MAPK signal specificity: the right place at the right time

Review

TRENDS in Biochemical Sciences

Vol.31 No.5 May 2006

MAPK signal specificity: the right place at the right time Leon O. Murphy1 and John Blenis2 1

Developmental and Molecular Pathways, Novartis Institutes for BioMedical Research, 250 Massachusetts Avenue, Cambridge, MA 02139, USA 2 Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, MA 02115, USA

Although the mechanisms that lead to activation of the Ras, extracellular-signal-regulated kinase mitogen-activated protein kinase (Ras/ERK-MAPK) signaling pathway have been studied intensively, the fundamental principles that determine how activation of ERK signaling can result in distinct biological outcomes have only recently received attention. Factors such as cell-surface receptor density, expression of scaffolding proteins, the surrounding extracellular matrix, and the interplay between kinases and phosphatases modulate the strength and duration of ERK signaling. Furthermore, the spatial distribution and temporal qualities of ERK can markedly alter the qualitative and quantitative features of downstream signaling to immediate early genes (IEG) and the expression of IEG-encoded protein products. As a result, IEG products provide a molecular interpretation of ERK dynamics, enabling the cell to program an appropriate biological response.

Dynamics of ERK signaling and downstream interpretation Hormonal activation of cell-surface receptors results in a cascade of events that promotes activation of Ras and, ultimately, dual phosphorylation and activation of the extracellular-signal-regulated kinases (ERKs) (Figure 1). Activated ERKs phosphorylate and activate downstream targets such as the family of w90-kDa ribosomal S6 kinases (RSKs), which consists of RSK1 to RSK4 and mitogen- and stress-activated kinases 1 and 2 [1]. ERK1 and ERK2 (hereafter ERK1/2) and RSK1/2 can also translocate into the nucleus where, together with nuclear members of the RSK family, they phosphorylate several important transcriptional regulators including Elk-1, CREB and histone H3 [2]. The result is rapid transcriptional activation of IEGs, including transcription factors that control the cell cycle and/or cell survival. Nearly all growth factors and cytokines activate ERK and RSK, induce their translocation to the nucleus, and promote activation of the same IEGs. Yet these growth factors and cytokines can evoke markedly different biological responses. How do specific cellular responses result from diverse extracellular stimuli that converge on the ERK/IEG signaling cassette? Here we propose that Corresponding author: Blenis, J. ([email protected]). Available online 17 April 2006

signal specificity is derived from the ability of IEGencoded protein products to respond to changes in the spatial and temporal dynamics of ERK activity. Why is signal duration important? More than ten years ago, Marshall [3] proposed a model to explain how the inactivation kinetics or signal duration of ERK could regulate a cell fate decision. The basis for this model came from several studies showing that sustained, but not transient, activation of ERK signaling preceded the differentiation of rat PC12 pheochromocytoma cells into sympathetic-like neurons. This correlation between the duration of ERK signaling and distinct cell behavior has been also documented in fibroblasts [4,5], macrophages [6] and T lymphocytes [7,8]. Thus, normal cells use ERK signal duration as a way to drive many biological processes such as cell-cycle control. Importantly, mutations in key pathway components that result in sustained ERK activation correlate with carcinogenesis. For example, the overall frequency of Ras mutations in cancer is w30%, but O90% of pancreatic cancers have mutations in K-Ras and individuals affected with Costello syndrome have a high incidence of H-Ras mutations and are predisposed to tumor formation [9]. Mutations in B-raf are found in 67% of melanomas and at a lower frequency in colon and thyroid cancers. Not surprisingly, small-molecule inhibitors of the Ras/ERK pathway have been intensely pursued for therapeutic purposes. Such agents could inactivate constitutive oncogenic signaling, thereby leading to an increase in apoptosis and/or limiting the tumorigenic potential of the cancer cells. Thus, in cancers in which ERK activity is required for proliferation and survival, such as melanomas, inhibition of this pathway might be particularly beneficial [10]. Receptor density and trafficking Studies using the PC12 model of neuronal differentiation provided the first evidence that receptor numbers can alter the outcome of ERK signaling. When treated with nerve growth factor (NGF), PC12 cells express neuronal markers and show neurite outgrowth. Such differentiation is preceded by sustained activation of ERK signaling for up to 24 h. When treated with epidermal growth factor (EGF) or insulin, by contrast, PC12 cells show a weak proliferative response and do not differentiate.

www.sciencedirect.com 0968-0004/$ - see front matter Q 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibs.2006.03.009

Review

TRENDS in Biochemical Sciences

Vol.31 No.5 May 2006

Transient signaling

269

Sustained signaling

Ligand Plasma membrane RTKs

P P

P

Ras GDP

GEF GAP

Ras

Raf

Raf

GTP

GTP MEK

Early endosome

Phosphatases

Ras

MKP-3, PTP-SL STEP, He-PTP

GEF GAP

Ras

P P

P

GDP

P P

P P

P

P

P P

P

MEK

Signaling?

ERK

ERK

Recycling endosome

P P P

P P

RSK

P

P

P

Sustained signaling?

RSK

P

?

