Portal de Periódicos da CAPES

Rsk-mediated phosphorylation and 14-3-3ε binding of Apaf-1 suppresses cytochrome c -induced apoptosis

2012; Springer Nature; Volume: 31; Issue: 5 Linguagem: Inglês

10.1038/emboj.2011.491

1460-2075

Jiyeon Kim, Amanda B. Parrish, Manabu Kurokawa, Kenkyo Matsuura, Christopher D. Freel, Joshua L. Andersen, Carrie E. Johnson, Sally Kornbluth,

ATP Synthase and ATPases Research

Article13 January 2012free access Rsk-mediated phosphorylation and 14-3-3ε binding of Apaf-1 suppresses cytochrome c-induced apoptosis Jiyeon Kim Jiyeon Kim Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Amanda B Parrish Amanda B Parrish Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Manabu Kurokawa Manabu Kurokawa Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Kenkyo Matsuura Kenkyo Matsuura Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Christopher D Freel Christopher D Freel Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Joshua L Andersen Joshua L Andersen Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Carrie E Johnson Carrie E Johnson Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Sally Kornbluth Corresponding Author Sally Kornbluth Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Jiyeon Kim Jiyeon Kim Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Amanda B Parrish Amanda B Parrish Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Manabu Kurokawa Manabu Kurokawa Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Kenkyo Matsuura Kenkyo Matsuura Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Christopher D Freel Christopher D Freel Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Joshua L Andersen Joshua L Andersen Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Carrie E Johnson Carrie E Johnson Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Sally Kornbluth Corresponding Author Sally Kornbluth Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA Search for more papers by this author Author Information Jiyeon Kim1,‡, Amanda B Parrish1,‡, Manabu Kurokawa1, Kenkyo Matsuura1, Christopher D Freel1, Joshua L Andersen1, Carrie E Johnson1 and Sally Kornbluth 1 1Department of Pharmacology and Cancer Biology, Duke University Medical Center, Durham, NC, USA ‡These authors contributed equally to this work *Corresponding author. Department of Pharmacology and Cancer Biology, Duke University Medical Center, C370 LSRC, Research Drive, DUMC Box 3813, Durham, NC 27710, USA. Tel.: +1 919 668 4729; Fax: +1 919 681 1005; E-mail: [email protected] The EMBO Journal (2012)31:1279-1292https://doi.org/10.1038/emboj.2011.491 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Many pro-apoptotic signals trigger mitochondrial cytochrome c release, leading to caspase activation and ultimate cellular breakdown. Cell survival pathways, including the mitogen-activated protein kinase (MAPK) cascade, promote cell viability by impeding mitochondrial cytochrome c release and by inhibiting subsequent caspase activation. Here, we describe a mechanism for the inhibition of cytochrome c-induced caspase activation by MAPK signalling, identifying a novel mode of apoptotic regulation exerted through Apaf-1 phosphorylation by the 90-kDa ribosomal S6 kinase (Rsk). Recruitment of 14-3-3ε to phosphorylated Ser268 impedes the ability of cytochrome c to nucleate apoptosome formation and activate downstream caspases. High endogenous levels of Rsk in PC3 prostate cancer cells or Rsk activation in other cell types promoted 14-3-3ε binding to Apaf-1 and rendered the cells insensitive to cytochrome c, suggesting a potential role for Rsk signalling in apoptotic resistance of prostate cancers and other cancers with elevated Rsk activity. Collectively, these results identify a novel locus of apoptosomal regulation wherein MAPK signalling promotes Rsk-catalysed Apaf-1 phosphorylation and consequent binding of 14-3-3ε, resulting in decreased cellular responsiveness to cytochrome c. Introduction Apoptosis is characterized by the death and elimination of individual cells within an organism while preserving the overall