Báo cáo khoa học: Fidelity and spatio-temporal control in MAP kinase (ERKs) signalling Delivered on 24 October 2002 at the 28th FEBS Meeting in Istanbul potx

THE SIR HANS KREBS LECTURE Fidelity and spatio-temporal control in MAP kinase (ERKs) signalling Delivered on 24 October 2002 at the 28th FEBS Meeting in Istanbul Jacques Pouysse ´ gur and Philippe Lenormand Institute of Signaling, Developmental Biology and Cancer Research, CNRS-UMR 6543, Centre Antoine Lacassagne, Nice, France The mitogen activated protein (MAP) kinase module: (Raf fi MEK fi ERKs) is central to the control of cell growth, cell differentiation and cell survival. The fidelity of signalling and the spatio-temporal activation are key deter- minants in generating precise biological responses. The fidelity is ensured by scaffold proteins – protein kinase ÔinsulatorsÕ – and by specific docking sites. The duration and the intensity of the response are in part controlled by the compartmentalization of the signalling molecules. Growth factors promote rapid nuclear translocation and persistent activation of p42/p44 MAP kinases, respectively and ERK2/ERK1, during the entire G1 period with an extinc- tion during the S-phase. These features are exquisitely con- trolled by the temporal induction of the MAP kinase phosphatases, MKP1–3. MKP1 and 2 induction is strictly controlled by the activation of the MAP kinase module providing evidence for an auto-regulatory mechanism. This negative regulatory loop is further enhanced by the capacity of p42/p44 MAPK to phosphorylate MKP1 and 2. This action reduces the degradation rate of MKPs through the ubiquitin–proteasomal system. Whereas the two upstream kinases of the module (Raf and MEK) remain cytoplasmic, ERKs (anchored to MEK in the cytoplasm of resting cells) rapidly translocate to the nucleus upon mitogenic stimula- tion. This latter process is rapid, reversible and controlled by the strict activation of the MAPK cascade. Following long- term MAPK stimulation, p42/p44 MAPKs progressively accumulate in the nucleus in an inactive form. Therefore we propose that the nucleus represents a site for ERK action, sequestration and signal termination. With the generation of knockdown mice for each of the ERK isoforms, we will illustrate that besides controlling cell proliferation the ERK cascade also controls cell differentiation and cell behaviour. Keywords: MAP kinases; MAPK-phosphatases; scaffolding proteins; nucleus; growth control; cell signalling. Introduction It is a great privilege for me to be invited to give this lecture in honour of one of the most emblematic and unique figures in Biochemistry. I first started out studying the regulation of metabolism of ÔexoticÕ sugars (Hexuro- nates) in Escherichia coli at a time when prokaryotic genetics was ÔexplodingÕ, confirming the extraordinary accuracy and complexity of metabolic pathways. Unfor- tunately, I never had the opportunity to meet Hans Krebs, however, I had the immense pleasure of starting my first postdoctoral training in 1971 with one of Krebs’s prominent students, Hans Kornberg (Professor of Bio- chemistry, Leicester at that time). I then turned my interest to growth control in mammalian cells, studying successively, cell surface glycopropteins, anaerobic glyco- lysis, pHi molecular control, growth factor action and MAP kinase signalling. p42/p44 MAP kinases (ERKs) belong to a major signal- ling module, conserved throughout evolution, that is activated in mammalian cells via stimulation of receptor tyrosine kinases, G-protein coupled receptors and integrins [1].These cell surface signals converge towards activation of the small G-protein, Ras that recruits the serine/threonine kinase, Raf to the membrane where it is fully activated by largely unknown mechanisms [2]. The signal is amplified via two downstream kinases, MAPKK or ERK kinase (MEK) and extracellular regulated kinase (ERK), that are activated uniquely via phosphorylation. MEK is phosphorylated on two serine residues by Raf, and then ERKs are dually phosphorylated on a tyrosine and threonine residue by MEK (sequence TEY). Amplification via this signalling cascade is such that activation of  5% of Ras molecules is sufficient to induce full activation of ERK [3]. In 1993, we were the first to report that the long-term activation of p42/44 MAP kinases is mandatory for cell cycle entry [4]. ERK activation provides an integrated pleiotypic response: it activates the transcription of many genes, via phosphorylation of transcription factors and Correspondence to J. Pouysse ´ gur, Institute of Signaling, Develop- mental Biology and Cancer Research, CNRS-UMR 6543, Centre Antoine Lacassagne, 33 Avenue de Valombrose, 06189 Nice, France. Fax: + 33 492 03 1225, Tel.: + 33 492 03 1222, E-mail: pouysseg@unice.fr Abbreviations: CDK, cyclin dependent kinase; Crm1, chromosomal region maintenance 1; EGF, epithelial growth factor; ERK, extra- cellular regulated kinase; JNK, c-jun N-terminal kinase; KSR, kinase suppressor of Ras; MAPK, mitogen activated protein kinase; MEK, MAPK of ERK kinase; MKP, MAP kinase phosphatase; MNK1, mitogen and stress kinase1; MP1, MEK partner 1; NGF, nerve growth factor; PEA15, phosphoprotein enriched in astrocytes 15 kDa; PI3K, phosphatidyl inositol 3 kinase; PP2A, protein phosphatase 2 A; Ste5, sterile 5. (Received 17 May 2003, accepted 6 June 2003) Eur. J. Biochem. 