The Tumor Suppressor DAP Kinase Is a Target of RSK-Mediated Survival Signaling

Transcription

1 Current Biology, Vol. 15, , October 11, 2005, 2005 Elsevier Ltd All rights reserved. DOI /j.cub The Tumor Suppressor DAP Kinase Is a Target of RSK-Mediated Survival Signaling Rana Anjum, 1 Philippe P. Roux, 1 Bryan A. Ballif, 1,2 Steven P. Gygi, 1,2 and John Blenis 1, * 1 Department of Cell Biology 2 Taplin Biological Mass Spectrometry Facility Harvard Medical School Boston, Massachusetts Summary The viability of vertebrate cells depends on a complex signaling interplay between survival factors and celldeath effectors. Subtle changes in the equilibrium between these regulators can result in abnormal cell proliferation or cell death, leading to various pathological manifestations [1, 2]. Death-associated protein kinase (DAPK) is a multidomain calcium/calmodulin (CaM)-dependent Ser/Thr protein kinase with an important role in apoptosis regulation and tumor suppression [3 6]. The molecular signaling mechanisms regulating this kinase, however, remain unclear. Here, we show that DAPK is phosphorylated upon activation of the Ras-extracellular signal-regulated kinase (ERK) pathway [7]. This correlates with the suppression of the apoptotic activity of DAPK. We demonstrate that DAPK is a novel target of p90 ribosomal S6 kinases (RSK) 1 and 2, downstream effectors of ERK1/2 [8 11]. Using mass spectrometry, we identified Ser-289 as a novel phosphorylation site in DAPK, which is regulated by RSK. Mutation of Ser- 289 to alanine results in a DAPK mutant with enhanced apoptotic activity, whereas the phosphomimetic mutation (Ser289Glu) attenuates its apoptotic activity. Our results suggest that RSK-mediated phosphorylation of DAPK is a unique mechanism for suppressing the proapoptotic function of this death kinase in healthy cells as well as Ras/Raf-transformed cells. Results and Discussion In HEK293E cells, overexpression of DAPK results in extensive apoptotic cell death by 24 hr posttransfection (Figure 1A, compare bar 1 to bar 5). We sought to determine whether activation of cell-survival pathways in DAPK-expressing cells would provide protection against cell death [7, 12]. The Ras-MAPK (mitogen-activated protein kinase) pathway has been shown to promote cell-survival signaling and is the major mechanism for protection against cell death in certain cell types [7]. Interestingly, we found that treatment of cells with the phorbol ester PMA, which activates the Ras-MAPK pathway but not the PI3K-Akt/PKB pathway, antagonized the death-inducing activity of DAPK, leading to increased survival of DAPK-expressing cells (Figure 1A). In contrast, when cells were pretreated with the MEK1/2/5 inhibitor U0126, there was enhanced cell *Correspondence: john_blenis@hms.harvard.edu death and PMA did not attenuate the apoptotic effects of DAPK (Figure 1A). We considered the possibility that DAPK was inhibited upon phosphorylation by kinases downstream of MEK in this cascade. Because RSK is a potential cell-survival kinase in this cascade [7], we used an antibody that recognizes the consensus phosphorylation motif, RXRXXpS/T, which recognizes substrates of several basophilic kinases within the AGC family such as RSK, as well as Akt and ribosomal S6 kinase 1 (S6K1) [13 15]. Using this antibody, we found that PMA stimulated basophilic site phosphorylation of endogenous (see Figure S1A in the Supplemental Data available with this article online) and overexpressed DAPK (Figure 1B). Importantly, we also found that phosphorylation of DAPK, similar to the death-protecting effect of PMA, was completely inhibited by U0126. These results indicated that activation of the Ras-MAPK cascade induced MEK-dependent phosphorylation of DAPK, suggesting that this cascade may inhibit apoptosis by directly regulating DAPK phosphorylation. To further characterize the pathways leading to DAPK phosphorylation, we stimulated cells with different agonists and analyzed them for DAPK phosphorylation. Stimulation of cells with either PMA or EGF led to robust activation of the ERK-MAPK pathway and DAPK phosphorylation, whereas incubation of cells with insulin, a PI3-kinase agonist, or with anisomycin, a p38-mapk agonist, did not alter DAPK phosphorylation (Figure 1C). Consistent with these results, PMA-induced phosphorylation of DAPK was completely inhibited by U0126 and BIM (bisindolylmaleimide I), indicating that PMA-stimulated DAPK phosphorylation requires MEK and protein kinase C (PKC), respectively. Furthermore, treatment of cells with wortmannin or rapamycin did not alter DAPK phosphorylation, again demonstrating that the PI3-kinase and mtor pathways are not necessary for these effects (Figure