Akt (protein kinase B) negatively regulates SEK1 by means of protein phosphorylation.

The protein serine-threonine kinase Akt mediates cell survival signaling initiated by various growth-promoting factors such as insulin. Here we report that SEK1 is a target of Akt in intact cells. Insulin inhibited the anisomycin-induced stimulation of both endogenous SEK1 and its substrate c-Jun N-terminal kinase (JNK), but not that of the upstream kinase MEKK1, in 293T cells. The inhibitory action of insulin on SEK1 or JNK1 activation was prevented by the phosphatidylinositol 3-kinase inhibitor LY294002. Expression of a constitutively active form of Akt also inhibited both SEK1 and JNK1 activation, but not that of MEKK1, in transfected 293T cells. Co-immunoprecipitation analysis revealed that endogenous Akt physically interacted with endogenous SEK1 in cells and that this interaction was promoted by insulin. In vitro and in vivo (32)P labeling indicated that Akt phosphorylated SEK1 on serine 78. The SEK1 mutant SEK1(S78A) was resistant to Akt-induced inhibition. Finally, activated Akt inhibited SEK1-mediated apoptosis, and this effect of Akt was prevented by overexpression of SEK(S78A). Taken together, these results suggest that Akt suppresses stress-activated signaling by targeting SEK1.

Akt, also known as protein kinase B, is a serine-threonine kinase that plays an important role in a variety of biological processes including cell survival, cell growth, gene expression, and oncogenesis (1). Various peptide growth factors, including insulin and insulin-like growth factor I, induce the activation of Akt as a result of their triggering of the phosphatidylinositol 3-kinase (PI3K) 1 signaling pathway (2)(3)(4). The binding of such growth factors to their cell surface receptors results in the recruitment of PI3K to the plasma membrane (4) and the consequent phosphorylation of phosphatidylinositol, which generates phosphatidylinositol 3-phosphate, phosphatidylinositol 3,4-bisphosphate, and phosphatidylinositol 3,4,5-trisphosphate (5,6). The interaction of these phosphoinositides with the pleckstrin homology domain of Akt is responsible for the translocation of this enzyme from the cytoplasm to the plasma membrane (1,5,7), where it is activated as a result of phosphorylation by 3Ј-phosphoinositide-dependent kinase 1 and 2 (2, 3, 8 -10). Akt activated by the growth factor receptor signaling promotes cell survival by phosphorylating and inactivating various pro-apoptotic proteins, including Bad, Forkhead family transcription factors, caspase-9, and ASK1 (4,(11)(12)(13)(14). Identification of additional target proteins of Akt should provide greater insight into the precise role of this kinase in the regulation of cell survival and apoptosis.
c-Jun N-terminal kinase (JNK), also known as stress-activated protein kinase (SAPK), is a member of the family of mammalian mitogen-activated protein kinases and mediates intracellular signaling associated with a variety of cellular functions including cell death (15)(16)(17). The JNK/SAPK pathway is activated by proinflammatory cytokines such as tumor necrosis factor-␣, and interleukin-1, as well as by cellular stress including UV irradiation, genotoxic stress, and heat (15,16). This pathway consists of three components: a MAP3K such as MEKK1, a MAP2K such as SEK1 (also known as JNKK1 or MKK4) or MKK7, and JNK/SAPK. Activated MEKK1 (or an equivalent MAP3K) phosphorylates and activates SEK1 or MKK7, which in turn phosphorylates and activates JNK/SAPK (15,16). Activated JNK/SAPK then phosphorylates substrates that include several nuclear transcription factors such as Elk-1, c-Jun, and ATF-2 (18 -20).
To clarify the molecular mechanisms by which Akt modulates intracellular signaling cascades, we have investigated the effects of Akt on the components of the JNK/SAPK pathway. We now show that SEK1 is a target of Akt. Activated Akt physically interacts with and phosphorylates SEK1, thereby suppressing SEK1-mediated signaling, in intact cells.

