Phosphorylation of EEA1 by p38 MAP kinase regulates μ opioid receptor endocytosis
2005; Springer Nature; Volume: 24; Issue: 18 Linguagem: Inglês
10.1038/sj.emboj.7600799
ISSN1460-2075
AutoresGaëtane Macé, Marta Miączyńska, Marino Zerial, Ángel R. Nebreda,
Tópico(s)Pancreatic function and diabetes
ResumoArticle1 September 2005free access Phosphorylation of EEA1 by p38 MAP kinase regulates μ opioid receptor endocytosis Gaëtane Macé Gaëtane Macé European Molecular Biology Laboratory, Heidelberg, Germany CNIO (Spanish National Cancer Center), Madrid, SpainPresent address: Institut Gustave Roussy-CNRS UPR2169, 94805 Villejuif Cedex, France Search for more papers by this author Marta Miaczynska Marta Miaczynska Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, GermanyPresent address: International Institute of Molecular and Cell Biology, Ks. Trojdena 4, 02-109 Warsaw, Poland Search for more papers by this author Marino Zerial Marino Zerial Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany Search for more papers by this author Angel R Nebreda Corresponding Author Angel R Nebreda European Molecular Biology Laboratory, Heidelberg, Germany CNIO (Spanish National Cancer Center), Madrid, Spain Search for more papers by this author Gaëtane Macé Gaëtane Macé European Molecular Biology Laboratory, Heidelberg, Germany CNIO (Spanish National Cancer Center), Madrid, SpainPresent address: Institut Gustave Roussy-CNRS UPR2169, 94805 Villejuif Cedex, France Search for more papers by this author Marta Miaczynska Marta Miaczynska Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, GermanyPresent address: International Institute of Molecular and Cell Biology, Ks. Trojdena 4, 02-109 Warsaw, Poland Search for more papers by this author Marino Zerial Marino Zerial Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany Search for more papers by this author Angel R Nebreda Corresponding Author Angel R Nebreda European Molecular Biology Laboratory, Heidelberg, Germany CNIO (Spanish National Cancer Center), Madrid, Spain Search for more papers by this author Author Information Gaëtane Macé1,2, Marta Miaczynska3, Marino Zerial3 and Angel R Nebreda 1,2 1European Molecular Biology Laboratory, Heidelberg, Germany 2CNIO (Spanish National Cancer Center), Madrid, Spain 3Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Germany *Corresponding author. CNIO (Spanish National Cancer Center), Melchor Fernández Almagro 3, 28029 Madrid, Spain. Tel.: +34 91 7328038; Fax: +34 91 7328033; E-mail: [email protected] The EMBO Journal (2005)24:3235-3246https://doi.org/10.1038/sj.emboj.7600799 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Morphine analgesic properties and side effects such as tolerance are mediated by the μ opioid receptor (MOR) whose endocytosis is considered of primary importance for opioid pharmacological effects. Here, we show that p38 mitogen-activated protein kinase (MAPK) activation is required for MOR endocytosis and sufficient to trigger its constitutive internalization in the absence of agonist. Further studies established a functional link between p38 MAPK and the small GTPase Rab5, a key regulator of endocytosis. Expression of an activated mutant of Rab5 stimulated endocytosis of MOR ligand-independently in wild-type but not in p38α−/− cells. We found that p38α can phosphorylate the Rab5 effectors EEA1 and Rabenosyn-5 on Thr-1392 and Ser-215, respectively, and these phosphorylation events regulate the recruitment of EEA1 and Rabenosyn-5 to membranes. Moreover, phosphomimetic mutation of Thr-1392 in EEA1 can bypass the requirement for p38α in MOR endocytosis. Our results highlight a novel mechanism whereby p38 MAPK regulates receptor endocytosis under physiological conditions via phosphorylation of Rab5 effectors. Introduction Despite its powerful analgesic properties, morphine is almost only administrated in cases of intensive pain because of the development of adverse side effects, including respiratory depression, tolerance and dependence. Morphine acts through a seven-transmembrane G protein-coupled receptor (GPCR), called μ opioid receptor (MOR) (Wang et al, 1994). Studies using opioid receptor knockout mice have demonstrated that MOR mediates all morphine biological