270, 3291–3299 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03707.x chromatin; it induces cyclin D1, a rate-limiting component of the G1 phase [5,6]; it stimulates protein synthesis via MNK1 and increases nucleotide synthesis (review in [7]). In a single cell, activation of the ERK pathway can lead to induction of antagonistic effects, e.g. in PC12 cells, both differentiation and cell proliferation require ERK activation [following nerve growth factor (NGF) or expithelial growth factor (EGF) stimulation, respectively]. In these cells, EGF causes transient activation of ERK, whereas NGF causes sustained activation, thus the duration of ERK activation specifies signal identity [8]. Similarly, we have observed in fibroblasts a correlation between the strength of mitogenic signalling and the duration of ERK stimulation. We have shown that non-mitogenic factors induce transient activa- tion of ERK (<60 min) that does not lead to cell cycle entry, whereas mitogens induce cell proliferation and concomitant long-term stimulation of ERKs (up to 6 h) [9]. Similarly, it has been shown that very potent ERK activation protects cells from apoptosis induced by anchor- age and serum removal [10], whereas moderate ERK activation is required to permit apoptosis induced by anchorage and serum removal [11]. Clearly, the ERK pathway must be tightly controlled in its duration of activation and subcellular localization to ensure the proper outcome of integrated biological responses such as cell proliferation, differentiation and survival. In this lecture, I will address some of the key questions of ERK signalling: (a) how is fidelity of p42/p44 MAPK signalling ensured? (b) how is ERK activity controlled in time and space and (c) why are there two ERK isoforms and do they have overlaping functions? Fidelity in signalling Scaffolding and docking sites provide the solution MAP kinase modules have evolved by gene duplication and several closely related modules, delivering specific biological responses, are coexpressed in a particular cell. This is well illustrated in Saccharomyces cerevisiae (Fig. 1). There is a high degree of homology between mammalian MAPK modules, both in their general organization and at the protein level, with a high percentage of similarity in the primary sequence of the different MAPKs (60% amino acid identity between ERKs, JNKs and p38 MAPKs). Further- more, the substrates of the three main mammalian MAPKs: ERK, p38 and JNK display similar phosphorylation consensus motifs: (T/S)P. How does the cell succeed in delivering specific biological responses, limiting therefore inappropriate crosstalk between the parallel MAPK mod- ules? How is signal fidelity built within these modules? Two mechanistic devices have emerged to enforce specificity. Scaffold proteins. These scaffolds create multienzyme complexes that bring together components of a single kinase cascade (review by [12]). These complexes insulate the module from activation by irrelevant stimuli and favour the rapid transmission of the signal through the cascade. Second, specific docking sites on MAPKs that serve for the binding of substrates, activators and regulators increase the fidelity and the efficiency of the enzymatic reactions. The most studied MAPK scaffolding protein is Ste5, from the yeast S. cerevisiae (reviewed in [13]). Through distinct regions, Ste5 binds simultaneously to Ste11 (MAPKKK), Ste7 (MAPKK) and Fus3 (one of the two MAPKs) but Ste5 binds weakly to the other MAPK, Kss1. Fus3 is preferentially implicated in the mating pheromone response, whereas Kss1 is primarily involved in the filamentous response. Fus3 and Kss1 share the same activators, Ste11 and Ste7, however, it seems that Kss1 may be better at transmitting low-level, long-duration, scaffold independent signalling, whereas Fus3 preferentially transmits scaffold- associated signalling (review by [14]). Contrary to that previously thought, scaffolded complexes are not stably assembled. During vegetative growth, recent work showed that the upstream activators Ste11 and Ste7 are predomin- antly cytoplasmic, while the scaffold Ste5 and the MAPK Fus3 are located both in the nucleus and in the cytoplasm and shuttle permanently between these two cellular com- partments [15]. In pheromone-treated cells, Ste11, Ste7 and Fus3 are colocalized with Ste5 to tips of mating projections. However, subsequently activated-Fus3 dissociates rapidly from this multiprotein complex to translocate to the nucleus. The role of the scaffolding protein, Ste5 in this signalling pathway is essential, as Ste5 that cannot transit via the nucleus is unable to localize to the cell periphery and is unable to activate the pathway [16]. This novel regulatory scheme may ensure that cytoplasmic Ste5 does not activate downstream kinases in the absence of pheromone. In mammalian cells, the identification and the role of scaffolding proteins, in particular for the ERK module, is not as well advanced. First, a two-hybrid screen, using MEK1 as a bait, identified MP1 (MEK Partner 1) as a scaffold protein that specifically binds MEK1 and ERK1 to the exclusion of MEK2 and ERK2, thereby enhancing the activation of ERK1 [17]. A partner of MP1, p14 was discovered recently and the MP1-p14 complex scaffolds MEK1 and ERK1 to the cytoplasmic surface of late endosomes lysosomes where P14 is localized [18]. Reduc- tion of MP1 or p14 protein levels by short interfering (si)RNA results in defective signal transduction [19]. Another and presumably more central scaffold protein for the ERK pathway is the KSR protein (kinase suppressor of Ras). KSR was first identified by genetic screening in Drosophila melanogaster and Caenorhabditis elegans as an activator of the Ras pathway as mutations in KSR resulted Fig. 1. MAPK modules and their associated functions in Saccharomyces cerevisiae. 3292 J. Pouysse ´ gur and P. Lenormand (Eur. J. Biochem. 270) Ó FEBS 2003 in attenuation of Ras-mediated signalling (reviewed in [20,21]). A mammalian homologue has been isolated that interacts directly with MEK and ERK via distinct domains while interaction with Raf appears to be indirect. KSR1 translocates from the cytoplasm to the cell membrane in response to growth factor treatment. This process is controlled by the serine/threonine kinase, C-TAK1 that phosphorylates KSR1 at a site that confers 14-3-3 binding, thus sequestering the KSR1 complex in the cytoplasm in the absence of stimulation [22]. In response to growth factors, the KSR1 S392 site is dephosphorylated by an unknown phosphatase, and KSR1 is liberated from 14-3-3 binding and translocates to the plasma membrane where it brings MEK and ERK in close vicinity to the active Raf signalling complex. Therefore, KSR1 seems to act as a scaffold protein to maintain specificity and ensure signalling through the ERK cascade. This notion has been beautifully demonstra- ted by siRNA-mediated KSR knockdown in Drosophila melanogaster [23]. For the JNK module, an interesting