Nuclear translocation

Ra s GT P

Lysosome Negative feedback

Golgi/ER

ERK RSK

P

Ets H3 MKP-1/2

IEG products degraded

STAT3

P

CREB

Immediate early gene response

P

P

Nuclear envelope

IEG products stablised Ti BS

Figure 1. Pathways of transient and sustained extracellular-signal-regulated kinase (ERK) activation. After ligands bind to the receptor tyrosine kinases (RTKs) and Ras is activated, members of the Raf family are recruited to membrane-associated activators. Raf then activates mitogen-activated protein kinase kinases 1 and 2 (MEK1/2), which in turn phosphorylate and dissociate from ERK1/2. Activated ERK1/2 phosphorylate targets at the membrane and in the cytoplasm, such as ribosomal S6 kinases (RSKs), and a portion of active ERK1/2 translocates into the nucleus by an unknown mechanism. Active RSK1/2 also translocate into the nucleus where, together with nuclear RSK isoforms (RSK3/4, MSK1/2) and ERK1/2, they phosphorylate and activate several nuclear targets such as the transcriptional regulators Ets, STAT, CREB and histone H3. This results in the rapid induction of immediate early genes (IEGs) including c-fos. If signaling remains active after the IEG-encoded protein products are translated, the sustained signal can lead directly to phosphorylation of the IEG products and can prolong their expression and activity in the nucleus for several hours. If signaling is transient, the IEG products are unstable and are degraded by the proteasome. Abbreviations: ER, endoplasmic reticulum; GAP, GTPase-activated protein; GEF, guanine nucleotide exchange factor; HePTP, hematopoietic PTP; MKP, MAPK phosphatase; PTP, protein tyrosine phosphatase; PTP-SL, STEP-like PTPase; STEP, striatal enriched PTP.

Importantly, EGF induces potent transient activation of the ERK pathway, whereas insulin causes weak activation (Box 1). Overexpression of either EGF receptors or insulin receptors, however, results in sustained ERK activation and differentiation [11,12]. In support of a hypothesis based on receptor signal strength, clonal isolates of PC12 cells expressing lower numbers of NGF receptors do not undergo differentiation in response to NGF [13]. Similarly, we have shown that

Box 1. Unanswered questions † What causes the rapid inactivation of ERK signaling? † What is the significance of Ras/ERK signaling at the ER and Golgi? † What is the role of MKPs in controlling ERK activity during the cell cycle? † As signal strength increases, what pathways become engaged that contribute to sustained ERK activity? www.sciencedirect.com

changing the amount of receptor occupancy by lowering agonist concentrations (thereby lowering signal strength) also alters the duration of ERK signaling and the proliferative response in fibroblasts [14]. Furthermore, a frequent hallmark of tumor-derived cells is the overexpression and/or amplification of cell-surface growth factor receptors, resulting in the aberrant activation of downstream signaling in the absence of extracellular ligands. In addition to the impact of receptor density on ERK signal strength and duration, the rate and extent of receptor internalization also contribute to ERK signaling. Thus, not only is receptor internalization an important control point for signal termination, but it can also result in further signaling by the internalized receptor–ligand complex from a different cellular location. Receptor internalization and trafficking are best understood for members of the EGF receptor family [15].

270

Review

TRENDS in Biochemical Sciences

The EGF receptors ErbB1, ErbB2, ErbB3 and ErbB4 show basal levels of internalization into endosomes, but only the EGF receptor shows accelerated internalization after ligand binding [16]. Ligand-bound EGF receptors are internalized via the dynamin/clathrin-mediated endocytic pathway into early endosomes, where they are either recycled back to the plasma membrane or processed for lysosomal degradation (Figure 1). The process underlying the decision to degrade or to recycle is not well understood, but it might be determined by the nature of the ligand that is bound. For example, receptors associated with transforming growth factor-a are recycled back to the plasma membrane, whereas those associated with EGF are destined for lysosomal degradation [17]. Ligand-bound EGF receptors internalize into endosomes at a much faster rate than unbound receptors [16], but the key events that accelerate internalization of EGF receptors are poorly understood. Several studies have indicated that the E3 ubiquitin ligase c-Cbl binds to and monoubiquitinates the C terminus of the EGF, plateletderived growth factor (PDGF), hepatocyte growth factor and colony-stimulating factor 1 receptors, thereby promoting the trafficking of these receptors to the lysosome [16,18]. The ErbB2 receptor does not bind c-Cbl, which could explain both its tendency to be recycled and its oncogenicity. The signals that control the choice between receptor recycling and degradation are unknown, but are likely to be chief determinants of signal duration, location and specificity. Internalized activated receptors can also signal from the endosome. For example, expression of an endocytosisdefective dynamin mutant that blocks internalization of the EGF receptor has been found to prevent maximal EGF-induced ERK activation [19]. Recent observations show that internalized EGF, PDGF and insulin receptors colocalize with the endoplasmic reticulum (ER)-associated protein tyrosine phosphatase 1B (PTP1B), and this colocalization acts as the inactivation step before receptor degradation [20–22]. EGF receptors destined for lysosomal degradation are trafficked to late endosomes, but even in these structures the active receptors might still be able to regulate ERK signaling. The ERK scaffold protein MP1, which is a partner protein of MAPK kinase 1 (MEK1), is recruited to late endosomes by the adaptor p14 and can regulate ERK activation during sustained signaling [23,24].

