Role of Apoptosis Signal-regulating Kinase 1 (ASK1) as an Activator of the GAPDH-Siah1 Stress-Signaling Cascade*

Background: Apoptosis signal-regulating kinase 1 (ASK1), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and seven in absentia homolog 1 (Siah1) are molecules associated with stress-signaling cascades. Results: Identification of Siah1 as a substrate of ASK1 for activation of the GAPDH-Siah1 signaling cascade. Conclusion: ASK1 triggers the GAPDH-Siah1 stress-signaling cascade. Significance: This study provides insight into crosstalk among cell stress-signaling cascades. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) plays roles in both energy maintenance, and stress signaling by forming a protein complex with seven in absentia homolog 1 (Siah1). Mechanisms to coordinate its glycolytic and stress cascades are likely to be very important for survival and homeostatic control of any living organism. Here we report that apoptosis signal-regulating kinase 1 (ASK1), a representative stress kinase, interacts with both GAPDH and Siah1 and is likely able to phosphorylate Siah1 at specific amino acid residues (Thr-70/Thr-74 and Thr-235/Thr-239). Phosphorylation of Siah1 by ASK1 triggers GAPDH-Siah1 stress signaling and activates a key downstream target, p300 acetyltransferase in the nucleus. This novel mechanism, together with the established S-nitrosylation/oxidation of GAPDH at Cys-150, provides evidence of how the stress signaling involving GAPDH is finely regulated. In addition, the present results imply crosstalk between the ASK1 and GAPDH-Siah1 stress cascades.


Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) plays roles in both energy maintenance, and stress signaling by forming a protein complex with seven in absentia homolog 1 (Siah1).
Mechanisms to coordinate its glycolytic and stress cascades are likely to be very important for survival and homeostatic control of any living organism. Here we report that apoptosis signalregulating kinase 1 (ASK1), a representative stress kinase, interacts with both GAPDH and Siah1 and is likely able to phosphorylate Siah1 at specific amino acid residues (Thr-70/Thr-74 and Thr-235/Thr-239). Phosphorylation of Siah1 by ASK1 triggers GAPDH-Siah1 stress signaling and activates a key downstream target, p300 acetyltransferase in the nucleus. This novel mechanism, together with the established S-nitrosylation/oxidation of GAPDH at Cys-150, provides evidence of how the stress signaling involving GAPDH is finely regulated. In addition, the present results imply crosstalk between the ASK1 and GAPDH-Siah1 stress cascades.
Survival of living organisms is dependent on homeostatic control of both energy maintenance and flexible response to environmental stressors (1)(2)(3)(4). In complex living creatures, which are comprised of multiple cells and organs, survival of each cell and elimination of damaged cells becomes important for homeostatic control (5)(6)(7)(8). Despite the importance of this fundamental concept, the molecular machinery underlying these mechanisms, and their coordination are not fully understood.
A major mechanism for supplying cellular energy is glycolysis, in which glucose is catabolized to pyruvic acid via several enzymatic reactions. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) 3 is a key enzyme in glycolysis, and so plays a major role in cellular energy supply (9). In addition to this classic concept, GAPDH is involved in many subcellular processes that include; DNA repair (10), membrane fusion, and transport (11), tRNA export (12), and cell death (13)(14)(15)(16)(17)(18)(19). The functional diversity of GAPDH is largely regulated by its subcellular localization and post-translational modifications (20). Recently studies have revealed that GAPDH can be oxidized and/or S-nitrosylated under stress conditions. Following this post-translational modification GAPDH is then translocated to the nucleus as a complex with Siah1, which has a strong nuclear localization signal (21). Only about 2% of total GAPDH, and a small pool of Siah1 participate in this mechanism, therefore a gain of function by these two molecules due to the specific posttranslational modifications, instead of their loss of functions, is crucial for this signaling cascade (21). In the nucleus, GAPDH modulates several proteins, in particular stimulating the catalytic activity of acetyltransferase p300/CREB-binding protein (CBP) that regulates transcription of various genes (22). Therefore, GAPDH may modulate homeostatic control by bridging energy supply (glycolytic pathway) to stress response (GAPDH-Siah1 cascade), which is finely regulated by post-translational modification (23)(24)(25)(26).