structure of the surrounding tissue (Zimmermann et al, 2001; Danial and Korsmeyer, 2004); deregulation of apoptosis is associated with degenerative disorders (excessive apoptosis) and neoplasia (too little apoptosis) (Thompson, 1995). Under apoptotic conditions, cells activate a family of cysteine proteases known as caspases, which cleave a host of proteins to facilitate cellular breakdown. Apoptosis can be initiated by two main pathways, extrinsic or intrinsic (Danial and Korsmeyer, 2004); prodeath stimuli triggering the intrinsic pathway converge on the mitochondria to induce release of cytochrome c to the cytoplasm. After translocation to the cytosol, cytochrome c is incorporated into a cell death complex known as the apoptosome, which serves as a platform for activation of the initiator caspase, caspase-9 (Liu et al, 1996; Kluck et al, 1997; Ow et al, 2008). To nucleate apoptosome formation, cytosolic cytochrome c binds the adaptor protein Apaf-1, inducing a conformational change and dATP hydrolysis on Apaf-1. Subsequent nucleotide exchange allows Apaf-1 to oligomerize into a large heptameric structure, which can then recruit and bind the zymogenic form of caspase-9 (Kim et al, 2005, 2008). This interaction, mediated by binding between caspase recruitment domains (CARDs) on caspase-9 and Apaf-1, promotes dimerization of caspase-9, leading to its activation by induced proximity (Pop et al, 2006; Riedl and Salvesen, 2007). The downstream effector caspases-3 and -7 are cleaved by active caspase-9, leading to their activation and the cleavage of a host of cellular substrates (Li et al, 1997; Inoue et al, 2009), which ultimately results in the dismantling of the dying cell. Cytochrome c-induced caspase activation is regulated by post-translational modification of apoptosome-associated proteins, and these can be both inhibitory and activating (Schafer and Kornbluth, 2006; Kurokawa and Kornbluth, 2009). The most intensively studied post-translational modifications occur on caspase-9 where phosphorylation is known to reduce its activation and/or cleavage. Phosphorylation of T125 on caspase-9, which was first identified as an Erk-catalysed modification, decreases caspase-9 activation and processing (Allan et al, 2003). More recently, the cyclin-dependent kinase cdk1, DYRK1A and p38α have been shown to phosphorylate caspase-9 at this site, reducing its activity and the downstream activation of caspase-3 (Allan and Clarke, 2007; Laguna et al, 2008; Seifert et al, 2008; Seifert and Clarke, 2009). Phosphorylation of cytochrome c, another core apoptosome component, has been reported to occur at Y48 and Y97, though the consequences of these modifications remain unclear (Lee et al, 2006; Yu et al, 2008). It has been reported that active protein kinase A (PKA) can inhibit caspase-9 activation and that PKA can directly phosphorylate Apaf-1 at least in vitro (Martin et al, 2005). Nevertheless, the physiological relevance of this phosphorylation has yet to be elucidated. The mitogen-activated protein kinase (MAPK) pathway promotes cell survival in response to various stimuli, including growth factors, serum and phorbol esters (Anjum and Blenis, 2008). Activation of the Raf-MEK pathway leads to activation of Erk1/2, which, in turn, results in direct phosphorylation and activation of the effector kinase, the p90 kDa ribosomal S6 kinase (Rsk). Rsk was first identified as a kinase that phosphorylated rpS6 in unfertilized Xenopus laevis eggs, and the MAPK/Rsk pathway is highly active during Xenopus oocyte maturation and mitosis (Erikson and Maller, 1985, 1989). Two Rsk isoforms (Rsk1 and Rsk2) have been characterized in Xenopus eggs, while there are four known human isoforms (Rsk1–4). Both Erk and Rsk have direct substrates that are involved in the regulation of cell survival; Rsk1 can phosphorylate the pro-apoptotic Bcl-2 family member, Bcl-xL/Bcl-2-associated death promoter (Bad) at S112, inhibiting Bad-mediated apoptosis. Both Rsk1 and Rsk2 have been shown to suppress the pro-apoptotic activity of DAPK through phosphorylation at S289 (Shimamura