Ras The GTPase Ras is activated downstream of many receptor tyrosine kinases and is essential for ERK activation. All three Ras isoforms (N-Ras, K-Ras and H-Ras) are expressed ubiquitously and are modified by farnesyl groups, resulting in their association with membranes. Until recently, it was widely assumed that the active GTP-bound Ras was located mainly at the plasma membrane in close proximity to activated cellsurface receptors. Several studies indicate, however, that this model is too simplistic and that specific Ras isoforms have an important role in spatial and temporal ERK signaling. www.sciencedirect.com

Vol.31 No.5 May 2006

A specific cysteine residue in the C terminus of each Ras isoform is farnesylated in the cytosol; the farnesylated Ras protein then associates with the ER, where the modified cysteine is methylated [25]. N- and H-Ras are additionally mono- and di-palmitoylated, respectively, at neighboring cysteines and transported to the plasma membrane via the secretory pathway. K-Ras bypasses the secretory pathway and integrates into the plasma membrane via a polybasic sequence adjacent to its farnesylated cysteine. Although all three Ras isoforms bind GTP and become activated at the plasma membrane within minutes of growth factor treatment, activation of H-Ras can also take place on the ER and Golgi in fibroblasts, PC12 and Jurkat T cells [26,27]. Interestingly, activation of H-Ras on the plasma membrane is more rapid than activation of Golgi-localized H-Ras, which is delayed but sustained by comparison [26]. In contrast to these studies, activation of glutamate receptors on hippocampal neurons results in a rapid redistribution of K- but not H-Ras from the plasma membrane to intracellular membranes including the Golgi [28]. Palmitoylation of N- and H-Ras can be reversed, resulting in their release from the plasma membrane and their relocalization to Golgi membranes [29]. This retrograde transport of N- and H-Ras occurs irrespective of their activation state [29], but the exact components of the retrograde pathway are currently unknown [30]. Interestingly, Src signaling via phospholipase Cg can activate H-Ras at the Golgi in a process that is independent of receptor tyrosine kinase activation at the plasma membrane [27]. Perhaps because of their proximity to the nucleus, the Golgi and ER signaling platforms are uniquely positioned to promote efficient ERK-dependent transcriptional responses long after the activation of cell-surface receptors (Box 1). Notably, in contrast to K-rasK/K mice, doubly deficient N-rasK/K/H-rasK/K mice develop normally and are viable [31–33]. Thus, assuming that there is no compensation from other Ras molecules in the N-rasK/K/H-rasK/K mice, these observations suggest that Ras signaling from the ER and Golgi is dispensable for normal development. Clearly, our understanding of the spatial and temporal aspects of Ras signaling remains limited. In an alternative view of how Ras family proteins lead to variations in ERK activation kinetics, different Ras pathways could couple to distinct receptors. Data obtained from PC12 cells suggest that the Ras relative Rap1b shows sustained activation after treatment with NGF but not EGF [34,35], whereas activation of Ras is transient with EGF or NGF. Thus, a combination of transient Ras and sustained Rap1b activation could account for the generation of distinct ERK kinetics. NGF-regulated sustained activation of ERK signaling seems to be mediated by NGFreceptor-specific recruitment of the adaptor FRS2 and the Rap exchange factor C3G [34]. The general role of Rap1b in mediating sustained ERK signaling remains, however, controversial [36]. Indeed, if Rap1b couples to the NGF receptor, then it is unclear why reducing levels of the NGF receptor results in transient activation of ERK signaling [13]. We propose that receptor density, in combination

Review

TRENDS in Biochemical Sciences

with recruitment of Rap1b or a similar regulator, could be responsible for generating distinct signal kinetics in PC12 cells. Modular interactions proximal to ERK Recent studies have established that activation of ERK1/2, phosphorylation of their downstream targets and their subsequent inactivation by phosphatases require distinct types of protein–protein or docking interaction. For MEK, RSK and some phosphatases, binding to ERK requires a ‘docking’ (D) domain class of docking site [37]; by contrast, for several physiological substrates, many of which are nuclear, binding to ERK requires a ‘docking site for ERK, [Phe/Tyr]-Xaa-[Phe/Tyr]Pro’ (DEF) domain [38,39]. Whereas D domains bind to a region in ERK known as the ‘common docking’ (CD) domain [40], DEF domains bind to a set of residues that are proximal to the catalytic pocket in ERK [41]. An additional region in ERK, called the kinase interaction motif (KIM), is also important for binding to MEK [42]. Although not yet proved, it is likely that the binding of various D-domain and DEF-domain partner proteins to ERK is in constant equilibrium and thus can be altered by changes in localization and/or the subcellular concentration of ERK and its partner proteins (see later). Consistent with this notion, the binding of RSK1/2 to ERK1/2 before growth factor stimulation is disrupted after activation and relocation of ERK from the cytoplasm to the plasma membrane and then to the nucleus, which exposes the CD domain for a new round of interaction with other D-domain targets [43,44]. Importantly, binding of ERK to D and DEF domains is not mutually exclusive [45] and several ERK targets such as ELK-1 and JunD contain both classes of docking site [46,47], suggesting that an ERK dimer can engage these domains simultaneously. The in vivo role of ERK signaling via D and DEF domains has been recently investigated [45]. Although early studies suggested that the CD domain and the KIM mediated efficient ERK activation [40,42], it is now known that the CD domain is not required in vivo [45]. It is, however, required for efficient activation of RSK in vivo [45], a finding consistent with previous results [43,48,49]. The induction of ERK target genes such as egr-1, and the phosphorylation of c-Fos, takes place independently of D-domain interactions but requires ERK signaling to DEF domains [45]. Importantly, it is possible to inhibit selectively either mode of docking without disrupting ERK-mediated phosphotransfer. As mentioned, phosphatases containing D domains can bind directly to ERK and can dephosphorylate threonine and tyrosine residues in its activation loop. These enzymes, known as MAPK phosphatases (MKPs) are members of the dual specificity phosphatase family of PTPs [50]. Some PTPs, such as striatal enriched PTP (STEP) and hematopoietic PTP (He-PTP), can also bind to ERK but dephosphorylate only phosphotyrosine [50]. The significance of mono- versus di-dephosphorylation might be that negative regulators such as He-PTP generate a primed pool of ERK that is preferentially activated by MEK. If D-domain-containing phosphatases are required for inactivation of ERK signaling, then disrupting this www.sciencedirect.com