Apoptosis signal-regulating kinase 1 (ASK1) is a member of the mitogen-activated protein kinase kinase kinase (MAPKKK) family. Although cellular substrates of ASK1 have not yet been fully studied, MAPKK4/7 (a kinase that phosphorylates JNK) and MAPKK3/6 that phosphorylates p38 are well-established substrates (27). ASK1 is activated in response to oxidative stress, endoplasmic reticulum stress, and other forms of cellular stress (28,29). In addition, ASK1 plays pivotal roles in a wide variety of cellular responses, which include, but are not limited to, apoptosis (30,31). Dysregulation of the ASK1 signaling pathway is closely linked to various diseases, such as cancer, cardiovascular diseases, diabetes, and neurodegenerative diseases including polyglutamine-induced neurodegeneration and Parkinson disease (29,(32)(33)(34)(35)(36)(37).
The primary focus of the present study is to elucidate regulatory mechanisms of the GAPDH-Siah1 pathway. Here we report that ASK1 phosphorylates Siah1 and critically modulates the GAPDH-Siah1 pathway via direct protein interaction.
Co-immunoprecipitation (co-IP)-Cells were lysed in immunoprecipitation (IP) buffer (50 mM Tris pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.1 mg/ml BSA, protease inhibitor mixture), co-IPed, and Western blotted as previously described (21,22). For sequential co-IP studies, the first co-IP reactions with an anti-HA antibody were eluted with HA peptide, elutes were subjected to a subsequent co-IP with anti-Myc antibody. The final co-IP was then subject to Western blot analysis with an anti-FLAG antibody.
Extraction of Nuclear and Cytoplasmic Proteins-Nuclear and cytoplasmic extracts were prepared using the Biovision nuclear/cytosol extraction kit according to the manufacturer's instructions.
In Vitro Binding Assays-For ASK1-Siah1 and ASK1-GAPDH in vitro binding assays with equal molar concentrations of GST-tagged-ASK1, Siah1, and His-tagged-GAPDH were incubated in binding buffer (0.1% Nonidet P-40, 0.5 mM DTT, 10% glycerol, 1 mM PMSF, and 2 g/ml aprotinin in PBS) for 2 h at 4°C. For measuring the effects of GAPDH on ASK1-Siah1 binding; 0, 1, and 3 (GAPDH:Siah1) molar concentrations of recombinant GAPDH and Siah1 were incubated in binding buffer, as mentioned above. To obtain recombinant GAPDH and Siah1 without a GST tag, GSH Sepharosebounded protein was released via thrombin digestion, dialyzed and purity analyzed by Western blot. All in vitro binding assays were done by GST pull-down via incubation with GSH-Sepharose beads (50% slurry) for 1 h, the samples were centrifuged at 4000 rpm for 1 min, washed three times in binding buffer, and resuspended in LDS sample buffer (Invitrogen) with 5% ␤-mercaptoethanol (Sigma) and then heated at 95°C for 5 min. Western blot analysis of the protein precipitates were done using anti-GAPDH, Siah1, and GST antibodies.
Statistical Analysis-Two-group analysis was performed by t test (paired or unpaired as appropriate). A value of p Ͻ 0.05 is considered significant. All data were obtained from the results of three or four independent experiments.

GAPDH and Siah1 Bind to ASK1 and Form a Ternary Complex in Cells-GAPDH-Siah1
and ASK1 have been reported independently to play roles in several pathological brain conditions, and are commonly shown to be key stress mediators (26,30,(41)(42)(43). Thus, we hypothesized that Siah1 and GAPDH might interact with ASK1 at the molecular level. To address this question, we examined mouse brain lysates, and we observed endogenous protein interactions of ASK1-Siah1 and ASK1-GAPDH by co-immunoprecipitation (Fig. 1A). These interactions were also seen and augmented in the presence of H 2 O 2 (1 mM, 30 min) in HEK293 cells (Fig. 1B).