effects, both beneficial and adverse (Matthes et al, 1996; Kieffer, 1999). Agonist stimulation of MOR induces the activation of Gi/o proteins, including signals responsible for the analgesic effect, and can also trigger receptor internalization (Sternini et al, 1996; Law et al, 2000), a process requiring G-protein-receptor kinase(s) (GRK)-mediated phosphorylation of the receptor and recruitment of β-arrestin2 protein. The receptor is endocytosed via clathrin-coated vesicles and transported to early endosomes (Keith et al, 1998; Ferguson, 2001). Two mechanisms have been envisaged to explain a possible role of MOR endocytosis in tolerance to morphine, defined as the decrease in drug efficacy with time (reviewed by Kieffer and Evans, 2002). Tolerance could be caused by a reduction in the number of MORs at the plasma membrane. Alternatively, MOR endocytosis could be a protective process against the development of tolerance, which would be consistent with the failure of morphine to trigger MOR endocytosis despite potently inducing tolerance in vivo (Finn and Whistler, 2001; He et al, 2002). The regulation of MOR endocytosis is, therefore, of primary importance for the understanding of opioid pharmacological effects. Endocytic trafficking of plasma membrane receptors is regulated by monomeric small GTPases of the Rab family (Zerial and McBride, 2001). Within this family, Rab5 is involved in endocytosis of several receptors, including GPCRs (reviewed by Seachrist and Ferguson, 2003). Rab5 coordinates multiple processes, such as the formation of clathrin-coated vesicles, their fusion with early endosomes and homotypic early endosome fusion, as well as motility of endosomes along microtubules (Zerial and McBride, 2001). As for other Rab GTPases, the activity of Rab5 is regulated via two overlapping cycles. First, Rab5 shuttles between the cytosol and the membrane, chaperoned by Rab GDP dissociation inhibitor (GDI) (Sasaki et al, 1990; Pfeffer et al, 1995). Specific delivery to the membrane is catalyzed by GDI displacement factors that promote the dissociation of the Rab–GDI complex (Sivars et al, 2003). Second, upon membrane delivery, Rab5 oscillates between the GDP-bound (inactive) and GTP-bound (active) forms (Rybin et al, 1996) regulated by GDP/GTP exchange factors (GEF) (Horiuchi et al, 1997) and GTPase-activating proteins (GAP) (Lanzetti et al, 2000). The fraction of GTP-bound Rab5 on the membrane is thus rate limiting for endosome dynamics (Rybin et al, 1996). This activity is exerted through the interaction with effector proteins, including Rabaptin-5, phosphatidylinositol 3-kinases (PI3-K; VPS34/p150 and p85α/p110β), EEA1 and Rabenosyn-5 (Zerial and McBride, 2001). EEA1, one of the best-characterized Rab5 effectors, is involved in tethering/docking and fusion of early endosomes (Mills et al, 1998; Simonsen et al, 1998; Christoforidis et al, 1999b). EEA1 displays a complex modular architecture consisting of an N-terminal C2H2 zinc-finger, which includes a Rab5-binding site, four heptad repeats and a C-terminal region containing a calmodulin-binding motif (IQ), a second Rab5-binding site and the C-terminal FYVE domain that specifically binds to phosphatidylinositol 3-phosphate (PI(3)P) (Mu et al, 1995; Patki et al, 1997; Simonsen et al, 1998). Rab5 plays a dual role in the recruitment of EEA1, by direct interaction with the N- and C-terminal binding sites as well as by association with VPS34, the PI3-K that generates PI(3)P (Christoforidis et al, 1999b). The mechanism of membrane recruitment of EEA1 is shared by Rabenosyn-5, which also contains a Rab5-binding domain and PI(3)P-binding FYVE finger (Nielsen et al, 2000). EEA1 and Rabenosyn-5 are both necessary and play complementary roles in Rab5-dependent endosome tethering and fusion (Nielsen et al, 2000). Other recently established regulators of endocytosis are the p38 mitogen-activated protein kinases (MAPK), a family of Ser-Thr kinases that can regulate numerous cellular responses (reviewed by Nebreda and Porras, 2000). The first evidence for a regulatory role of this signalling pathway in endocytosis was provided by the finding that, in the stress-induced response, p38α can phosphorylate Rab GDI, enhancing its activity in retrieving