RTK/Ras/MAPK signaling^*

Receptor Tyrosine Kinase (RTK)/Ras GTPase/MAP kinase (MAPK) signaling pathways are used repeatedly during metazoan development to control many different biological processes (Schlessinger, 2000). Mutations affecting RTK/Ras/MAPK signaling cause many human syndromes and diseases, including cancer (Malumbres and Barbacid, 2002). Studies in C. elegans have provided much insight into the basic framework of this canonical type of Ras pathway, its regulation, how it elicits cell-type specific responses, and how it interacts with other signaling pathways (for recent reviews see Moghal and Sternberg, 2003; Sundaram, 2004; Tan and Kim, 1999; Wang and Sternberg, 2001).

C. elegans Ras is called LET-60 (Han and Sternberg, 1990). LET-60 Ras acts downstream of at least two different RTKs, LET-23 (related to the Epidermal Growth Factor Receptor or EGFR; Aroian et al., 1990) and EGL-15 (related to the Fibroblast Growth Factor Receptor or FGFR; DeVore et al., 1995). The only known role of LET-60 Ras is to stimulate a MAPK cascade consisting of the kinases LIN-45 (Raf; Han et al., 1993), MEK-2 (MEK; Church et al., 1995; Kornfeld et al., 1995; Wu et al., 1995) and MPK-1 (ERK/MAPK; Lackner et al., 1994; Wu and Han, 1994). LET-60 Ras signaling is required for multiple developmental events, the best studied of which is vulval induction (see Vulval development). Genetic screens based on various let-60 mutant phenotypes have identified many generally-acting “core” components of the Ras pathway as well as numerous regulators or targets of the pathway (Figure 1; Table 1). Indeed, many important Ras pathway genes were first identified in the worm.

A general model for the core C. elegans RTK/Ras/MAPK signaling pathway (and various tissue-specific variations) is shown in Figure 1 (Moghal and Sternberg, 2003; Schlessinger, 2000). Upon growth factor binding, an RTK such as LET-23 or EGL-15 dimerizes and autophosphorylates its C-terminal region. The resulting phospho-tyrosine residues serve as docking sites for adaptor proteins such as SEM-5 (Grb2) or SOC-1 (similar to Gab1). These adaptors recruit the Guanine Nucleotide Exchange Factor SOS-1 to activate the small GTPase LET-60 Ras. LET-60-GTP then binds to LIN-45 Raf and promotes its stable association with the plasma membrane and/or endomembranes, where other events then activate LIN-45 kinase activity (Chong et al., 2003). The scaffold protein KSR may assist in LIN-45 activation, but also promotes further signal transmission by bringing together different components of the MAPK cascade (Morrison and Davis, 2003). LIN-45 phosphorylates and activates MEK-2, MEK-2 phosphorylates and activates MPK-1, and MPK-1 then phosphorylates and either activates or inactivates various target proteins. In many cases MPK-1 may move into the nucleus to phosphorylate transcription factors such as the Ets domain protein LIN-1, thus leading to changes in gene expression.

This model draws on a large body of data from multiple systems. C. elegans genetics has been most useful for identifying the genes involved in particular signaling events and determining their order of action and cellular focus. In some cases, physical interactions and/or phosphorylation events have been demonstrated for the worm proteins (e.g., Jacobs et al., 1999; Sieburth et al., 1998; Wu et al., 1995), but in many cases such interactions are inferred based on biochemical studies of related proteins in vertebrate cells. Because the core Ras pathway (and much of its regulation) appears highly conserved between C. elegans and vertebrates (Table 1), current models draw on the combined data from these different systems.

Ras signaling is not required for cellular viability in C. elegans (Yochem et al., 1997), but it is required for organismal viability and for many different developmental processes. Because of the widespread roles of Ras signaling during development, mutations affecting the Ras pathway can cause many different pleiotropic defects (Figure 2).

Lethality: Ras signaling promotes the excretory duct cell fate, and mosaic analysis suggested that loss of this one specific cell can account for the zygotic lethality of let-60 null mutants (Yochem et al., 1997). The excretory duct cell is required for osmoregulation (Nelson and Riddle, 1984). let-60 loss-of-function mutants, which lack the excretory duct cell, die as “rod-like” larvae with a fluid-filled appearance (Figure 2). let-60 gain-of-function mutants sometimes have two excretory duct cells (Yochem et al., 1997). Ras signaling may have other (maternally-rescued) essential roles in addition to excretory duct formation, since mutations in let-23/EGFR, egl-15/FGFR and ptp-2/SHP-2 cause distinct “scrawny” and lethal defects that are rescued by constitutive forms of LET-60 Ras (DeVore et al., 1995; Gutch et al., 1998; Koga and Ohshima, 1995).

