Effectors and signalling to the nucleus

A number of major breakthroughs in the 1980s and 1990s substantially increased our appreciation of the true versatility of signalling events initiated by GPCRs. The fact that different ion channels can act as direct GPCR effectors (acting immediately downstream of Ga-GTP or G^y subunits) will be discussed in the next Chapter. Another key advance came with the appreciation that at least some members of the GPCR superfamily can affect cellular activity at the level of gene transcription (Collins et al. 1992). Here, the mitogen-activated protein (MAP) kinases will be considered as an example of a nuclear signalling pathway responsive to GPCR activation, and the regulation of the transcription factor cyclic AMP response element-binding protein (CREB) by cell-surface receptors (including GPCRs) will also be discussed in a neuronal context.

5.4.1 GPCRs and mitogen-activated protein (MAP) kinases

Many papers published over the past 10-15 years have brought together a compelling case for GPCRs possessing the potential to bring about phenotypic changes in the cell, including oncogenic transformation (Dhanasekaran et al. 1995; Gutkind 1998). Thus, activation of an array of GPCRs, which couple preferentially to different G proteins, has been shown to impact upon cell fate in a variety of cell-types (Gutkind 1998) and these effects can often be mimicked by expression of constitutively active Gsa, Gi a, Gqa, or G12/13 a mutant proteins (Dhanasekaran et al. 1995). An important, highly conserved, family of protein kinases, collectively termed the MAP kinases, have been established as important relays linking signals arriving at the cell-surface to cellular changes at the level of transcription and translation (Davis 1993; Schaeffer and Weber 1999). At the core of MAP kinase pathways is a cassette of three kinases acting in series (MAPKKK ^ MAPKK ^ MAPK; see Fig. 5.1). The most studied of these involves dual phosphorylation of extracellular signalregulated protein kinases (ERK1/2; also known as p44MAPK and p42MAPK, respectively) at threonine and tyrosine residues (in the TEY-motif) by the MAPKKs, MEK1/2, which in turn are dually phosphorylated at serine/threonine residues by the MAPKKK, Raf-1 (Kolch 2000). At least three additional MAP kinase pathways have been recognized in

RTK/ Neurotrophin receptor wevwfiww ¿yuyyififiMft

Grb2-mSOS

Ras-GDP

GPCR

RTK/ Neurotrophin receptor

GPCR

GPCR

RTK/ Neurotrophin receptor

GPCR

Grb2-mSOS

Ras-GRF CalDAG-GEFII

Crk, C3G

Ras-GDP

Ras-GTP

Ras-GRF CalDAG-GEFII

Ras-GTP

Three kinase module

Raf-1

Crk, C3G

Rap1-GTP

Rap1-GTP

Rap1-GDP

CAMP-GEFI/II CalDAG-GEFI

Rap1-GDP

Cytosolic effectors

Cytosolic effectors

Three kinase module

Cytoplasm

Cytoplasm

Nucleus

p90RSK

p90RSK

Nucleus

CREB

Fig. 5.1 Receptor-dependent regulation of ERK activity in neurons.

mammalian systems, the c-Jun N-terminal protein kinases (JNK1/2/3, also known as stress-activated protein kinases, SAPK1a/p/y), the p38 MAP kinases (p38a/p/y/8) and ERK5 (Schaeffer and Weber 1999; Kolch 2000). Here, we will focus on the pathways linking GPCRs to ERK1/2, firstly in a general context and then with specific reference to neuronal function.