Rab5 from the membrane, with the consequent loss of EEA1 from early endosomes (Cavalli et al, 2001). Endocytosis of AMPA receptors as well as phagolysosome biogenesis has been shown to be modulated by p38α (Fratti et al, 2003; Huang et al, 2004). In this study, we discovered a new mechanistic role for p38α MAPK in endocytosis, centered on the regulation of membrane recruitment of Rab5 effectors. Results MOR endocytosis requires p38α MAPK As long-term exposure to morphine has been reported to correlate with p38 MAPK activation (Ma et al, 2001; Singhal et al, 2002), we investigated the function of this signalling pathway in MOR endocytosis. We established an HEK293 cell line expressing GFP-tagged MOR. These cells were stimulated by Damgo ([Tyr-DAla-Gly-MePhe-Gly-ol]enkephalin), a specific agonist of MOR, and activation of p38 MAPKs was examined by Western blotting using phospho-specific antibodies. In agreement with the absence of detectable endogenous MOR in these cells (data not shown), we found that Damgo was only able to induce p38 MAPK phosphorylation in HEK293 cells expressing GFP-MOR (Figure 1A). In contrast, p38 phosphorylation induced by UV treatment was independent of GFP-MOR expression. By immunoprecipitation followed by in vitro kinase assay, we confirmed that Damgo induced p38α MAPK activation (Figure 1B). The activation of p38α by Damgo was rapid and transient, with the kinase activity peaking after 5 min of stimulation. Quantitative assessment indicated that p38α activity levels upon Damgo stimulation were 8–16% of those obtained in UV-treated cells. Figure 1.p38 MAPK activation is necessary and sufficient for MOR endocytosis. (A) Western blots of lysates from HEK293 cells transfected with GFP-MOR or the vector, either unstimulated (C) or after stimulation with Damgo for 5 min (D) or UV. (B) Kinase activity in p38α immunoprecipitates from HEK293 cells transfected or not with GFP-MOR (+ and −, respectively) and either left untreated (Control) or stimulated with UV or Damgo. (C) Plasma membrane-associated GFP-MOR in HEK293 cells, either expressing MKK6DD or incubated with Damgo for 30 min in the presence or absence of SB203580, as indicated. GFP-MOR was detected by avidin pulldown of biotinylated membrane proteins followed by Western blot using a GFP antibody. The histogram represents the quantification of the GFP-MOR band (pulldown versus total cell lysate) using the Odyssey Imaging system. The experiment was repeated three times. (D) Damgo binding assays in wt or p38α−/− MEFs and p38α−/− MEFs expressing exogenous p38α, either unstimulated, after 30 min Damgo stimulation or transfected with MKK6DD. Results are expressed as the percentage of binding in wt MEFs and represent the mean±s.e.m. of six experiments. (E) Western blots of lysates from HEK293 cells expressing GFP-MOR alone (vector) or together with MKK6DD before (−) and after (+) 5 min of Damgo stimulation. The arrow indicates the overexpressed MKK6DD. (F) Western blots of lysates from wt and p38α−/− immortalized MEFs. (G) Western blots of lysates from wt and p38α−/− MEFs either transfected or not with MKK6DD (indicated by an arrow). Download figure Download PowerPoint MOR endocytosis was investigated in GFP-MOR-expressing HEK293 cells by biotinylation of intact cells after Damgo stimulation, followed by the purification of biotinylated membrane proteins using avidin beads. In this assay, we detected a decrease in the amount of GFP-MOR associated with plasma membrane after 30 min of Damgo stimulation (Figure 1C), which was not due to degradation of the MOR receptor (Supplementary Figure S1). MOR downregulation was confirmed by radioligand binding assays using 3H-labelled Damgo, which showed an ∼40% decrease in Damgo-specific plasma membrane binding sites after 30 min of stimulation (see below). Given the Damgo-induced p38 MAPK activation, we investigated the requirement for p38 MAPK activity in MOR internalization. First, the p38α and p38β inhibitor SB203580 strongly impaired the internalization of GFP-MOR in Damgo-stimulated cells (Figure 1C). Second, we used mouse embryonic fibroblasts (MEFs) from p38α−/− mice, which lack the most abundant p38 MAPK family member (Adams et al, 2000). In agreement with previous