Hyperactivation of egl-15/FGFR or let-60/Ras can also lead to lethality with a “Clear”, fluid-filled appearance (Kokel et al., 1998; Schutzman et al., 2001). Mosaic analysis suggests that the Clear phenotype is caused by hypodermal defects (Huang and Stern, 2004).

Vulval defects: Ras signaling cooperates with Notch and Wnt signaling to promote hermaphrodite vulval development (see Vulval development). let-60 loss-of-function mutants lack a vulva (Vulvaless or Vul phenotype; Figure 2) whereas let-60 gain-of-function mutants have extra vulval tissue (Multivulva or Muv phenotype).

Uterine defects: Ras signaling promotes the uterine uv1 fate, which is important for establishing a proper vulval-uterine connection (Chang et al., 1999). In the absence of uv1, hermaphrodites cannot lay eggs.

P11 and P12 ectoblast defects: Ras signaling cooperates with Wnt signaling to promote the P12 ectoblast cell fate (Fixsen et al., 1985; Jiang and Sternberg, 1998). The P11 and P12 ectoblasts are neighboring cells that give rise to slightly different types of hypodermal and neuronal descendants (Sulston and Horvitz, 1977; Hermaphrodite cell fate specification). Reduced Ras signaling causes P12-> P11 cell fate transformations (Figure 2), while increased signaling causes P11-> P12 fate transformations.

Male spicule defects: Ras signaling promotes male spicule fates (Chamberlin and Sternberg, 1994; Male development). Reduced Ras signaling causes spicule defects and prevents males from mating.

Sex myoblast migration defects: Ras signaling helps to specify the proper endpoint of sex myoblast migration (Sundaram et al., 1996). In let-60 loss-of-function mutants, sex myoblasts adopt a broadened range of final positions. This defect can affect egg-laying.

Axon guidance defects: Ras signaling controls the paths of certain ventral cord neurons relative to the ventral midline (Bulow et al., 2004). Whereas in wild-type these neurons extend axons along one side of the midline, in let-60 loss-of-function mutants the neurons wander across the midline.

Sterility due to pachytene exit (Pex) defects: Ras signaling is required for progression through the pachytene stage of meiosis (Church et al., 1995; Somatic sex determination). let-60 loss-of-function mutants are sterile because germ cells arrest in pachytene (Figure 2).

Olfaction defects: Ras signaling is required for sensitivity to volatile attractants (Hirotsu et al., 2000). Whereas wild-type C. elegans will chemotax toward volatile attractants such as isoamylalcohol and diacetyl, let-60 loss-of-function mutants fail to chemotax towards such attractants.

Resistance to Microbacterium nematophilum-induced swelling: ksr-1, lin-45, mek-2 and mpk-1 are required for the swelling response to infection by M. nematophilum (Nicholas and Hodgkin, 2004). Interestingly, let-60 does not appear to be important for this response, suggesting that a bacterial toxin may directly activate the Raf/MEK/ERK cascade.

The vast majority of studies on RTK/Ras/MAPK signaling have focused on the role of the Ras pathway in promoting vulval development (see Vulval development), and many known components of the pathway have been identified through forward genetic screens for mutants with Vul or Muv mutant phenotypes (Ferguson and Horvitz, 1985). Some components of the pathway have been identified based on other mutant phenotypes such as sex myoblast migration defects (Stern and Horvitz, 1991) or germline meiosis defects (Church et al., 1995; Ohmachi et al., 2002). Some have been identified through reverse genetic approaches (Dutt et al., 2004; Gutch et al., 1998; Kamikura and Cooper, 2003; Yoo et al., 2004). Finally, some core Ras pathway components as well as a large set of other genes that influence Ras signaling have been identified through genetic suppressor or enhancer screens (Sternberg and Han, 1998). Many modifier genes have essentially wild-type null phenotypes (Table 1), indicating a significant and surprising amount of redundancy amongst Ras pathway regulators.

