HIGHLIGHTED ARTICLE| INVESTIGATION

Ras-Dependent Cell Fate Decisions Are Reinforced bythe RAP-1 Small GTPase in Caenorhabditis elegans

Neal R. Rasmussen,* Daniel J. Dickinson,† and David J. Reiner*,1

*Institute of Biosciences and Technology, College of Medicine, Texas A&M University, Houston, Texas 77030 and †Department ofMolecular Biosciences, University of Texas, Austin, Texas 78705

ORCID IDs: 0000-0003-2651-2584 (D.J.D.); 0000-0002-0344-7161 (D.J.R.)

ABSTRACT The notoriety of the small GTPase Ras as the most mutated oncoprotein has led to a well-characterized signaling networklargely conserved across metazoans. Yet the role of its close relative Rap1 (Ras Proximal), which shares 100% identity between theircore effector binding sequences, remains unclear. A long-standing controversy in the field is whether Rap1 also functions to activatethe canonical Ras effector, the S/T kinase Raf. We used the developmentally simpler Caenorhabditis elegans, which lacks the extensiveparalog redundancy of vertebrates, to examine the role of RAP-1 in two distinct LET-60/Ras-dependent cell fate patterning events:induction of 1� vulval precursor cell (VPC) fate and of the excretory duct cell. Fluorescence-tagged endogenous RAP-1 is localized toplasma membranes and is expressed ubiquitously, with even expression levels across the VPCs. RAP-1 and its activating GEF PXF-1function cell autonomously and are necessary for maximal induction of 1� VPCs. Critically, mutationally activated endogenous RAP-1 issufficient both to induce ectopic 1�s and duplicate excretory duct cells. Like endogenous RAP-1, before induction GFP expression fromthe pxf-1 promoter is uniform across VPCs. However, unlike endogenous RAP-1, after induction GFP expression is increased inpresumptive 1�s and decreased in presumptive 2�s. We conclude that RAP-1 is a positive regulator that promotes Ras-dependentinductive fate decisions. We hypothesize that PXF-1 activation of RAP-1 serves as a minor parallel input into the major LET-60/Ras signalthrough LIN-45/Raf.

KEYWORDS Ras; Rap1; Raf; PDZ-GEF; CRISPR

THE founder of the Ras superfamily of small GTPases, Ras,is the most mutated oncoprotein in cancer (COSMIC v84;

http://cancer.sanger.ac.uk/cosmic; Papke and Der 2017).With diverse functions throughout cell biology, members ofthe Ras superfamily are conserved across Metazoa. Withingiven subfamilies, the core effector binding sequences aretypically identical among Caenorhabditis elegans, Drosophila,and mammals, suggesting functional conservation of GTPaseinteractions with downstream effectors (Reiner and Lundquist2016), many of which are also conserved across Metazoa.

Small GTPases are typically membrane-bound throughlipid prenylation and processing of their C-termini(Hancock et al. 1990; Prior and Hancock 2012; Reiner and

Lundquist 2016). They function as molecular switches, cy-cling between GTP-bound (active) and GDP-bound (inactive)states. Thus, activity of small GTPases is controlled by GEFs(guanine nucleotide exchange factors), which displace GDPto allow GTP loading, and GAPs (GTPase activating pro-teins), which stimulate the otherwise inefficient intrinsicGTPase activity that hydrolyzes GTP to GDP (Wennerberget al. 2005).

The Rap (Ras proximal) subfamily comprises GTPasesclosely related to Ras itself. Rap1 (previously known asK-Rev1) shares identical core effector binding sequences withRas, while Rap2 diverges at one residue in the core effectorbinding sequence (Supplemental Material, Figures S1 andS2). The Rap subfamily shares some GEFs and GAPs with theRas subfamily, but also has some GEFs and GAPs that arespecific to Raps (Raaijmakers and Bos 2009; Gloerich and Bos2011). Historically, Rap1 has mostly been implicated in theregulation of cell-cell junctions, which is generally not con-sidered to be a functional site of Ras action (Caron et al. 2000;Reedquist et al. 2000; Knox and Brown 2002; Bos 2005). Yet