To test whether or not ASK1, Siah1 and GAPDH interact, we performed a sequential co-IP on cell lysates from HEK293 cells expressing ASK1, Siah1, and GAPDH: we observed ASK1-Siah1 and ASK1-GAPDH interaction are induced by treatment with H 2 O 2 (Fig. 2, upper panel). Previous studies have reported stimulation of ASK1 activity by H 2 O 2 (44). To determine if H 2 O 2 stimulated ASK1 activity may induce complex formation we measured ASK1 activity via phosphorylation of ASK1 at Thr-845 and demonstrated that ASK1 activity correlated with ASK1-Siah1-GAPDH complex formation (Fig. 2, lower panel). These results suggest that ASK1, Siah1, and GAPDH form a complex in response to extracellular stressors.
Direct ASK1-Siah1 Interaction and Modulation by GAPDH in Vitro-GAPDH and Siah1 are known to bind directly (21). To characterize the interaction of Siah1 and GAPDH with ASK1 we purified recombinant proteins for in vitro binding studies. Incubation of recombinant Siah1 together with gluta-FIGURE 1. Siah1 and GAPDH bind ASK1. A, Siah1-ASK1 and GAPDH-ASK1 binding in mouse brain extracts. Mouse brain extracts were immunoprecipitated with anti-Siah1 or anti-GAPDH antibodies and analyzed by Western blot with anti-ASK1, GAPDH, and Siah1 antibodies. Input is total cell lysates for immunoprecipitation (IP). B, Siah1-ASK1 binding and GAPDH-ASK1 binding is induced by stress in HEK293 cells. HEK293 cells were treated with 1 mM H 2 O 2 for 30 min, and protein extracts were immunoprecipitated with anti-Siah1, anti-GAPDH, or anti-ASK1 antibodies and analyzed by Western blot with anti-ASK1, GAPDH, and Siah1 antibodies. Input is total cell lysates for IP. thione S-transferase (GST) or GST-tagged-ASK1 (amino acids 1-940) demonstrated that Siah1 binds directly to the N-terminal region of ASK1 (Fig. 3A). In contrast, incubation of recombinant GAPDH together with GST or GST-tagged-ASK1 failed to demonstrate that GAPDH directly binds to ASK1 in vitro (data not shown). Given that stress has been demonstrated to induce direct binding of GAPDH and Siah1 (21), we hypothesized that GAPDH may augment ASK1-Siah1 binding. To determine if GAPDH modulated ASK1-Siah1 binding we performed in vitro binding assays with recombinant Siah1 and ASK1 (amino acids 1-940) in the presence of increasing amounts of GAPDH. These studies demonstrated that a three molar equivalent of GAPDH augmented ASK1-Siah1 direct binding (Fig. 3B). ASK1 (amino acids 1-940) contains the ASK1 kinase domain (649 -946) (27).