set of JIP proteins (JNK-Interacting-Protein) has been identified. These JIP proteins function by aggregating components of a JNK module (including MLK, MKK7, and JNK) [24]. Interestingly JIP-1 has recently been shown to also bind to the MAP kinase phosphatase MKP-7 indicating that JIP-1 scaffold protein modulates JNK signalling via association with both protein kinases and protein phosphatases [25]. Therefore, from the unique properties of Ste5, JIP and KSR, an emerging concept arises: scaffolding proteins are not only insulators between homologous signalling mod- ules, but they play an important role as regulators of the subcellular localization and modulation of the kinase signal intensity. Docking sites. All MAPK members phosphorylate their substrates on the consensus (T/S)P sequence and many potential substrates contain this minimal motif (review by [26]). Therefore MAPKs, presumably like all enzymes, have acquired specific docking sites to specify interactions with relevant substrates. These docking sites also contribute to increase the local concentration of the kinase, hence favouring substrate phosphorylation. The key residue of the ERK docking site is composed of a cluster of acidic residues on the C-terminus of the kinase that is remarkably conserved from C. elegans to humans. This acidic cluster, also called CD (for common docking) is not only found in ERKs but in all MAPK members [27,28]. Data from the three-dimensional structure of ERK indicate that the common docking site is localized on the opposite side of the kinase respective to the catalytic cleft, thus, substrates must dissociate from the docking site to be phosphorylated, indicating that association via kinase docking sites is highly dynamic [29]. Docking sites on ERK interacting proteins have been identified on substrates, activators, scaffolding proteins and phosphatases. On interactor proteins, docking sites are constituted by a cluster of positively charged amino acids (D-domain), that interact on the same negatively charged ERK docking site. This implies that interaction of these proteins with ERK are mutually exclusive, thereby providing a molecular mechanism for the sequential activation and inactivation of ERK. The specificity of the ERK interaction with proteins may not be determined solely by the negatively charged ERK motif, as interchanging (by mutation) of the docking site present on ERK by the docking site present on p38 MAPK still allows the binding of MEK to ERK while no binding of MKK6 (the upstream activator of p38 MAPK) can be detected [27]. It is now thought that the docking region on ERK is contained in a docking groove, where several interacting motifs cooperate to confer strong and specific binding for each MAPK-interacting molecules [28,29]. The spacing and organization of these different motifs on the different MAPK interacting proteins is a feature that may account for the differential MAPK specificities observed [30]. Furthermore, in yeast, both the MKK–MAPK dock- ing interaction and binding to the scaffolding protein, Ste5 make mutually reinforcing contributions to efficiently conduct mating pheromone signalling [31]. Moreover, the phosphorylation state of partners can also modulate the affinity of the interaction. For example, the association of ERK with its substrate, Elk1, is enhanced upon ERK activation [32], whereas, interaction of ERK with its activator, MEK, is reduced upon activation of the signalling cascade [33]. Interestingly, crosstalk with other signalling pathways can be mediated by regulating docking interactions. On the matter of protein substrates, there are two classes of docking sites, the D-domain (cluster of positively charged amino acids) and the FXFP motif, whose binding pocket on ERK remains to be determined. A systematic study of docking sites on Elk1 indicates that the D-domain, and the FXFP motif form a flexible modular system that has two functions [34]. First, the affinity of a substrate for ERK can be regulated by the number, type, position and arrangement of these docking sites. Second, docking sites can direct phosphorylation of specific (S/T)P residues [29,34]. The discovery of these kinase docking sites has provided new tools to deregulate the ERK signalling cascade. The first example based on these interactions was the trapping of active ERKs in the cytoplasm by overexpression of an inactive form of MKP3, which possesses a specific ERK docking site. MKP3 is a cytoplasmic MAPK phosphatase, therefore, its overexpression was able to retain ERKs in the cytoplasm upon mitogenic stimulation [35]. A second example was provided by microinjection into the nucleus of a peptide corresponding to the ERK binding site of MEK. This action led to the disruption of the association of ERK/MEK in the nucleus hence significantly inhibiting the MEK driven export of ERK out of the nucleus [36]. Similarly, microinjection into the nucleus of a peptide corresponding to the ERK binding site on p90rsk, has been shown to disrupt interaction between ERK and nuclear phosphatases thus increasing active ERK in the nucleus [37]. This latter experiment confirms that several interacting proteins act via highly homologous docking sites as a peptide corresponding to the sequence of an ERK substrate can impede ERK association with phosphatases. Spatio-temporal control of ERK activity Schematically, mitogenic stimulation of G0-arrested cells elicits biphasic ERK activation. After an initial burst of activation (30–60 min) that varies with the cell type and the strength of the stimulus, there is a prolonged activation peaking from 2–h poststimulation, finally this activation Ó FEBS 2003 MAPK (ERK) signalling module (Eur. J. Biochem. 