S1B). These data suggests that basophilic kinases regulated by the Ras- MAPK pathway are likely involved in DAPK phosphorylation stimulated by PMA and exclude the involvement of other basophilic kinases including Akt, MSK, and S6K. To further establish the link between Ras signaling and DAPK phosphorylation, we transfected activated versions of Ras (Ras61L), B-Raf (ER- B-Raf), and MEK (MEK-DD) in cells overexpressing DAPK [16, 17]. These oncogenic signaling proteins potently stimulated growth-factor- and phorbol-ester-independent phosphorylation of DAPK (Figures 1D and 1E), suggesting that activation of the Ras-MAPK pathway is sufficient to induce DAPK phosphorylation. Finally, we also observed a reduction in DAPK phosphorylation by PMA by partially knocking down endogenous ERK1/2 (Figure S1C). The Ras-MAPK pathway activates the RSK family of proteins in response to growth factors and phorbol esters [8]. Activation of RSK coincides with oncogenic transformation [18], stimulation of G0/G1 transition [19 21], and differentiation of PC12 cells [10]. To determine whether RSK could directly phosphorylate DAPK, we

2 DAP Kinase Is a Target of RSK 1763 Figure 1. Inhibition of Cell Death and Phosphorylation of DAPK by the Activation of Ras-MAPK Pathway (A) HEK293E cells were transfected with either the vector or Flag-DAPK construct, along with EGFP plasmid (10:1 ratio), serum-starved 10 hr after transfection, and treated with DMSO alone or PMA (50 ng/ml) with or without U0126 (10 M) or UO126 alone. After 24 hr, cells were fixed and stained with activated anti-caspase-3 antibody. To determine the percent of apoptotic cells, we scored 100 GFP-positive cells in six different fields for apoptotis by counting cells positive for caspase-3 staining. All experiments were repeated at least three times, and the results are represented as means ± standard error. (B) HEK293E cells, transfected with either the vector or Flag-DAPK, were serum-starved for 18 hr and then stimulated with 50 ng/ml PMA for further 15 min, either alone or after pretreatment of cells with U0126 (10 M) for 30 min. The lysates were immunoprecipitated with anti-flag antibody, and DAPK phosphorylation on RXRXXS/T motif was analyzed after immunoblotting with Akt-pSub antibody [α-(p)dapk]. The cell lysates were also immunoblotted for total protein levels (α-flag) and ERK1/2 phosphorylation. (C) DAPK-transfected cells were serum-starved and stimulated with either 50 ng/ml PMA (15 min), 50 ng/ml EGF (15 min), 100 nm insulin (30 min), or 2 M anisomycin (30 min). DAPK was immunoprecipitated from the cell lysates and immunoblotted as in (B), for DAPK phosphorylation, total DAPK levels (α-flag), Akt phosphorylation [α-(p)akt (S473)], (P)-p38, and ERK1/2. (D) HEK293E cells were transfected with DAPK and cotransfected with vector, MEK-DD, or activated Ras (Ras61L). The vector-cotransfected cells were stimulated with PMA (50 ng/ml) for 15 min, after pretreatment of cells with 10 M U0126 (30 min), where indicated. Cell extracts were immunoprecipitated and immunoblotted as in (C). (E) NIH3T3 fibroblasts expressing a conditionally active form of B-Raf were transfected with DAPK. Cells were serum-starved for 18 hr and left untreated or treated with 1 M 4-hydroxy tamoxifen (4-HT) for the indicated time periods for the induction of B-Raf [16]. Cells were immunoprecipitated and immunoblotted as in (C). performed an in vitro kinase assay with immunoprecipitated DAPK as substrate. The kinase-inactive DAPK (K42A) was used in this experiment to avoid potential 32 P incorporation through possible DAPK autophosphorylation. DAPK was phosphorylated in vitro by activated RSK1 but not by kinase-inactive RSK1, and RSK-dependent DAPK phosphorylation was inhibited by UO126 (Figure 2A). To test whether RSK can phosphorylate DAPK in vivo, we coexpressed RSK1 with DAPK in HEK293E cells, which resulted in an 8-fold increase in phosphorylation of DAPK upon stimulation with PMA (Figure 2B). Moreover, when coexpressed in cells, kinase-inactive RSK1 did not increase DAPK phosphorylation (Figure 2B), suggesting that RSK kinase activity is necessary for phosphorylation of DAPK in cells. There was no effect on DAPK phosphorylation when cells were coexpressed with either Akt or MSK2 (Figure S2A). We also observed that both RSK1 and RSK2 were able to phosphorylate DAPK in HEK293E cells (Figure S2B). Interestingly, a constitutively active form of RSK1 (myrrsk1) [8, 22], when coexpressed with DAPK, strongly induced DAPK phosphorylation (Figure 2C). RSK has been shown to be associated with most of its physiological targets [8]. In the present study, we also observed a stimulation-independent interaction of RSK1 with DAPK (Figure 2D).