studies (Nitsche and Pintar, 2003), endogenous MOR was expressed in wild-type (wt) MEFs and only slightly reduced in p38α−/− MEFs (Figure 1F), as detected by Western blotting and radioligand binding assays (Figure 1D, compare lanes 1 and 4). MEFs (wt and p38α−/−) were incubated with Damgo and plasma membrane expression of MOR was quantified in radioligand binding assays. Damgo stimulation decreased by about 70% the number of surface Damgo-binding sites in wt MEFs. In contrast, Damgo failed to induce any detectable endocytosis of endogenous MOR in p38α−/− MEFs, whereas expression of exogenous p38α in these cells restored endocytosis to about the same level as in wt cells (Figure 1D). Using primary MEFs derived from three different sets of littermate embryos, we confirmed that Damgo induced MOR internalization in wt but not in p38α−/− cells (data not shown). Altogether, the results demonstrate that Damgo-induced MOR endocytosis requires p38α MAPK. Specific activation of p38 MAPKs induces constitutive MOR endocytosis Having shown the requirement for p38 MAPK, we next investigated whether its activation was sufficient to induce MOR endocytosis. HEK293 cells expressing GFP-MOR were transfected to coexpress MKK6DD, a constitutively active form of the p38 MAPK activator MKK6 (Alonso et al, 2000). MKK6DD expression resulted in higher levels of p38 MAPK phosphorylation both in untreated and Damgo-stimulated cells, without affecting MOR expression (Figure 1E). Interestingly, MKK6DD-expressing cells showed reduced levels of GFP-MOR plasma membrane expression in the absence of Damgo stimulation (Figure 1C). Moreover, stimulation with Damgo for 30 min resulted in the almost complete loss of GFP-MOR plasma membrane in MKK6DD-expressing cells (Figure 1C). By radioligand binding assays, we confirmed that expression of MKK6DD reduced by 40–70% the number of Damgo-binding sites on the plasma membrane in wt MEFs (Figure 1D) and increased Damgo-induced MOR endocytosis from 50 to 90% in GFP-MOR-expressing HEK293 cells (Figure 2A, lanes 2 and 5). The level of p38 MAPK phosphorylation in HEK293 cells transfected with MKK6DD varied between experiments, but there was good correlation between MKK6DD-induced p38 MAPK phosphorylation and MOR endocytosis (Figure 1C and E versus Figure 2A and B). Figure 2.Specific activation of p38 MAPK stimulates morphine-induced MOR endocytosis. (A) Damgo binding assays in HEK293 cells expressing GFP-MOR alone (Vector) or together with MKK6DD, either unstimulated (Control) or after 30 min Damgo or morphine stimulation. The results are expressed as the percentage of binding in the untreated cells expressing GFP-MOR. The experiment was repeated three times. (B) Western blots of lysates from HEK293 cells expressing GFP-MOR alone or together with MKK6DD, nonstimulated (C) and stimulated with Damgo for 5 min (D) or with morphine for the indicated times. (C) Plasma membrane-associated levels of MOR in SH-SY5Y cells treated with Damgo or morphine, either alone or together with H2O2 in the presence or absence of SB203580, as indicated. MOR was visualized as in Figure 1C, but using a MOR antibody for the Western blot. The experiment was repeated twice. Download figure Download PowerPoint To confirm the specificity of the above effect, we prevented the stimulatory activity of p38 MAPK using chemical inhibitors. Treatment of MKK6DD-expressing cells with SB203580 (Figure 1C) or PD169316 (data not shown) significantly increased the amount of GFP-MOR at the plasma membrane. Furthermore, expression of MKK6DD (Figure 1G) was unable to induce MOR endocytosis in p38α−/− MEFs (Figure 1D, lane 6). These observations indicate that the constitutive endocytosis of MOR induced by MKK6DD is dependent on p38α activity. Previous studies have reported that morphine is not able to trigger MOR endocytosis (Whistler and von Zastrow, 1998) and we obtained the same result in HEK293 cells expressing GFP-MOR (Figure 2A). We found that, in contrast to Damgo, short-term exposure to morphine was not able to induce p38 MAPK activation (Figure 2B). We next investigated whether constitutive activation of p38 MAPKs could overcome the inability of morphine to trigger MOR endocytosis. Interestingly, morphine-induced