270) 3293 gradually diminishes and ERK activity is reduced almost to basal levels at the end of the G1 phase of the cell cycle. This activity remains very low along the S phase, whereas, a burst of ERK activity appears at the G2/M transition [38]. Considering that dephosphorylation of either the threonine or the tyrosine residue within the ERK activation loop TEY motif is sufficient for total kinase inactivation [39], numer- ous phosphatases could be implicated in the two phases of inactivation: the rapid initial phase and the slower and delayed one. The serine/threonine specific phosphatase, PP2A has been implicated in the first inactivation of ERK observed within minutes of NIH-3T3 cell stimulation [40] and of Xenopus oocytes stimulation [41]. The remaining phospho- tyrosine residue must be removed by a constitutive phosphatase. Several related tyrosine specific phosphatases such as PTP-SL, STEP, He-PTP/LC-PTP show a good specificity towards ERKs [30,42,43]. However, these cytosolic tyrosine phosphatases present a restricted expres- sion pattern and thus the ubiquitously expressed phos- phatase(s) that may play the same role in most cells is (are) not yet identified. Interestingly, a cytosolic Droso- phila tyrosine phosphatase, PTP-ER, related to the tyrosine phosphatases mentioned above, plays an import- ant role in down-regulating ERK activation during Drosophila eye development [44]. It is not known, however, if PTP-ER plays a major role in the inactivation of the first peak of ERK activation. The delayed phase of ERK inactivation is dependent on protein synthesis, indicating that neosynthesized phospha- tases are required [40,45]. Furthermore, these phosphatases have a tyrosine specificity as they are inhibited by vanadate treatment [37,40,41]. The phosphatases that fullfil these criteria are the MAPK phosphatases (MKPs). MKPs belong to the dual specificity phosphatases family (DUSP) as they are capable of dephosphorylating both the tyrosine and the threonine residues of MAPKs (reviewed in [46,47]). Indeed, we have demonstrated that MKPs are good candidates for setting the low steady-state activity of ERKs. In fact, as shown in Fig. 2, ERK activity itself induces an autocontrol mechanism. We first showed that exclusive activation of p42/p44MAPKs is sufficient to induce the immediate early genes, mkp1 and mkp2 [45]. Second, we established that MKP1 and MKP2 proteins are direct substrates of ERKs and third that these MKPs, when phosphorylated, are less sensitive to rapid degradation by the ubiquitine-proteasomal system [48]. Indeed, MKP1 is phosphorylated on serine 359 and serine 364 by ERK [48], which does not modify phosphatase activity, but increases their half life, reinforcing the negative feed back autocontrol (Fig. 2). Finally, MKP3 [49] and MKP1 [50] are catalyti- cally activated upon ERK binding to their N-terminal non-catalytic moiety. Catalytic activation of MKP1 and of MKP3 occurs by binding of ERK via the classical docking site. Hence, substrate-specificity is ensured by two means: protein–protein interaction and catalytic activation of the phosphatase. The precise role of each MKPs in vivo is not yet understood. Clearly, expression of some MKPs is restricted to specific subcellular compartments, cytoplasm (MKP3) or nucleus (MKP1 and 2) that must impinge on the range of available substrates. It is probable that there is some degree of redundancy between MKPs as invalidation of the mkp1 gene did not affect mouse physiology [51]. More work is required to assess the functional role of individual MKPs in vivo. Temporal compartimentalization of the ERK module ERK nuclear translocation is a key event in signalling. In resting cells, Raf, MEK and ERKs are cytoplasmic. Following mitogenic stimulation, intracellular redistribution of ERK occurs in two phases. First, there is an immediate ERK nuclear translocation that can be visualized, in particular with antibodies specific for the phosphorylated and active form of ERKs, by immunofluorescence as soon as 2 min [37]. The pool of ERK protein progressively accumulates in the nucleus after several hours of mitogenic stimulation (3–6 h depending of the cell type), depleting the cytoplasmic compartment. This process of ERK nuclear accumulation is reversible and follows the time-course of ERK inactivation. If the activation of the ERK module is maintained (ER-Raf construct, activatable by tamoxifen), ERKs remain in the nucleus [52]. Non-mitogenic stimuli induce the initial nuclear entry but fail to trigger the nuclear accumulation of ERKs [52,53]. Similarly, when the fate of cells is differentiation, only differentiating signals trigger the nuclear accumulation of ERK observed after several hours of stimulation [8]. This nuclear translocation is a key event in ERK signalling. This was demonstrated by an experiment designed to retain active ERK in the cytoplasm [35]; under these conditions, fibrobasts fail to replicate their DNA. Alternatively, forcing an active form of ERK into the nucleus of fibroblasts promotes oncogenic transformation [54]. Physiologically, the cytoplasmic retention of ERK may play a critical role in maintaining a differentiated phenotype in some cell types. For example, increased expression of the protein, Phosphoprotein Enriched in Astrocytes 15kD (PEA15) traps ERK in the cytoplasm of astrocytes and Fig. 2. Schematic model illustrating the auto-regulation of the ERK module. MAP kinase phosphatase 1 and 2 (MKP1/2), products of the immediate early genes mkp1 and mkp2, are directly induced via the activation of p42/p44 MAP kinases, providing a progressive retro- inhibition of ERKs. In addition, MKP1/2 are directly phosphorylated by ERKs, increasing their stability and therefore reinforcing the retrocontrol [48]. 3294 J. Pouysse ´ gur and P. Lenormand (Eur. J. Biochem. 