3 Current Biology 1764 Figure 2. RSK Interacts with and Phosphorylates DAPK In Vitro and In Vivo (A) HEK293E cells were transfected with either HA-RSK1 or kinase-inactive RSK1 (RSK1 / ), serum-starved for 18 hr, and treated with PMA (50 ng/ml) for 15 min, with or without pretreatment of cells with U0126 (10 M) for 30 min, and immunoprecipitated with α-ha antibodies. Kinase-inactive DAPK (K42A)-expressing cells were separately lysed, and DAPK was immunoprecipitated with α-flag antibodies. These immunoprecipitates were incubated with RSK immunoprecipitates in a kinase reaction containing [γ- 32 P] ATP for 10 min. The samples were subjected to SDS-PAGE, and the gel was autoradiographed. The lysates were immunoblotted for protein levels, RSK expression, and ERK1/2. (B) Cells were transfected with DAPK and cotransfected with either wild-type RSK1 or the kinase-inactive RSK1 (RSK1 / ). After serum starvation for 18 hr, cells were treated with PMA (50 ng/ml) for 15 min and lysed. DAPK was immunoprecipitated and immunoblotted for phosphorylation on RXRXXS/T motifs [(P)DAPK]. Lysates were immunoblotted for levels of transfected proteins and ERK1/2. (C) HEK293E cells were transfected with DAPK and cotransfected with either wild-type RSK1 or myristoylated RSK1. Cells were serumstarved, stimulated, immunoprecipitated, and immunoblotted as in (B). (D) HEK293E cells were transfected with Flag-DAPK and cotransfected with HA-RSK1. Cells were serum-starved for 18 hr followed by stimulation with PMA (50 ng/ml) for 15 min. The lysates were immunoprecipitated with α-flag antibody and immunoblotted for DAPK phosphorylation. The coimmunoprecipitated RSK was immunoblotted with α-ha antibody. The lysates were also immunoblotted for total protein levels, RSK expression, and ERK1/2. (E) HEK293 cells were transfected with either the mock or RSK1/2 sirna and cotransfected with Flag-DAPK. Cells were starved, stimulated, immunoprecipitated, and immunoblotted as in (D). To further validate the role of RSK1/2 in DAPK phosphorylation, we eliminated RSK1/2 expression by partially knocking down endogenous RSK1/2 with sirna. This resulted in a dramatic decrease in DAPK phosphorylation (Figure 2E), indicating that both RSK1 and RSK2 play a role in regulating DAPK phosphorylation. Taken together, these results demonstrate that DAPK is a specific target of RSK. To identify potential RSK phosphorylation sites in DAPK, tryptic peptides of immunopurified DAPK were subjected to LC-MS/MS/MS analysis with a Fourier transform ion cyclotron resonance (FTICR)/ion trap hybrid mass spectrometer (Figure 3A). One tryptic phosphopeptide, found uniquely from PMA-treated cells, showed phosphorylation at Ser-289, a residue lying in a minimal target site for RSK. The high-mass accuracy measurement of this phosphopeptide in the FTICR cell (0.9 ppm) and the MS 3 spectrum of the precursor ion less a neutral loss of phosphoric acid permitted an unequivocal identification of the phosphorylation site (Figure 3B). The identified phosphorylation site does not lie in a perfect RXRXXS/T consensus sequence reported to be recognized by the anti-rxrxxps/t motif antibody. Therefore, to determine whether this phosphospecific antibody was indeed recognizing DAPK phosphorylated at Ser-289, we generated a phosphorylation-deficient mutant of DAPK (Ser289Ala). Mutation of Ser-289 to alanine almost completely inhibited PMA-induced DAPK phosphorylation (Figure 3C), indicating that Ser- 289 is the predominant site phosphorylated by the activation of Ras-MAPK pathway. Thus, the anti-rxrxxps/ T motif antibody appears to recognize similar, but not identical, basophilic sequences. We also generated point mutations of several other RXRXXS/T sites in DAPK, but we did not see any effect on PMA-induced phosphorylation (data not shown). Moreover, we did not detect these sites in the MS analysis, notwithstanding our conducting of a targeted approach as done for Ser-289. To address the physiological relevance of the RSK-

