Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), LipofectAMINE PLUS reagent, anti-Xpress, horseradish peroxidase-conjugated anti-rabbit, and anti-mouse IgGs were purchased from Invitrogen (Carlsbad, CA, USA). Enhanced chemiluminescence (ECL) reagents and [γ-³²P]ATP were purchased from Perkin-Elmer Life and Analytical Sciences (Waltham, MA, USA). Glutathione-Sepharose 4B and Protein A-Sepharose were obtained from Amersham Biosciences (Piscataway, NJ, USA). Anti-phosphoASK1 (Thr-845), anti-phosphoJNK, and anti-Dyrk1A antibodies were purchased from Cell Signaling Technology (Beverly, MA, USA). Anti-HA, anti-ASK1, anti-JNK, anti-Hsp90, mouse immunoglobulins (IgGs) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Anti-Dyrk1A antibody was purchased from Abnova Corporation (Taipei City, Taiwan). Anti-phosphotyrosine antibody was purchased from Upstate Biotechnology (Lake Placid, NY, USA). Anti-Flag antibodies and all other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA). Mammalian constructs encoding wild-type and kinase-inactive rat Dyrk1A tagged with hemagglutinin (HA) (pSVL-HA-Dyrk1A WT and K188R) were a kind gift from W. Becker (Institut fur Pharmakologie und Toxikologie, Universitatsklinikum der RWTH, Germany). Plasmids encoding 6xHis-Xpress-tagged wild-type and K188R mutant Dyrk1A (pcDNA4/HisMax-Dyrk1A WT and K188R) were generated as described previously (Park et al., 2007). Constructs encoding HA-tagged forms of ASK1, ASK1ΔN, ASK1ΔC, ASK1-NT, and Flag-ASK1 was kindly provided by E.J. Choi (Korea University, Seoul, Korea).

Human embryonic kidney 293 (HEK293) cells were maintained in DMEM containing 10% FBS and 100 unit/ml penicillin-streptomycin. The cells were transfected with various expression vectors using LipofectAMINE PLUS reagent, according to the manufacturer's protocols. In order to prepare cell lysates, the cells were rinsed twice with ice-cold phosphate-buffered saline, and then solubilized in lysis buffer (50 mM Tris, pH 7.5, containing 1.0% Nonidet P-40, 150 mM NaCl, 10% glycerol, 1 mM Na₃VO₄, 1 µg/ml leupeptin, 1 µg/ml aprotinin, 1 mM EGTA, 1 mM EDTA,10 mM NaF, and 0.2 mM phenylmethylsulfonyl fluoride). The cells were scraped, and the supernatants were collected after centrifugation for 20 min at 14,000×g at 4℃.

One microgram of suitable antibodies was incubated with 0.5 to 1 mg of cell extracts prepared in cell lysis buffer overnight at 4℃. Fifty microliter of a 1:1 suspension of Protein A-Sepharose beads was added and incubated for 2 h at 4℃ with gentle rotation. After beads were pelleted and washed extensively with cell lysis buffer, the immunocomplexes were dissociated by boiling in SDS-PAGE sample buffer, separated onto SDS-PAGE gel, and transferred to a nitrocellulose membrane (Millipore, Japan). Membranes were then blocked in TBST buffer (20 mM Tris, pH 7.5, 137 mM NaCl, and 0.1% Tween 20) plus 5% nonfat dry milk for 1 h at room temperature, and incubated overnight at 4℃ in TBST buffer with 3% nonfat dry milk and the appropriate primary antibodies. Membranes were washed three times in TBST, and then incubated with appropriate secondary IgG-coupled horseradish peroxidase antibodies for 1 h at room temperature. The membranes were washed three times with TBST and visualized with ECL reagent.

After cells were harvested and lysed in lysis buffer, the protein extracts (~500 µg) were incubated with appropriate antibody for overnight at 4℃. The immunocomplexes were mixed with protein A-Sepharose beads, and washed three times in lysis buffer. For JNK assay, the kinase reaction was carried out at 30℃ for 30 min in 20 µl of kinase buffer (25 mM HEPES, pH 7.4, 25 mM glycerophosphate, 25 mM MgCl₂, 100 µM sodium orthovanadate) containing 10 µM ATP, 5 µCi of [γ-³²P] ATP, and 5 µg GST-c-Jun as a substrate. For Dyrk1A kinase assay, immunocomplexes were resuspended in the kinase reaction buffer (20 mM Tris-HCl, pH 7.5, 20 mM MgCl₂) containing 10 µM ATP, 5 µCi [γ-³²P]ATP, and incubated for 15 min at 30℃ in the presence of 5 µg of GST-ASK1 as a substrate. The reactions were stopped by adding SDS-sample buffer, and analyzed by SDS-PAGE followed by autoradiography.