Vol.31 No.5 May 2006

271

interaction should result in prolonged ERK activation; however, only subtle changes in the inactivation kinetics of ERK have been observed in vivo ([40,51]; J. MacKeigan and J. Blenis, unpublished). It is possible that, in addition to the CD domain that binds phosphatases, ERK has other interaction surfaces for these regulators – the KIM, for example – that can compensate for a loss of CD interactions. Although it has been recognized for many years that MKPs such as MKP1/2 are encoded by IEGs that are transcriptionally upregulated in response to growth factors [52,53], it is unclear why ERK signaling is not completely inactivated in these situations [39] (Box 1). A likely explanation for this observation is that the strength of the MKP1/2 negative feedback is outbalanced by the still active upstream inputs to ERK. Because overactivation of Ras/ERK signaling in some contexts can result in senescence, cell-cycle arrest and/or apoptosis [54–56], it is also possible that induction of MKP1/2 might dampen ERK activity during the G1/S transition. In support of this idea, treatment with cycloheximide results in hyperactivation of ERK signaling 1–2 h after initial growth factor stimulation, indicating that de novo synthesis of gene products such as MKP1/2 is involved in reducing the strength and duration of ERK signaling [57]. It is important to remember, however, that in situations in which growth factors induce extremely transient activation of ERK signaling [39,58], the MKP1/2 gene products appear after signaling is terminated, indicating that these factors are not the key molecules that promote rapid inactivation. Furthermore, because MKP1/2 are localized to and dephosphorylate ERK in the nucleus [59,60], they are unlikely to be involved in the mechanisms that inactivate cytoplasmic ERK signaling. In these situations, constitutively expressed MKP3 might be responsible for inactivating ERK signaling. MKP3 shows selectivity for ERK1/2 over the stress-activated MAPKs [50], and its specific activity increases after docking to ERK1/2 [61]. In addition to the MKPs, the ratio of cytoplasmic to nuclear active ERK can be regulated by genes encoding Sef (similar expression to Fgf) [62], a transmembrane protein that is thought to localize to the Golgi. Sef prevents dissociation of the MEK–ERK complex, thereby inhibiting translocation of ERK to the nucleus; however, ERK can still signal to cytoplasmic targets [63]. Notably, additional negative feedback control of ERK activity can be mediated by the Sprouty family of proteins [64], as has been recently reviewed elsewhere [65]. In some cases, the docking of ERK to binding partners containing D domains can be regulated by phosphorylation in or around the D domain. For example, disruption of ERK binding to RSK1/2 occurs through ERK-mediated phosphorylation of the Ser749 residue that lies proximal to the C-terminal D domain in RSK, and mutation of this residue confers constitutive binding [44]. In T cells, ERK is maintained in an inactive state owing to its constitutive binding to the He-PTP phosphatase. On phosphorylation of the D domain in He-PTP by protein kinase A, however, ERK is released and is activated by MEK [66]. In primary striatal neurons, the STEP phosphatase is inactive owing

Review

TRENDS in Biochemical Sciences

to phosphorylation of its D domain. After treatment with glutamate, which activates N-methyl D -aspartate receptors and leads to an increase in intracellular Ca2C, STEP is rapidly dephosphorylated by the Ca2C-dependent phosphatase calcineurin and then physically interacts with ERK1/2, resulting in dephosphorylation of its activation-loop tyrosine [67]. Interestingly, STEP has been reported to localize to the ER, lending further support to the idea that sustained ERK signaling emanates from intracellular membrane compartments and not the plasma membrane. Downstream sensing of transient and sustained signaling Although the importance of signal duration and strength is generally accepted, we have only recently gained a detailed understanding of how the dynamics of ERK signaling are interpreted at the molecular level. It was previously assumed that nuclear signaling by ERK would underlie signal specificity; however, growth factors that trigger transient and sustained signaling induce similar patterns of IEG expression [39,68–70]. Thus, differences in the initial transcription program alone cannot account for cell fate decisions that occur downstream of ERK. Crucially, the expression kinetics of IEG-encoded protein products, such as c-Fos, is marked in response to agonists that induce sustained rather than transient ERK and RSK activation [14,39]. c-Fos protein is very unstable (half-life, t1/2z30 min) and will accumulate only if its C terminus is phosphorylated under conditions of sustained ERK activation (Figure 2). Because c-Fos is an important component of the dimeric AP-1 transcription factor, an increase in its stability results in greater promoter occupancy and expression of target genes [39] (L.O. Murphy and J. Blenis, unpublished) and in cellular transformation [71,72]. Accordingly, expression of the late-response gene Fra-1, a target of c-Fos [73], is sustained only after prolonged expression of c-Fos [14]. Thus, the behavior of c-Fos after transient or sustained ERK signaling enables the cell to distinguish among agonists that induce different activation kinetics. Therefore, by its ability to exist in unstable and stabilized states, the c-Fos transcription factor can function as a ‘sensor’ for ERK activation dynamics. In addition to increasing stability, the C-terminal modification of c-Fos enhances the process of ERKmediated phosphorylation of Thr325 and Thr331 within its C terminus [39]. These phosphorylation events are dependent on a DEF domain in c-Fos, and the integrity of this domain is required for cellular transformation [39]. Thus, ERK- and RSK-regulated phosphorylation at the extreme C terminus serves to ‘prime’ c-Fos for ERK docking and for further phosphorylation at Thr325 and Thr331 (Figure 2). Interestingly, the priming phosphorylation sites are located in a region of c-Fos that has been characterized as a transrepression domain [74], and they are replaced by an irrelevant sequence in the oncogenic homolog p55 v-Fos. Although v-Fos has the same stability as phosphorylated c-Fos [71,72,75], it does not contain a functional DEF domain because this is also replaced by an www.sciencedirect.com