EXPERIMENTAL PROCEDURES
Cell Culture and Transfection-Human embryonic kidney 293T and human cervical carcinoma HeLa cells were routinely maintained at 37°C in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. For DNA transfection, cells were plated at a density of 2 ϫ 10 6 cells/100-mm dish, grown overnight, and transfected for 48 h with appropriate expression vectors with the use of LipofectAMINE (Invitrogen) or by the calcium phosphate method (21). The transfected cells were then treated with the indicated agents and harvested for further experiments.
Immunoblot Analysis-Cell lysates were subjected to centrifugation at 12,000 ϫ g for 10 min at 4°C, and the resulting soluble fraction was subjected to SDS-PAGE. The separated proteins were transferred to a nitrocellulose membrane, which was then incubated for 1 h at room temperature with Tris-buffered saline (pH 7.4) containing 0.1% Tween 20 and 5% nonfat dried milk. The blots were probed with various antibodies including rabbit polyclonal antibodies to Akt (New England Biolabs) and a mouse monoclonal antibody to SEK1 (PharMingen). Immune complexes were detected with horseradish peroxidase-conjugated secondary antibodies to rabbit or mouse IgG (Amersham Biosciences, Inc.) and an enhanced chemiluminescence system (Pierce).
Immunocomplex Kinase Assays-Cultured cells were harvested in a lysis buffer (22), and the resulting cell lysates were subjected to centrifugation at 12,000 ϫ g for 10 min at 4°C. The soluble fraction was subjected to immunoprecipitation with appropriate antibodies. The resulting immunoprecipitates were rinsed three times with lysis buffer and then twice with 20 mM Hepes buffer (pH 7.4). Immunocomplex kinase assays were performed by incubating the immunoprecipitates for 30 min at 30°C with 2 g of substrate protein in 20 l of a kinase reaction buffer (22). Phosphorylated substrates were resolved by SDS-PAGE, and protein phosphorylation was quantified with the use of a Fuji phosphorimager. Mouse monoclonal antibodies used for immunocomplex kinase assays included those to JNK1 (PharMingen), to SEK1 (PharMingen), to MEKK1 (Santa Cruz Biotechnology), to the hemagglutinin epitope (HA) tag (Roche Molecular Biochemicals), and to the Flag epitope tag (Stratagene). Protein concentration was determined by the Bradford method (Bio-Rad).
Kinase Assay for GST-SEK1-The kinase assay for GST-SEK1 activity was performed as described (23,24). Briefly, 293T cells were transiently transfected for 48 h with pEBG-SEK1, a mammalian expression vector encoding GST-SEK1. Cell lysates were solubilized with 1% Triton X-100, and the solubilized fraction was applied to glutathione-Sepharose beads. GST fusion protein eluted from the beads was assayed for SEK1 activity by incubation for 30 min at 30°C with 2 g of GST-JNK3/SAPK␤(K55R) in 20 l of a kinase reaction buffer, as described for the immunocomplex kinase assays.
Site-directed Mutagenesis-Site-directed mutagenesis of mouse SEK1 was performed with a QuikChange kit (Stratagene). The SEK1(S78A) mutant cDNA was generated with the oligonucleotide 5Ј-GAGACTGAGAACACACgcCATTGAGTCATCAGGAAAAC-3Ј (mismatches with the wild-type SEK1 template are indicated by lowercase letters).
Metabolic Labeling with 32 P-293T cells were transferred to phosphate-free Dulbecco's modified Eagle's medium (Invitrogen) containing [ 32 P]orthophosphate (100 Ci/ml) and were incubated for 3 h. Where indicated, the cells were treated with 100 nM wortmannin for 30 min and subsequently with 100 nM insulin for 20 min. The cells were lysed and subjected to immunoprecipitation with a mouse monoclonal anti-SEK1 antibody. The immunoprecipitates were subjected to SDS-PAGE on a 12% gel, and phosphorylation of SEK1 was examined with a Fuji phosphorimager.
