If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
* This work was supported by the “Institut National de la Santé et de la Recherche Médicale” (Inserm), Biomed Grant CEE BMH4 CT97–2107, “Association pour la Recherche contre le Cancer” Grant 9872, “La Ligue contre le Cancer,” and the “Associazione Italiana per la Ricerca sul Cancro.”The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ Recipient of a postdoctoral fellowship from Biomed. ¶ Present address: EMI 00–09, IFR50, Faculté de Médecine Pasteur, 06107 Nice Cedex 02, France. ** Postdoctoral fellow of Inserm (Poste vert) and a recipient of a fellowship from the “Fondazione Italiana per la Ricerca sul Cancro.” ‡ These two authors contributed equally to this work.
Glial cell line-derived neurotrophic factor (GDNF) plays a crucial role in rescuing neural crest cells from apoptosis during their migration in the foregut. This survival factor binds to the heterodimer GDNF family receptor α1/Ret, inducing the Ret tyrosine kinase activity. ret loss-of-function mutations result in Hirschsprung's disease, a frequent developmental defect of the enteric nervous system. Although critical to enteric nervous system development, the intracellular signaling cascades activated by GDNF and their importance in neuroectodermic cell survival still remain elusive. Using the neuroectodermic SK-N-MC cell line, we found that the Ret tyrosine kinase activity is essential for GDNF to induce phosphatidylinositol 3-kinase (PI3K)/Akt and ERK pathways as well as cell rescue. We demonstrate that activation of PI3K is mandatory for GDNF-induced cell survival. In addition, evidence is provided for a critical up-regulation of the ERK pathway by PI3K at the level of Raf-1. Conversely, Akt inhibits the ERK pathway. Thus, both PI3K and Akt act in concert to finely regulate the level of ERK. We found that Akt activation is indispensable for counteracting the apoptotic signal on mitochondria, whereas ERK is partially involved in precluding procaspase-3 cleavage. Altogether, these findings underscore the importance of the Ret/PI3K/Akt pathway in GDNF-induced neuroectodermic cell survival.
enteric nervous system
extracellular signal-regulated kinase
glial cell line-derived neurotrophic factor
GDNF family receptor α1
wild type Ret
Ser765 → Pro Ret
RetWT-transfected SK-N-MC cells
whole cell lysate(s)
mitochondrial transmembrane potential
terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling
polyacrylamide gel electrophoresis
Embryonic development of the enteric nervous system (ENS)1 requires proliferation, migration, differentiation, and survival of the neural crest cells. The glial cell line-derived neurotrophic factor (GDNF), produced by the enteric mesenchyme, has emerged as a key molecule for this process (
). Compelling evidence has helped to determine that signals by GDNF-activated Ret are essential for the survival of neural crest cells that colonize the gut. Indeed, germ line loss-of-function ret mutations that impair either its expression or its intrinsic TK are involved in the Hirschsprung's disease, a frequent development defect (1 in 5000 births) characterized by the loss of the enteric ganglia (for a review, see Ref.
) results in intestinal aganglionosis. The notion that Ret is absolutely required for rescue of neural crest cells came from the recent demonstration thatret-deficient mice develop an aganglionic phenotype because of the apoptosis of crest-derived precursors as they enter the foregut (
Although critical to the normal ENS development, the intracellular signaling cascades used by the GDNF-induced Ret activity to block neuroectodermic cell apoptosis still remain elusive. In several neuronal cell lines, it has been observed that GDNF induces the activation of phospholipase C-γ, c-Src, phosphatidylinositol 3-kinase (PI3K), and extracellular signal-regulated kinase (ERK; also referred as mitogen-activated protein kinase) pathways (
). Altogether, this illustrates that the implication of ERK and/or PI3K pathways in the mediation of cell survival is critically dependent of the cytokines but also of the biological systems examined. Three recent studies have suggested the involvement of PI3K in the GDNF-induced survival of dopaminergic, sympathetic, and spinal cord motor neurons (
), although the PI3K targets and their role in the mediation of this antiapoptotic effect are still unknown. However, the fact that the development of the enteric neurons and not of these GDNF-responsive neurons is impaired inret-deficient mice supports the idea that the mechanisms underlying the survival of the crest-derived cells might be different from those of other origins.