ASK1 Phosphorylates Siah1-Given that Siah1 was determined to directly bind within the kinase domain of ASK1, we hypothesized that Siah1 might be a novel substrate of ASK1 phosphorylation. To investigate whether ASK1 could phosphorylate Siah1, we conducted phosphorylation studies using recombinant proteins in vitro. Phosphorylation of GST-tagged Siah1, but not GST, was observed when Siah1 was incubated with the constitutively active kinase domain of ASK1 (amino acids 649 -946) (Fig. 4A). Since ASK1 has been reported to phosphorylate substrates carrying the (S/TXXXS/T) consensus motif (45,46), we examined the amino acid sequence of Siah1 and identified four potential phosphorylation sites which we designated M1, M2, M3, and M4, respectively (Fig. 4B). To characterize the ASK1 phosphorylation sites on Siah1, we generated mutants of Siah1 with point mutations (threonine to alanine substitution) in each consensus sequence and examined how these mutations affected phosphorylation of Siah1 by ASK1 via in vitro kinase assays (Fig. 4C). Amino acid substitution in M2 led to significant decreases in the phosphorylation of Siah1 by ASK1. However, Siah1 phosphorylation by ASK1 was reduced the most when we introduced mutations at both M2 and M4 (M2ϩM4) (Fig. 4C). These data indicate that Thr-70/ Thr-74, and Thr-235/Thr-239 together are the critical sites in Siah1 phosphorylated by ASK1. We next addressed whether the phosphorylation of Siah1 was induced by ASK1 in cells. To address this question we transfected HEK293 cells with a constitutively kinase activity-positive (CA) ASK1 (amino acids 649 -1375) together with WT Siah1 or M2ϩM4 mutant Siah1 lacking ASK1 phosphorylation sites. The phosphorylation levels of Siah1 were examined by using an anti-Myc antibody after enrichment of proteins with phosphorylation at threonine residues (pThr proteins) by immunoprecipitation with an antiphosphothreonine antibody. We observed a marked increase in the signal from WT Siah1, but not M2ϩM4 mutant Siah1, only when co-transfected with CA-ASK1, indicating that Siah1 is phosphorylated at M2ϩM4 sites by ASK1 (Fig. 4D). At the same time, GAPDH binding with Siah1 was diminished by the

. ASK1 directly binds to Siah1 and is modulated by GAPDH in vitro.
A, recombinant GST-ASK1 (aa 1-940) or GST were incubated with recombinant Siah1 in vitro and subjected to GST pull-down, followed Western blot with an anti-Siah1 antibody. Input is the starting material for immunoprecipitation (IP). The input lanes were probed as follows: GST-Siah1 with an anti-Siah1 antibody; GST-ASK1 and GST with an anti-GST antibody. The arrow indicates GST-ASK1 (1-940). B, GAPDH facilitates ASK1-Siah1 binding in vitro. ASK1-Siah1 interaction was assessed by incubating recombinant GST-ASK1 (aa 1-940), or GST with recombinant Siah1 and GAPDH protein at 0, 1, or 3 (GAPDH:Siah1) molar concentrations, followed by to GST pull-down. Precipitates were analyzed by Western blot with an anti-Siah1 antibody. Input is the starting material for IP. ASK1-Siah1 binding was quantified by densitometric analyses (t test; *, p Ͻ 0.05 versus Siah1).

ASK1 as an Activator of GAPDH-Siah1 Signaling
replacement of the key threonine residues (the phosphorylation sites of Siah1 by ASK1) to alanine (Fig. 4E): these results suggest thatASK1 kinase activity is crucial for the binding of GAPDH and Siah1, that is, the activation of the GAPDH-Siah1 stress-signaling cascade.
ASK1 Augments GAPDH-Siah1 Binding in Cells-We next addressed whether ASK1 modulates GAPDH-Siah1 signaling. Thus, we introduced wild-type (WT) ASK1, kinase-dead (KD) ASK1 that was generated with one amino acid substitution at 709 (K709M), and constitutively kinase activity-positive (CA) ASK1 (see above) in cells, respectively (39). We then examined how these distinct forms of ASK1 affected GAPDH-Siah1 bind-ing. We previously reported that sodium nitroprusside (a nitric oxide donor) could affect GAPDH-Siah1 binding by S-nitrosylation of GAPDH (21), which was used as a reference of the binding change. Introduction of WT ASK1 dramatically augmented GAPDH-Siah1 binding, which was considerably greater than the change elicited by sodium nitroprusside in total cell lysates (Fig. 5, lower panel). When kinase activity of ASK1 was selectively reduced (KD ASK1), such augmentation was also reduced. Consistent with this observation, introduction of CA ASK1 also dramatically augmented GAPDH-Siah1 binding. In all the conditions, the levels of GAPDH-Siah1 binding were normalized by the levels of Myc-Siah1 (Fig. 5, upper panel). Note, the levels of Siah1 are affected by expression of different types of ASK1 constructs. These results suggest that ASK1, especially its kinase activity, is critical in augmenting GAPDH-Siah1 protein binding.