270) Ó FEBS 2003 blocks cell proliferation, whereas, genetic deletion of PEA15 increases astrocyte proliferation [55]. Regulation of ERK signalling by cytoplasmic trapping may be a frequent phenomenon as it was shown recently that b-arrestin associates with ERK and enhances the ERK cytoplasmic activity while inhibiting ERK mediated tran- scription [56]. ERK nuclear accumulation and inactivation. This ERK translocation process produced a surprise when we double-labelled cells either with anti-ERK protein or with antibodies directed against active ERKs (phosphoERK- antibodies). Clearly, this double-labelling revealed that the ERK protein pool that accumulated in the nucleus became inactivated with time [37]. As seen in Fig. 3, at the peak of ERK nuclear accumulation (3 h in hamster fibroblasts), virtually no active phospho-ERK was detect- able in the nucleus. We demonstrated by an independent approach that the capacity of nuclear ERKs to phos- phorylate nuclear substrates (HIF-1a)atthistimeof stimulation (3 h of FCS) was severely blunted [37]. However, short inhibition of tyrosine phosphatases with vanadate, fully reactivated phospho-ERK in the nucleus and maximally phosphorylated the nuclear substrate, HIF-1a [37]. When activation of the ERK pathway is transient, ERKs rapidly exit out of the nucleus, however, during sustained activation of the module, ERKs remain in the nucleus in an inactive form as shown in Fig. 3 [37,57]. The nuclear accumulation of ERK in the nucleus requires the ERK-dependent transcriptional induction of short-lived nuclear anchoring proteins [57]. The identity of these nuclear anchors remains elusive, however, the use of anti-phospho-ERK antibodies provided new clues in understanding this nuclear accumulation of ERK. MKP1 and MKP2 are the best candidates for inactiva- tion of ERK in the nucleus as: (a) they are induced by activation of the ERK pathway; (b) they are localized in the nucleus; (c) they possess ERK docking sites and (d) they are inhibited by tyrosine phosphatase specific inhibitors [46,47]. Furthermore, MKP1 and MKP2 may participate in the nuclear anchoring of ERKs as these proteins present all the characteristics of ERK nuclear anchors, MKP1/2 are induced by ERK activation [45], and are short-lived nuclear proteins whose expression is virtually abolished within 1 h upon traductional or translational block [48]. The use of RNA interference to abrogate expression of each MKP isoforms, may help to provide quick answers to these questions. The mechanisms of ERK nuclear import and export are still largely unknown. These protein kinases do not possess any of the common nuclear import sequences (NLS) and previous work has shown that ERKs cross the nucleopore by passive diffusion [58]. It has been shown that ERK associates with MEK in the cytoplasm of resting cells via their docking sites, an interaction that is reduced dramati- cally upon activation of the MEK/ERK signalling pathway, thus allowing ERK to translocate to the nucleus [36]. Clearly, activation of the pathway is essential as blocking MEK activity abrogates ERK nuclear translocation [52]; however, phosphorylation mutants of ERK still translocate to the nucleus in response to cell stimulation [53,59,60]. Phosphorylation-dependent dimerization of ERK has also Fig. 3. Long-term activation of p42/p44 MAPKs induces their nuclear accumulation in a dephosphorylated and inactive form. Resting CCL39 hamster fibrobasts (left panel) were stimulated for 3 h with 10% fetal bovine serum (FCS) (middle and right panels).The green immunoflurescence indicates the location of the proteins, ERK1 and 2, whereas the red immunofluorescence indicates the dual phosphorylated and active forms of ERK1 and 2. The right panel shows confocal images. Reproduced from reference [37]. Ó FEBS 2003 MAPK (ERK) signalling module (Eur. J. Biochem. 270) 3295 been proposed to explain ERK nuclear entry. Indeed, ERK- b-galactosidase fusion proteins are unable to enter the nucleus when ERK dimerization motifs are mutated [58]. However, ERK1 dimerization mutants expressed in erk1 null mouse fibroblasts present the same time course of nuclear translocation as wild type ERK1 (P, Lenormand and J, Pouyssegur, unpublished results). As discussed previously, recent work on the scaffolding protein, KSR indicates that KSR participates in the regulation of the subcellular localization of kinase cascade components. However, the contribution of KSR in the release of ERK from the cytoplasmic complex has not yet been established. Recent work in Drosophila may provide clues in under- standing this phenomenon. Indeed, deletion or mutations in the Drosophila importin a homologue, DIM-7 or mutations in the importin b homologue Ketel, reduce the nuclear localization of D-ERK, the ERK Drosophila homologue. Interestingly, DIM-7 associates with phosphorylated D-ERK which should allow a better understanding of how ERK can interact with the active import machinery while lacking a classical NLS [61]. Another point of interest is the demonstration of direct binding of ERK to nucleo- pore complex [60]. In that case, ERK transport across the nucleopore would be propelled by Brownian motion. It has been shown in permeabilized mammalian cells that ERK associates directly with the nucleopore complex and trans- locates to the nucleus independently of soluble factors and ATP. Furthermore, ERK binds in vitro to an FG repeat region of nucleoporin CAN/Nup214. Altogether, the relative contribution of these different mechanisms in conducting ERK across the nuclear mem- brane remains to be determined. Several studies have clearly established the continuous shuttling of ERK between the cytoplasm and nucleus. When quiescent cells are treated with leptomycin B, that blocks Crm1-dependent nuclear export, ERK appears within minutes in the nucleus [37]. This occurs in the total absence of ERK activation as it is not impeded by prior treatment with the MEK inhibitor, U0126 [52]. The export of ERK from the nucleus has remained as enigmatic as the import as ERK1 and ERK2 protein sequences do not show motifs homologous to a nuclear export sequence (NES). However, blocking active nuclear export with leptomycin B triggers the nuclear accumulation of ERK and MEK [37,62]. In the presence of leptomycin B, addition of growth factors for 5 min is sufficient to mobilize the entire cytoplasmic pool of ERK in the nucleus. This result stresses the rapid and constant ERK cytoplasmic/nuclear shuttling. We believe that MEK, with its built-in export sequence, might be at the heart of this shuttling mechanism. Although MEK always appears in the cytoplasm, due to its very efficient NES, MEK is the Ôdriving exporting-forceÕ of nuclear inactivated ERKs. In summary, ERKs oscillate between two high affinity complexes in separate cellular compartments. In resting cells ERKs are associated with the Ôactivating centerÕ, the Raf-MEK cytoplasmic complex. Upon long mitogenic treatment, ERKs are sequestered in the nucleus, closely