights into the cellular responses to oxidative stress and the fine molecular interplay that balances between pro-survival and pro-death signalling. Because oxidative stress can cause modification of proteins, DNA and lipids that is likely to compromise normal cellular functions, this MST3-dependent mechanism might serve to prevent genetically damaged cells from continued survival.
activator protein 1
activating transcription factor
forkhead box O
germinal centre kinase
haem oxygenase 1
c-Jun N-terminal kinase
mitogen-activated protein kinase
mouse embryonic fibroblast
manganese superoxide dismutase
mammalian sterile 20-like kinase
reactive oxygen species
short hairpin RNA
Tris-buffered saline containing 0.1% Tween 20
tumour necrosis factor
zinc protoporphyrin IX
We thank Dr Jan Gruber and Dr Li Lei for helpful discussions.
This work was supported by the Institute of Molecular and Cell Biology, Agency for Science, Technology and Research (A*STAR), Singapore.