Vol.31 No.5 May 2006

P

P

c-Fos (aa300–380) FTYP ERK

P P

ERK,RSK

Docking Priming

ERK activity

272

Transformation

Sustained signaling

Fos, FosB, Fra-2, Jun IEG products

Myc, JunB, Fra-1 Myc, Fra-1

G0

S

Signal propagation Ti BS

Figure 2. Signal propagation by IEG-encoded protein products. During sustained ERK signaling, C-terminal phosphorylation of c-Fos by ERK and RSK increases the stability of c-Fos during G1 and enables ERK to dock at the DEF domain and to regulate phosphorylation of Thr325 and Thr331 in c-Fos [39]. Other IEG products such as Fra-2 and c-Myc are similarly regulated by persistent ERK signaling, enabling propagation and interpretation of the sustained activity [14].

irrelevant sequence (L.O. Murphy and J. Blenis, unpublished). On the basis of these observations, we propose that the DEF domain and phosphorylation of Thr325 and Thr331 are required to drive c-Fos into an activated state (during sustained signaling) and that the unique C terminus in v-Fos adopts a conformation that mimics the same activated state. Notably, unless it is constitutively expressed, the c-Fos sensor is never exposed to the maximal level of ERK activity detected in cells immediately after the addition of growth factors, because endogenous c-Fos is expressed minimally in most nongrowing cells during the initial phase of ERK activation. As c-Fos accumulates, however, it becomes hyperphosphorylation during the submaximal phase of ERK signaling in G1. We suspect that the priming phosphorylation sites act as a gating mechanism to enable efficient hyperphosphorylation via the DEF domain only if and when signaling reaches a threshold strength and duration. An apparent problem with this model is that fibroblasts that are deficient in c-fos proliferate normally [76]. This paradox is more apparent than real, however, because additional Fos-like IEG-encoded sensors of ERK can compensate for the lack of c-Fos expression [14]. Notably, fibroblasts deficient in both c-fos and fosB are compromised in their ability to undergo DNA synthesis [77].

Review

TRENDS in Biochemical Sciences

Vol.31 No.5 May 2006

273

speculate that the switch-like behavior seen in Xenopus oocytes is conserved in this pathway but is located at the level of IEG induction and/or phosphorylation [78] (Figure 3). The identification of signaling mechanisms that control the ERK–IEG switch-like behavior deserves attention (Box 1).

Although FosB does not have a DEF domain motif, it contains similar C-terminal priming residues and a phospho-mimetic aspartic acid in a position identical to Thr325 in c-Fos, which drives transformation [39]. In addition to FosB, the Fos-related transcription factors Fra-1 and Fra-2, and other IEG products, contain functional DEF domains and associated phosphorylation sites and, similar to c-Fos, propagate signals after growth factor treatment [14]. Notably, the induction kinetics of IEG products such as c-Fos, c-Myc, JunB and Fra-1 is nonidentical, resulting in a cascade of potential IEG-encoded ERK sensors that coordinates progression through G1 (Figure 2). Importantly, not all ‘sustained’ ERK signals are created equally. In response to different agonists, ERK signals sustained 4 h after stimulation can be quantitatively different (e.g. 20% versus 40% maximal), leading to differential effects on the expression, phosphorylation and biological activity of IEG sensors during G1. This observation might explain, in part, why some mitogens are more potent than others in triggering DNA synthesis even though they induce similar IEGs. Thus, the sensitivity that IEG-encoded sensors show for ERK signal strength and duration, coupled with the fact that the transcriptional induction kinetics of IEGs is staggered, represents one of the principal specificity determinants downstream of ERK signaling (Figure 2). Although in mammalian cells ERK activation is graded in response to an increasing stimulus [78,79], induction and hyperphosphorylation of the ERK sensor c-Fos is ultrasensitive [78]. In contrast to mammalian cells, ERK activation in Xenopus oocytes can be bistable or irreversible owing to a strong positive feedback loop regulated by the Mos kinase [80]. Although the lack of a similar feedback loop in fibroblasts explains why ERK activity is graded in mammalian cells, we

Concluding remarks Sustained activation of ERK signaling is involved in cellcycle progression, cellular transformation and differentiation. The stabilization of IEG-encoded transcription factors takes place downstream of sustained signaling and results in an increase in target gene expression. Stabilization of IEG products containing DEF domains also might indirectly affect the dynamics of nuclear and/or cytoplasmic ERK signaling. The localization and subcellular concentration of ERK will dictate with which of the many binding partners the kinase will interact. Similarly, an increase in concentration of a binding partner, such as an IEG-encoded DEF-domain-containing sensor, will shift the equilibrium towards DEF binding and away from other classes of partner such as D-domaincontaining phosphatases. We propose that an increase in the abundance of nuclear IEG-encoded DEF domain targets after ERK activation will result in nuclear retention of active ERK, via DEF binding, and will sustain its activity in the nucleus by protecting it from nuclear MKPs (Figure 3). In support of this model, inhibition of protein synthesis with cycloheximide, which inhibits the expression of IEG products, decreases nuclear localization of ERK in serum-stimulated CCL39 fibroblasts [81]. Thus, we propose that IEG-encoded sensors have the ability not only to interpret ERK signal strength and duration but Ras

ERK

IEG induction ERK-mediated positive feedback to IEGs • IEG product phosphorylation • Stabilization of IEG product • Ultrasensitivity