This prompted us to use in this study the human neuroectodermic SK-N-MC cell line, because it presents the advantage of expressing constitutively GFRα1 but not Ret (
). In these conditions, SK-N-MC cells could be stably transfected with either the wild type (WT) or the kinase-dead H13 ret cDNA. This model is suitable to dissect the signaling events induced by GDNF and to address their involvement in the survival of crest-derived cells. We demonstrate that the Ret-dependent activation of PI3K/Akt is essential for SK-N-MC cell rescue. Interestingly, we found that ERK is finely regulated by PI3K and Akt and that it does contribute to the ability of GDNF to prevent procaspase-3 cleavage. Altogether, our data underscore the critical role played by the PI3K/Akt pathway in tight control of ERK signal and GDNF-induced cell rescue. Implications of these findings in the understanding of how GDNF supports ENS development are discussed.
The full-length ret cDNA coding for the short isoform (Ret9; 1072 amino acids) was cloned into the pAlter vector (Promega), and site-directed mutagenesis of Ser765 by Pro (H13 mutation, TCC → CCC) was performed as already described (
). Both the WT and the H13 mutated retcDNAs were then subcloned into the XbaI site of the pRc/CMV expression vector (Invitrogen). The plasmids for the following cDNA constructs have been described previously: Δp85PI3K (a deletion mutant of p85PI3K, which is unable to bind to p110PI3K, thus preventing specifically its activation (
) was cultured in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (Life Technologies, Inc.), 2 mm pyruvate, 2 mml-glutamine, 100 units/ml penicillin, and 100 μg/ml streptomycin (complete medium). To generate SK-N-MC cell lines expressing either the WT or the H13 mutated Ret, 150-mm diameter plates of SK-N-MC cells were stably transfected with 10 μg of corresponding ret plasmid using the polyethyleneimine (Sigma) precipitation method. 48 h later, the transfected cells were selected in complete medium containing 0.8 mg/ml G418 (Geneticin; Life Technologies). After 2 weeks, at least 100 G418-resistant clones were picked, expanded, and assayed for Ret expression by anti-Ret Western blotting (C19; Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The SK-N-MC cell lines that express either the WT or the H13 mutated Ret are referred throughout as SK-RetWT and SK-RetH13 cells, respectively.
For all experiments, cells were grown to 70% confluence, serum-starved for 6–16 h in fresh Dulbecco's modified Eagle's medium supplemented with 0.1% bovine serum albumin (A7030; Sigma), and stimulated with GDNF (+ in Figs. 1, 2B, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, Figure 10; 100 ng/ml; Genentech). When either wortmannin (PI3K inhibitor (
), 50 μm; Biomol Research Laboratories) was used, it was added to the starvation medium (for 60 and 90 min, respectively) prior to GDNF activation. As controls, cells were incubated with Me2SO (excipient; 1:5000) or left untreated (− in Figs. 1, 3, and Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, Figure 10).
Cell Lysis and Immunoprecipitation
Cells were washed with PBS and solubilized in Nonidet P-40 lysis buffer (50 mm Tris-HCl, pH 7.5, 150 mm NaCl, 1% Nonidet P-40, 1 mm Na3VO4, 10 mm β-glycerophosphate, 10 mm sodium fluoride, 2 mm EDTA, 1 μm aprotinin, 25 μm leupeptin, 1 μm pepstatin, 2 mm phenylmethylsulfonyl fluoride). 1 mg of precleared whole cell lysates (WCL) was immunoprecipitated with 2 μg of either anti-Ret, anti-phosphotyrosine (Tyr(P); Upstate Biotechnology, Inc.), anti-Shc (Upstate Biotechnology), or anti-p85PI3K(Upstate Biotechnology) antibody, for either 4 h (kinase assays) or 16 h at 4 °C, as described (
WCL (50 μg) or immunoprecipitates were separated on 9% SDS-PAGE, electroblotted onto a polyvinylidene fluoride membrane (Immobilon-P; Millipore Corp.) and analyzed by Western blotting with either anti-Ret (1:2000), biotinylated anti-Tyr(P) (Upstate Biotechnology; 1:10000), anti-phospho-Ser473 Akt (anti-P Akt; 1:1000; New England Biolabs) or the indicated monoclonal or polyclonal specific antibodies, as described (
). Where indicated, the immunoblots were stripped by 30-min incubation at 50 °C in 67 mm Tris-HCl, pH 6.7, 2% SDS, 100 mm β mercaptoethanol and reprobed.