ERK1/2 were initially shown to be downstream targets of growth factor receptor tyrosine kinases (RTKs) (Schlessinger 1993), capable of phosphorylating specific cell proteins, including a variety of transcription factors (Davis 1993; Kolch 2000). RTK activation leads to dimerization and autophosphorylation of specific tyrosine residues within the intracellular domains of the receptor to create docking sites for src-homology-2 (SH2) domain-containing proteins. One such protein, Grb2 (growth factor receptor binding protein-2) binds to specific intracellular tyrosine-phosphorylated sites (e.g. 740Y in the PDGF-ß receptor). The tethered Grb2 acts as an adaptor for the guanine nucleotide exchange factor (GEF), mSOS (mammalian 'Son-of-Sevenless' protein), with which it is constitutively associated. mSOS in turn recruits p21ras and facilitates GTP-loading of Ras leading to its activation. Ras-GTP binds to a number of proteins including Raf-1 leading to the translocation of this MAPKKK to the plasma membrane and activation of the Raf-1-MEK1/2-ERK1/2 cascade (Fig. 5.1). It is likely that endocytosis of the complex, in whole or in part, is a prerequisite for efficient ERK activation by at least some RTKs (Vieira et al. 1996). ERK1/2 activation leads to the proline-directed serine/threonine phosphorylation of specific cytoplasmic substrate proteins, and often to nuclear translocation, where ERK1/2 can directly affect transcription through direct phosphorylation of ternary complex factors (e.g. Elk-1), or indirectly through activation of other protein kinases (e.g. p90RSK—see next section) that, in turn, regulate transcription factor activity (Davis 1993; Treisman 1996).

Initial investigations of how GPCRs link to MAPK/ERK activation, which often involved transient expression of GPCRs, MAPKs and putative intermediary components in nonneuronal cell-lines, suggested that traditional G protein-effector-second messenger pathways might directly regulate components of the MAPK cascade in either a Ras-dependent or -independent manner (Alblas et al. 1993; Howe and Marshall 1993; Kolch et al. 1993; Offermanns et al. 1993; van Corven et al. 1993), with both Ga- and/or ßy-subunits being implicated as transducers of MAPK signalling (Gutkind 1998). For many cell systems it could be demonstrated that ßy-subunits play the major role in determining Ras-dependent ERK activity downstream of G;/o-coupled GPCRs, whereas both Gq/11a- and ßy-subunits have been implicated downstream of Gq/n-coupled GPCRs in mediating both Ras-dependent and -independent ERK activation (Crespo et al. 1994; Faure et al. 1994; Koch et al. 1994; Hawes et al. 1995).

Attempts to delineate the sequence of proteins linking GPCR/G protein activation to the MAPK/ERK pathway have yielded a wide range of potential signalling intermediates. A common feature of many of these proteins is their ability to act as adaptors linking GPCRs to the Ras activation sequence utilized by RTKs. Thus, a number of GPCRs have been shown to promote the phosphorylation of the adaptor protein Shc leading to Grb2-mSOS recruitment (Ohmichi etal. 1994; van Biesen etal. 1995; Chen etal. 1996). GPCRs have also been shown to 'transactivate' some RTKs (Daub etal. 1996; Luttrell etal. 1997; Herrlich etal. 1998). This process was initially considered to occur independently of ligand activation of the RTK, with the receptor undergoing ligand-independent phosphorylation (by non-receptor tyrosine kinases (PTKs)) to allow it to act as a signalling scaffold. However, more recent data have now shown that GPCR activation may stimulate metalloproteinases capable of releasing RTK ligands (Prenzel et al. 1999). In addition to RTKs acting as signalling 'scaffolds' for the GPCR-ERK pathway a number of other protein scaffolds have been identified. These include integrins (Schlaepfer et al. 1994; Slack 1998; Litvak et al. 2000; Short et al. 2000) and the GPCRs themselves.