MOR endocytosis was stimulated by the activation of p38 MAPKs upon expression of MKK6DD. Quantification by radioligand binding assays showed that upon morphine stimulation up to 75% of GFP-MOR was endocytosed after 30 min in cells expressing MKK6DD (Figure 2A, lane 6). We next extended our results to the neuroblastoma cell line SH-SY5Y, which can differentiate into cells with a neuronal phenotype and express endogenous MOR. In these neuronal cells, p38 MAPK activation was required for Damgo-induced MOR endocytosis (Figure 2C). Moreover, treatment of SH-SY5Y cells with 50 μM H2O2, a stress that moderately activates p38 MAPK (Supplementary Figure S2), increased the rate of Damgo-induced MOR endocytosis and enabled also MOR endocytosis upon morphine treatment (Figure 2C). Taken together, our results show that constitutive activation of p38 MAPKs is sufficient to trigger MOR endocytosis, even in the absence of agonist stimulation or in the presence of morphine, which normally fails to induce MOR internalization. Rab5 is involved in Damgo induced-MOR endocytosis The small GTPase Rab5 is a key regulator of transport in the early endocytic pathway (Zerial and McBride, 2001) and previous work has shown that a dominant-negative Rab5 mutant (Rab5-N133I) restores membrane expression of a MOR mutant, which is normally not detected at the plasma membrane (Li et al, 2001). To determine whether p38 MAPK requires Rab5 activity in the regulation of MOR endocytosis, we transiently transfected HEK293 cells expressing GFP-MOR, either alone or together with MKK6DD, with the dominant-negative mutant Rab5S34N or with RN-tre, a Rab5 GAP (Lanzetti et al, 2000). We found that either Rab5S34N or RN-tre markedly inhibited the constitutive (triggered by MKK6DD) as well as the Damgo-induced GFP-MOR internalization (Figure 3A, compare lane 2 with lanes 6 and 10, and lanes 3 and 4 with lanes 7 and 8, and 11 and 12), without affecting p38 MAPK activation (Figure 3B). These results implicated Rab5 in the stimulation of MOR internalization by p38α MAPK. Figure 3.Rab5 is required for MOR endocytosis. (A) Damgo binding assays in HEK293 cells expressing GFP-MOR alone or together with MKK6DD and transfected with Rab5S34N or RN-tre before and after Damgo stimulation. The values are normalized to those of untreated GFP-MOR cells and are expressed as the mean±s.e.m. of three experiments. (B) Western blots of lysates from HEK293 cells expressing GFP-MOR alone (Vector) or together with the indicated proteins, before (−) and after (+) Damgo stimulation for 5 min. (C) Damgo binding assays of wt and p38α−/− MEFs transfected with Rab5Q79L and myc-tagged p38α or the vectors alone and then incubated with SB203580, as indicated. The results are expressed as the percentage of binding in the corresponding control-vector cells. The experiment was repeated three times. (D) Western blots of lysates from wt and p38α−/− MEFs transfected with Rab5Q79L, myc-tagged p38α or the vectors alone, before and after Damgo stimulation for 5 min, as indicated. Download figure Download PowerPoint Can the constitutively active mutant Rab5Q79L bypass the requirement for p38 MAPK in Damgo-induced MOR internalization? The effect of Rab5Q79L on MOR endocytosis was compared in wt and p38α−/− MEFs (Figure 3C). Expression of Rab5Q79L significantly decreased the level of plasma membrane-associated MOR only in wt but not in p38α−/− cells (Figure 3C, compare lanes 2 and 5). The stimulation of MOR internalization observed in wt cells was abrogated upon treatment with SB203580 (Figure 3C, lane 3). Moreover, transfection of p38α into p38α−/− cells was able to rescue the ability of Rab5Q79L to induce constitutive (Damgo-independent) MOR endocytosis (Figure 3C, lane 7). Importantly, we did not observe any changes in the levels of p38 MAPK phosphorylation upon Rab5Q79L expression, either alone or in combination with Damgo (Figure 3D). Thus, the failure of Rab5Q79L to trigger MOR endocytosis in p38α−/− cells indicates an essential requirement for p38α in Rab5Q79L activity. Regulation of EEA1 activity by p38α MAPK phosphorylation Stress-induced activation of p38α can result in the phosphorylation of Rab GDI, enhancing its activity in retrieving Rab5 from the