Analysis of ERK Activation
Analysis by Western Blotting
Western blots were generated as described above but developed with an affinity-purified rabbit polyclonal antibody that specifically recognizes the dually Thr202/Tyr204-phosphorylated, active form of ERK kinases (anti-phospho-ERK; 1:2000; New England Biolabs). Equal loadings of proteins were verified by reprobing the blot with Erk1 (Santa Cruz Biotechnology) antibody.
Analysis by Elk-dependent Gene Transcription
Once phosphorylated, ERKs translocate to the nucleus, where they phosphorylate and activate the transcription factor Elk-1 (reviewed in Ref.
). Hence, to assess the activation of the ERK pathway, we also measured the Elk-1-driven transcription of luciferase reporter gene using the PathDetect Elk-1 trans-reporting system (Stratagene) following the manufacturer's instructions. Briefly, SK-RetWT cells (0.8 106/six-well dish) were transiently transfected using the polyethyleneimine reagent with 1 μg of firefly luciferase reporter pFR-LUC plasmid and 0.1 μg of pFA2-ELK-1 trans-activator plasmid. The contribution of signaling molecules to the activation of the ERK pathway was then tested by cotransfection with 0.5 μg of vectors encoding either dominant negative or positive mutants. Transfection efficiencies were normalized by measuring the activity of a cotransfected pRL-TK plasmid (0.01 μg) that expresses Renilla luciferase (dual luciferase reporter assay system; Promega). Following transfection, cells were left for 24 h in serum-free medium, and GDNF (100 ng/ml) was added for 18 h. 48 h after transfection, dual luciferase activities (firefly and Renilla) were measured in the same sample with a Berthold luminometer. The firefly luciferase activity was adjusted by dividing with the corresponding Renilla luciferase activity.
Pull-down Experiment to Detect Activated Ras
Ras activity was measured by a nonradioactive method based on the capability of the active GTP-bound Ras to bind to the Ras binding domain of Raf-1 (
). Briefly, SK-RetWT cells were stimulated with GDNF for 20 min and lysed in buffer containing 1% Triton X-100 and 1% N-octyl-d-glucopyranoside. 15 μg of recombinant GST-Raf-1 Ras binding domain previously coupled to glutathione-Sepharose (Amersham Pharmacia Biotech) were added to ∼750 μg of protein extract. Protein complexes were allowed to form for 2 h at 4 °C. Precipitates were washed three times with lysis buffer with N-octyl-d-glucopyranoside and once with PBS and analyzed by Western blotting with an anti-pan-Ras antibody (Calbiochem), as described above.
Raf-1 Kinase Assay
Cells were stimulated with GDNF for 20 min, lysed (in Nonidet P-40 buffer), and immunoprecipitated with 2 μg of anti-Raf-1 (Upstate Biotechnology) antibodies for 4 h. The Raf-1 kinase activity in immunoprecipitates was measured in vitroby a coupled kinase assay. In this assay, the ERK cascade was reconstituted in vitro; the activity of the immunoprecipitated Raf-1 was measured by its ability to stimulate a recombinant nonphosphorylated MEK1 (400 ng), the activation of which was then revealed by phosphorylation of recombinant p42ERK2 (Raf-1 kinase cascade assay; Upstate Biotechnology). After a 30-min incubation at 30 °C in the presence of Mg2+ and [γ-32P]ATP (Amersham Pharmacia Biotech), kinase reactions were stopped by the addition of reduced Laemmli buffer, resolved onto 12.5% SDS-PAGE, and visualized by autoradiography. Immunoprecipitation of equal amounts of Raf-1 was verified by immunoblotting. Control reactions, in which either MEK1 or Raf-1 immunoprecipitates were omitted, showed no ERK phosphorylation.