An involvement of PTKs, such as Src (and other Src family proteins such as Yes, Fyn, and Lyn), Pyk2 and focal adhesion kinase (FAK) has also been shown for a number of G;/o- and Gq/11-coupled GPCRs (Dikic et al. 1996; Luttrell et al. 1996; Della Rocca et al. 1997). The linkage between G proteins and PTKs in some cases can be clearly defined; for example, Pyk2 is a Ca2+/calmodulin- and PKC-regulated enzyme (Lev et al. 1995; Dikic et al. 1996). A further (indirect) link between Gßy and PTKs has been shown for at least some members ofthe class 1 PI3Ks (Hawes etal. 1996; Lopez-Ilasaca etal. 1997), suggesting a Gßy-PI3K-Src sequence. Finally, a direct Gßy interaction has been proposed for some PTKs (Gutkind 1998). Another fascinating development has been the finding that some GPCRs may link to ERK signalling via a different G protein subpopulation to that utilized in their coupling to classical effectors. Thus, Daaka and colleagues have shown that ß2-adrenoceptors, expressed in HEK293 cells, couple to a Src- and Ras-dependent ERK pathway via ßy-subunits released by Gi protein activation. These data led to the proposal of a model where the ß2-adrenoceptor-Gs protein interaction becomes rapidly uncoupled and the receptor, instead of desensitizing with respect to all G protein interactions, 'switches' to activate Gi proteins (Daaka et al. 1997). It is noteworthy that there are a number of examples of GPCRs, considered to be Gq/11 or Gs-coupled that link to ERK via pertussis toxin-sensitive Gi/0 proteins (Ferraguti etal. 1999; Soeder etal. 1999), and therefore it is possible that ERK activation arising through GPCR switching may be a widespread phenomenon.

The finding that covalent modification of GPCRs, normally associated with receptor desensitization, could instead switch the GDP/GTP exchange activity of the receptor from one subpopulation of G proteins to another led to investigations of whether proteins previously considered only to be associated with the uncoupling of GPCRs (i.e. G protein-coupled receptor kinases and arrestins) might fulfil other functions (Table 5.1). Using the HEK293-ß2-adrenoceptor model system it was shown that the 'desensitized' ß2-adrenoceptor/ß-arrestin1 complex recruits Src (Luttrell et al. 1999). In addition, the clathrin-adaptor function of ß-arrestin was shown to be essential, as internalization of the ß2-adrenoceptor/ß-arrestin1-Src complex was necessary for ERK activation (Luttrell et al.

1999). A comparable GPCR/ß-arrestin2 scaffold has been reported to recruit the MAP-KKK/MAPKK/MAPK components ASK1/MKK4/JNK3 (McDonald etal. 2000). Endocytosis of the GPCR as a crucial aspect of ERK signalling has subsequently been shown for a subset of GPCRs that stimulate ERK activity (Pierce et al. 2001) and there appears to be receptor subtype variation in whether endocytosis of the entire signalling complex (DeFea et al.

2000), or only a terminal (MEK/ERK) portion (Kranenburg et al. 1999) is essential. As a variant on this theme, some GPCRs are reported to be able directly to recruit Src providing a scaffold independent of ß-arrestin for building ERK activation complexes, or to activate other signalling pathways (Cao et al. 2000; Heuss and Gerber 2000). Direct, G protein-independent GPCR recruitment of Src family PTKs is believed to depend on proline-rich motifs present in the 3rd intracellular loop or C-terminal tails of only a subset of GPCRs (e.g. D4-dopamine, mGlu1a and mGlu5a receptors (Pierce etal. 2001; Hermans and Challiss

The work outlined above has highlighted the fact that GPCR phosphorylation may switch G protein-effector signalling, rather than simply arresting it. Furthermore, the demonstration that ß-arrestin isoforms may act as scaffolds to assemble different proteins that regulate distinct MAPK pathways, together with the fact that GPCR/ß-arrestin isoform specificity has been reported (Penn et al. 2001), leads to the intriguing possibility that receptor phosphorylation and selective ß-arrestin recruitment may contribute to defining the (MAPK) pathways activated downstream of a particular GPCR. The ability of ß-arrestins to act as scaffolds for assembling specific MAPK signalling complexes is reminiscent of the role of Ste5 in S. cerevisiae (Schaeffer and Weber 1999), and further potential scaffolding proteins are being identified in mammalian systems, including JNK-interacting (JIP) and MEK partner (MP) proteins and MKK4, suggesting a further mechanism of imposing specificity and control to MAPK signalling (Hagemann and Blank 2001).