membrane (Cavalli et al, 2001). However, the phosphorylation of Rab GDI by p38 MAPK and the consequent acceleration in the cycle of Rab5 membrane association/dissociation cannot, at least single-handedly, explain the effects of p38 MAPK on MOR endocytosis. The observation that increased levels of Rab5:GTP cannot promote endocytosis in p38α−/− cells argues that downstream effectors or regulators of Rab5 may be phosphorylated by p38 MAPK. To test this hypothesis, we screened for substrates of p38α among the Rab5 effectors and regulators, including the Rabaptin-5–Rabex-5 complex, EEA1, Rabenosyn-5 and the GAP RN-tre (TrH domain) (Zerial and McBride, 2001). Indeed, we found that the C-terminus of EEA1 (residues 1257–1411) and the full-length Rabenosyn-5 are both phosphorylated by p38α MAPK (Figure 4A and E). To identify the p38α phosphorylation site on EEA1, we first mutated several Ser and Thr residues located in the C-terminus to Ala. Mutation of Thr-1392, which is a canonical Ser/Thr-Pro MAPK phosphorylation site, to Ala abolished p38α-mediated phosphorylation of the EEA1 C-terminus (Figure 4A), whereas analogous mutations of Ser-1394 and Ser-1395 did not affect phosphorylation by p38α (data not shown). Using a phospho-threonine antibody that recognizes EEA1 phosphorylated on Thr-1392 (Figure 4A, right panel), we confirmed the phosphorylation of endogenous EEA1 upon activation of p38 MAPKs by MKK6DD both in HEK293 cells and in SH-SY5Y neuronal cells (Figure 4B). Importantly, Damgo stimulation also induced the phosphorylation of endogenous EEA1, which was mediated by p38 MAPK, as indicated by the inhibitory effect of SB203580 in HEK293 cells expressing GFP-MOR and in the neuronal cell line SH-SY5Y (Figure 4C). However, EEA1 phosphorylation was not observed upon morphine stimulation (Figure 4C, left panel), providing a good correlation between p38 MAPK activation, EEA1 phosphorylation and MOR endocytosis. Interestingly, the C-terminus of EEA1 could efficiently be phosphorylated in vitro by 10-fold lower amounts of p38α than required for phosphorylating Rab GDI (Figure 4D). Figure 4.p38α MAPK phosphorylates EEA1 and Rabenosyn-5. (A) In vitro kinase assay using MKK6DD-activated p38α (+) or MKK6DD alone (−) and the GST-fused C-terminus of EEA1 (amino acids 1257–1411) wt or with the mutation T1392A (10 μg each) as substrates (left panels). Western blot using a phospho-Thr antibody of GST-EEA1 (1257–1411) wt or T1392A (1 μg) after phosphorylation with MKK6DD-activated p38α or MKK6DD alone (right panels). (B) Endogenous EEA1 was immunoprecipitated from MKK6DD-expressing HEK293 cells (left panel) or SH-SY5Y cells (right panel) and blotted with antibodies against EEA1 or phospho-Thr. (C) HEK293 or SH-SY5Y cells were treated as indicated and endogenous EEA1 was immunoprecipitated and blotted with antibodies against EEA1 or phospho-Thr. (D) In vitro kinase assay using the indicated amounts of MKK6DD-activated GST-p38α and either 10 μg GST-EEA1 (1257–1411) or 15 μg His-GDI. (E) In vitro kinase assay using MKK6DD-activated p38α and the GST-Rabenosyn-5 wt or with the mutation S215A (10 μg each). (F) Endogenous Rabenosyn-5 was immunoprecipitated from SH-SY5Y cells expressing MKK6DD and blotted with antibodies against Rabenosyn-5 or phospho-Ser. Download figure Download PowerPoint The C-terminal fragment of EEA1 contains the PI(3)P-binding FYVE domain (Stenmark et al, 2002). Based on the EEA1 crystal structure (Dumas et al, 2001), the p38 MAPK phosphorylation site lies within the FYVE domain dimer interface. The N-terminus of Rabenosyn-5 also includes an FYVE domain containing two p38 MAPK canonical phosphorylation sites (Nielsen et al, 2000; Stenmark et al, 2002). Consistently, mutation of Ser-215 to Ala abolished p38 MAPK-mediated phosphorylation of Rabenosyn-5 (Figure 4E), indicating that this is the major p38 MAPK phosphorylation site. Moreover, using a phospho-serine antibody, we confirmed the phosphorylation of endogenous Rabenosyn-5 upon activation of p38 MAPKs by MKK6DD in the neuronal cell line SH-SY5Y (Figure 4F). Since the FYVE domain of EEA1 is required for homodimerization, endosomal membrane localization and, consequently, activity in endosome fusion (Stenmark et al, 1996; Callaghan