One especially productive modifier screen involved looking for suppressors of the let-60 gain-of-function Muv phenotype. This screen identified alleles of the downstream kinase genes lin-45/Raf (Hsu et al., 2002), mek-2/MEK (Kornfeld et al., 1995; Wu et al., 1995) and mpk-1/ERK (Lackner et al., 1994; Wu and Han, 1994), revealing that LET-60 Ras signals through the MAPK cascade. This screen also identified numerous positive regulators of this cascade such as cdf-1 (Jakubowski and Kornfeld, 1999), ksr-1 (Kornfeld et al., 1995; Sundaram and Han, 1995), sur-6 (Sieburth et al., 1999), sur-7 (Yoder et al., 2004), and sur-8 (Sieburth et al., 1998; see below), as well as a key downstream factor, sur-2 (Singh and Han, 1995; see below). Another very productive type of screen has been to look for enhancers of lethal or vulval defects in the background of a very mildly affected or “phenotype-less” mutant. For example, screens for enhancers of lin-45 lethality identified the downstream factors eor-1 and eor-2 (Howard and Sundaram, 2002; Rocheleau et al., 2002), while screens for enhancers of the gap-1 Muv phenotype identified the MAPK phosphatase lip-1 (Berset et al., 2001). These types of genetic modifier screens have allowed the identification of genes with very subtle null phenotypes.

The power of genetic screening approaches in C. elegans is reflected in the large number of conserved genes whose involvement in RTK/Ras/MAPK signaling was first discovered in the worm; these include ark-1 (Hopper et al., 2000), cdf-1/ZnT1 (Jakubowski and Kornfeld, 1999), ksr-1 (Kornfeld et al., 1995; Sundaram and Han, 1995), lin-2/CASK, lin-7 and lin-10/Mint (Ferguson and Horvitz, 1985), sem-5/Grb2 (Clark et al., 1992), sli-1/Cbl (Jongeward et al., 1995), sur-2 (Singh and Han, 1995), sur-5 (Gu et al., 1998), sur-7 (Yoder et al., 2004), and sur-8/soc-2 (Selfors et al., 1998; Sieburth et al., 1998).

The C. elegans genome contains twenty-eight predicted RTKs (Popovici et al., 1999), only a few of which have been characterized mutationally (see Genomic overview of protein kinases). Of these characterized RTKs, only the EGF receptor LET-23 and the FGF receptor EGL-15 are known to signal positively through Ras/MAPK (see below). The ephrin receptor VAB-1 negatively regulates MAPK activation during oocyte maturation (Miller et al., 2003). Notably, the insulin-like RTK DAF-2 does not appear to signal through Ras/MAPK (G. Ruvkun, personal communication), but instead signals through a PI3-kinase/Akt pathway (see Signaling in the immune response).

Signaling by the LET-23 and EGL-15 RTKs can account for many but not all LET-60 Ras-dependent developmental events. The EGF-related ligand LIN-3 signals through LET-23/EGFR and LET-60 to control specification of hermaphrodite vulval and uterine fates, male spicule cell fates, the P12 ectoblast cell fate, and most likely the excretory duct cell fate (Aroian and Sternberg, 1991; Chang et al., 1999; Figure 1A). The FGF-related ligand EGL-17 signals through EGL-15/FGFR and LET-60 to control sex myoblast migration (Figure 1B), while the FGF-related ligand LET-756 signals through EGL-15 and LET-60 to promote axon guidance and viability (Bulow et al., 2004; Burdine et al., 1997; DeVore et al., 1995; Goodman et al., 2003; Roubin et al., 1999; Sundaram et al.,1996). In addition, LET-60 acts independently of any known RTK to control olfaction (Hirotsu et al., 2000) and the progression of germline meiosis (Church et al., 1995; Figure 1C). Furthermore, LIN-45, MEK-2 and MPK-1 act independently of known RTKs or LET-60 to mediate responses to Microbacterium nematophilum infection (Nicholas and Hodgkin, 2004).

In some cases the LET-23 and EGL-15 RTKs can also act through LET-60 Ras-independent pathways. LIN-3 and LET-23 control oocyte maturation via PLC gamma and an inositol polyphosphate signaling pathway (Bui and Sternberg, 2002; Clandinin et al., 1998; Yin et al., 2004). EGL-17 and EGL-15 may act through additional, unknown pathways to control sex myoblast migration and osmoregulation (DeVore et al., 1995; Schutzman et al., 2001; Sundaram et al., 1996).