Copyright © 2018 by the Genetics Society of Americadoi: https://doi.org/10.1534/genetics.118.301601Manuscript received April 6, 2018; accepted for publication September 15, 2018;published Early Online September 26, 2018.Supplemental material available at Figshare: https://doi.org/10.25386/genetics.7130846.1Corresponding author: Center for Translational Cancer Research, Institute ofBiosciences and Technology, Texas A&M Health Science Center, 2121 W.Holcombe Blvd., Houston, TX 77030. E-mail: dreiner@ibt.tamhsc.edu

Genetics, Vol. 210, 1339–1354 December 2018 1339

because of their identical core effector binding sequences, ithas long been hypothesized that Ras and Rap1 share effectorsand may interact in signaling networks, while this has notbeen suggested for Rap2. Unfortunately, an early experimen-tal artifact in cell culture experiments clouded the role ofRap1 relative to Ras and its interaction with the canonicalRas effector, the Raf Ser/Thr kinase. Activated Rap1 trans-fected into cells inhibited oncogenic transformation andblocked activation of the Ras-Raf-MEK-ERK canonical MAPkinase cascade (Kitayama et al. 1989, 1990; Sakoda et al.1992; Cook et al. 1993), yet failed to interfere with ERKactivation in more physiologically relevant conditions(Zwartkruis et al. 1998). This observation led to the probablyerroneous model that Rap1 was a competitive inhibitor ofRas. Because this observation was only observed under condi-tions of high ectopic expression (Kitayama et al. 1989), a plau-sible explanation is that overexpressed activated Rap1sequestered Raf to cell-cell junctions where Ras-Raf signalingis typically not functional, thus diminishing activation of Raf-MEK-ERK signal by endogenous Ras. Yet the role of Rap1relative to Ras-Raf signaling has remained murky since, de-spite certain conditions in which Rap1 appears to contribute toRaf activation in cultured mammalian cells (York et al. 1998).

However, several lines of evidence support Rap1 contrib-uting to tumorigenic growth through activation of Raf, sug-gesting that Ras and Rap1may function in parallel to activateRaf. Rap1 can oncogenically transform mammalian tissueculture cells (Altschuler and Ribeiro-Neto 1998). Noncanon-ical putative activating mutations in Rap1 have also beenassociated with Kabuki syndrome (Bögershausen et al.2015), part of the RASopathy spectrum of heritable disordersassociated with inappropriate weak activation of the Ras-Rafsignaling axis (Aoki et al. 2016). Although typically not mu-tated itself as an oncogene, rare mutations in Rap1 have beenobserved in various cancers (COSMIC v84; Gyan et al. 2005).The paucity of such activating mutations may be because ofthe strong role of Rap1 in assembly and maintenance of cell-cell junctions, which may counter tumorigenic growth if notspatially controlled. A comparison to Rap1 may be drawnwith the Rho GTPase family member, Rac. Oncogenic Racmutations are rare and are mostly found in melanoma(Sergent 1990; Krauthammer et al. 2012). Yet while mostlynot mutated to drive cancer, probably because of its centralrole in control of the cytoskeleton, cell morphogenesis, andmigration, inappropriate upstream activation of Rac can stillcontribute to tumorigenesis (Lindsay et al. 2011; Srijakotreet al. 2017). This loss of regulation vs. mutational activationhas also been observed with Rap1, with the loss of a numberof RapGAPs, thus implicating them as tumor suppressors(McLaughlin et al. 2013; Maertens and Cichowski 2014;Zhao et al. 2015).