+

+

DEF–domain-mediated positive feedback to ERK • ERK nuclear retention • Protection from MKPs

IEG products

IEG products stabilized

Biological outcome Figure 3. Dynamics of ERK activation and IEG expression during sustained signaling. Two possible feedback loops exist between the ERK pathway and IEG-encoded protein products. In the first, signal strength and duration are interpreted by IEG products such as c-Fos and translated into ultrasensitive changes in the stability of IEG products [78]. In the second, stabilized nuclear IEG products retain activated ERK in the nucleus via their DEF domains and promote sustained nuclear signaling by ERK. www.sciencedirect.com

274

Review

TRENDS in Biochemical Sciences

also to feedback positively to ERK by sustaining its presence and signaling in the nucleus. Recent advances have revealed mechanisms that control temporal and spatial aspects of ERK signaling by modulating receptor activity, localization of Ras, Raf and phosphatases, and signaling in the nucleus where many ERK effectors are located. Differences in ERK kinetics and localization are then interpreted by several IEG products with DEF domains that physically interact with activated ERK. Many cancers are known to have manipulated these pathway control points to their advantage. In particular, activating mutations in Raf and overexpression of IEGs such as c-fos, c-jun and c-myc are a hallmark of several malignancies. In such cases, therapeutic targeting of the ERK/DEF domain signaling axis could be beneficial. Acknowledgements We thank Ben Neel, Chris Dimitri, Greg Hoffman and Jeff MacKeigan for critical comments on the manuscript. This work was supported by a grant from the National Institutes of Health (RO1CA46595).

References 1 Roux, P.P. and Blenis, J. (2004) ERK and p38 MAPK-activated protein kinases: a family of protein kinases with diverse biological functions. Microbiol. Mol. Biol. Rev. 68, 320–344 2 Sharrocks, A.D. (2001) The ETS-domain transcription factor family. Nat. Rev. Mol. Cell Biol. 2, 827–837 3 Marshall, C.J. (1995) Specificity of receptor tyrosine kinase signaling: transient versus sustained extracellular signal-regulated kinase activation. Cell 80, 179–185 4 Cowley, S. et al. (1994) Activation of MAP kinase kinase is necessary and sufficient for PC12 differentiation and for transformation of NIH 3T3 cells. Cell 77, 841–852 5 Mansour, S.J. et al. (1994) Transformation of mammalian cells by constitutively active MAP kinase kinase. Science 265, 966–970 6 Whalen, A.M. et al. (1997) Megakaryocytic differentiation induced by constitutive activation of mitogen-activated protein kinase kinase. Mol. Cell. Biol. 17, 1947–1958 7 Dustin, M.L. and Chan, A.C. (2000) Signaling takes shape in the immune system. Cell 103, 283–294 8 Sharp, L.L. et al. (1997) The influence of the MAPK pathway on T cell lineage commitment. Immunity 7, 609–618 9 Aoki, Y. et al. (2005) Germline mutations in HRAS proto-oncogene cause Costello syndrome. Nat. Genet. 37, 1038–1040 10 Solit, D.B. et al. (2005) BRAF mutation predicts sensitivity to MEK inhibition. Nature 439, 358–362 11 Dikic, I. et al. (1994) PC12 cells overexpressing the insulin receptor undergo insulin-dependent neuronal differentiation. Curr. Biol. 4, 702–708 12 Traverse, S. et al. (1994) EGF triggers neuronal differentiation of PC12 cells that overexpress the EGF receptor. Curr. Biol. 4, 694–701 13 Schlessinger, J. and Bar-Sagi, D. (1994) Activation of Ras and other signaling pathways by receptor tyrosine kinases. Cold Spring Harb. Symp. Quant. Biol. 59, 173–179 14 Murphy, L.O. et al. (2004) A network of immediate early gene products propagates subtle differences in mitogen-activated protein kinase signal amplitude and duration. Mol. Cell. Biol. 24, 144–153 15 Le Roy, C. and Wrana, J.L. (2005) Clathrin- and non-clathrinmediated endocytic regulation of cell signalling. Nat. Rev. Mol. Cell Biol. 6, 112–126 16 Wiley, H.S. (2003) Trafficking of the ErbB receptors and its influence on signaling. Exp. Cell Res. 284, 78–88 17 Yarden, Y. (2001) The EGFR family and its ligands in human cancer. signalling mechanisms and therapeutic opportunities. Eur. J. Cancer 37(Suppl. 4), S3–S8 18 Peschard, P. and Park, M. (2003) Escape from Cbl-mediated downregulation: a recurrent theme for oncogenic deregulation of receptor tyrosine kinases. Cancer Cell 3, 519–523 www.sciencedirect.com