In a previous report we have shown that the addition of etoposide to embryonic hippocampal H19-7 cells causes the activation of Dyrk1A, which facilitates the interaction between Hip-1 and caspase-3 and then leads to the cell death (Kang et al., 2005). In addition to etoposide, we firstly examined whether Dyrk1A becomes activated upon the stimulation with other toxic stimuli, such as tumor necrosis factor-α (TNFα) and hydrogen peroxide (H₂O₂). After HEK293 cells were incubated with hydrogen peroxide or TNFα, cell lysates were subjected to immunoprecipitation with anti-Dyrk1A, followed by immunoblotting with anti-phosphotyrosine. As shown in Fig. 1, the auto-phosphorylation of Dyrk1A was significantly increased upon the exposure to these two stimuli, reaching the maximum value of Dyrk1A activity at 2 hr and sustained thereafter until 4 hr of post-stimuli treatment. The quantification of the phospho-Dyrk1A bands revealed a 14.92- and 1.81-fold increase of Dyrk1A activity in response to TNFα and H₂O₂, respectively (Fig. 1). These results suggest that Dyrk1A becomes activated in response to toxic stimuli.

Based on the finding that intracellular JNK signaling pathway is activated by exposure with pro-inflammatory cytokines and cellular stress stimuli, such as etoposide and H₂O₂ (Ip and Davis, 1998), we next examined whether Dyrk1A is linked to JNK signaling pathway, and if it occurs, how they interact together in a biochemical and functional mode. After HEK293 cells were mock-transfected or transfected with Xpress-Dyrk1A for 24 hr followed by treatment with etoposide, cell lysates were immunoblotted with anti-phosphoJNK. As shown in Fig. 2A, while the addition of etoposide caused JNK activation, transfection of wild type Dyrk1A led to a synergistic increase in the level of phospho-JNK in a dose-dependent manner. To further check whether Dyrk1A causes the JNK activation, in vitro kinase assay was performed. HEK293 cells were transfected a plasmid encoding either Xpress-tagged wild type Dyrk1A (Xpress-Dyrk1A) or its kinase-deficient mutant (Xpress-Dyrk1A-K188R), or HA-tagged kinase-defective SEK1, the upstream activator of JNK (HA-SEK mutant), and/or pre-incubated with JNK inhibitor SP600125 for 24 hr. The cells were then left unstimulated or treated with etoposide for 1 hr, and the total cell lysates were immunoprecipitated with anti-JNK followed by in vitro kinase assay by using bacterial recombinant c-Jun fused to GST as a substrate. As shown in Fig. 2B, the phosphorylation of GST-c-Jun was enhanced upon the stimulation with etoposide. Furthermore, the expression of Dyrk1A resulted in an enhancement of phosphorylated c-Jun levels, compared with the cells stimulated by etoposide alone (Fig. 2B). Meanwhile, marked reduction of phospho-c-Jun levels was observed under the inhibition of Dyrk1A arising from the transfection of kinase-deficient Dyrk1A mutant. These results suggest that Dyrk1A activates JNK signaling pathway. This finding was further supported by the observation that the levels of phospho-c-Jun were decreased by JNK inhibitor (SP600125) or kinase-deficient SEK mutant type (Fig. 2B), which is present upstream component of apoptotic JNK pathway.

To further characterize how Dyrk1A is linked to JNK pathway, we examined the possibility that Dyrk1A directly interacts with JNK. After HEK293 cells were transfected with Xpress-Dyrk1A or/and Flag-JNK alone or together, cell lysates were immunoprecipitated with anti-Flag followed by immunoblotting with anti-Xpress. As shown in Fig. 3, the anti-Flag-JNK immunoprecipitates did not include Xpress-tagged Dyrk1A, suggesting that Dyrk1A does not bind with JNK directly, whereas it is related to JNK signaling pathway.

Like as other MAP kinase family member, ERK and p38, JNK-mediated signaling pathway consists of three components, such as upstream MAP3K (ASK1), MAP2K (MKK4/7 or SEK1), and JNK/SAPK, (JNK1 and JNK2) (Kanamoto et al., 2000). Next we examined whether Dyrk1A interacts with ASK1. After HEK293 cells were transfected with HA-ASK1 or/and Xpress-Dyrk1A alone or together, cell lysates were immunoprecipitated with anti-HA and immunoblotted with anti-Xpress. As shown in Fig. 4A, when the cells were co-transfected with Dyrk1A and ASK1, Dyrk1A binds to ASK1. We further investigated whether the kinase activity of Dyrk1A is necessary for its binding to ASK1. As shown in Fig. 4B, HEK293 cells were transfected with a plasmid encoding either Xpress-tagged wild type or kinase-defective Dyrk1A alone or together with Flag-ASK1. Co-immunoprecipitation assay revealed that ASK1 interacts with kinase-deficient Dyrk1A as well as with wild type (Fig. 4B). These data suggest that the kinase activity of Dyrk1A appears not to be required for the interaction between Dyrk1A and ASK1.