Apoptotic Cell Death-HeLa cells were transfected with pEGFP together with expression vectors encoding the indicated proteins. At 48 h of transfection, the cells were fixed with 0.25% glutaraldehyde and stained with 4,6-diamidino-2-phenylindole dihydrochloride (DAPI). The DAPIstained nuclei in GFP-positive cells were analyzed for apoptotic morphology by fluorescence microscopy. The percentage of apoptotic cells was calculated as the number of GFP-positive cells with apoptotic nuclei divided by the total number of GFP-positive cells.

RESULTS
Akt Suppresses the JNK/SAPK Signaling Pathway-To investigate possible effects of Akt on the JNK/SAPK signaling pathway, we first examined whether stress-induced activation of JNK in 293T cells was influenced by pretreatment of the cells with insulin, which activates the PI3K-Akt pathway (2)(3)(4). Exposure of 293T cells to anisomycin induced a marked increase in the activity of JNK1, and insulin pretreatment inhibited this effect of anisomycin (Fig. 1A). The inhibitory action of insulin on JNK1 activation was blocked by the PI3K inhibitors wortmannin (Fig. 1A) and LY294002 (data not shown). Furthermore, overexpression of a constitutively active myristoylated Akt (Akt-CA; Ref. 25) resulted in inhibition of anisomycin-induced JNK1 activation (Fig. 1B). These results suggest that the PI3K-Akt pathway mediates the inhibitory action of insulin on JNK1 activation.
SEK1 Is a Target of Akt-The JNK signaling cascade consists of JNK and upstream kinases that include MAP2K such as SEK1/JNKK/MKK4 and MAP3K such as MEKK1 (15, 16, 26 -28). To identify the possible target (or targets) of Akt in the JNK pathway, we first transfected 293T cells with expression vectors encoding Akt-CA and Flag-tagged MEKK1 (Fig. 1C). Exposure of the transfected cells to anisomycin resulted in activation of MEKK1, and the anisomycin-induced MEKK1 activation was not inhibited by overexpression of Akt-CA. In contrast, Akt-CA inhibited the activation of JNK1 induced by ⌬MEKK1, a constitutively active form of MEKK1 (Fig. 1D). Akt-CA also inhibited the ⌬MEKK1-induced stimulation of a transcription-stimulating activity of c-Jun (data not shown). Phosphorylation of c-Jun by JNK increases the transcriptionstimulating activity of c-Jun (29). Together, these results suggest that Akt inhibits the activity of a component downstream of MEKK1 in the JNK/SAPK signaling pathway. We next examined the effect of Akt on SEK1 activity in 293T cells transfected with plasmid vectors expressing HA-Akt-CA and GST-SEK1. Exposure of the cells to anisomycin resulted in an increase in SEK1 activity, and this effect was inhibited by overexpression of Akt-CA ( Fig. 2A). Overexpressed Akt-CA also inhibited SEK1 activation induced by other cellular stresses including UV light, sorbitol, and hydrogen peroxide (data not shown). In comparison, overexpressed Akt-CA did not affect anisomycin-induced MKK7 activation (Fig. 2B). The effect of insulin on the activities of endogenous SEK1 and MEKK1 in 293T cells was also investigated (Fig. 3). Insulin prevented the anisomycin-induced increase in SEK1 activity, and the inhibitory effect of insulin on SEK1 activation was blocked by LY294002 (Fig. 3A). In contrast, insulin had no effect on the activation of MEKK1 (Fig. 3B). Collectively, these results suggest that SEK1 might be a target of the Akt signaling.
Insulin Promotes the Physical Interaction of Akt with SEK1 in Intact Cells-Given that our results suggest that SEK1 is a target of Akt, we next investigated whether these two proteins interact physically in intact cells. 293T cells were cotransfected with vectors encoding HA-tagged wild-type Akt (HA-Akt) and GST-SEK1, and were then subjected to co-immunoprecipitation analysis (Fig. 4A). Immunoblot analysis using anti-GST antibody of HA immunoprecipitates from the transfected cells revealed that HA-Akt physically associated with GST-SEK1 in the cells. Furthermore, exposure of the cells to insulin promoted the physical interaction between HA-Akt and GST-SEK1. We also examined whether endogenous Akt and SEK1 could interact in intact cells (Fig. 4B). Immunoblot analysis using anti-SEK1 antibody of the Akt immunoprecipitates indicated that insulin treatment induced the physical association of two endogenous Akt and SEK1 in 293T cells. Conversely, immunoblot analysis using anti-Akt antibody of the SEK1 immunoprecipitates also showed the insulin-induced increase in the interaction between the two endogenous proteins.