associated to the neosynthesized ÔMKP- inactivating centerÕ away from the site of activation. By this mechanism we propose that mammalian cells operate the termination of the MAPK signal, a condition required to trigger the appropriate biological response. Knockdown of erk1 and erk2 genes in mice In previous experiments in which the biological functions of p42/p44 MAP kinases have been addressed (antisense or expression of dominant-negative MEK or ERK), both isoforms have been inactivated [63]. So far, the pharmaco- logical inhibition of ERK1 and ERK2 relies on MEK specific inhibitors that invariably blunt the activation of ERK1 and ERK2 [64]. Therefore, the specific role of the two ERK isoforms is still an entirely open question. In mammals, ERK1 and ERK2 are expressed ubiquitously, although the expression level could vary in different tissues. These two protein kinases are highly similar (overall 84% identity at the amino acid level, and up to 90% identity when the short N-terminal stretch is not taken into account) and, in vitro, both isoforms present apparently the same substrate specificity and the same time course of activation. Interestingly however, ERK1 and ERK2 do not share an identical pattern of compartimentalization as illustrated from the work of HuberÕs group [19]. Thus, a pressing question is what are their specializations and do they have overlaping functions? A way to address this issue is to produce single ERK invalidating mutations in mice. From our published work, and work in progress, it is clear that isoform-specific invalidation in mice provides contrasting results (Fig. 4). First ERK1 –/– mice are viable, fertile and of normal size [63]. Clearly, in these animals, in which we only found a defect in thymocyte terminal differentiation, ERK2 can compensate for most of the functions of ERK1. On the contrary, disruption of the Erk2 locus leads to embryonic lethality early in mouse development after the implantation stage. Erk2 mutant embryos failed to form the ectoplacental cone and extraembryonic ectoderm, which give rise to mature trophoblast derivatives in the foetus (Sylvain Meloche, Institute of Clinical Research, Montreal; personal communication). In these embryos, ERK1 cannot compen- sate for the loss of ERK2, thus, specific functions of these isoforms remain to be discovered. Alternatively, ERK1 is not expressed in some cells or at such low levels compared to ERK2 that it cannot provide the strength of activation required for embryonic survival. Finally, during the course of these studies it became apparent that mice disrupted in the Erk1 locus were more actively displaying facilitated striatal-mediated learning and memory [65]. This unexpected behaviour revealed that ERK1 ablation led to an increase in the temporal activation of ERK2. The exact mechanism of the interplay between the two isoforms is not understood, but this finding indicates Fig. 4. Phenotypes of ERK1 and ERK2 null mice. 3296 J. Pouysse ´ gur and P. Lenormand (Eur. J. Biochem. 270) Ó FEBS 2003 that a little alteration in the intensity and temporal activation of ERKs could have a profound effect in animal physiology. In conclusion, the ERK MAP kinase module, reported in early1990,hasbeenshowntoplayacentralrolein signalling growth, differentiation and survival from inver- tebrates to humans. A sophisticated autocontrol mechanism associated with nuclear/cytoplasmic shuttling ensure the intensity and temporal modulation of ERK activity in response to various hormonal, growth factor and extracel- lular matrix stimuli. Cancer and many other diseases are simply the reflection of alterations in this fine tuning mechanisms. Despite intense research efforts worldwide, questions concerning this ERK module remain unanswered: are there other scaffolding proteins in mammalian cells? What are the exact roles of KSR1 and KSR2? Do they represent the basis for two separate and competing modules for MEK1/MEK2 and ERK1/ERK2? What is the identity of the ERK nuclear anchoring complex and which MAPK phosphatases are essential in terminating the ERK signal? We anticipate that the siRNA knockdown approach with inducible vectors will greatly facilitate the investigation of these questions. Acknowledgements We thank Drs Gilles Page ` s, Fergus McKenzie, Anne Brunet, Jean- Marc Brondello and the regretted Veronique Volmat for their unique contribution to the MAP kinase project and all the members of the laboratory for helpful discussion. I am particularly grateful to Dr C. Brahimi-Horn for carefull reading of the manuscript. References 1. Widmann, C., Gibson, S., Jarpe, M.B. & Johnson, G.L. (1999) Mitogen-activated protein kinase: conservation of a three-kinase module from yeast to human. Physiol. Rev. 79, 143–180. 2. Kerkhoff, E. & Rapp, U.R. (2001) The Ras-Raf relationship: an unfinished puzzle. Adv. Enzyme Regul. 41, 261–267. 3. Hallberg, B., Rayter, S.I. & Downward, J. (1994) Interaction of Ras and Raf in intact mammalian cells upon extracellular stimu- lation. J. Biol Chem. 269, 3913–3916. 4. Page ` s, G., Lenormand, P., L’Allemain, G., Chambard, J C., Me ´ loche, S. & Pouysse ´ gur, J. (1993) Mitogen-activated protein kinases p42mapk and p44mapk are required for fibroblast cell proliferation. Proc. Natl. Acad. Sci. USA 90, 8319–8323. 5. Lavoie, J.N., L’Allemain, G., Brunet, A., Muller, R. & Pouysse- gur, J. (1996) Cyclin D1 expression is regulated positively by the p42/p44MAPK and negatively by the p38/HOGMAPK pathway. J. Biol Chem. 271, 20608–20616. 6. Lavoie, J.N., Rivard, N., L’Allemain, G. & Pouyssegur, J. (1996) A temporal and biochemical link between growth factor-activated MAP kinases, cyclin D1 induction and cell cycle entry. Prog. Cell Cycle Res. 2, 49–58. 7. Whitmarsh, A.J. & Davis, R.J. (2000) A central control for cell growth. Nature. 403, 255–256. 8. Traverse,S.,Seedorf,K.,Paterson,H.,Marshall,C.,Cohen,P.& Ullrich, A. (1994) EGF triggers neuronal differentiation of PC12 cells that overexpress the EGF receptor. Curr. Biol. 4, 694–701. 9. Kahan, C., Seuwen, K., Me ´ loche, S. & Pouysse ´ gur, J. (1992) Coordinate, biphasic activation of p44 mitogen activated protein kinase and S6 kinase by growth factors in hamster fibroblasts. J. Biol. Chem. 267, 13369–13375. 