et al, 2000; Anjum et al, 2005). More recently, it was demonstrated that Rsk1 and Rsk2 promote proteasomal degradation of the pro-apoptotic BH3 protein, Bim, in cooperation with Erk 1/2 (Dehan et al, 2009). In the current study, we report a novel mode of apoptosome regulation, wherein Rsk kinase can directly phosphorylate Apaf-1 at two sites. Binding of the small adaptor protein 14-3-3ε at one of the sites (phospho-S268) impeded the interaction between cytochrome c and Apaf-1, the initial step of apoptosome formation, thereby inhibiting Apaf-1 oligomerization and caspase activation. Accordingly, mutation of S268 to Ala diminished the interaction between 14-3-3ε and Apaf-1, restoring the ability of Apaf-1 to interact with cytochrome c even in the presence of high Rsk activity. Moreover, knockdown of 14-3-3ε using RNAi overrode the inhibition of cytochrome c-mediated caspase activation induced by Rsk. Collectively, these data suggest a novel mechanism of apoptosome inhibition mediated by Rsk and 14-3-3ε, reducing cytochrome c-mediated apoptosome activation and apoptosis. Results Defective apoptosome formation in Xenopus mitotic extracts is mediated by Rsk phosphorylation of Apaf-1 Given the conserved nature of the apoptotic pathway and the ease of obtaining cytosolic extract without mitochondrial disruption, the Xenopus egg extract system has served as a powerful tool for probing underlying mechanisms of apoptotic regulation (Kluck et al, 1997). By utilizing Xenopus egg extracts, our laboratory has previously shown that extracts arrested at the metaphase of second meiosis (also known as cytostatic factor-arrested extracts) are refractory to cytochrome c, compared with interphase extracts. This cytochrome c resistance was traced to high activity of MAPK signalling pathways as cytochrome c sensitivity was restored by depletion of MEK or by treatment of extracts with the MEK inhibitor U0126 (Tashker et al, 2002). We noted that the inhibition of cytochrome c-induced caspase activation was also observed in mitotic egg extracts generated by addition of non-degradable cyclin B to interphase extracts; when exogenous cytochrome c was added, caspase activation was significantly lower in mitotic extracts than in interphase extracts prepared from the same eggs (Figure 1A). Moreover, both the cdk inhibitor, roscovitine (ROS) and the MEK inhibitor, U0126 relieved the mitotic inhibition of caspase activation (Figure 1A), suggesting that a kinase downstream of Cdc2 and MEK1/2 might be involved in the observed resistance to cytochrome c-induced caspase activation in mitotic extracts. Seeking a mechanistic explanation for these observations, we analysed specific features of apoptosome formation in interphase and mitotic extracts. To examine the critical interaction between Apaf-1 and cytochrome c, we incubated agarose-bound cytochrome c (hereafter, referred to as cytochrome c beads) in either interphase or mitotic extract and found that significantly less Apaf-1 was affinity precipitated from the mitotic extract than from interphase extract (Figure 1B); interphase and mitotic extracts contain comparable amounts of Apaf-1 (Figure 1C). Figure 1.Rsk phosphorylation of Apaf-1 delays apoptosome formation in mitotic Xenopus egg extract. (A) Cytosolic interphase (S) or mitotic (M) extract was made from Xenopus eggs, and the inhibitors roscovitine (ROS) and U0126 were used to treat cytosolic extract at 1.2 mM and 200 μM, respectively. Purified exogenous cytochrome c was added to the extracts at 2 ng/μl and incubated for 60 min. The colorimetric substrate Ac-DEVD-pNA was used to quantitate caspase activity. Data shown are the mean±s.e.m. of four independent experiments. (B) Agarose-bound cytochrome c (cytochrome c beads) was incubated in S or M extract for 20 min. Bead-bound proteins were resolved by SDS–PAGE and immunoblotted with anti-Apaf-1 and anti-cytochrome c antibodies. (C) The indicated volumes of S or M Xenopus egg extract were subjected to SDS–PAGE and immunoblotted with anti-Apaf-1 or anti-actin antibodies. (D) Cytosolic S or M extract was