Activation of the p38 signaling pathway by heat shock involves the dissociation of glutathione S-transferase Mu from Ask1.

Despite the importance of the stress-activated protein kinase pathways in cell death and survival, it is unclear how stressful stimuli lead to their activation. In the case of heat shock, the existence of a specific mechanism of activation has been evidenced, but the molecular nature of this pathway is undefined. Here, we found that Ask1 (apoptosis signal-regulating kinase 1), an upstream activator of the stress-activated protein kinase p38 during exposure to oxidative stress and other stressful stimuli, was also activated by heat shock. Ask1 activity was required for p38 activation since overexpression of a kinase dead mutant of Ask1, Ask1(K709M), inhibited heat shock-induced p38 activation. The activation of Ask1 by oxidative stress involves the oxidation of thioredoxin, an endogenous inhibitor of Ask1. A different activation mechanism takes place during heat shock. In contrast to p38 induction by H(2)O(2), induction by heat shock was not antagonized by pretreatment with the antioxidant N-acetyl-l-cysteine or by overexpressing thioredoxin and was not accompanied by the dissociation of thioredoxin from Ask1. Instead, heat shock caused the dissociation of glutathione S-transferase Mu1-1 (GSTM1-1) from Ask1 and overexpression of GSTM1-1-inhibited induction of p38 by heat shock. We concluded that because of an alternative regulation by the two distinct repressors thioredoxin and GSTM1-1, Ask1 constitutes the converging point of the heat shock and oxidative stress-sensing pathways that lead to p38 activation.

Heat shock affects all proteins and structures but nevertheless produces a highly specific stress response aimed at protecting the cells and re-establishing homeostasis. In addition to the well characterized transcriptional activation of the genes coding for heat shock proteins (1-3), within minutes heat shock activates a major signal transduction pathway involving the stress-activated protein kinase p38 and leading to the phosphorylation of heat shock protein 27 (HSP27) (4,5). Phosphorylation of HSP27 activates a protective function, which may result from the known phosphorylation-modulated function of the protein at the level of the actin microfilaments (6 -8) or from other described protective activities, either as a chaperone (9 -11) or as an inhibitor of apoptotic processes (12)(13)(14). Activation of the p38 pathway and phosphorylation of HSP27 occurs within minutes at elevated temperature and constitutes a very tightly regulated response (15). After a mild heat shock, cells becomes refractory to reinduction of p38 activity by a second heat shock but remained fully responsive to reinduction by other stresses, cytokines, or growth factors (15). The specificity of this desensitization reinforces the existence of a highly specific heat shock-sensing pathway upstream of p38. Despite its importance for cell survival, the signaling components and the molecular mechanism leading to heat shock-induced p38 activation are unknown.
Little is known about the mechanisms of activation of the stress-sensitive pathways. In the case of UV light and hyperosmotic shock, activation of the stress-activated protein kinase JNK 1 is triggered by an activation of the receptors for epidermal growth factor, tumor necrosis factor (TNF) ␣, and interleukin-1 (16). Alterations of receptor conformation by energy absorption or physical perturbation of the cell surface are thought to be the initial triggering events causing the clustering and internalization of these receptors and the subsequent subversion of signaling pathways normally used by growth factors or cytokines (16). In the case of oxidative stress, the sensing mechanism seems to act at the level of Ask1 (apoptosis signal-regulating kinase-1). Ask1 is a MAP kinase kinase kinase that can activate the MAP kinase kinases 3 and 6 leading to the activation of p38, or the MAP kinase kinases 4 and 7 leading to the activation of JNK (17). The redox regulatory protein thioredoxin (Trx) acts as the oxidative stress sensor for this cascade (18). Under normal conditions, Trx in the reduced state binds to and inhibits Ask1. Upon oxidative stress, oxidation of Trx triggers its dissociation from Ask1, allowing the activation of Ask1 and the subsequent activation of downstream kinases.