Next we examined whether the mutual interaction between endogenous Dyrk1A and endogenous ASK1 still occurs. After HEK293 cell lysates were immunoprecipitated with mouse anti-Dyrk1A, followed by immunoblotting with anti-ASK1 and anti-Dyrk1A, Dyrk1A well interacts with ASK1 endogenously (Fig. 4C). Immunoprecipitation of the same lysates with mouse monoclonal anti-ASK1 antibody with followed by immunoblotting with anti-Dyrk1A in a reverse order produced the same binding band (Fig. 4C). As a control, the immunoprecipitation with mouse preimmune anti-IgG produced no band with anti-Dyrk1A or anti-ASK1, confirming the validity of assay (Fig. 4C). This data confirms that the mutual interaction between ASK1 and Dyrk1A is not an artifact arising from the DNA transfection, and still occurs inside cells.

To examine whether the interaction between Dyrk1A and ASK1 is being affected by toxic stimuli, including H₂O₂, TNFα, or etoposide, HEK293 cells were incubated with each stimuli for 1 hr. When cell lysates were immunoprecipitated with anti-ASK1, followed by immunoblotting with anti-Dyrk1A, the binding of Dyrk1A to ASK1 was significantly increased by the exposure with these toxic stimuli (Fig. 4D). While the addition of TNFα caused the most remarkable increased binding of Dyrk1A to ASK1, H₂O₂ and etoposide had less effect (Fig. 4D). These results suggest that toxic stimuli facilitate interaction between Dyrk1A and ASK1.

To identify the specific Dyrk1A-binding region within ASK1, several ASK1 constructs encoding its deletion fragments, such as HA-tagged wild type ASK1, ASK1-ΔN, ASK1-ΔC, and ASK1-NT, were generated (Fig. 5A). After HEK293 cells were transfected with expression plasmid encoding HA-tagged wild type ASK1, ASK1ΔN, ASK1ΔC, or ASK1-NT alone or together with Xpress-Dyrk1A, cell lysates were then subjected to immunoprecipitation with anti-HA. Immunoblotting of anti-HA-ASK1 complexes with anti-Xpress revealed that Dyrk1A still binds to ASK1-ΔN as well as the full length of ASK1 (Fig. 5B). However, Dyrk1A did not interact with ASK1-ΔC and ASK1-NT (Fig. 5B). These results suggest that the C-terminal domain of ASK1 spanning 937~1145^th amino acid is critical for the interaction with Dyrk1A.

Based on the finding that Dyrk1A is linked to JNK signaling pathway and interacts with ASK1, we next examined how Dyrk1A is linked to ASK1 and/or JNK signaling cascades. To firstly determine whether the activation of Dyrk1A is mediated by ASK1, HEK293 cells were mock-transfected or transfected with Flag-tagged kinase defective ASK1 for 24 hr. After treatment with TNFα, cell lysates were then immunoprecipitated with anti-Dyrk1A, followed by immunoblotting with anti-phosphotyrosine. As shown in Fig. 6A, the exposure of TNFα caused the activation of Dyrk1A as well as JNK, and the presence of kinase-defective ASK1 resulted in the significant reduction of JNK activity, suggesting that ASK1 is present upstream of JNK. Unlike to JNK, the levels of phospho-Dyrk1A levels were not changed remarkably by kinase-inactive ASK1, compared to the cells in the absence of ASK1 mutants (Fig. 6A). When the same experiments were performed under the condition of H₂O₂- and etoposide-induced cell death, the similar activation pattern of JNK and Dyrk1A was observed in HEK293 cells, whereas these two stimuli caused much less activation of Dyrk1A than TNFα (Fig. 6A). However, the presence of kinase-defective ASK1 did not affect the activation of Dyrk1A induced by H₂O₂ and etoposide (Fig. 6A). Taken together, these results suggest that ASK1 does not affect Dyrk1A activity.

We next examined whether Dyrk1A acts as an upstream regulator of ASK1-JNK signaling. After HEK293 cells were co-transfected with HA-ASK1 alone or together with kinase-defective Xpress-Dyrk1A-K188R, the activation pattern of ASK1 and JNK under the condition of toxin-induced cell death was determined by immunoblot assay with anti-phospho ASK1 and anti-phospho JNK. As shown Fig. 6B, when the cells were co-transfected with ASK1 plus kinase-defective Dyrk1A, the activation of ASK1 and JNK was significantly inhibited by TNFα, H₂O₂, and etoposide, compared with the cells transfected with ASK1 alone. These data indicated that Dyrk1A as a upstream regulator positively modulates ASK1-JNK signaling during toxin-induced cell death.

To examine Dyrk1A directly phosphorylates ASK1, in vitro kinase assay was performed. After transfected with a plasmid encoding Xpress-Dyrk1A for 24 h, HEK293 cells were stimulated with etoposide. Then, cell lysates were immunoprecipitated with anti-Xpress antiserum, followed by in vitro kinase assay by using bacterial recombinant ASK1^750-1013 or ASK1^1014-1374 fragment fused to GST as a substrate. As shown in Fig. 7, both of these two ASK1 fragments were well phosphorylated by Dyrk1A. These results suggest that Dyrk1A directly phosphorylates the C-terminal domain of ASK1.