incubated with cytochrome c (0 or 2 ng/μl) and incubated for 45 min at 25°C. Gel filtration chromatography was then performed, and one tenth of each fraction was resolved by SDS–PAGE and immunoblotted for Apaf-1. (E) Cytochrome c beads were used to affinity precipitate Apaf-1 from S or M extract pretreated with DMSO, 1.2 mM ROS or 200 μM U0126. Bead-bound proteins were resolved by SDS–PAGE, and immunoblotted with anti-Apaf-1 and anti-cytochrome c antibodies (top). Extracts were analysed by immunoblotting for Rsk2, phopho-Erk (pErk), and actin (bottom). (F) His–Apaf-1 linked to nickel agarose (His–Apaf-1) was incubated in S or M extract with γ-32P-ATP. Beads were then washed and resolved by SDS–PAGE. Incorporation of 32P was assessed by autoradiography. Note that all lanes come from the same gel. (G) His–Apaf-1 was subjected to an in-vitro kinase assay with active Erk, Rsk2 or heat-inactivated Rsk2 (Rsk-HI). Proteins were incubated with γ-32P-ATP, and the incorporation of 32P was detected by autoradiography. (H) Cytochrome c-induced caspase activation was analysed as in (A) using cytosolic S or M extract pretreated with DMSO, 200 μM U0126, 4 mM SL0101 or 100 μM BI-D1870. Data shown are the mean±s.e.m. of three independent experiments. Figure source data can be found in Supplementary data. Download figure Download PowerPoint If the interaction between cytochrome c and Apaf-1 is impaired in mitotic egg extract, we would also expect to observe reduced apoptosome formation. To evaluate this, we examined Apaf-1 oligomerization by performing gel filtration chromatography on cytosolic extract. In the absence of cytochrome c, both interphase and mitotic extracts contained Apaf-1 in lower molecular-weight fractions where it exists as a monomer (Figure 1D). However, upon addition of cytochrome c to the interphase extract, a proportion of the Apaf-1 shifted into higher molecular-weight fractions (fractions 21–24). This recruitment of Apaf-1 into the apoptosomal fractions was not detected in the mitotic extract, implying that Apaf-1 fails to efficiently oligomerize during mitosis even in the presence of cytochrome c (Figure 1D). Importantly, when 35S-labelled in-vitro translated procaspase-9 T125A (refractory to Erk phosphorylation) was added to each fraction following gel filtration, cleavage products of in-vitro translated procaspase-9 were only detectable in the interphase extract supplemented with cytochrome c, indicating the formation of functional Apaf-1 oligomers (Supplementary Figure S1). Since both ROS and U0126 restored the sensitivity of mitotic Xenopus egg extract to cytochrome c (Figure 1A), we wished to determine whether these two kinase inhibitors could restore the interaction between Apaf-1 and cytochrome c in mitotic extracts. As shown in Figure 1E, in comparison with mitotic extract alone, treatment with either ROS or U0126 enhanced the amount of Apaf-1 affinity precipitated by cytochrome c beads. It thus seemed likely that kinase activity downstream of Cdc2 and MEK1/2 was required to inhibit apoptosome formation in mitotic extract, leading us to examine whether the phosphorylation of Apaf-1 might be involved. His-tagged Apaf-1 recombinant protein incubated in egg extracts with γ-32P-ATP was well phosphorylated in mitotic, but not in interphase extract (Figure 1F). We next investigated whether Apaf-1 could be directly phosphorylated by Erk and/or Rsk since they are common kinases downstream of Cdc2 and MEK1/2 in Xenopus mitotic signalling (Yue and Ferrell, 2004). As shown in Figure 1G, Rsk, but not its activator, Erk, was able to directly phosphorylate Apaf-1, suggesting that Rsk phosphorylation of Apaf-1 might be responsible for the decreased apoptosome formation in mitotic extract. In support of this hypothesis, treating mitotic extract with an Rsk inhibitor (SL0101 or BI-D1870) restored sensitivity of the extract to cytochrome c (Figure 1H). Furthermore, immunodepleting Rsk from extracts enhanced recruitment of endogenous Apaf-1 to cytochrome c beads (Supplementary Figure S2). Since Rsk is a member of AGC kinase