Here we show that Ask1 is also activated during heat shock and that this activation is responsible for p38 activation. However, heat shock activation of Ask1 does not proceed by a redox-dependent mechanism as shown for oxidative stress. Instead, a new mechanism of Ask1 activation is described involving the heat shock-induced dissociation from Ask1 of a recently identified inhibitor of Ask1, glutathione S-transferase Mu1-1 (GSTM1-1) (19). It is concluded that the alternative regulation of Ask1 by the redox-sensitive repressor Trx or the heat-sensitive repressor GSTM1-1 defines the converging point of the heat shock and oxidative stress-sensing pathway leading to p38 activation.

EXPERIMENTAL PROCEDURES
Materials-[␥-32 P]ATP (3000 Ci/mmol) was purchased from PerkinElmer Life Sciences. H 2 O 2 , N-acetyl-L-cysteine (NAC), and myelin basic protein (MBP) were from Sigma. Protein A-Sepharose was from Amersham Biosciences. Chemicals for electrophoresis were obtained from Bio-Rad and Fisher.
Antibodies-HA.11 is a mouse monoclonal antibody recognizing the YPYDVPDYA peptide sequence from human influenza hemagglutinin protein (Covance Research Products, Philadelphia, PA). 9E10 is a mouse monoclonal antibody recognizing the EQKLISEEDL peptide sequence from the human c-Myc protein. 9E10 was prepared from hybridoma cells (American Type Culture Collection). Monoclonal anti-human thioredoxin was obtained from Serotec (Missisauga, ON, Canada). All other antibodies used are polyclonal antibodies raised in rabbit. Anti-p38 recognizes the C-terminal sequence PPLQEEMES of murine p38 (7). Antibody against phosphorylated p38 was obtained from New England Biolabs (Beverly, MA). Anti-Ask1 is a novel antibody developed for this study. It was raised in rabbits against the C-terminal sequence of human Ask1 protein (KAIIDFRNKQT) as described before for anti-p38 (7). Anti-GSTM1-1 is an affinity-purified polyclonal antibody (20).
Cell Culture and Treatments-Chinese hamster CCL39 and human Hela cells were cultivated in Dulbecco's modified Eagle's medium containing 2.2 g/l NaHCO 3 and 4.5 g/l glucose, and supplemented with 5 or 10% fetal bovine serum, respectively. Cultures were maintained at 37°C in a 5% CO 2 humidified atmosphere. Exponentially growing cells (10 6 cells per 60 ϫ 15-mm culture dish plated the day before the experiment) were used for all treatments. For heat shock treatment, the dishes were sealed with parafilm and immersed into a circulating water bath thermoregulated at 44 Ϯ 0.05°C for the indicated period of time. All other inducers used were added directly into culture medium, and the cells were maintained at 37°C for the duration of treatments. NAC was used after adjusting the pH to 7.4 with NaOH.
Kinase Activity Assay-After treatments, cells were scraped and extracted in lysis buffer containing 20 mM Tris-HCl, pH 7.5, 12 mM ␤-glycerophosphate, 150 mM NaCl, 5 mM EGTA, 1 mM Na 3 VO 4 , 10 mM NaF, 1% Triton X-100, 0.5% deoxycholate, 20 g/ml aprotinin, 3 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. The extracts were vortexed and centrifuged at 17,000 ϫ g for 12 min at 4°C. The clarified supernatants were frozen on dry ice and stored at Ϫ80°C. The further steps were carried out at 4°C. To assay Ask1 activity, an equal volume of Ask1 cell lysate normalized for Ask1 protein was incubated with 10 l of anti-Ask1 antibody for 1 h and harvested with 15 l of protein A-Sepharose 50% v/v in lysis buffer. After 30 min, the samples were centrifuged for 15 s and washed twice with 300 l of 20 mM Tris-HCl, pH 7.5, containing 250 mM NaCl, 5 mM EGTA, 1% Triton X-100, 2 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. This was followed by two washes with 20 mM Tris-HCl, pH 7.5, containing 5 mM EGTA, 2 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride. Immunoprecipitate was directly used for the kinase assays. The assay was carried out in 20 l of kinase buffer (100 M ATP, 3 Ci of [␥-32 P]ATP, 20 mM Tris-HCl, pH 7.5, 20 mM MgCl 2 , and 6.5 g of MBP). Ask1 activity was assayed for 12 min at 30°C and was stopped by the addition of 10 l of SDS-PAGE loading buffer. Kinase activity was evaluated by measuring incorporation of the radioactivity into MBP after resolution by SDS-PAGE and quantification using a Phos-phorImager (Molecular Dynamics).