Reverse genetic approaches have identified several factors important for ligand processing and secretion. The ligand LIN-3 exists in both trans-membrane (locally-acting) and diffusible forms (Hill and Sternberg, 1992; Thomas et al., 1990); generation of the diffusible form in vulval cells appears to require cleavage by the Rhomboid ortholog ROM-1 (Dutt et al., 2004). The ligand EGL-17 is secreted via a mechanism that requires the lipoprotein receptor-related proteins LRP-1 and LRP-2 and the Disabled-related adaptor DAB-1 (Kamikura and Cooper, 2003).

Factors that regulate LET-23 RTK trafficking and localization are important modulators of signaling strength. A complex consisting of three PDZ-domain proteins, LIN-2, LIN-7 and LIN-10, positively regulates signaling by localizing LET-23 to the basolateral membrane of vulval precursor cells, adjacent to the source of LIN-3 ligand (Kaech et al., 1998; Simske et al., 1996). The E3 ubiquitin ligase SLI-1 (Cbl; Jongeward et al., 1995; Yoon et al., 1995), the clathrin adaptors UNC-101 (Lee et al., 1994) and DPY-23 (Yoo et al., 2004), and the sorting nexin LST-4 (Yoo et al., 2004) all negatively regulate signaling, probably by promoting LET-23 endocytosis and/or degradation.

RTKs transduce signals by autophosphorylation and subsequent binding to phospho-tyrosine binding adaptor proteins (Schlessinger, 2000). LET-23 seems to act primarily through the adaptor SEM-5 (Clark et al., 1992), while EGL-15 acts through both SEM-5 and the adaptor SOC-1 (Schutzman et al., 2001). Factors that regulate RTK phosphorylation status could obviously modulate signaling strength. For example, the receptor tyrosine phosphatase CLR-1 negatively regulates EGL-15 RTK activity, most likely through direct dephosphorylation of key tyrosine residues (Kokel et al., 1998). On the other hand, the cytosolic tyrosine phosphatase PTP-2 (Shp-2) positively regulates LET-23 and EGL-15 RTK signaling (Gutch et al., 1998; Schutzman et al., 2001), and the tyrosine kinase ARK-1 (Ack-related) negatively regulates LET-23 RTK signaling (Hopper et al., 2000). The substrates and mechanisms of action of PTP-2 and ARK-1 are still unclear, but these proteins can bind to the adaptors SOC-1 and SEM-5, respectively, and appear to act at a step immediately downstream of the RTKs.

As is the case for other small GTPases (see Small GTPases), the activity of LET-60 Ras is controlled by Guanine Nucleotide Exchange Factors (GEFs), which activate Ras by stimulating conversion of Ras-GDP to Ras-GTP, and by GTPase activating proteins (GAPs), which inactivate Ras by stimulating conversion of Ras-GTP to Ras-GDP (Figure 3). The GEF SOS-1 appears necessary for most Ras-mediated developmental events (Chang et al., 2000). The GAPs GAP-1 and GAP-2 negatively regulate Ras signaling during vulval development and excretory duct development, respectively (Hajnal et al., 1997; Hayashizaki et al., 1998). Another negative regulator of Ras is SUR-5, a protein of unknown function that resembles acetyl coenzyme A synthetases (Gu et al., 1998).

Genetic modifier screens have identified a large set of genes that promote or inhibit signaling through the MAPK cascade, but that are not individually required for normal development in most tissues. In general, the roles of these genes can only be detected as suppressor or enhancer effects in appropriate double mutant combinations. Raf/MEK/ERK signaling is apparently subject to many different levels of regulation that individually have modest effects on signaling strength.

Positively-acting gene products in this category include so-called “scaffold” proteins - CNK-1 (Rocheleau et al., 2005), KSR-1 (Kornfeld et al., 1995; Sundaram and Han, 1995), KSR-2 (Ohmachi et al., 2002), and SUR-8/SOC-2 (Selfors et al., 1998; Sieburth et al., 1998) - that bind to one or more core components of the signaling cascade and are thought to bring different components together at the right place and time (Morrison and Davis, 2003). KSR-1 and KSR-2 are individually required for some developmental processes, but redundantly required for many others (Ohmachi et al., 2002; Figure 1). Also in this category is SUR-6, a PR55/B-type regulatory subunit of Protein Phosphatase 2A (Sieburth et al., 1999), which has been suggested to dephosphorylate and activate both Raf and KSR (Kao et al., 2003; Ory et al., 2003). Finally, the zinc transporter proteins CDF-1 and SUR-7 also positively regulate signaling, suggesting that high levels of intracellular zinc inhibit Ras signaling (Bruinsma et al., 2002; Yoder et al., 2004).