FIG. 4. Insulin enhances a physical interaction between Akt and SEK1 in intact cells. A, 293T cells were transiently transfected with expression vectors encoding GST-SEK1 and HA-Akt, as indicated.
At 48 h of transfection, the cells were untreated or treated with 100 nM insulin for 20 min. Cell lysates were subjected to immunoprecipitation with anti-HA antibody, and the resulting immunoprecipitates were subjected to immunoblot analysis with anti-GST antibody. Cell lysates were also immunoblotted with antibodies to GST and to HA. B, 293T cells were incubated for 20 min in the absence or presence of 100 nM insulin, and then the cell lysates were subjected to immunoprecipitation with anti-Akt (left panel) or anti-SEK1 antibody (right panel), respectively. The resulting immunoprecipitates were subjected to immunoblot analysis with anti-SEK1 (left panel) or anti-Akt antibody (right panel). Cell lysates were also immunoblotted with antibodies to SEK1 and Akt, as indicated.
sulin-treated 293T cells catalyzed the phosphorylation of purified recombinant GST-SEK1(K129R). SEK1(K129R), a kinaseinactive SEK1 mutant, was used in the Akt phosphorylation reaction to exclude the possibility of SEK1 autophosphorylation. In comparison, the HA-Akt immunoprecipitates from insulin-treated cells did not phosphorylate GST, GST-c-Jun-(1-79), or GST-JNK3(K55R) (Fig. 5A). Furthermore, with the use of site-directed mutagenesis, we showed that replacement of Ser 78 of SEK1 with alanine prevented the in vitro phosphorylation of the recombinant protein by HA-Akt immunoprecipitates prepared from insulin-treated cells (Fig. 5B).
We next examined whether Akt phosphorylates SEK1 in vivo. Insulin increased the extent of phosphorylation of endogenous SEK1 protein in 293T cells metabolically labeled with [ 32 P]orthophosphate, and this effect of insulin was blocked by wortmannin (Fig. 5C). We next transiently transfected 293T cells with vectors encoding HA-Akt-CA and either GST-SEK1(K129R) or GST-SEK1(S78A), and then metabolically labeled the transfected cells with [ 32 P]orthophosphate. Expression of Akt-CA resulted in an increase in the extent of phosphorylation of GST-SEK1(K129R), but not in that of GST-SEK1(S78A), in the transfected cells (Fig. 5D). SEK1(S78A) Is Resistant to Akt-induced Inhibition-We next determined whether replacement of Ser 78 of SEK1 with alanine affected the inhibitory action of Akt on SEK1 activity. Exposure of 293T cells transfected with a vector encoding GST-SEK1(S78A) to anisomycin induced the stimulation of the kinase activity of the recombinant protein (Fig. 6). The anisomycin-stimulated kinase activity of GST-SEK1(S78A) was neither inhibited by expression of Akt-CA (Fig. 6A), nor by insulin (Fig.  6B). The Ser 78 residue of SEK1 thus appears to be essential for the inhibitory action of Akt.
We next investigated whether the Akt-mediated phosphorylation of SEK1 on Ser 78 could affect an interaction between SEK1 and its substrate JNK1 in intact cells. 293T cells were transfected with a combination of plasmid vectors expressing GST-SEK1, JNK1-Flag, and HA-Akt-CA, and a physical interaction between GST-SEK1 and JNK1-Flag in the transfected cells was examined (Fig. 7A). Immunoblot analysis using anti-Flag antibody of the GST pull-down precipitates showed a physical interaction between GST-SEK1 and JNK1-Flag in the transfected cells. The interaction between GST-SEK1 and

SEK1 Phosphorylation by Akt
JNK1-Flag was abolished in the cells cotransfected with HA-Akt-CA (Fig. 7A). In contrast, HA-Akt-CA failed to suppress the interaction between GST-SEK1(S78A) and JNK1-Flag in the cotransfected cells (Fig. 7B). These data suggest that the Akt-directed phosphorylation of SEK1 on Ser 78 interferes to the interaction between SEK1 and JNK, thereby inhibiting SEK1-catalyzed JNK phosphorylation.