To assay endogenous or transfected GST-tagged p38 activity, cells were lysed directly in SDS-PAGE loading buffer. Proteins were separated by SDS-PAGE on 10% acrylamide gels. Under these conditions, the transfected tagged protein is distinguished from the endogenous protein due to its slower migration. Proteins were then transferred onto nitrocellulose as previously described (6). After reacting the membranes with anti-phospho-p38 antibody, proteins were detected using an enhanced chemiluminescence (ECL) detection kit (Pierce). Equal loading of the kinase on different lanes was verified by immunoblotting with anti-p38 antibody.
Coimmunoprecipitation Assay-After treatments, cells were scraped and extracted in coimmunoprecipitation buffer containing 20 mM Tris-HCl, pH 7.5, 10% glycerol, 150 mM NaCl, 10 mM EDTA, 1 mM Na 3 VO 4 , 1% Triton X-100, 0.5% deoxycholate, 20 g/ml aprotinin, and 1 mM phenylmethylsulfonyl fluoride. The extracts were vortexed and centrifuged at 17,000 ϫ g for 12 min at 4°C. The clarified supernatants were incubated with anti-Trx or anti-GSTM1-1 antibody for 1 h, and the immune complexes were harvested with 10 l of protein A-Sepharose 50% suspension in coimmunoprecipitation buffer. After 30 min, samples were centrifuged for 15 s and washed at least three times with 300 l of coimmunoprecipitation buffer. Proteins were separated using SDS-PAGE, transferred onto nitrocellulose, and immunoblotted with the appropriate antibody.

Ask1
Is an Essential Component for Heat Shock-mediated p38 Activation-Heat shock strongly activates the stress-activated protein kinase p38, but the upstream signaling pathway leading to this activation is not known. The MAP kinase kinase kinase Ask1 has been shown to be an upstream activator of p38 during exposure to several stressful stimuli (17, 19, 24 -27). We tested whether heat shock might also activate Ask1. Endogenous Ask1 was immunoprecipitated from CCL39 cells at various times after shifting the temperature to 44°C. Ask1 activity was determined using MBP as substrate. Ask1 activity increased in a time-dependent manner upon heating (Fig. 1A). To determine whether Ask1 was an obligatory component of the heat shock-induced p38 activation pathway, we examined the effect of overexpressing Ask1(K709M), a catalytically inactive mutant of Ask1, on p38 activation. Cells were transfected with expression vectors encoding HA-tagged Ask1(K709M) and GST-tagged p38 and exposed to heat shock or other known activators of p38. p38 activity was measured in transfected cells using a phosphospecific p38 antibody (Fig. 1B, left panels). Western blot analysis confirmed equivalent levels of expression of GST-tagged p38 and increasing levels of expression of HA-Ask1(K709M) (Fig. 1B, middle and right panels). Expression of Ask1(K709M) inhibited heat shock-induced p38 activation in a dose-dependent manner. The inhibitory effect was specific. In agreement with previous reports (18,21), Ask1(K709M) blocked the activation by H 2 O 2 ; however, it had no effect on the activation of p38 by hyperosmotic shock (sorbitol) or sodium arsenite. These findings indicate that Ask1 is an essential upstream activator of p38 in response to heat shock. The specificity of the inhibitory effect indicates that Ask1 mediates only a subset of the stimuli that activates p38.
