Ras effector switching as a developmental strategy

Abstract

Organisms pattern and specify cell fates with remarkably high fidelity and robustness, and cancer may be considered in part to be a disease of fate specification gone awry. During C. elegans vulval development an initial EGF signal prompts Ras to activate its canonical effector pathway, Raf-MEK-ERK, to induce a primary cell, which subsequently signals its 2 neighbors via Notch to develop as secondary cells. We have shown that Ras signaling through an alternate effector pathway, RalGEF-Ral, antagonizes Ras-Raf pro-primary signaling. Ras-RalGEF-Ral instead promotes secondary fate in support of Notch. We validated a previous model that EGF can also contribute to secondary fate, and argue that Ras-RalGEF-Ral mediates this EGF pro-secondary activity. Ras-Raf-MEK-ERK signaling was previously shown to be extinguished from secondary cells by secondary-specific expression of MAP kinase phosphatase, and we found that Ral expression is transcriptionally restricted to secondary cells. Thus during vulval development Ras switches effectors from Raf to RalGEF to promote divergent and mutually antagonistic cell fates, perhaps mirroring divergent effector usage in Ras-dependent tumors with differential pharmacological responsiveness.

Introduction

Roughly thirty percent of all tumors contain mutationally activated Ras, making it the most frequently mutated oncoprotein.¹ Historically, the Raf S/T kinase was considered the canonical Ras effector in oncogenic transformation. Subsequently, other effectors have been shown to mediate Ras oncogenesis, including RalGEFs, exchange factors for Ral GTPases, Ras relatives.² Yet the consequences of Ras-dependent RalGEF-Ral activation in oncogenesis are poorly understood.

Ras has also been characterized as a major signal transduction molecule in genetic model organisms. In development of the Drosophila R7 photoreceptor and the C. elegans vulva, Ras-Raf-MEK-ERK signaling was established as a paradigm of developmental fate induction.^3–5 The vulva is derived from a developmental equivalence group of 6 vulval precursor cells (VPCs), lying in an anterior-posterior line, with potential to form 1 of 3 fates: primary (1°), secondary (2°) or tertiary (3°). EGF secreted by the gonadal anchor cell patterns the vulva to a highly reproducible sequence (3°-3°-2°-1°-2°-3°), with the 1° and 2° cells executing distinctive cell division lineages to form the vulva while 3° cells divide once and fuse with the surrounding epithelium. LET-60/Ras activation of the Raf-MEK-ERK kinase cascade inhibits HNF and Ets transcription factors, which in turn inhibit 1° fate through as yet unidentified transcriptional client genes. The presumptive 1° cell secretes DSL ligands that induce its 2 neighbors via a highly conserved LIN-12/Notch pathway to assume 2° fate.^6,7 Together, the EGF-Ras-Raf pro-1° signal and the Notch pro-2° signal are necessary and sufficient to pattern the 2°-1°-2° vulval fates. However, the astonishing vulval patterning fidelity and environmental robustness⁸ probably require an additional layer of checks and balances to generate a resilient and practically error-free signaling system.

Pro-1° and pro-2° pathway crosstalk ensures that presumptive 1° and 2° fates are mutually exclusive. The LIN-12/Notch pathway induces transcription of a suite of genes, the function of some of which is to inhibit pro-1° signaling in presumptive 2° cells.^9–12 For example, the LIN-12/Notch transcriptional target lip-1 encodes a MAP kinase phosphatase that inactivates ERK pro-1° signaling. Conversely, Ras-Raf pro-1° signal promotes internalization and degradation of LIN-12/Notch in presumptive 1° cells, thereby abrogating the Notch pro-2° signal.^13,14 While naïve VPCs are presumably receptive to all signals, interplay of early inductive pathways reinforces initial patterning by a process of signal exclusion or quenching.