PI Kinase Assay
Cells were treated with GDNF or left unstimulated for 15 min prior to Nonidet P-40 cell lysis. The PI kinase assay was then carried out on anti-Tyr(P) (Upstate Biotechnology) immunoprecipitates with 10 μCi of [γ-32P]ATP and freshly prepared phosphatidylinositol (PI; Sigma) as substrate, as previously described (
). The radiolabeled phosphatidylinositol phosphate obtained was analyzed by thin layer chromatography using CHCl3/CH3OH/NH4OH/H2O (8.6:7.6:1:1.4, v/v/v/v) as migration buffer. In the case of the anti-Tyr(P) immunoprecipitates from SK-RetWT cells, PI kinase analysis was also performed in the presence of 100 nm wortmannin.
Induction and Detection of Apoptosis
Cells were treated with GDNF for 1 h before being either incubated with anisomycin (10 μg/ml; Sigma) or camptothecin (1 μm; Sigma) or exposed to ultraviolet B irradiation (200 mJ/cm2). 2–4 h after the induction of apoptosis, both adherent cells (recovered after trypsinization) and nonadherent cells (present in the culture medium) were combined and washed in PBS. Then three different parameters describing characteristic features of apoptotic cells were measured as discussed below.
Breakdown of Mitochondrial Transmembrane Potential (ΨM)
Early in apoptosis, the ΨMdecreases because of the opening of mitochondrial pores through which cytochrome C and apoptosis-inducing factor are released into the cytoplasm. To discriminate between living cells (high ΨM) and apoptotic cells (low ΨM), the above cell preparations were stained with DiOC6 (3,3′-dihexyloxadicarbocyanine; 40 ng/ml; Molecular Probes, Inc., Eugene, OR), a lipophilic cationic fluorochrome that accumulates in mitochondria with an uptake directly proportional to ΨM. Propidium iodide (5 μg/ml, Sigma) was then added to stain the cells with damaged plasma membrane (late apoptotic and necrotic cells) before the analysis by flow cytometry (FACScan Becton Dickinson apparatus). Hence, the cell population with low fluorescence intensity with both DiOC6(green fluorescence) and propidium iodide (red fluorescence) was the early apoptotic cells and scored among 10,000 gated events (using CellQuest software (Becton Dickinson)).
Caspase-3 (Cpp32) is a cysteine protease that plays a key role in the execution of the apoptotic program (
). It is synthesized as an inactive proenzyme of 32 kDa that is cleaved in cells undergoing apoptosis. Therefore, procaspase-3 cleavage was analyzed by Western blotting of WCL using an anti-procaspase-3 antibody (1:2000; Transduction Laboratories, Lexington, KY).
The degradation of genomic DNA into small oligonucleosomal fragments is a late hallmark of cells that succumb to apoptosis but not to necrosis. Since they possess a 3′-OH termini, these fragments can be labeled in cells by the terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) method. Briefly, the above cell preparations were fixed with 4% paraformaldehyde for 30 min at 4 °C and permeabilized with 70% ethanol at −20 °C for at least 30 min. After washing, the presence in apoptotic cells of ladder fragments was detected by labeling their termini with bromoconjugated dUTP and exogenous terminal deoxynucleotidyltransferase (APO-BrDu kit; Pharmingen) for 1 h at 37 °C. Subsequently, the reaction was blocked, and anti-bromo-dUTP fluorescein isothiocyanate was added for 30 min at room temperature before the analysis by flow cytometry.
It has been recently established that GDNF plays a crucial role in promoting the survival of neural crest-derived precursors during the ENS development (
). The importance of a functional Ret TK in the mediation of GDNF-induced neural crest cell survival has recently been evidenced by the apoptosis of enteric crest-derived precursors in ret-deficient mice (
). The aims of this study were thus to ascertain the requirement of Ret for the mediation of GDNF-induced survival in neuroectodermic cells and to dissect the signaling pathways involved in this process. To address this issue, we used the human neuroectodermic SK-N-MC cell line as a cellular model, since it presents the advantage of expressing constitutively GFRα1 but not Ret (
). In these conditions, SK-N-MC cells were transfected with either the WT or the kinase-dead H13 ret cDNA.