Activation of Ask1 by H 2 O 2 in Heat-desensitized Cells-We previously reported that following a first heat shock treatment cells become temporarily desensitized to further p38 activation by heat shock but remain fully responsive to p38 activation by cytokines, growth factors, and stresses (15). This suggests the existence of a specific mechanism for p38 activation by heat shock. We examined whether this homologous heat desensitization process also affected Ask1 activity. Cells were first exposed to a 20-min heat shock at 44°C and then exposed 7 h later to a second heat shock or to a H 2 O 2 treatment. As previously demonstrated for p38 (15) and confirmed here ( Fig. 2A), a heat shock pretreatment desensitized cells for the activation of Ask1 by a second heat shock but had no effect on H 2 O 2induced Ask1 activation (Fig. 2B). These results suggested that heat shock might activate Ask1 by a mechanism different from that used by H 2 O 2 .
Heat Shock Activation of the p38 Pathway Is Redox-insensitive and Is Not Affected by Thioredoxin-Ask1 is known to be responsive to the cellular redox state. We therefore investigated whether a perturbation in the redox state was a necessary event for heat shock activation of the p38 pathway. Cells were treated for 60 min with increasing concentration of the antioxidant NAC prior to exposure to heat shock or H 2 O 2 treatment. Pretreatment with NAC at concentrations higher than 10 mM inhibited H 2 O 2 activation of p38 but had no effect on the stimulation induced by heat shock (Fig. 3A). Thus, whereas generation of reactive oxygen metabolites is an essential event upstream of p38 activation in response to H 2 O 2 , oxidative stress is not a key triggering element for the induction of p38 in response to heat shock.
We next investigated if Trx, an endogenous inhibitor of Ask1, could block p38 activation by heat shock. CCL39 cells were transfected with expression vectors encoding GST-tagged p38 and Myc-tagged Trx and exposed to heat shock or H 2 O 2 treatments. p38 activity was measured using a phosphospecific p38 antibody (Fig. 3B, upper panels). Equivalent levels of expression of GST-tagged p38 and increasing levels of expression of Myc-tagged Trx were confirmed by Western blot analysis (Fig.  3B, middle and bottom panels). Expression of Trx inhibited H 2 O 2 activation of GST-tagged p38 in a dose-dependent manner but had no effect on heat shock activation of GST-tagged p38. We examined the effect of heat shock treatment on the in vivo interaction between Ask1 and Trx. Hela cells were transfected with an expression vector encoding HA-tagged Ask1-WT or with an empty vector and exposed to heat shock or H 2 O 2

FIG. 2. Activation of p38 and Ask1 by H 2 O 2 in heat-desensitized cells.
A, CCL39 cells were submitted (ϩ) or not (Ϫ) to a pre-heat shock (Pre-HS) of 20 min at 44°C and then exposed 7 h later to a second heat shock for 0 (untreated, Ctl) or 20 min (HS) at 44°C, or treated with H 2 O 2 (5 mM, 15 min). After treatments, extracts were prepared and subjected to SDS-PAGE/Western blot analysis (WB) using antibodies specific for phospho-38 (p38-p, upper panel) and total p38 (p38, lower panel). B, same as in A except that the extracts were prepared and processed to determined endogenous Ask1 activity in immune complexes using MBP as substrate. Proteins were then separated by SDS-PAGE, and kinase activity was visualized by autoradiography of the 32 P-labeled substrate.

FIG. 3. NAC pretreatment or overexpression of thioredoxin inhibits p38 activation by H 2 O 2 but not by heat shock.