Another level of patterning was identified in elegant studies that implied the existence of an EGF vulval patterning gradient. EGF or EGF receptor signaling dose was varied and isolated VPCs at varying distances from the anchor cell were analyzed, and the results suggested that strong EGF signal, detected by the VPCs closest to the anchor cell, induced 1° fate, weaker EGF or more distal VPCs led to 2° fate, and the most distal VPCs became 3°, thereby describing a functional EGF gradient.^15–18 However, the mechanism of EGF gradient signaling was unknown. The demonstration that EGFR was necessary only for 1° fate and not 2° fate in conjunction with LIN-12/Notch being necessary and sufficient for 2° fate discouraged further investigation of the EGF gradient.^19–23

Ras-RalGEF Antagonizes Ras-Raf during Vulval Patterning

C. elegans expresses the single H/N/K-Ras ortholog LET-60 (humans have 3), the single Raf LIN-45 (humans have 3), the single Ras-regulated RalGEF RGL-1 (humans have 4) and the single Ral RAL-1 (humans have 2), greatly simplifying interpretation of signaling relationships among these proteins. We found that neither loss nor activation of RGL-1 or RAL-1 caused visible defects in otherwise wild-type animals, indicating that RalGEF-Ral is not central in vulval fate patterning. However, we found that RGL-1-RAL-1 loss conferred robust phenotypes in sensitized backgrounds where 1° fate is hyper-induced, causing enhancement of vulval hyper-induction.²⁴ Conversely, ectopically activated RAL-1(Q75L) or RalGEF-selective LET-60(G₁2V,E37G) suppressed vulval hyper-induction. In null or reduced function mutants for downstream transcription factors that inhibit 1° induction, loss of LET-60/Ras, RGL-1 or RAL-1 enhanced hyper-induction, arguing that the Ras-RalGEF-Ral signaling module antagonizes Ras-Raf pro-1° signaling in a parallel signal. Revealing an antagonistic Ras-mediated signal also resolves a long-standing conundrum in the field, which is why activated Ras cannot maximally induce 1° vulval fate.

In the absence of EGF signal, mutationally activated Ras-RalGEF-Ral drove 2° fate. Weak ectopic EGF or weakly activated EGF receptor induction of 2° fate was also dependent on Ras and Ral, consistent with Ras-RalGEF-Ral mediating part of the EGF pro-2° signal. Finally, loss of RAL-1 very weakly suppressed a reduced function mutation in LIN-12/Notch. This last effect was subtle, suggesting that the contribution of Ras-RalGEF-Ral under normal conditions is modest, and that the signaling module exists to increase robustness of vulval development under adverse conditions. Taken together, our results argue that the alternate Ras-RalGEF-Ral pro-2° signal antagonizes the Ras-Raf pro-1° signal, thus reinforcing the mutual antagonism of 1° and 2° vulval fates²⁴ and precluding intermediate or ambiguous vulval fates that could compromise vulval function under adverse conditions.¹⁰

The apparent mechanism of Ras effector switching is 2-fold. First, as described above, 2° specific Notch-dependent expression of the LIP-1 MAP kinase phosphatase should quench Ras-Raf-MEK-ERK signaling in presumptive 2° cells, thus excluding the Ras-Raf signal.⁹ Additionally, we found that a transcriptional reporter of the ral-1 promoter driving GFP expression was, like lip-1, initially expressed in all VPCs then rapidly restricted to presumptive 2° cells after induction.²⁴ The presence of concentrated LIN-12/Notch response elements in the lip-1 promoter suggest how that gene is spatially regulated in response to the sequential EGF-Notch signal,¹¹ but such elements are lacking in the ral-1 promoter, leaving open the question of how ral-1 expression is spatially regulated. But our results lead to a revised model where the contribution of the EGF gradient to vulval fate is superimposed on the sequential induction of EGF-Ras-Raf pro-1° signaling leading to Notch-dependent pro-2° signaling (Fig. 1). RAL-1 signaling output in 2° cells is unknown.

Figure 1.

Ras effector switching between competing vulval cell types. Presumptive 3° cells, which assume non-vulval fate, are not shown. Color coding: Green molecules can promote either 1° or 2° fate, blue molecules promote 1° fate, red molecules promote 2° fate, gray molecules are functionally inactivated (solid line) or transcriptionally excluded (dotted line for Ral).