Negatively acting gene products in this category include the G-protein coupled receptor SRA-13 and its Gα target GPA-5, which may modulate Ras signaling in response to environmental conditions such as food availability (Battu et al., 2003). Also in this category are the kinase PAR-1 (which is thought to modulate KSR localization; Kao et al., 2003; Muller et al., 2001; Yoder et al., 2004), the MPK-1-binding protein LST-1 (Yoo et al., 2004), and the MAP kinase phosphatase LIP-1 (Berset et al., 2001).

It is important to note that some regulatory proteins may influence Raf/MEK/ERK signaling indirectly, by affecting neighboring cells. For example, the Gαq protein EGL-30 and the voltage-gated calcium channel EGL-19 promote vulval induction, but function in neurons and muscle, respectively (Moghal et al., 2003). Similarly, SRA-13 and GPA-5 influence vulval induction but it is unknown whether these genes function in vulval cells or in neurons (Battu et al., 2003). Also, the zinc transporter CDF-1 can influence vulval induction when expressed in either the vulva or the intestine (Bruinsma et al., 2002). How tissues such as the intestine, neurons and muscle can influence Raf/MEK/ERK activity in vulval cells is still unclear.

No single downstream target of MPK-1 can account for all of the effects of RTK/Ras/MAPK signaling. Rather, different tissues seem to require different subsets of potential targets, and the availability of certain targets may control tissue-specific responses.

One widely important MPK-1 target is the Ets domain transcription factor LIN-1 (Beitel et al., 1995; Figure 1A). LIN-1 antagonizes Ras signaling and appears to be negatively regulated by MPK-1 phosphorylation (Jacobs et al., 1998). However, some evidence suggests an additional positive role in Ras signaling (Howard and Sundaram, 2002). Like other Ets domain factors (Yordy and Muise-Helmericks, 2000), LIN-1 may have both transcriptional repressor and transcriptional activator functions.

Four widely important positive factors are the Mediator subunit SUR-2, the BTB/Zinc finger protein EOR-1, and the novel nuclear proteins LIN-25 and EOR-2 (Howard and Sundaram, 2002; Singh and Han, 1995; Tuck and Greenwald, 1995; Figure 1A). None of these proteins are known to be direct targets of MPK-1, but their functions are important for downstream cellular responses. SUR-2 is a conserved component of the Mediator complex, which links certain sequence-specific DNA binding proteins (such as Ets proteins) to the general RNA Polymerase II transcriptional machinery (Boyer et al., 1999; Stevens et al., 2002). SUR-2 and LIN-25 appear to function together, and they have strong effects on vulval development and weaker effects on excretory duct and P12 development (Nilsson et al., 1998; Nilsson et al., 2000; Singh and Han, 1995; Tuck and Greenwald, 1995). EOR-1 is related to known transcriptional activators and repressors (Barna et al., 2002; Collins et al., 2001). EOR-1 and EOR-2 appear to function together, and they have moderate effects on excretory duct and P12 development and weaker effects on vulval development (Howard and Sundaram, 2002). EOR-1 and EOR-2 appear to act redundantly with SUR-2 and LIN-25 (Howard and Sundaram, 2002).

Several transcription factors are required for cell-type specific responses to RTK/Ras/MAPK signaling, and are candidate MPK-1 targets. These include the forkhead transcription factor LIN-31 (Miller et al., 1993; Tan et al., 1998), which both promotes and inhibits vulval development, the Hox protein LIN-39 (Clark et al., 1993; Eisenmann et al., 1998; Maloof and Kenyon, 1998), the zinc-finger protein SEM-4 (Grant et al., 2000) and the GATA factors EGL-18 and ELT-6 (Koh et al., 2004; Koh et al., 2002), which promote vulval development, and the Hox protein EGL-5 (Chisholm, 1991; Jiang and Sternberg, 1998), which promotes P12 development. Ectopic expression of LIN-31 or LIN-39 can cause other tissues to adopt vulval-like characte