A, exponentially growing CCL39 cells were pretreated for 60 min with Nacetyl-L-cysteine (NAC, 0-30 mM) and then exposed to a 20-min heat shock at 44°C (HS), to H 2 O 2 (5 mM, 15 min) or left untreated (Ctl). Extracts were prepared and subjected to SDS-PAGE/Western blot analysis (WB) using antibodies specific for phospho-p38 (p38-p, upper panel) and total p38 (p38, lower panel). B, CCL39 cells were co-transfected with 3 g of pEBGp38-GST together with varying concentrations of pCMV5myc-Trx (Myc-TRX, 0-12 g). 48 h after transfection, cells were left untreated (Ctl) or were exposed to a 20-min heat shock at 44°C (HS) or to H 2 O 2 (5 mM, 15 min). After treatments, extracts were prepared and analyzed for the expression of phosphorylated GST-p38 (p38-p, upper panels), total GST-p38 (p38, middle panels) and Myc-Trx (Myc, bottom panels). C, Hela cells were transfected with 3 g of PCDNA3-HA-Ask1-WT or empty vector (Ϫ). 48 h after transfection, cells were left untreated (C), exposed to a 20-min heat shock at 44°C (HS), or treated with H 2 O 2 (P, 5 mM, 15 min). After treatments, the cell lysates were immunoprecipitated (IP) with anti-Trx antibody, and the immunopellets were subjected to SDS-PAGE/Western blot analysis using anti-HA antibody to detect the presence of HA-Ask1-WT (upper panel). The expression of HA-Ask1 (middle panel) and endogenous Trx (bottom panel) in the total soluble cell lysates was verified by immunoblotting with anti-HA and anti-Trx antiserum, respectively. treatment. Endogenous Trx was immunoprecipitated with monoclonal anti-human Trx antibody, and the immunocomplexes were analyzed by immunoblot with anti-HA antibody (Fig. 3C, upper panel). Western blot analyses confirmed equivalent levels of expression of HA-tagged Ask1-WT and endogenous Trx (Fig. 3C, middle and bottom panels). An association between Trx and Ask1 could be demonstrated under control conditions. H 2 O 2 treatment, but not heat shock, caused a dissociation of Trx from Ask1.
Heat Shock Induces the Dissociation of GSTM1-1 from Ask1-GSTM1-1 was recently identified as another endogenous inhibitor of Ask1 activity (19). We tested if GSTM1-1 could possibly modulate p38 activation by heat shock. CCL39 cells were transfected with expression vectors encoding GSTtagged p38 and HA-tagged GSTM1-1(Y6F) and exposed to heat shock, H 2 O 2 , or sorbitol treatment. GST-tagged p38 phosphorylation was measured using a phosphospecific p38 antibody (Fig. 4A, left panels). Equivalent levels of expression of GSTtagged p38 and increasing levels of expression of HA-tagged GSTM1-1(Y6F) were confirmed by Western blot analysis (Fig.  4A, middle and right panels). Expression of GSTM1-1 inhibited in a dose-dependent manner heat shock-induced activation and, to a lesser extent, H 2 O 2 -induced activation of p38. It had no effect on sorbitol activation of p38. This result was consist-ent with the fact that GSTM1-1 is an inhibitor of Ask1 and that Ask1 was not involved in the activation of p38 by sorbitol (Fig.  1B). We next examined the effect of heat shock treatment on the interaction of Ask1 with GSTM1-1 in vivo. Cells were transfected with expression vector encoding HA-tagged Ask1-WT and exposed to heat shock or H 2 O 2 treatment. Endogenous GSTM1-1 was immunoprecipitated with polyclonal anti-GSTM1-1 antibody, and the immunocomplexes were analyzed by immunoblot using anti-HA antibody (Fig. 4B). Western blot analyses confirmed equivalent levels of expression of HA-tagged Ask1-WT and endogenous GSTM1-1 (Fig. 4B, middle and bottom panels). The immunoblot data revealed an association between GSTM1-1 and Ask1 under control conditions. Heat shock treatment, but not H 2 O 2 exposure, disrupted this interaction.