family and many AGC kinases have redundant functions, we wished to determine if other AGC kinase(s) could also be involved in inhibiting the Apaf-1 and cytochrome c interaction. As shown in Supplementary Figure S3, we used a panel of AGC kinase inhibitors to assess binding of Apaf-1 to cytochrome c beads. Most of the inhibitors (except for the PKC inhibitor Ro-31-8220, which has been shown to also function as an Rsk inhibitor) could not restore the binding of Apaf-1 to cytochrome c (Alessi, 1997). This finding implies that within the AGC kinase family, the Rsk kinase is most likely to be the relevant modulator of Apaf-1-cytochrome c binding. It has been reported that T125 on caspase-9 is modified by multiple kinases including Erk; phosphorylation at this site leads to inhibition of caspase-9 activity and/or processing. To examine whether Rsk can also phosphorylate caspase-9, GST-tagged recombinant wild type or T125 mutant (T125A) caspase-9 prodomain was incubated in vitro with Rsk1, Rsk2 or Erk in the presence of γ-32P-ATP. As shown in Supplementary Figure S4, Rsk was unable to phosphorylate the caspase-9 prodomain in vitro, suggesting that Rsk-mediated inhibition of apoptosome formation in Xenopus mitotic extracts likely results from Rsk phosphorylation of Apaf-1, rather than that of caspase-9. Rsk phosphorylates Apaf-1 and inhibits its binding to cytochrome c in mammalian cells Rsk kinase is activated in a subset of cancers, either through activation of the MAPK pathway or through Rsk gene amplification (Berns et al, 2004; Clark et al, 2005; Smith et al, 2005; Thakur et al, 2007; Kang et al, 2010). Thus, we next examined whether Rsk phosphorylation of Apaf-1 might contribute to the inhibition of cytochrome c-induced caspase activation in mammalian cells under circumstances where Rsk kinase activity is upregulated. After overexpressing constitutively active myristoylated-Rsk1 (myr-Rsk1) or vector alone, 293T cells were serum starved to minimize the activation of the endogenous MAPK pathway. Cytosolic lysates from these cells were then utilized for a kinase assay with recombinant His–Apaf-1 protein as a substrate. As shown in Figure 2A, recombinant Apaf-1 was well phosphorylated in the myr-Rsk1 expressing lysates. This constitutive Rsk activity also decreased the sensitivity of the lysates to cytochrome c, consistent with our data in Xenopus egg extracts (Figure 2B). Figure 2.Rsk phosphorylates Apaf-1 in mammalian cells and reduces sensitivity to cytochrome c. (A, B) 293T cells were transfected with expression plasmids encoding constitutively active myristoylated-Rsk1 (myr-Rsk) or empty vector as control. Cells were incubated in complete media for 24 h and then serum starved for additional 24 h. Cytosolic lysates were then prepared. (A) His–Apaf-1 was incubated with γ-32P-ATP in cytosolic lysates. Samples were resolved by SDS–PAGE, and incorporation of 32P was detected by autoradiography (top). Lysates were analysed by immunoblotting for total Rsk1, phopho-S6 (pS6) and actin (bottom). (B) Exogenous cytochrome c (10 ng/μl) and dATP (1 mM) were added to cytosolic lysates. Samples were incubated for 30 min, and caspase activity was assayed as previously described; the experiment was repeated at least three times and shown is a representative assay profile. (C–E) 293T cells were left untreated or treated with PMA (25 ng/ml) and/or SL0101 (100 μM) for 30 min at 37°C as indicated. Cytosolic lysates were prepared following 24 h of serum starvation. (C) His–Apaf-1 was incubated with γ-32P-ATP in cytosolic lysates, and the incorporation of 32P was assessed by autoradiography. (D) Cytochrome c-induced caspase activation was measured following incubation of cytosolic lysates with cytochrome c (10 ng/μl) and dATP (1 mM); the experiment was repeated at least three times and shown is a representative assay profile. (E) Cytochrome c beads were incubated in cytosolic lysates. Bead-bound proteins were resolved by SDS–PAGE and immunoblotted with anti-Apaf-1 and cytochrome c antibodies. (F–H) 293T cells were transfected with control siRNA or Rsk1 