It is not clear that Ras-Raf and Ras-RalGEF have a comparable relationship in all systems.²⁵ Ras collaborates with Notch in pancreatic ductal adenocarcinomas and Ral proteins play important roles in these tumors.^26–29 The details of the Ras-Raf versus Ras-RalGEF relationship is unclear in Drosophila R7 development, another paradigmatic cell patterning model. Some genetic evidence points to the Ras relative Rap activating RalGEF in Drosophila, but data are also consistent with the parallelism we discovered, and resolution of the question awaits application of more modern fly experimental tools.³⁰ Data from other Drosophila studies may have foreshadowed our findings, but could not be interpreted or tested at the time.^31,32

Why Switch Effectors?

The multiplicity of Ras effectors and differential effector usage in different tumor cell lines or primary tumors³³ suggests the possibility of an effector switching mechanism, and Ras effector switching has been the subject of theoretical discussions at past meetings. However, prior to our study strong evidence supporting a switch has been lacking. In most developmental genetic analyses of Ras signaling, including the vulva, a single main effector had been implicated.^34–36 Yet generally alternate effectors have not been studied in the same process, and perhaps the use of switching is more common than we know.

An obvious use of multiple effectors is to execute diverse signaling outcomes in different tissues as a consequence of Ras activation. For example, perhaps 1 effector promotes cell proliferation and another, survival. Yet our vulval study raises the novel possibility that effects are switched as a developmental strategy to prevent contradictory signals within the same cell and instead function to diverge cell signaling from the default in a developmental equivalency group. The former case could be termed pre-programmed effector utilization, in contrast to effector switching in the latter case.

I propose a different purpose for effector switching. Yes, Ras-RalGEF-Ral pro-2° activity is likely to increase developmental fidelity and robustness, but it may do so in a manner other than actual pro-2° signaling in support of Notch. Perhaps it is equally important to keep Ras otherwise occupied so presumptive 2° cells do not experience conflicting signals. In this manner, Ras effector switching from Raf to RalGEF is a Ras-Raf quenching mechanism comparable to and redundant with LIP-1 expression. So I speculate that a function of future Ras effector-switching may be to keep Ras occupied in a modestly constructive and non-harmful activity in cellular milieus where Ras signaling through a different effector might be dangerous to the development or wellbeing of the organism.

A metaphor for this hypothesis can be found in electronics, where electrical termination, usually through a high-resistance terminal cap, is frequently used for SCSI or coaxial cables to avoid signal reflection. In signal transduction, signal terminators may be important to avoid spurious noise in the system that could increase error rate and decrease vitality, and the best method for doing so may be to harness Ras signaling constructively for a specific fate (here, 2°).

Regulation of Ras Effector Switching

Potential mechanisms for Ras effector switching exist, but their natures are ill defined. Perhaps the best example is the C. elegans protein SOC-2/SUR-8 (human Shoc2), which was independently discovered as a suppressor of activated LET-60/Ras in the vulva and activated FGF receptor (via LET-60/Ras) in fluid homeostasis.^37,38 The human SOC-2 ortholog Shoc2 was subsequently found to function as a Ras-Raf scaffold^39,40 or an M-Ras-Raf scaffold.⁴¹ Like components of the Ras-Raf pathway, Shoc2 can be mutationally activated to confer Noonan Syndrome congenital defects.⁴² Regulation of SOC-2, either transcriptional or post-transcriptional, can be envisioned as a Ras-Raf effector switch, either promoting or curtailing Raf utilization.

An additional possible mechanism of effector switching is regulation of subcellular compartment by post-translational modification. Emerging evidence argues that intracellular signals modify the C-terminal end of Ras family small GTPases, thereby altering subcellular compartment occupancy, and thus effector encounters and signaling consequences.^43–48

Conclusion

Ras effector switching can play a critical role in signal transduction, yet we are probably just scratching the surface of this exciting field. Ras effector usage differs widely from cell line to cell line or primary tumor to primary tumor. It is tempting to draw a parallel to the striking heterogeneity of tumor response to drugs; perhaps evolving effector usage of Ras and other signaling systems contributes to such differential pharmaceutical response of cancers. It behooves us to understand better how and why different Ras effectors are utilized in both normal development and in pathogenic conditions like cancer.

Extra View to: Zand TP, Reiner DJ, Der CJ. Ras effector switching promotes divergent cell fates in C. elegans vulval patterning. Dev Cell. 2011;20:84–96. doi: 10.1016/j.devcel.2010.12.004.