Ask1 Is an Obligatory Component of the Specific Heat-sensitive p38
Signaling Pathway-The stress-activated protein kinase p38 is activated by a number of different stimuli including growth factors and cytokines, but also by many different stressing agents or conditions such as exposures to physical or chemical DNA damaging agents, cytoskeleton disrupting drugs, hypo-or hyperosmotic shock, shear stress, reoxygenation following hypoxia, oxidative stress, and heat shock (7, 15, 28 -37). The fact that so many different stressing agents can activate the p38 pathway suggests the existence of distinct sensing pathways that converge on an upstream activator capable of integrating different signals. Using classical desensitization experiments, we recently demonstrated the existence of such a specific heat shock-sensing pathway upstream of p38, distinct from that used by other stresses such as hyperosmotic stress or H 2 O 2 (15). Here we showed that this sensing pathway converges on Ask1, a MAP kinase kinase kinase that also mediates p38 activation by TNF␣, cisplatin, and H 2 O 2 (17-19, 21, 27, 38). Two lines of evidence indicate that Ask1 is an obligatory component of the heat-specific pathway leading to p38 activa-FIG. 5. Ask1 is located at the converging point of distinct stress-sensing pathways. Distinct pools of Ask1 signaling complexes allow Ask1 to be activated by different stimuli. In the case of H 2 O 2 , its activation involves the oxidation-mediated dissociation of thioredoxin (TRX), an endogenous inhibitor of Ask1. In the case of TNF a similar mechanism has been described. Reactive oxygen species produced in response to TNF cause the oxidation-mediated dissociation of Trx from Ask1, enabling the binding of TRAF2 to Ask1 and its activation. During heat shock a dissociation of the complex formed by glutathione Stransferase Mu1-1 (GSTM1-1) and Ask1 occurs, leading to the activation of Ask1 and downstream kinases. The mechanism that leads to this dissociation remains to be determined. It may involve the release of a lipophilic molecule that competes with Ask1 for binding with GSTM1-1 following heat shock. Because of an alternative regulation by distinct repressors, thioredoxin and GSTM1-1, Ask1 constitutes the converging point of the heat shock and oxidative stress-sensing pathways that lead to p38 activation (see text for details).
Distinct Molecular Mechanisms Mediate Heat Shock and H 2 O 2 -induced Activation of Ask1-We recently reported that the signaling elements downstream of Ask1 are desensitized homologously by heat shock (15). Here we showed that a priming heat shock treatment also inhibits the activation of Ask1 by a subsequent heat shock, but does not affect activation of Ask1 by H 2 O 2 (Fig. 2B). This result implied that there exists a heat-specific activation mechanism operating at the level or upstream of Ask1 that is distinct from that used by H 2 O 2 . Such distinct mechanisms of Ask1 activation are presented in Fig. 5.
H 2 O 2 activates the Ask1-p38 module by the oxidation of the redox-sensing protein Trx (18). In non-stressed cells Trx binds to Ask1, an association that keeps Ask1 in an inactive form. This interaction is dependent on the redox status of Trx. Oxidation of Trx dissociates the complex, allowing the activation of Ask1 by oligomerization and autophosphorylation (26). Recent evidence also suggests the participation of an unidentified kinase in this process (39). A similar mechanism is involved in TNF-induced activation of Ask1 (38,40). Reactive oxygen species produced in response to TNF cause the oxidation-mediated dissociation of Trx from Ask1, enabling the binding of TNF receptor-associated factor 2 (TRAF2) to Ask1 and its activation (Fig. 5). Activation of Ask1 could be blocked in both cases with free-radical scavengers, including the overexpression of Trx (38,40).
In contrast to H 2 O 2 -induction, activation of p38 by heat shock was not antagonized by pretreatment with the antioxidant NAC or by overexpression of Trx (Fig. 3, A and B). Furthermore, in contrast to H 2 O 2 , heat shock did not cause a dissociation of Trx from Ask1 (Fig. 3C). Instead, we found that heat shock-induced activation of Ask1 involves the modulation of GSTM1-1, a potent endogenous inhibitor of Ask1 activity (19). It was previously proposed that GSTM1-1 could inhibit Ask1 activity by binding to the kinase and preventing its oligomerization. The modulation of Ask1 activity by GSTM1-1 was hypothesized to be the results of modulations in the level of expression of GSTM1-1. In particular it was suggested that induction of GSTM1-1 expression after stress could participate in a homeostatic mechanism to block further Ask1 induction, thereby protecting the cells from Ask1-induced apoptosis (19). Our results imply a more direct regulation of GSTM1-1 inhibitory activity by heat shock. We showed that heat shock causes the release of GSTM1-1 from Ask1 (Fig. 4B). The dissociation of GSTM1-1 from Ask1 is not induced by H 2 O 2 treatment suggesting that the dissociation is not a consequence of Ask1 activation. Instead, the dissociation of GSTM1-1 likely triggers Ask1 activation since overexpression of exogenous GSTM1-1 inhibited p38 activation by heat shock in a dose-dependent manner (Fig. 4A).