and 2 targeted siRNA. Forty-eight hours post transfection, cells were serum starved for 24 h prior to PMA treatment. Cytosolic lysates were then prepared. (F) His–Apaf-1 was incubated with γ-32P-ATP in lysates to examine phosphorylation of Apaf-1 by autoradiography (top). Lysates were analysed by immunoblotting for Rsk1, Rsk2 and actin (bottom). (G) Caspase activity was measured as described in (B) for control and Rsk1 and 2 knockdown (Rsk KD) cells; data shown are representative profiles (n=3). (H) Cytosolic lysates were incubated with cytochrome c beads, and the amount of bound Apaf-1 was examined by immunoblotting. Figure source data can be found in Supplementary data. Download figure Download PowerPoint Phorbol 12-myristate 13-acetate (PMA) is known to activate the MAPK pathway. When recombinant His-tagged Apaf-1 protein was incubated in lysates from PMA-treated 293T cells, Apaf-1 phosphorylation was markedly increased, as measured by γ-32P-ATP incorporation (Figure 2C). More importantly, the Rsk inhibitor SL0101 reduced the PMA-induced phosphorylation of Apaf-1, suggesting that this phosphorylation is at least in part dependent on Rsk activation (Figure 2C). Caspase activity of cytosolic lysates from these cells demonstrated that PMA treatment also reduced the responsiveness of the lysates to cytochrome c (Figure 2D). In agreement with this finding, cytochrome c beads precipitated less Apaf-1 from the PMA-treated lysates (Figure 2E). To further confirm that Rsk kinase activity was required for the PMA-induced inhibition of the apoptosome, RNAi knockdown of Rsk1 and 2, two prominent isoforms of the kinase in mammalian cells, was performed. As shown in Figure 2F, silencing Rsk1 and 2 ablated phosphorylation of recombinant His-Apaf-1 protein added to cell lysates. More importantly, Rsk knockdown restored responsiveness to cytochrome c in lysates from the PMA-treated cells (Figure 2G; compare with Figure 2D where PMA-treated cells were less sensitive to cytochrome c than controls, while knockdown of Rsk rendered them more sensitive than controls). These data were consistent with an observed increased interaction between Apaf-1 and cytochrome c under these conditions (Figure 2H). The MAPK-Rsk pathway is elevated in many cancers as a result of constitutively active Ras. Thus, we examined whether Rsk is responsible for reduced caspase activation in cells stably expressing oncogenic Ras12V. As shown in Supplementary Figure S5, RNAi knockdown of Rsk1 and 2 in Ras12V cells substantially increased responsiveness of cytosolic lysates to cytochrome c. Taken together, our results suggest that in mammalian cells, as we observed in Xenopus egg extracts, Rsk can phosphorylate Apaf-1 and inhibit its ability to bind to cytochrome c. Apaf-1 is phosphorylated in PC3 prostate cancer cells It has been reported that Rsk protein levels are elevated in some prostate cancers, where it can play a role in proliferation and prostate cancer progression. Although Apaf-1 and caspase-9 levels were the same in 293T cells and the prostate cancer cell line PC3 (ABP, JK and SK, unpublished observations), recombinant His–Apaf-1 protein was phosphorylated more robustly upon incubation in PC3 cell lysates than in lysates from 293T cells (Figure 3A). Interestingly, the phosphorylation levels of Apaf-1 in lysates from PMA-treated PC3 cells remained unchanged even following serum starvation. This result is consistent with the previous reports that in PC3 cells, Rsk is highly active regardless of Erk activation (Clark et al, 2005). Importantly, compared with lysates prepared from 293T cells, lysates from PC3 cells exhibited marked resistance to caspase activation upon addition of exogenous cytochrome c (Figure 3B). Treatment of PC3 cells with the Rsk inhibitor, SL0101, was able to enhance the sensitivity of the resulting cell lysates to cytochrome c, allowing for caspase activation (Figure 3C). A similar (but even more significant) result was observed following suppression of both Rsk1 and 2 in PC3 cells using RNAi; Apaf-1 phosphorylation was reduced following Rsk