Despite the considerable progress that has been made in mapping the biochemical pathways GPCRs can utilize to couple to ERK activation, it has proved to be very difficult to draw up general rules. There are a number of reasons for this. First, the pathway linking a GPCR to ERK activation appears to be highly dependent on cell background and it has been shown on many occasions that it is possible for a specific GPCR to couple to ERK1/2 activation via different pathways in different cell-types (Duckworth and Cantley 1997; Della Rocca et al. 1999). A further difficulty arises from the fact that a single GPCR might activate multiple parallel or overlapping pathways capable of linking to ERK activation (Duckworth and Cantley 1997; Della Rocca et al. 1999; Blaukat et al. 2000) and these pathways may show an interdependency or redundancy with respect to both the magnitude and duration of the consequent ERK activation. In vivo it is likely that such apparent anarchy does not reign and instead specificity is achieved through a variety of mechanisms. These might include the expression of only a subset of intermediary components in a particular subtype and com-partmentation (e.g. through signalling complex assembly on component-defining scaffolds) within the cell.

5.4.2 GPCR and ERK signalling in the CNS

The biochemical elucidation of the RTK-Ras-ERK pathway (Schlessinger 1993) was quickly followed by demonstrations that other signalling events could also trigger its activation to regulate a wide range of cell outcomes in a many tissues, including the CNS (Ghosh and Greenberg 1995; Grewal et al. 1999). In parallel with studies showing ERK activation by GPCRs, studies in neuronal cell-lines (notably PC12 rat pheochromocytoma cells) demonstrated Ras/ERK activation by Ca2+-influx (through voltage-dependent or receptor-dependent mechanisms), suggesting that changes in neuronal excitability might also regulate this pathway (Rosen et al. 1994; Rusanescu et al. 1995; Finkbeiner and Greenberg 1996). Alongside this initial functional demonstration it was also shown that many of the components of the Ras/ERK pathway are enriched in the mammalian CNS, where they exhibit an unequal regional and sub-cellular distribution, and are expressed at highest levels in brain regions capable of undergoing long-term changes in synaptic efficacy (Finkbeiner and Greenberg 1996; Impey etal. 1999).

Although it has come to be generally accepted that MAPK/ERK signalling is crucial for an array of neuronal functions, and considerable progress has been made with respect to implicating ERK regulation by neurotrophins/growth factors, and ionotropic receptor- and voltage-dependent mechanisms, relatively few studies have been performed with respect to GPCR inputs to the ERK pathway in neurons (see for example Roberson et al. 1999; Watanabe et al. 2000). This is surprising as the progress that has been made strongly suggests that metabotropic mechanisms can directly regulate MAPK/ERK activation and therefore influence neuronal survival, differentiation and plasticity (Finkbeiner and Greenberg 1996; Impey et al. 1999; Sweatt 2001). For example, cyclic AMP has been shown to regulate the transcription factor Elk-1 by an ERK-dependent mechanism in PC12 cells (Vossler et al. 1997). The pathway leading to ERK activation involves the small GTPase Rap1 that preferentially activates the neuronal MAPKKK, B-Raf (Fig. 5.1). Subsequent work showed that PC12 cell differentiation, mediated by the neurotrophin, nerve growth factor (NGF), also requires ERK activation downstream of Rap1/B-Raf (York et al. 1998). Furthermore, the sustained nature of the ERK activation mediated by Rap1/B-Raf may 'programme' the cell to differentiate, rather than to proliferate (Marshall 1995).

Table 5.2 Guanine nucleotide exchange factors: links between neuronal GPCR-second messenger signalling and the ERK pathway

Guanine nucleotide exchange factor (GEF)

Upstream regulators

Small GTPase target

mSOS

Pyk2/Src/Shc/Grb2

Ras

Ras-GRF (CDC25Mm)

GPy, Ca2+/calmodulin

Ras

CalDAG-GEF-II (Ras-GRP)

DAG, Ca2+

Ras(Rap1)

CalDAG-GEF-III

Rap1, Rap2

C3G

FRS2, Crk cyclic AMP/PKA

Rap1

cAMP-GEF-I (Epac1)

cyclic AMP

Rap1

cAMP-GEF-II

cyclic AMP

Rap1

CalDAG-GEF-I

Ca2+, DAG

Rap1

Ras/Rap1 guanine nucleotide exchange factors (GEFs) reported to be expressed in the CNS. mSOS and C3G are regulated by growth factor/neurotrophin-regulated tyrosine kinase/adaptor activities. However, many GEFs are directly or indirectly regulated by changes in second messenger and/or Ca2+ concentration, providing a mechanism by which GPCR regulation of adenylyl cyclase/phospholipase C activities can influence ERK pathways downstream of Raf-1/B-Raf.