10. Le Gall, M., Chambard, J.C., Breittmayer, J.P., Grall, D., Pouyssegur, J. & Van Obberghen-Schilling, E. (2000) The p42/p44 MAP kinase pathway prevents apoptosis induced by anchorage and serum removal. Mol. Biol. Cell. 11, 1103–1112. 11. Zugasti, O., Rul, W., Roux, P., Peyssonnaux, C., Eychene, A., Franke, T.F., Fort, P. & Hibner, U. (2001) Raf-MEK-Erk cascade in anoikis is controlled by Rac1 and Cdc42 via Akt. Mol. Cell Biol. 21, 6706–6717. 12. Whitmarsh, A.J. & Davis, R.J. (1998) Structural organization of MAP-kinase signaling modules by scaffold proteins in yeast and mammals. Trends Biochem. Sci. 23, 481–485. 13. Elion, E.A. (1995) STE5: a meeting place for MAP kinases and their associates. Trends Cell Biology 5, 322–327. 14. Pryciak, P.M. (2001) MAP kinases bite back. Dev. Cell. 1, 449–451. 15. van Drogen, F., Stucke, V.M., Jorritsma, G. & Peter, M. (2001) MAP kinase dynamics in response to pheromones in budding yeast. Nat. Cell Biol. 3, 1051–1059. 16. Mahanty, S.K., Wang, Y., Farley, F.W. & Elion, E.A. (1999) Nuclear shuttling of yeast scaffold Ste5 is required for its recruit- ment to the plasma membrane and activation of the mating MAPK cascade. Cell 98, 501–512. 17. Schaeffer, H.J., Catling, A.D., Eblen, S.T., Collier, L.S., Krauss, A. & Weber, M.J. (1998) MP1: a MEK binding partner that enhances enzymatic activation of the MAP kinase cascade. Science. 281, 1668–1671. 18. Wunderlich, W., Fialka, I., Teis, D., Alpi, A., Pfeifer, A., Parton, R.G., Lottspeich, F. & Huber, L.A. (2001) A novel 14-kilodalton protein interacts with the mitogen-activated protein kinase scaf- fold mp1 on a late endosomal/lysosomal compartment. J. Cell Biol. 152, 765–776. 19. Teis, D., Wunderlich, W. & Huber, L.A. (2002) Localization of the MP1-MAPK scaffold complex to endosomes is mediated by p14 and required for signal transduction. Dev Cell. 3, 803–814. 20. Therrien, M., Chang, H.C., Solomon, N.M., Karim, F.D., Was- sarman, D.A. & Rubin, G.M. (1995) KSR, a novel protein kinase required for RAS signal transduction. Cell 83, 879–888. 21. Morrison, D.K. (2001) KSR: a MAPK scaffold of the Ras path- way? J. Cell Sci. 114, 1609–1612. 22. Muller, J., Ory, S., Copeland, T., Piwnica-Worms, H. & Morrison, D.K. (2001) C-TAK1 regulates Ras signaling by phosphorylating the MAPK scaffold, KSR1. Mol. Cell. 8, 983–993. 23. Roy, F., Laberge, G., Douziech, M., Ferland-McCollough, D. & Therrien, M. (2002) KSR is a scaffold required for activation of the ERK/MAPK module. Genes Dev. 16, 427–438. 24. Yasuda, J., Whitmarsh, A.J., Cavanagh, J., Sharma, M. & Davis, R.J. (1999) The JIP group of mitogen-activated protein kinase scaffold proteins. Mol. Cell Biol. 19, 7245–7254. 25. Willoughby, E.A., Perkins, G.R., Collins, M.K. & Whitmarsh, A.J. (2003) The JNK-interacting protein-1 scaffold protein targets MAPK phosphatase-7 to dephosphorylate JNK. J. Biol. Chem. 278, 10731–10736. 26. Lewis, T.S., Shapiro, P.S. & Ahn, N.G. (1998) Signal transduction through MAP kinase cascades. Adv. Cancer Res. 74, 49–139. 27. Tanoue, T., Adachi, M., Moriguchi, T. & Nishida, E. (2000) A conserved docking motif in MAP kinases common to substrates, activators and regulators. Nat. Cell Biol. 2, 110–116. 28. Bardwell, L. & Thorner, J. (1996) A conserved motif at the amino termini of MEKs might mediate high affinity interaction with the cognate MAPKs. Trends in Biochem. Sci. 21, 373–374. 29. Tanoue, T., Maeda, R., Adachi, M. & Nishida, E. (2001) Identi- fication of a docking groove on ERK and p38 MAP kinases that regulates the specificity of docking interactions. EMBO J. 20, 466–479. 30. Tarrega, C., Blanco-Aparicio, C., Munoz, J.J. & Pulido, R. (2002) Two clusters of residues at the docking groove of Ó FEBS 2003 MAPK (ERK) signalling module (Eur. J. Biochem. 270) 3297 mitogen-activated protein kinases differentially mediate their functional interaction with the tyrosine phosphatases PTP-SL and STEP. J. Biol. Chem. 277, 2629–2636. 31. Sharrocks, A.D., Yang, S.H. & Galanis, A. (2000) Docking domains and substrate-specificity determination for MAP kinases. Trends Biochem. Sci. 25, 448–453. 32. Yang, S.H., Yates, P.R., Whitmarsh, A.J., Davis, R.J. & Shar- rocks, A.D. (1998) The Elk-1 ETS-domain transcription factor contains a mitogen-activated protein kinase targeting motif. Mol. Cell. Biol. 18, 710–720. 33. Fukuda, M., Gotoh, Y. & Nishida, E. (1997) Interaction of MAP kinase with MAP kinase kinase: its possible role in the control of nucleocytoplasmic transport of MAP kinase. EMBO J. 1901–08. 34. Fantz, D.A., Jacobs, D., Glossip, D. & Kornfeld, K. (2001) Docking sites on substrate proteins direct extracellular signal- regulated kinase to phosphorylate specific residues. J. Biol. Chem. 276, 27256–27265. 35. Brunet, A., Roux, D., Lenormand, P., Dowd, S., Keyse, S. & Pouyssegur, J. (1999) Nuclear translocation of p42/p44 mitogen- activated protein kinase is required for growth factor-induced gene expression and cell cycle entry. EMBO J. 18, 664–674. 36. Fukuda, M., Gotoh, I., Gotoh, Y. & Nishida, E. (1996) Cyto- plasmic localization of MAP kinase kinase directed by its N-terminal, leucin-rich amino acid sequence, which acts as a nuclear export signal. J. Biol. Chem. 271, 20024–20028. 37. Volmat, V., Camps, M., Arkinstall, S., Pouysse ´ gur, J. & Lenormand, P. (2001) The nucleus, a site for signal termination by sequestration and inactivation of p42/p44 MAP kinases. J. Cell. Sci. 114, 3433–3443. 38. Roberts, E.C., Shapiro, P.S., Nahreini, T.S., Pages, G., Pouysse- gur, J. & Ahn, N.G. (2002) Distinct cell cycle timing requirements for extracellular signal-regulated kinase and phosphoinositide 3-kinase signaling pathways in somatic cell mitosis. Mol. Cell Biol. 22, 7226–7241. 39. Posada, J. & Cooper, J.A. (1992) Requirements for phosphory- lation of MAP kinase during meiosis in xenopus oocytes. Science. 255, 212–215. 40. Alessi, D., Gomez, N., Moorhead, G., Lewis, T., Keyse, S.M. & Cohen, P. (1995) Inactivation of p42 MAP kinase by protein phosphatase 2A and a protein tyrosine phosphatase, but not CL100, in various cell lines. Curr. Biol. 5, 283–295. 