et al, 1999), we first investigated whether Thr-1392 phosphorylation by p38α modulated the membrane recruitment of EEA1. The C-terminus of EEA1 either wt or containing the nonphosphorylatable mutation T1392A were in vitro translated and their ability to bind to isolated endosomal membranes was analyzed. Indeed, the EEA1 mutant T1392A was less efficiently recruited to endosomal membranes compared to wt EEA1 (Figure 5A). Quantitative analysis (Supplementary Figure S3) indicated that the recruitment of EEA1-T1392A was about 50% of the wt EEA1 value, both in the presence and absence of exogenous Rab5:GDI (the latter reflecting binding by the endogenous Rab5). The fact that EEA1-T1392A is still capable of membrane binding, albeit at reduced levels, is in agreement with multiple molecular interactions targeting EEA1 to endosomes (Simonsen et al, 1998; McBride et al, 1999; Lawe et al, 2003). Figure 5.Phosphorylation of Thr-1392 regulates EEA1 subcellular localization. (A) In vitro recruitment of 35S-labelled EEA1 (1257–1411) wt and T1392A to early endosomes in the absence or in the presence of Rab5:GDI complex or GDI alone. Proteins bound to endosomal membranes were detected by autoradiography and Western blot. Syntaxin-13 was used as a membrane marker. (B) MEFs (wt or p38α−/−) were transiently transfected with full-length EEA1 either wt, T1392A or T1392D, fixed and stained with antibodies to EEA1. Scale bar represents 20 μm. (C) Western blots of membrane preparations from wt or p38α−/− MEFs transiently transfected with full-length EEA1 either wt, T1392A or T1392D. The histogram represents the mean±s.e.m. of two experiments. Download figure Download PowerPoint In order to confirm in intact cells the results obtained using purified endosomal membranes, we transiently expressed full-length wt EEA1 and the corresponding mutants T1392A or T1392D (to mimic p38α phosphorylation) in wt and p38α−/− MEFs and determined their localization by both immunofluorescence microscopy analysis (Figure 5B) and subcellular fractionation (Figure 5C). In agreement with the in vitro experiments, EEA1-T1392A expressed in either wt or p38α−/− MEFs was also less efficiently recruited to endosomal membranes than wt EEA1. Interestingly, the T1392D mutation rescued to a large extent the membrane localization of EEA1. These results suggest that phosphorylation of Thr-1392 is required for the efficient recruitment of EEA1 to endosomal membranes in vivo. Surprisingly, in p38α−/− cells expressing wt EEA1, we observed small clusters of endosomes corresponding to tightly packed and tethered individual endosomes (Figure 5B, panel 4, inset). We estimated a frequency of about 14 clusters per p38α−/− cell versus not more than 1 per wt cell. Such a phenotype is reminiscent of the effects induced by the lack of EEA1 function in the conversion from membrane tethering to fusion (Christoforidis et al, 1999a; Lawe et al, 2002). The T1392D mutation also partially restored EEA1 fusion activity, as assessed by a decreased number of clustered endosomes (Figure 5B, panel 6; about 6 clusters/cell). Altogether, our results suggest that p38α MAPK-mediated phosphorylation enhances both EEA1 membrane localization and fusion activity. We then investigated whether the Damgo-induced EEA1 phosphorylation correlated with membrane recruitment. Interestingly, Damgo increased the membrane recruitment of EEA1 in wt but not p38α−/− cells, indicating that Damgo-induced EEA1 membrane localization is strictly p38α dependent (Figure 6A). Moreover, incubation with SB203580 for 1 h reduced by about 25% the levels of membrane-localized EEA1 in wt cells, suggesting that p38α contributes to the regulation of membrane recruitment of EEA1 also in nonstimulated cells. Importantly, membrane recruitment of both EEA1 and Rabenosyn-5 was also increased, in a p38 MAPK-dependent manner, upon Damgo treatment of SH-SY5Y neuronal cells (Figure 6B). Figure 6.p38 MAPKs regulate EEA1 membrane localization. (A) Western blots of membrane preparations from wt or p38α−/− MEFs transfected with EEA1wt and treated with 1 μM Damgo or 10 μM SB203580 for 30 min and 1 h, respectively. The histogram represents the quantification of two experiments. (B) Western blots of
Referência(s)