We provide here several lines of evidence that the activity of the Ret TK is essential for the GDNF-antiapoptotic signaling. Indeed, we show that GDNF did rescue the neuroectodermic SK-N-MC cells that coexpressed Ret and GFRα1 but not the cells that harbored GFRα1 alone or together with the RetH13 mutant. Instead, expression of this mutant was associated with an increased sensitivity to anisomycin-induced apoptosis, as was previously reported for a kinase-dead mutant of the insulin-like growth factor I receptor (
). Consistently, we found that GDNF stimulation led to Ret autophosphorylation, assembly of phosphotyrosine-dependent signaling complexes, and the subsequent activation of ERK and PI3K pathways over several hours in SK-RetWT cells. By contrast, binding of GDNF to GFRα1 or to RetH13-GFRα1 complex was unable to induce signal transduction. Therefore, we conclude that the absence of a functional Ret TK abrogates the ability of GDNF to trigger both signaling and survival of neuroectodermic cells. Conversely, recent studies have evidenced the ability of GDNF to promote cell survival (
). These differences might explain why germ line loss-of-function Ret mutations specifically affect the enteric ganglia and highlight the importance of studying the signaling pathways critical for GDNF-induced neural crest cell survival in a relevant neuroectodermic model.
We found that GDNF stimulated the activity of PI3K and association of autophosphorylated p170ret with p85PI3K in SK-RetWT cells. The cytoplasmic domain of Ret contains one tyrosine (Tyr981) that matches the consensus YXXM (Y981RLM) motif that can accommodate the SH2 domains of p85PI3K. However, several recent reports have identified Tyr1062 on Ret as the tyrosine critical for PI3K activation and recruitment. Interestingly, Tyr1062 is the Ret binding site for Shc that nucleates the multiprotein complex (Ret-Shc-Gab-PI3K) (
), inhibition of PI3K activity by wortmannin pretreatment of SK-N-MC cells abolished both the ability of GDNF to trigger Akt and to prevent apoptosis, underscoring the essential role played by PI3K in the mediation of the GDNF-induced cell survival. We provide here the first direct demonstration that Akt activation is mandatory for GDNF-induced cell survival as expression of a dominant negative form of Akt blocked GDNF-induced survival. This notion was strengthened by the fact that expression of the activated Gag-Akt construct was sufficient by itself to protect cells from anisomycin-induced apoptosis. Growing evidence is accumulating showing that Akt promotes cell survival by phosphorylating and inactivating components of the apoptotic machinery such as BAD (
). Of these, the possible antiapoptotic mechanisms of downstream Akt in GDNF-stimulated cells are the protection of mitochondrial integrity, the ensuing inhibition of procaspase cleavage (as we have shown here), the up-regulation of Bcl-2 (
). These data suggest that the ERK pathway may contribute, like the PI3K/Akt pathway, to the GDNF-induced neuroectodermic cell survival. However, it has been demonstrated that inhibition of PI3K but not of ERK abrogates GDNF-induced survival of dopaminergic, sensory, and spinal cord motor neurons (
). These findings clearly indicate that PI3K is necessary and sufficient to mediate the GDNF-antiapoptotic effect. However, they do not rule out the possibility that ERK may cooperate with PI3K/Akt pathway to fully protect cells against apoptosis. Indeed, it has been recently demonstrated that activation of the ERK pathway by an oncogenic form of Ret is responsible for the rescue of PC12, a cell line of neural crest origin (
). Thus, the possibility that ERK could participate to the GDNF-induced neuroectodermic cell survival still remains open.
We found that GDNF stimulation of SK-RetWT cells resulted in a sustained ERK activation with a kinetic paralleling that of Akt. Moreover, GDNF-induced survival and Akt and ERK activations were wortmannin-sensitive. We therefore sought to determine whether ERK could behave as the downstream target that propagates the PI3K/Akt-dependent survival signals. Inhibition of GDNF-induced ERK activation by PI3K inhibitors (Δp85, wortmannin, and Ly294002) supported this possibility. Along this line, PI3K inhibitors did not modify the level of tyrosine phosphorylation of Ret and Shc, nor did they impair the GDNF-induced Ras activation, indicating that GDNF did activate the classical Shc/Grb2/SOS → Ras → ERK pathway in a PI3K-independent manner. However, we observed that, in SK-RetWT cells, GDNF exerted a potent PI3K-dependent activating effect on Raf-1 activity, in accord with previous data obtained with insulin-like growth factor I, interleukin-8, or integrin in myoblasts, neutrophils, or COS-7 cells (
). Our findings demonstrate that the Shc/Grb2/SOS → Ras and the PI3K pathways are both involved in the activation of Raf-1 and thus of ERK in GDNF-stimulated neuroectodermic cells.