Akt Activation Attenuates SEK1-mediated Apoptosis-We next examined whether Akt, by phosphorylating SEK1 on Ser 78 , could suppress SEK1-mediated apoptosis (Fig. 8). Transfection of HeLa cells with a vector encoding ⌬MEKK1, a constitutively active mutant of MEKK1, resulted in an increase in apoptotic cell death, and this effect was inhibited by co-expression of SEK1(K129R), a dominant negative mutant of SEK1 (Fig. 8A). These results thus indicate that activated MEKK1 induces apoptosis in HeLa cells through MEKK1-SEK1 signaling. The ⌬MEKK1-induced cell death was also inhibited by co-expression of Akt-CA. Furthermore, the protective effect of Akt-CA against ⌬MEKK1-induced apoptosis was abolished by co-expression of SEK1(S78A), which alone did not induce apoptosis. Finally, we examined the effect of insulin on UVinduced apoptosis (Fig. 8B). Exposure of HeLa cells to UV resulted in an increase in apoptosis, and this effect was blocked by expression of SEK1(K129R). These results suggest that SEK1 contributes to UV-induced apoptosis in these cells. Pretreatment of cells with insulin also inhibited UV-induced apoptosis, and this effect of insulin was antagonized by expression of either a dominant-negative mutant of Akt (Akt-DN) or SEK1(S78A). Taken together, these results suggest that the phosphorylation of SEK1 on Ser 78 is important in the mechanism by which Akt inhibits SEK1-mediated apoptosis. DISCUSSION We have shown that Akt inhibits the activation of SEK1, suggesting that SEK1 is an intracellular target of Akt. The negative regulation of SEK1 by Akt is further supported by our observation that endogenous Akt, when activated, physically interacts with endogenous SEK1 in intact cells. Furthermore, Akt phosphorylates SEK1 both in vitro and in vivo. Site-specific mutagenesis data indicated that the phosphorylation of mouse SEK1 by Akt occurs on Ser 78 , a residue that is located within a consensus phosphorylation sequence (RXRXX(S/T)) for this kinase (4,7,30). Taken together, our data indicate that activated Akt physically associates with and phosphorylates SEK1, thereby inhibiting SEK1 activation and, consequently, the JNK/SAPK signaling pathway.
In this study we demonstrate that Akt-mediated SEK1 phos- phorylation on Ser 78 results in a decrease in the binding of SEK1 to JNK1 in cells. Thus, inhibition of the interaction between SEK1 and JNK may be a possible mechanism by which Akt-mediated SEK1 phosphorylation inhibits SEK1 activity. Other possibilities could be also proposed. For instance, Ser 78 phosphorylation might directly inhibit the catalytic mechanism of SEK1, or interfere with the activation of SEK1 by a MAP3K. Further studies are needed to clarify these possibilities.
The JNK/SAPK signaling pathway mediates many cellular events including cell death (15)(16)(17). Recent studies have suggested that Akt negatively regulates the JNK signaling pathway and JNK-mediated apoptosis (31)(32)(33)(34). However, the molecular mechanism by which Akt suppresses the JNK signaling pathway has remained unclear. Activated Akt has been shown recently to phosphorylate and inhibit ASK1, a MAP3K that stimulates JNK and p38 mitogen-activated protein kinase signaling pathways (14). Our data now suggest that SEK1 is another target of Akt in the stress-activated protein kinase pathways. Thus, Akt may tightly regulate stress-activated signals through phosphorylation of both SEK1 and ASK1 in the JNK/SAPK signaling cascade. In summary, our study demonstrates that the inhibition of SEK1 by Akt may be an integral component of the mechanism by which Akt functions as a survival factor and as a negative regulator of the stress-activated signals.