Vol.31 No.5 May 2006

19 Vieira, A.V. et al. (1996) Control of EGF receptor signaling by clathrinmediated endocytosis. Science 274, 2086–2089 20 Boute, N. et al. (2003) Dynamics of the interaction between the insulin receptor and protein tyrosine-phosphatase 1B in living cells. EMBO Rep. 4, 313–319 21 Romsicki, Y. et al. (2004) Protein tyrosine phosphatase-1B dephosphorylation of the insulin receptor occurs in a perinuclear endosome compartment in human embryonic kidney 293 cells. J. Biol. Chem. 279, 12868–12875 22 Haj, F.G. et al. (2002) Imaging sites of receptor dephosphorylation by PTP1B on the surface of the endoplasmic reticulum. Science 295, 1708–1711 23 Wunderlich, W. et al. (2001) A novel 14-kilodalton protein interacts with the mitogen-activated protein kinase scaffold MP1 on a late endosomal/lysosomal compartment. J. Cell Biol. 152, 765–776 24 Teis, D. et al. (2002) Localization of the MP1–MAPK scaffold complex to endosomes is mediated by p14 and required for signal transduction. Dev. Cell 3, 803–814 25 Bivona, T.G. and Philips, M.R. (2003) Ras pathway signaling on endomembranes. Curr. Opin. Cell Biol. 15, 136–142 26 Chiu, V.K. et al. (2002) Ras signalling on the endoplasmic reticulum and the Golgi. Nat. Cell Biol. 4, 343–350 27 Bivona, T.G. et al. (2003) Phospholipase Cg activates Ras on the Golgi apparatus by means of RasGRP1. Nature 424, 694–698 28 Fivaz, M. and Meyer, T. (2005) Reversible intracellular translocation of KRas but not HRas in hippocampal neurons regulated by Ca2C/calmodulin. J. Cell Biol. 170, 429–441 29 Rocks, O. et al. (2005) An acylation cycle regulates localization and activity of palmitoylated Ras isoforms. Science 307, 1746–1752 30 Meder, D. and Simons, K. (2005) Cell biology. Ras on the roundabout. Science 307, 1731–1733 31 Johnson, L. et al. (1997) K-ras is an essential gene in the mouse with partial functional overlap with N-ras. Genes Dev. 11, 2468–2481 32 Koera, K. et al. (1997) K-ras is essential for the development of the mouse embryo. Oncogene 15, 1151–1159 33 Esteban, L.M. et al. (2001) Targeted genomic disruption of H-ras and N-ras, individually or in combination, reveals the dispensability of both loci for mouse growth and development. Mol. Cell. Biol. 21, 1444–1452 34 Sasagawa, S. et al. (2005) Prediction and validation of the distinct dynamics of transient and sustained ERK activation. Nat. Cell Biol. 7, 365–373 35 York, R.D. et al. (1998) Rap1 mediates sustained MAP kinase activation induced by nerve growth factor. Nature 392, 622–626 36 Bos, J.L. (1998) All in the family? New insights and questions regarding interconnectivity of Ras, Rap1 and Ral. EMBO J. 17, 6776–6782 37 Tanoue, T. and Nishida, E. (2003) Molecular recognitions in the MAP kinase cascades. Cell. Signal. 15, 455–462 38 Jacobs, D. et al. (1999) Multiple docking sites on substrate proteins form a modular system that mediates recognition by ERK MAP kinase. Genes Dev. 13, 163–175 39 Murphy, L.O. et al. (2002) Molecular interpretation of ERK signal duration by immediate early gene products. Nat. Cell Biol. 4, 556–564 40 Tanoue, T. et al. (2000) A conserved docking motif in MAP kinases common to substrates, activators and regulators. Nat. Cell Biol. 2, 110–116 41 Lee, T. et al. (2004) Docking motif interactions in MAP kinases revealed by hydrogen exchange mass spectrometry. Mol. Cell 14, 43–55 42 Robinson, F.L. et al. (2002) Identification of novel point mutations in ERK2 that selectively disrupt binding to MEK1. J. Biol. Chem. 277, 14844–14852 43 Richards, S.A. et al. (2001) Characterization of regulatory events associated with membrane targeting of p90 ribosomal S6 kinase 1. Mol. Cell. Biol. 21, 7470–7480 44 Roux, P.P. et al. (2003) Phosphorylation of p90 ribosomal S6 kinase (RSK) regulates extracellular signal-regulated kinase docking and RSK activity. Mol. Cell. Biol. 23, 4796–4804 45 Dimitri, C.A. et al. (2005) Spatially separate docking sites on ERK2 regulate distinct signaling events in vivo. Curr. Biol. 15, 1319–1324

Review

TRENDS in Biochemical Sciences

46 Vinciguerra, M. et al. (2004) Differential phosphorylation of c-Jun and JunD in response to the epidermal growth factor is determined by the structure of MAPK targeting sequences. J. Biol. Chem. 279, 9634–9641 47 Fantz, D.A. et al. (2001) Docking sites on substrate proteins direct extracellular signal-regulated kinase to phosphorylate specific residues. J. Biol. Chem. 276, 27256–27265 48 Gavin, A.C. and Nebreda, A.R. (1999) A MAP kinase docking site is required for phosphorylation and activation of p90rsk/MAPKAP kinase-1. Curr. Biol. 9, 281–284 49 Smith, J.A. et al. (1999) Identification of an extracellular signalregulated kinase (ERK) docking site in ribosomal S6 kinase, a sequence critical for activation by ERK in vivo. J. Biol. Chem. 274, 2893–2898 50 Tonks, N.K. and Neel, B.G. (2001) Combinatorial control of the specificity of protein tyrosine phosphatases. Curr. Opin. Cell Biol. 13, 182–195 51 Bott, C.M. et al. (1994) The sevenmaker gain-of-function mutation in p42 MAP kinase leads to enhanced signalling and reduced sensitivity to dual specificity phosphatase action. FEBS Lett. 352, 201–205 52 Charles, C.H. et al. (1993) The growth factor-inducible immediateearly gene 3CH134 encodes a protein-tyrosine-phosphatase. Proc. Natl. Acad. Sci. U. S. A. 90, 5292–5296 53 Lau, L.F. and Nathans, D. (1985) Identification of a set of genes expressed during the G0/G1 transition of cultured mouse cells. EMBO J. 4, 3145–3151 54 Woods, D. et al. (1997) Raf-induced proliferation or cell cycle arrest is determined by the level of Raf activity with arrest mediated by p21Cip1. Mol. Cell. Biol. 17, 5598–5611 55 Sewing, A. et al. (1997) High-intensity Raf signal causes cell cycle arrest mediated by p21Cip1. Mol. Cell. Biol. 17, 5588–5597 56 Lundberg, A.S. et al. (2000) Genes involved in senescence and immortalization. Curr. Opin. Cell Biol. 12, 705–709 57 Cook, S.J. et al. (1997) Regulation of mitogen-activated protein kinase phosphatase-1 expression by extracellular signal-related kinasedependent and Ca2C-dependent signal pathways in Rat-1 cells. J. Biol. Chem. 272, 13309–13319 58 Chung, J. et al. (1991) Coordinate regulation of pp90rsk and a distinct protein-serine/threonine kinase activity that phosphorylates recombinant pp90rsk in vitro. Mol. Cell. Biol. 11, 1868–1874 59 Volmat, V. et al. (2001) The nucleus, a site for signal termination by sequestration and inactivation of p42/p44 MAP kinases. J. Cell Sci. 114, 3433–3443 60 Brondello, J.M. et al. (1995) Constitutive MAP kinase phosphatase (MKP-1) expression blocks G1 specific gene transcription and S-phase entry in fibroblasts. Oncogene 10, 1895–1904 61 Camps, M. et al. (1998) Catalytic activation of the phosphatase MKP-3 by ERK2 mitogen-activated protein kinase. Science 280, 1262–1265 62 Furthauer, M. et al. (2002) Sef is a feedback-induced antagonist of Ras/MAPK-mediated FGF signalling. Nat. Cell Biol. 4, 170–174 63 Torii, S. et al. (2004) Sef is a spatial regulator for Ras/MAP kinase signaling. Dev. Cell 7, 33–44