knockdown, and caspase activation in response to cytochrome c was restored (Figure 3D and E). Since the Rsk inhibitor SL0101 has been proposed as a potential chemotherapeutic (Clark et al, 2005; Smith et al, 2005), we also wished to determine whether pharmacological Rsk inhibition might induce basal cell death levels and/or boost the efficacy of chemotherapeutic agents (here, docetaxel) in PC3 cells. To this end, we treated PC3 cells with SL0101 alone or in combination with docetaxel (DTX). While there was a moderate increase in apoptotic cell death induced by SL0101 alone, we observed a significant enhancement of cell death when cells were co-treated with docetaxel and SL0101 (Figure 3F). Importantly, Apaf-1 knockdown using RNAi markedly reduced the synergy of DTX and SL0101 in inducing cell death, suggesting that the suppression of cell death by Rsk was exerted at least in part through Apaf-1 phosphorylation in PC3 cells (Figure 3F and G). Collectively, these data suggest that Rsk is responsible for both phosphorylation of Apaf-1 and resistance to cytochrome c-induced caspase activation, conferring a significant degree of protection from death stimuli in PC3 prostate cancer cells. Figure 3.Reduced caspase activation in prostate cancer cells is associated with strong phosphorylation of Apaf-1 by Rsk. (A) 293T and PC3 cells were serum starved for 24 h, and PC3 cells were either treated, or left untreated, with PMA. 293T and PC3 cytosolic lysates were prepared, and His–Apaf-1 was added with γ-32P-ATP to examine the phosphorylation status of Apaf-1. Incorporation of 32P was assessed by autoradiography. (B) Exogenous cytochrome c (10 ng/μl) and dATP (1 mM) were added to the cytosolic lysates prepared in (A) for 20 min, and caspase activity was assayed as previously described; the experiment was repeated at least three times and shown is a representative assay profile. (C) PC3 cells were incubated with DMSO or SL0101 (100 μM) for 3 h. Cytosolic lysates were prepared, and cytochrome c-induced caspase activation was analysed; the experiment was repeated at least three times and shown is a representative assay profile. (D) Rsk1 and 2 were silenced by a consecutive RNAi transfection in PC3 cells. PC3 cells were replated 24 h after the first transfection for the second transfection. Forty-eight hours after the second transfection, cells were harvested, and cytosolic lysates were prepared. His–Apaf-1 was incubated in these lysates with γ-32P-ATP to examine the phosphorylation of Apaf-1 by autoradiography (left). Immunoblotting was performed for Rsk1, Rsk2 and actin (right). (E) Cytosolic lysates from PC3 cells, transfected as in (D), were incubated with exogenous cytochrome c to examine caspase activity; the experiment was repeated at least three times and shown is a representative assay profile. (F) Apaf-1 expression in PC3 cells was silenced by siRNA transfection. Twenty-four hours post transfection, cells were treated with 50 μM SL0101 or DMSO for 24 h then treated with 100 nM docetaxel (DTX) or DMSO for an additional 24 h. Apoptotic cell death was assayed by Annexin V staining and FACS analysis. Results shown are mean±s.e.m. of four independent experiments (*P<0.05; NS, not significant). (G) Knockdown of Apaf-1 was confirmed by immunoblotting for Apaf-1 and actin as control. Figure source data can be found in Supplementary data. Download figure Download PowerPoint Rsk phosphorylates Apaf-1 at S268 Rsk recognizes a consensus motif defined as RXRXXS*/T* (phosphorylation sites are indicated with asterisks). Analyses of Rsk-consensus motifs on Apaf-1 using protein sequence annotation databases (GPS 2.1: gps.biocuckoo.org, ELM: elm.eu.org and Scansite: scansite.mit.edu) identified two target residues of particular interest, S268 and S357 (Figure 4A). To determine whether either of these sites was phosphorylated by Rsk, we mutated each of these two serines to alanine in full-length Apaf-1 and examined phosphorylation of the mutant recombinant proteins, Apaf-1 (S268A) and Apaf-1 (S357A). Lysates from 293T cells overexpressing myr-Rsk1 an