This heat shock-induced dissociation of GSTM1-1 from Ask1 is reminiscent of H 2 O 2 and TNF-␣-induced dissociation of Trx from Ask1 (18,21,40). A similar mechanism of activation has also been described for JNK. Association between GST-pi and JNK is disrupted by oxidative stress caused by the formation of GST-GST dimers and multimers, and the dissociation leads to enhanced-JNK activity (41). It is tempting to suggest that similar to Trx, which acts as a redox-sensitive repressor of Ask1, GSTM1-1 acts as a heat-sensitive repressor of Ask1, and its dissociation following heat shock leads to Ask1 activation.
The mechanisms that lead to the dissociation of GSTM1-1 from Ask1 and whether it can also be induced by stresses other than heat shock remain to be determined. The release of GSTM1-1 from Ask1 as well as inhibition of its activity were demonstrated using GSTM1-1(Y6F), a mutant with no glutathione S-transferase activity (19), indicating that this activity was not involved in the regulation of Ask1. GST activity results in the conjugation of glutathione to dangerous electrophilic compounds. In addition to their catalytic activity GST proteins have an ill-defined role as ligandins, being capable of binding a number of small lipophilic molecules such as steroids and their metabolites (42,43). It is possible that during heat shock such ligandin-binding molecules are released or produced in such a way that they compete with Ask1 for binding to GSTM1-1 (Fig.  5). Like the release of Trx from Ask1 in the case of H 2 O 2 , the release of GSTM1-1 during heat shock may then allow subsequent Ask1 activation by oligomerization. However, the release of this inhibitor may not be sufficient for activation of Ask1. In the case of TNF the participation of the activator TRAF2 is required for Ask1 activation (38,40). Likewise, Ask1 activation by heat shock may necessitate the contribution of a heat-induced activator in addition to its dissociation from GSTM1-1.
Ask1, a Tightly Regulated Kinase Integrating Two Stresssignaling Pathways-As a result of this alternative regulation exerted by Trx and GSTM1-1, Ask1 can constitute the converging point of the heat shock and oxidative stress-sensing pathways leading to p38 activation. The nature of this dual regulation remains to be determined. Since heat shock and H 2 O 2 cause the release of only one of the inhibitors, it could be suggested that neither Trx nor GSTM1-1 can by itself inhibit Ask1 and thus that the release of only one is sufficient for activation. A simpler hypothesis, however, is that both Trx and GSTM1-1 are by themselves fully efficient as inhibitors, but that there exist in the cells different pools of Ask1 at equilibrium, some bound to Trx and responsive to oxidative stress and others bound to GSTM1-1 and responsive to heat shock (Fig. 5). This is possible because both inhibitors bind to the N-terminal region of Ask1 (18,19) and thus may compete for the same region. Considering the large number of regulatory molecules known to bind Ask1 (18, 19, 25, 38, 40, 44 -53), several distinct pools of these regulatory complexes might exist. Indeed, one example already reported in the literature is the existence of a complex composed of Ask1, MKK4, and JNK3 that are held together by the scaffold protein ␤-arrestin-2. This complex mediates Ask1 and JNK3 activation in response to stimulation of guanine nucleotide-binding protein-coupled receptors (54). The position of Ask1 at the converging point of several sensing pathways might be particularly important in cell physiology in the light of the recently proposed pivotal role of Ask1 in determining cell fate between survival and apoptosis (55).