4 DAP Kinase Is a Target of RSK 1765 Figure 3. Ser-289 Is the Major PMA-Regulated Phosphorylation Site in DAPK (A) HEK293E cells were transfected with Flag-DAPK and treated as shown. A part of the sample was analyzed for DAPK phosphorylation and ERK1/2 levels. The remainder of the sample was subjected to SDS-PAGE and analyzed by liquid-chromatography tandem MS (LC-MS/MS). (B) LC-MS/MS/MS identifies phosphorylation of Ser-289 as an in vivo phosphorylation site on DAPK. MS 1 analysis (left) was performed in the FTICR cell, and the base-peak chromatogram for the entire run is shown above an extracted chromatogram for the m/z range from to encompassing the calculated mass, , for the previously observed phosphopeptide, KA(pS)AVNMEK. The observed mass differed from the calculated mass by less than 1 ppm. Targeted MS 2 (center) of the precursor ion and subsequent MS 3 (right) of the precursor ion less the neutral loss of phosphoric acid produced spectra unambiguously identifying the peptide containing phosphorylated serine 289. (C) HEK293E cells were transfected with either wild-type (WT) or the S289A mutant of DAPK, serum starved, and stimulated with PMA (50 ng/ml) for 15 min. DAPK was immunoprecipitated and immunoblotted for (P)DAPK, DAPK levels, total ERK1/2 levels, and ERK1/2 phosphorylation. mediated phosphorylation of DAPK, we determined whether phosphorylation of DAPK affected its ability to induce cell death. To test this, we transfected cells with wild-type or the phosphorylation-deficient DAPK mutant (Ser289Ala) and monitored their ability to induce cell death by 24 hr. Overexpression of the wild-type DAPK resulted in about 30% cell death. Interestingly, mutation of Ser-289 to alanine induced a dramatic increase in proapoptotic activity of DAPK, resulting in enhanced cell death (Figure 4A). Cells showed about a 50% enhancement in cell death over the wild-type protein. By about 48 hr, most of the cells expressing DAPK- Ser289Ala detached from the substratum. Conversely, mutation of Ser-289 to glutamic acid reduced the proapoptotic activity of DAPK. Ser-289 lies in the calmodulin binding domain (Figure 4B). Binding of calcium-activated calmodulin to the calmodulin binding domain, during the onset of apoptosis, has been shown to alter the conformation of DAPK so that the protein is relieved from autoinhibition [3]. On the basis of this model, it is possible that phosphorylation of Ser-289 in growing cells may hinder the binding of calmodulin so that DAPK retains the autoinhibited conformation. We have not been able to measure significant changes in the catalytic activity of the wild-type or the mutant DAPK proteins when MLC (myosin light chain) was used as a substrate (data not shown). It is possible that Ser289Ala mutation may affect the catalytic activity of DAPK toward specific intracellular substrates, and this may play a role in cell death. Contrary to our findings, a recent report by Chen et al. showed that activation of ERK in DAPK expressing human erythroblastic (D2) cells has a proapoptotic rather than a survival function [23]. These contradictory findings may be attributed in part to the cell-type differences. In addition, it is well established that very potent activation of Ras can lead to cell death [24], although the molecular basis for this remains unclear. Future work will undoubtedly reveal how this molecular switch between survival and death is made by the ERK/RSK/ DAPK protein-kinase module. Our results support a model wherein activation of