ASK1-induced GAPDH-Siah1 Binding in Cells Is
Independent of p38 and JNK Signaling-We considered the possibility that increased GAPDH-Siah1 binding could be affected by JNK and p38, two key kinases downstream of ASK1 (27). To test this idea, we used specific kinase inhibitors (SB203580 specific for p38 and SP600125 specific for JNK) and examined the effects on GAPDH-Siah1 binding in the presence of exogenous WT ASK1. Neither SB203580 nor SP600125 affected ASK1-GAPDH, ASK1-Siah1, and GAPDH-Siah1 interactions (Fig. 6,  A and B), suggesting that ASK1-induced GAPDH-Siah1 binding occurs independent to p38/JNK signaling. ASK1 Augments Nuclear Translocation of GAPDH and p300 Acetylation in Cells-The major event of activated GAPDH-Siah1 stress signaling is translocation of GAPDH (21). Thus, we questioned whether ASK1 might facilitate the nuclear translocation, and tested the effects of WT or KD ASK1. We observed robust levels of nuclear translocation of GAPDH in the presence of exogenous WT ASK1, whereas introduction of KD ASK1 did not elicit significant levels of GAPDH translocation (Fig. 7A). We further tested whether this translocation by ASK1 was related to phosphorylation of Siah1. In the presence of ASK1 the complex of GAPDH and mutant Siah1 (M2ϩM4 mutant; lacking ASK1 phosphorylation sites) displayed significantly reduced nuclear translocation compared with the complex of WT Siah1 and GAPDH (Fig. 7B). We then questioned whether these Siah1 mutations (M2ϩM4) critically affect nuclear functions in ASK1triggered GAPDH-Siah1 stress signaling. Thus, we examined GAPDH-p300 binding and acetylation of p300, which have been established as good functional indicators in the nucleus (22). Western blot analysis revealed a significant decrease in p300-GAPDH binding and acetylation of p300, which can be interpreted as a key consequence of reduction in nuclear translocation of GAPDH with the mutant Siah1 (Fig. 7C).
To address functional outcome downstream of P300-GAPDH, we applied H 2 O 2 (1 mM, 30 min) to HEK293 cells and examined expression of PUMA, which is known as a representative target of P300-GAPDH activation (22). Under these conditions we observed a significant reduction of PUMA levels in the cells that expressed M2ϮM4 mutant Siah1 deficient in ASK1 phosphorylation sites, compared with the cells with wild-type Siah1 (Fig. 8). These results suggest that phosphorylation of Siah1 by ASK1 is likely to play a key role in the GAPDH-Siah1 signaling cascade and subsequent functional effects in the nucleus.

DISCUSSION
Cellular signaling in response to stressors is crucial in homeostatic control and survival of all organisms. Thus, crosstalk among signaling cascades and their finely tuned regulation are expected. In the present study, we show that ASK1, a representative stress kinase, regulates GAPDH-Siah1 signaling. We demonstrate that ASK1 binds with GAPDH and Siah1:   JANUARY 2, 2015 • VOLUME 290 • NUMBER 1 JOURNAL OF BIOLOGICAL CHEMISTRY 61 these bindings are augmented under oxidative stress and are likely to be the basis of this signal network. We have identified Siah1 is a novel substrate of ASK1, with Thr-70/Thr-74 and Thr-235/Thr-239, as the critical phosphorylation sites on Siah1. Phosphorylation by ASK1 was found to increase GAPDH-Siah1 binding, its nuclear translocation, and subsequent acetylation of nuclear P300 and PUMA expression.