41. Sohaskey, M.L. & Ferrell, J.E. Jr (1999) Distinct, constitutively active MAPK phosphatases function in Xenopus oocytes: implications for p42 MAPK regulation in vivo. Mol. Biol. Cell. 10, 3729–3743. 42. Oh-hora, M., Ogata, M., Mori, Y., Adachi, M., Imai, K., Kosugi, A. & Hamaoka, T. (1999) Direct suppression of TCR-mediated activation of extracellular signal-regulated kinase by leukocyte protein tyrosine phosphatase, a tyrosine-specific phosphatase. J. Immunol. 163, 1282–1288. 43. Pettiford, S.M. & Herbst, R. (2000) The MAP-kinase ERK2 is a specific substrate of the protein tyrosine phosphatase HePTP. Oncogene. 19, 858–869. 44. Karim, F.D. & Rubin, G.M. (1999) PTP-ER, a novel tyrosine phosphatase, functions downstream of Ras1 to downregulate MAP kinase during Drosophila eye development. Mol. Cell. 3, 741–750. 45. Brondello, J.M., Brunet, A., Pouyssegur, J. & McKenzie, F.R. (1997) The dual specificity mitogen-activated protein kinase phosphatase-1 and – 2 are induced by the p42/p44MAPK cascade. J. Biol. Chem. 272, 1368–1376. 46. Camps, M., Nichols, A. & Arkinstall, S. (2000) Dual specificity phosphatases: a gene family for control of MAP kinase function. Faseb J. 14, 6–16. 47. Keyse, S.M. (2000) Protein phosphatases and the regulation of mitogen-activated protein kinase signalling. Curr. Opin. Cell Biol. 12, 186–192. 48. Brondello, J.M., Pouyssegur, J. & McKenzie, F.R. (1999) Reduced MAP kinase phosphatase-1 degradation after p42/ p44MAPK-dependent phosphorylation. Science. 286, 2514–2517. 49. Camps, M., Nichols, A., Gillieron, C., Antonsson, B., Muda, M., Chabert, C., Boschert, U. & Arkinstall, S. (1998) Catalytic acti- vation of the phosphatase MKP-3 by ERK2 mitogen-activated protein kinase. Science. 280, 1262–1265. 50. Slack, D.N., Seternes, O.M., Gabrielsen, M. & Keyse, S.M. (2001) Distinct binding determinants for ERK2/p38alpha and JNK map kinases mediate catalytic activation and substrate selectivity of map kinase phosphatase-1. J. Biol Chem. 276, 16491–16500. 51. Dorfman, K., Carrasco, D., Gruda, M., Ryan, C., Lira, S.A. & Bravo, R. (1996) Disruption of the erp/mkp-1 gene does not affect mouse development: normal MAP kinase activity in ERP/MKP- 1-deficient fibroblasts. Oncogene. 13, 925–931. 52. Volmat, V. & Pouyssegur, J. (2001) Spatiotemporal regulation of the p42/p44 MAPK pathway. Biol. Cell. 93, 71–79. 53. Lenormand, P., Sardet, C., Page ` s, G., L’Allemain, G., Brunet, A. & Pouysse ´ gur, J. (1993) Growth Factors Induce Nuclear Trans- location of MAP Kinases (p42maPk and p44mapk) but not of Their Activator MAP Kinase Kinase (p45mapkk) in Fibroblasts. J. Cell. Biol. 122, 1079–1089. 54. Robinson, M.J., Stippec, S.A., Goldsmith, E., White, M.A. & Cobb, M.H. (1998) A constitutively active and nuclear form of the MAP kinase ERK2 is sufficient for neurite outgrowth and cell transformation. Curr. Biol. 8, 1141–1150. 55. Formstecher, E., Ramos, J.W., Fauquet, M., Calderwood, D.A., Hsieh, J.C., Canton, B., Nguyen, X.T., Barnier, J.V., Camonis, J., Ginsberg, M.H. & Chneiweiss, H. (2001) PEA-15 mediates cyto- plasmic sequestration of ERK MAP kinase. Dev. Cell. 1, 239–250. 56. Tohgo, A., Pierce, K.L., Choy, E.W., Lefkowitz, R.J. & Luttrell, L.M. (2002) beta-Arrestin scaffolding of the ERK cascade enhances cytosolic ERK activity but inhibits ERK mediated transcription following angiotensin AT1a receptor stimulation. J. Biol. Chem. 277, 9429–9436. 57. Lenormand, P., Brondello, J.M., Brunet, A. & Pouyssegur, J. (1998) Growth factor-induced p42/p44 MAPK nuclear translo- cation and retention requires both MAPK activation and neo- synthesis of nuclear anchoring proteins. J. Cell Biol. 142, 625–633. 58. Adachi, M., Fukuda, M. & Nishida, E. (1999) Two co-existing mechanisms for nuclear import of MAP kinase: passive diffusion of a monomer and active transport of a dimer. EMBO J. 18, 5347–5358. 59. Khokhlatchev, A.V., Canagarajah, B., Wilsbacher, J., Robinson, M., Atkinson, M., Goldsmith, E. & Cobb, M.H. (1998) Phos- phorylation of the MAP kinase ERK2 promotes its homo- dimerization and nuclear translocation. Cell. 93, 605–615. 60. Matsubayashi, Y., Fukuda, M. & Nishida, E. (2001) Evidence for existence of a nuclear pore complex-mediated, cytosol- independent pathway of nuclear translocation of ERK MAP kinase in permeabilized cells. J. Biol. Chem. 276, 41755–41760. 61. Lorenzen, J.A., Baker, S.E., Denhez, F., Melnick, M.B., Brower, D.L. & Perkins, L.A. (2001) Nuclear import of activated D-ERK by DIM-7, an importin family member encoded by the gene moleskin. Development. 128, 1403–1414. 62. Adachi, M., Fukuda, M. & Nishida, E. (2000) Nuclear export of MAP kinase (ERK) involves a MAP kinase kinase (MEK)- dependent active transport mechanism. J. Cell Biol. 148, 849–856. 63. Page ` s, G., Guerin, S., Grall, D., Bonino, F., Smith, A., Anjuere, F., Auberger, P. & Pouyssegur, J. (1999) Defective thymocyte maturation in p44 MAP kinase (Erk 1) knockout mice. Science. 286, 1374–1377. 3298 J. Pouysse ´ gur and P. Lenormand (Eur. J. Biochem. 270) Ó FEBS 2003 64. Kohno, M. & Pouyssegur, J. (2003) Pharmacological inhibitors of the ERK signaling pathway: application as anticancer drugs. Prog.CellCycleRes. 5, 219–224. 65. Mazzucchelli, C., Vantaggiato, C., Ciamei, A., Fasano, S., Pakhotin,P.,Krezel,W.,Welzl,H.,Wolfer,D.P.,Pages,G., Valverde,O.,Marowsky,A.,Porrazzo,A.,Orban,P.C., Maldonado, R., Ehrengruber, M.U., Cestari, V., Lipp, H.P., Chapman, P.F., Pouyssegur, J. & Brambilla, R. (2002) Knockout of ERK1 MAP kinase enhances synaptic plasticity in the striatum and facilitates striatal-mediated learning and memory. Neuron. 34, 807–820. Ó FEBS 2003 MAPK (ERK) signalling module (Eur. J. Biochem. 270) 3299 . THE SIR HANS KREBS LECTURE Fidelity and spatio-temporal control in MAP kinase (ERKs) signalling Delivered on 24 October 2002 at the 28th FEBS Meeting in Istanbul Jacques Pouysse ´ gur and. by the serine/threonine kinase, C-TAK1 that phosphorylates KSR1 at a site that confers 14-3-3 binding, thus sequestering the KSR1 complex in the cytoplasm in the absence of stimulation [22]. In. extinc- tion during the S-phase. These features are exquisitely con- trolled by the temporal induction of the MAP kinase phosphatases, MKP1–3. MKP1 and 2 induction is strictly controlled by the