The molecular mechanism responsible for the PI3K activation of Raf-1 remains unknown. One possibility would be that PI3K promotes the endocytosis of activated Ret-Ras complex, to bring it in contact with Raf, as suggested by York et al. (
). Alternatively, a downstream PI3K target might relay an activating signal to Raf-1. In this regard, Akt and Rac appear as possible candidates. Rac is a downstream target of PI3K, known to modulate Raf activity via the serine/threonine kinase PAK (
). By the use of a dominant negative Rac mutant, we could demonstrate that Rac contributes to the activating PI3K effect on ERK. Due to the crucial effect played by Akt in the mediation of the GDNF rescuing effect, one might expect that Akt could also intervene in the mediation of the PI3K up-regulation of the ERK pathway. Surprisingly, expression of activated Gag-Akt construct failed to induce Elk, as reported in fibroblasts (
), and instead significantly decreased the GDNF-induced Elk activity. This raised the possibility that, although GDNF activated via PI3K both Akt and ERK activities, Akt attenuated the ERK pathway. This was supported by the fact that inhibition of the endogenous Akt, by the expression of two dominant negative Akt constructs (ST− Akt or K → A Akt), resulted in a significant increase of the GDNF-induced Elk activity. While this work was in progress, it was shown that Akt inhibits the insulin-like growth factor I-induced ERK activation in differentiated myotubes by the phosphorylation of Raf-1 on Ser259, which is in agreement with our results (
). However, an important distinction is that in this latter model PI3K also inhibits the ERK pathway. Intriguingly, in their myoblast precursors, PI3K is (as described herein) required for the efficient ERK activation (
). To our knowledge, our study provides the first evidence that Akt and PI3K could play opposite roles in the control of the ERK pathway in the same cellular context. Such a tight control of ERK led us to believe that the steady state level of ERK is critical for GDNF-induced neuroectodermic cell survival. Consistently, a significant reduction of the GDNF protective effect against anisomycin-induced procaspase-3 cleavage was observed upon PD98059 pretreatment. But at odds with these data, PD98059 did not prevent GDNF from rescuing SK-RetWT cells from apoptosis induced by different types of cell death inducers (anisomycin, camptothecin, or ultraviolet B (data not shown)).
Altogether, these findings demonstrate that the activation of the Ret → PI3K → Akt pathway was mandatory for GDNF-induced survival of neuroectodermic cells (Fig. 11). In addition, activation of the ERK pathway participated in the GDNF-protective effect by blocking procaspase-3. Interestingly, while PI3K amplified the ERK pathway by increasing the activation of Raf-1, activation of Akt in turn attenuated ERK. In this regard, it seems that both PI3K and Akt function in concert to finely regulate the level of the ERK activity. Considering the multiple steps intervening during the colonization of the foregut by crest-derived precursors, this sophisticated ERK homeostasis might have important consequences on the GDNF-induced cell fate, inasmuch as the duration and the extent of ERK activation have been shown to be crucial in the final cell decision between growth and differentiation (reviewed in Refs.
). This novel antagonistic effect of Akt and PI3K on the ERK pathway should have widespread implications, since these three signaling molecules are simultaneously activated in response to all growth and differentiation factors studied so far. Further studies of the GDNF signaling in neuroectodermic cells should provide important insights into the understanding of how subtle quantitative differences between PI3K, Akt, and ERK activities could lead to distinct cell responses.
We gratefully acknowledge Genentech for providing recombinant GDNF. We are indebted to A. Grima and R. Grattery for illustration work. We acknowledge Dr. Kasuga for providing the plasmid encoding the dominant negative mutant of p85PI3K (Δp85), Dr. Klippel for the dominant positive mutant of p110PI3K, Dr. Coffer for expression vectors encoding either the dominant negative (ST− Akt, K → A Akt) or the dominant positive (Gag-Akt) mutants of Akt, and Dr. Herrmann for the plasmid encoding the GST-Raf-1 Ras binding domain fusion protein. We are grateful to Dr. Tanti for providing reagents.