ASK1 as an Activator of GAPDH-Siah1 Signaling
Thus far, a specific post-translational modification of GAPDH (S-nitrosylation/oxidation at Cys-150) had been established as an initial trigger of GAPDH-Siah1 signaling (21). In the present study we show that specific post-translational modifications of Siah1, the other partner of this complex, can also be a trigger of this cascade (Fig. 9). ASK1 is activated in response to various stressors, such as oxidative stress and endoplasmic reticulum stress, it is likely that Siah1 is subsequently phosphorylated and mediates stress signaling by forming a complex with GAPDH. Thus, this study establishes the notion that the GAPDH-Siah1 cascade is activated by more than one mechanism in the presence of various cellular stressors. This cascade is also inhibited by at least two mechanisms: both inter-action with a cytosolic protein GOSPEL (47) and a set of deprenyl-related compounds (40). It is likely that the GAPDH-Siah1 cascade, which is crucial for homeostatic control, has multiple FIGURE 7. ASK1 augments nuclear translocation of GAPDH and p300 acetylation in cells. A, ASK1 augments nuclear translocation of GAPDH in a kinasedependent manner. Effects of ASK1 kinase activity on nuclear translocation of GAPDH were examined in HEK293 cells expressing HA-tagged WT or KD ASK1, followed by cellular fractionation and Western blot quantification of GAPDH nuclear/cytoplasmic ratios. GAPDH signals were normalized by LDH (a cytoplasmic marker) and P84 (a nuclear marker) (t test; ***, p Ͻ 0.001 versus Mock). B, M2ϩM4 phosphorylation sites of Siah1 by ASK1 are critical to induce nuclear translocation of GAPDH. HEK293 cells expressing HA-WT ASK1 together with Myc-tagged WT or M2ϩM4 mutant Siah1 were subjected to subcellular fractionation. Nuclear/cytoplasmic ratio of GAPDH or Siah1 was analyzed in cells expressing WT or M2ϩM4 mutant. The ratios of both endogenous GAPDH and exogenous myc-Siah1 were reduced in cells expressing M2ϩM4 mutant compared with those expressing WT Siah1 (t test; *, p Ͻ 0.05 versus WT-Siah1). C, M2 and M4 phosphorylation sites of Siah1 by ASK1 are required for acetylation of p300 and p300-GAPDH interaction. Lysates of HEK293 cells expressing HA-WT ASK1 together with Myc-WT Siah1 or M2ϩM4 mutant Siah1 were immunoprecipitated with an anti-p300 antibody and analyzed by Western blot with acetyl-p300 (Lys-1499)/CBP (Lys-1535) (p300/CBP Ac ) and GAPDH antibodies. Input is endogenous p300 and GAPDH in total cell lysates. Western blots were quantified by densitometric analyses (t test; *, p Ͻ 0.05 and **, p Ͻ 0.01 versus WT Siah1). mechanisms that regulate its initiation in both positive and negative ways. It will be interesting to clarify how these two distinct post-translational modifications (oxidation of GAPDH and phosphorylation of Siah1) are coordinated under different types of stress. Given that both ASK1 and GAPDH-Siah1 cascades are major routes of stress signaling, an important future question is to understand how GAPDH-Siah1 can also influence ASK1.
Here we demonstrate that ASK1 and GAPDH-Siah1 co-mediate stress signaling and up-regulate p300. It is known that p300 has multiple functions in different conditions: for example, in the heart p300 activation can lead to heart hypertrophy mediated by myocyte enhancer factor-2 (MEF2) (48), whereas in the brain p300 affects memory function via cAMP response element-binding (CREB) (49). Roles for stressors in these conditions are appreciated, but further molecular mechanisms remain to be elucidated. Thus, new generation of conditional knock-out mice or inducible transgenic models targeting these molecules will be crucial to test context-dependent crosstalk of ASK1 and GAPDH-Siah1. As far as we are aware, such mice are not available at present. Better understanding of the crosstalk between these two stress cascades (ASK1 and GAPDH-Siah1) may provide a more integrated and comprehensive picture of how our body responds to stressors in a context-dependent fashion and how disturbances of such mechanisms may lead to pathological conditions.