Vol.31 No.5 May 2006

275

64 Hanafusa, H. et al. (2002) Sprouty1 and Sprouty2 provide a control mechanism for the Ras/MAPK signalling pathway. Nat. Cell Biol. 4, 850–858 65 Kim, H.J. and Bar-Sagi, D. (2004) Modulation of signalling by Sprouty: a developing story. Nat. Rev. Mol. Cell Biol. 5, 441–450 66 Saxena, M. et al. (1999) Crosstalk between cAMP-dependent kinase and MAP kinase through a protein tyrosine phosphatase. Nat. Cell Biol. 1, 305–311 67 Paul, S. et al. (2003) NMDA-mediated activation of the tyrosine phosphatase STEP regulates the duration of ERK signaling. Nat. Neurosci. 6, 34–42 68 Greenberg, M.E. et al. (1985) Nerve growth factor and epidermal growth factor induce rapid transient changes in proto-oncogene transcription in PC12 cells. J. Biol. Chem. 260, 14101–14110 69 Greenberg, M.E. and Ziff, E.B. (1984) Stimulation of 3T3 cells induces transcription of the c-fos proto-oncogene. Nature 311, 433–438 70 Fambrough, D. et al. (1999) Diverse signaling pathways activated by growth factor receptors induce broadly overlapping, rather than independent, sets of genes. Cell 97, 727–741 71 Chen, R.H. et al. (1996) Phosphorylation of c-Fos at the C-terminus enhances its transforming activity. Oncogene 12, 1493–1502 72 Okazaki, K. and Sagata, N. (1995) The Mos/MAP kinase pathway stabilizes c-Fos by phosphorylation and augments its transforming activity in NIH 3T3 cells. EMBO J. 14, 5048–5059 73 Schreiber, M. et al. (1997) Structure and chromosomal assignment of the mouse Fra-1 gene, and its exclusion as a candidate gene for OC (osteosclerosis). Oncogene 15, 1171–1178 74 Ofir, R. et al. (1990) Phosphorylation of the C terminus of Fos protein is required for transcriptional transrepression of the c-fos promoter. Nature 348, 80–82 75 Curran, T. et al. (1984) Viral and cellular Fos proteins: a comparative analysis. Cell 36, 259–268 76 Brusselbach, S. et al. (1995) Cell proliferation and cell cycle progression are not impaired in fibroblasts and ES cells lacking c-Fos. Oncogene 10, 79–86 77 Brown, J.R. et al. (1998) Fos family members induce cell cycle entry by activating cyclin D1. Mol. Cell. Biol. 18, 5609–5619 78 Mackeigan, J.P. et al. (2005) Graded mitogen-activated protein kinase activity precedes switch-like c-Fos induction in mammalian cells. Mol. Cell. Biol. 25, 4676–4682 79 Whitehurst, A. et al. (2004) Stimulus-coupled spatial restriction of extracellular signal-regulated kinase 1/2 activity contributes to the specificity of signal-response pathways. Mol. Cell. Biol. 24, 10145–10150 80 Ferrell, J.E., Jr. and Machleder, E.M. (1998) The biochemical basis of an all-or-none cell fate switch in Xenopus oocytes. Science 280, 895–898 81 Lenormand, P. et al. (1998) Growth factor-induced p42/p44 MAPK nuclear translocation and retention requires both MAPK activation and neosynthesis of nuclear anchoring proteins. J. Cell Biol. 142, 625–633

Elsevier.com – Dynamic New Site Links Scientists to New Research & Thinking Elsevier.com has had a makeover, inside and out. As a world-leading publisher of scientific, technical and health information, Elsevier is dedicated to linking researchers and professionals to the best thinking in their fields. We offer the widest and deepest coverage in a range of media types to enhance crosspollination of information, breakthroughs in research and discovery, and the sharing and preservation of knowledge. Visit us at Elsevier.com.

Elsevier. Building Insights. Breaking Boundaries www.sciencedirect.com