5 Current Biology 1766 Figure 4. Phosphorylation of DAPK on Ser- 289 Antagonizes Its Proapoptotic Activity (A) Graph showing the relative apoptotic activities of the vector, WT DAPK, S289A, and S289E DAPK. HEK293E cells were transfected with the vector, WT DAPK, DAPK S289A, or DAPK S289E along with the EGFP plasmid. The apoptotic cells were quantified as in Figure 1A. All experiments were repeated at least three times, and the results are represented as means ± standard error. The protein levels of WT DAPK and the mutants are shown. (B) Schematic representation of the structure of DAPK, showing various domains and the PMA-regulated RXRXXS/T site, Ser-289. The homology alignment showing conservation of Ser-289 among different mammalian species is also shown. Ras-MAPK pathway negatively regulates the cell-death function of DAPK. In addition to Myt1, DAPK represents the only other kinase known to be regulated downstream of RSK, extending the kinase cascade regulated by the Ras proto-oncoprotein [25]. DAPK also represents the second described tumor suppressor, in addition to TSC2 [15], demonstrated to be antagonized by activated RSK. When Ras is improperly regulated, as in 30% of human cancers, activation of ERK and RSK, as well as subsequent inactivation of DAPK and TSC2, likely contribute to the full tumorigenic phenotype. Phosphorylation of DAPK by RSK may also be the mechanism used to restrain the apoptotic activity of DAPK in healthy cells. Future studies aimed at resolving the crystal structure of DAPK along with the regulatory calmodulin binding domain would provide insight into the role of regulatory phosphorylation in DAPK function. Supplemental Data Supplemental Data include two figures and Experimental Procedures and are available with this article online at current-biology.com/cgi/content/full/15/19/1762/dc1/. Acknowledgments We thank Ruey-Hwa Chen of National Taiwan University, Taiwan for Flag-DAPK cdna construct and Gregory Hoffman, Jessie Hanrahan, and Andrew Yoon Choo for critical reading of the manuscript. This work was supported by National Institutes of Health Grants CA46595 to J.B. and HG00041 to S.P.G., a postdoctoral fellowship from American Heart Association (AHA) to R.A., and a postdoctoral fellowship from the International Human Frontier Science Program Organization to P.P.R. Received: June 13, 2005 Revised: August 17, 2005 Accepted: August 18, 2005 Published: October 11, 2005 References 1. Thompson, C.B. (1995). Apoptosis in the pathogenesis and treatment of disease. Science 267, Danial, S.S., and Korsmeyer, S.J. (2004). Cell death: Critical control points. Cell 116, Cohen, O., Feinstein, E., and Kimchi, A. (1997). DAP-kinase is a Ca2+/calmodulin-dependent, cytoskeletal-associated protein kinase, with cell death-inducing functions that depend on its catalytic activity. EMBO J. 16, Shohat, G., Spivak-Kroizman, T., Cohen, O., Bialik, S., Shani, G., Berrisi, H., Eisenstein, M., and Kimchi, A. (2001). The proapoptotic function of death-associated protein kinase is controlled by a unique inhibitory autophosphorylation-based mechanism. J. Biol. Chem. 276, Inbal, B., Cohen, O., Polak-Charcon, S., Kopolovic, J., Vadai, E., Eisenbach, L., and Kimchi, A. (1997). DAP kinase links the control of apoptosis to metastasis. Nature 390, Raveh, T., Droguett, G., Horwitz, M.S., DePinho, R.A., and Kimchi, A. (2001). DAP kinase activates a p19arf/p53-mediated apoptotic checkpoint to suppress oncogenic transformation. Nat. Cell Biol. 3, Ballif, B.A., and Blenis, J. (2001). Molecular mechanisms mediating mammalian mitogen-activated protein kinase (MAPK) kinase (MEK)-MAPK cell survival signals. Cell Growth Differ. 