Ras/Rap1 guanine nucleotide exchange factors (GEFs) reported to be expressed in the CNS. mSOS and C3G are regulated by growth factor/neurotrophin-regulated tyrosine kinase/adaptor activities. However, many GEFs are directly or indirectly regulated by changes in second messenger and/or Ca2+ concentration, providing a mechanism by which GPCR regulation of adenylyl cyclase/phospholipase C activities can influence ERK pathways downstream of Raf-1/B-Raf.

Ras and Rap1 require GEFs to facilitate GTP for GDP exchange, and GTPase activating proteins (GAPs) to stimulate GTP hydrolysis (Fig. 5.1). Efforts to identify GEF and GAP activities in the CNS (and other tissues) have yielded a number of novel proteins (Grewal etal. 1999; Bos etal. 2001), many of which can be directly or indirectly regulated by mediators downstream of GPCRs (see Table 5.2). Thus, the neuronal Ras-GEF, Ras-GRF (also known as CDC25Mm) is directly regulated by Ca2+ /calmodulin and G^y-subunits (Farnsworth et al. 1995; Mattingly and Macara 1996) and transgenic knockout experiments have implicated it in different forms of synaptic plasticity, perhaps those especially requiring metabotropic receptor inputs (Brambilla et al. 1997). Another Ras-GEF, CalDAG-GEFII (also known as RasGRP) possesses, as its name implies, binding domains for both Ca2+ and DAG (Ebinu et al. 1998). In addition to the likely modulation by GPCRs of Ras-GTP loading, there is also the potential for regulation of Ras-GTPase activity. Thus, the brain-specific RasGAP, SynGAP is associated with post-synaptic density (PSD) complexes and is negatively regulated by Ca2+/calmodulin-dependent kinase II activity (Chen etal. 1998).

A number of Rap-GEFs and Rap-GAPs have also been reported, with many again showing a neuron-specific expression, and binding motifs that allow direct or indirect regulation by second messengers (Fig. 5.1; Table 5.2). The first Rap-GEF to be identified was C3G (Crk, SH3-domain-binding guanine-nucleotide releasing factor), which appears to play a similar role to mSOS in linking neurotrophin receptor activation to GTP-loading of Rap1, and is regulated upstream by FRS2-Crk adaptors (Kao etal. 2001). In contrast, cyclic AMP binding directly regulates the Rap1-GEFs, Epac1/cAMP-GEFI and cAMP-GEFII (de Rooij etal. 1998; Kawasaki et al. 1998) and these proteins are likely to account for cyclic AMP dependent, but PKA-independent activation of the Rap1/B-Raf/ERK pathway (Grewal et al. 1999; Bos et al. 2001). Finally, CalDAG-GEFI and CalDAG-GEFIII are directly regulated by Ca2+ and DAG and can facilitate guanine nucleotide exchange on Rap1 (and 2, and some other small GTPases; Bos et al. 2001). Although little is presently know about the regulation of neuronal Rap-GAP activities, there is evidence for G protein regulation of both Rap1-GAPI and Rap-GAPII (Jordan etal. 1999; Bos etal. 2001).

The identification of families of GEF/GAPs that regulate Ras/Rap1 GTP-loading/GTP hydrolysis through changes in Ca2+, cyclic AMP or DAG levels (Table 5.2) strongly suggests a signalling pathway through which Gs- and Gq-coupled GPCRs can regulate ERK activity in neurons. Emerging experimental data suggest that GPCRs are likely to utilize these pathways (Guo et al. 2001) and RTK/PTK-dependent pathways (Rosenblum et al. 2000; Peavy et al. 2001) to regulate ERK activity in the CNS.

0 0

Post a comment