12, Roux, P.P., and Blenis, J. (2004). ERK and p38 MAPK-activated protein kinases: A family of protein kinases with diverse biological functions. Microbiol. Mol. Biol. Rev. 68, Blenis, J. (1993). Signal transduction via the MAP kinases: Proceed at your own RSK. Proc. Natl. Acad. Sci. USA 90, Nguyen, T.T., Scimeca, J.C., Filloux, C., Peraldi, P., Carpentier, J.L., and Van Obberghen, E. (1993). Co-regulation of the mitogen-activated protein kinase, extracellular signal-regulated kinase 1, and the 90-kDa ribosomal S6 kinase in PC12 cells. Distinct effects of the neurotrophic factor, nerve growth factor, and the mitogenic factor, epidermal growth factor. J. Biol. Chem. 268, Shimamura, A., Ballif, B.A., Richards, S.A., and Blenis, J. (2000). Rsk1 mediates a MEK-MAP kinase cell survival signal. Curr. Biol. 10, Datta, S.R., Brunet, A., and Greenberg, M.E. (1999). Cellular survival: A play in three Akts. Genes Dev. 13, Manning, B.D., Tee, A.R., Logsdon, M.N., Blenis, J., and Cantley, L.C. (2002). Identification of the tuberous sclerosis complex-2 tumor suppressor gene product tuberin as a target of the phosphoinositide 3-kinase/akt pathway. Mol. Cell 10, Tee, A.R., Anjum, R., and Blenis, J. (2003). Inactivation of the tuberous sclerosis complex-1 and -2 gene products occurs by phosphoinositide 3-kinase/Akt-dependent and -independent phosphorylation of tuberin. J. Biol. Chem. 278, Roux, P.P., Ballif, B.A., Anjum, R., Gygi, S.P., and Blenis, J. (2004). Tumor-promoting phorbol esters and activated Ras inactivate the tuberous sclerosis tumor suppressor complex via p90 ribosomal S6 kinase. Proc. Natl. Acad. Sci. USA 101, Pritchard, C.A., Samuels, M.L., Bosch, E., and McMahon, M. (1995). Conditionally oncogenic forms of the A-Raf and B-Raf protein kinases display different biological and biochemical properties in NIH 3T3 cells. Mol. Cell. Biol. 15, Chou, M.M., and Blenis, J. (1996). The 70 kda S6 kinase com-

6 DAP Kinase Is a Target of RSK 1767 plexes with and is activated by the Rho family G proteins Cdc42 and Rac1. Cell 85, Erikson, E., Stefanovic, D., Blenis, J., Erikson, R.L., and Maller, J.L. (1987). Antibodies to Xenopus egg S6 kinase II recognize S6 kinase from progesterone- and insulin-stimulated Xenopus oocytes and from proliferating chicken embryo fibroblasts. Mol. Cell. Biol. 7, Blenis, J. (1993). Signal transduction via the MAP kinases: Proceed at your own RSK. Proc. Natl. Acad. Sci. USA 90, Chen, R.H., Chung, J., and Blenis, J. (1991). Regulation of pp90rsk phosphorylation and S6 phosphotransferase activity in Swiss 3T3 cells by growth factor-, phorbol ester-, and cyclic AMP-mediated signal transduction. Mol. Cell. Biol. 11, Chen, R.H., and Blenis, J. (1990). Identification of Xenopus S6 protein kinase homologs (pp90rsk) in somatic cells: Phosphorylation and activation during initiation of cell proliferation. Mol. Cell. Biol. 10, Roux, P.P., Richards, S.A., and Blenis, J. (2003). Phosphorylation of p90 ribosomal S6 kinase (RSK) regulates extracellular signal-regulated kinase docking and RSK activity. Mol. Cell. Biol. 23, Chen, C.H., Wang, W.J., Kuo, J.C., Tsai, H.C., Lin, J.R., Chang, Z.F., and Chen, R.H. (2005). Bidirectional signals transduced by DAPK-ERK interaction promote the apoptotic effect of DAPK. EMBO J. 24, Bar-Sagi, D.A. (2001). Ras by any other name. Mol. Cell. Biol. 21, Palmer, A., Gavin, A.C., and Nebreda, A.R. (1998). A link between MAP kinase and p34(cdc2)/cyclin B during oocyte maturation: p90(rsk) phosphorylates and inactivates the p34(cdc2) inhibitory kinase Myt1. EMBO J. 17,