Studies in Drosophila suggest that Rap1 is necessary formaximal induction of the Raf-MEK-ERK MAP kinase cas-cade, perhaps in parallel to Ras activation (Mishra et al.2005; Mavromatakis and Tomlinson 2012). Yet also inDrosophila, results with Rap1 have been contradictory

(Baril et al. 2014). TheDrosophilaRoughenedmutation, orig-inally thought to be a hypermorphic allele, was subsequentlyinferred to be a dominant-negative mutation (Hariharanet al. 1991; Mavromatakis and Tomlinson 2012). ThatDrosophila Rap1 is an essential gene further complicates in-vestigation of its in vivo function with regard to activation ofRaf during development. Consequently, there is great benefitto studying the role of Rap1 in a simple developmental sys-tem where Rap1 is not required for viability, and where thereis less concern about genetic redundancy. Thus, we are in-vestigating the role of Rap1 signaling in C. elegans, whereRap1 loss-of-function mutants are viable, and there is no re-dundancy in gene subfamilies compared to mammals (mam-mals possess three Ras-, two Rap1-, three Rap2-, and threeRaf-encoding genes).

C. elegans encodes RAP-1 and RAP-2, which correspond tomammalian Rap1 and Rap2, respectively (C. elegans RAP-3,which has a nonconservative change in a critical residue inthe core effector binding sequence, is likely to be a nematode-specific, functionally divergent Rap; Figure S1; Reiner andLundquist 2016). Consistent with their roles in promotingcell-cell junctions in Drosophila and mammalian cells, C. ele-gans RAP-1 and RAP-2 are redundant for larval molting andepithelial integrity, the double mutant causing lethality, aphenotype that is echoed by the deletion of the RapGEF,PXF-1 (Pellis-van Berkel et al. 2005). For assembly of cadherin-based cell-cell junctions in C. elegans, RAP-1 functionsredundantly with the Ras family small GTPase, RAL-1 (Ras-like; Frische et al. 2007). Yet thus far there has been noexamination of the role of RAP-1 in developmental pattern-ing of the C. elegans vulval precursor cell (VPC) fates, a sys-tem where the sole C. elegans Ras ortholog, LET-60, plays acentral role. Therefore, we investigated the role of RAP-1 inVPC fate patterning.

The six equipotent VPCs (P3-8.p) are induced by EGFsecreted by the anchor cell (AC) in the somatic gonad, suchthat the VPC nearest the AC (typically P6.p) is induced toassume 1� fate (Figure 1; Sternberg 2005). LET-23/EGFRreceives the EGF signal, and the 1�-promoting signal is trans-duced by a canonical Ras-Raf-MEK-ERK (C. elegans LET-60-LIN-45-MEK-2-MPK-1) MAP kinase cascade; this signal isnecessary and sufficient for induction of 1� fate (Sundaram2013). Induced presumptive 1� cells express DSL ligands(Chen and Greenwald 2004), which signal neighboring VPCsvia the LIN-12/Notch receptor to assume 2� fate. LIN-12/Notch is necessary and sufficient for 2� fate (Greenwaldet al. 1983; Greenwald and Kovall 2013).

Development of the wild-type 3�-3�-2�-1�-2�-3� pattern ofVPC fates occurs with 99.8% accuracy (Braendle and Felix2008). During VPC fate patterning, cells are initially speci-fied, then become committed to their fate (Sternberg 2005).During this process, the VPC signaling network is at leastpartially reprogrammed, perhaps contributing to final com-mitment and fidelity. For example, after initial induction,LIN-12/Notch is internalized and degraded in presumptive1� cells (Shaye and Greenwald 2002, 2005), thus precluding

1340 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

2�-promoting signal in a cell that is specified to 1� fate. Con-versely, presumptive 2� cells express the LIN-12/Notch tran-scriptional client, LIP-1/ERK phosphatase (Berset et al. 2001;Yoo et al. 2004), thereby prohibiting 1�-promoting signalingin cells that are specified as 2�. Additionally, expression of asuite of other LIN-12/Notch transcriptional client genes isaltered after initial induction (Berset et al. 2001, 2005; Yooet al. 2004; Yoo and Greenwald 2005; Zhang and Greenwald2011), supporting the idea of network reprogramming. Weobserved further evidence consistent with VPC reprogram-ming with ral-1 promoter expression, initially expressed inall VPCs, subsequently being excluded from presumptive1� cells, thereby prohibiting 2�-promoting signaling in fromcells that are specified as 1� (Reiner 2011; Zand et al. 2011).Most of this reprogramming at the transcriptional level oc-curs prior to the first VPC division. Although we do not knowthe precise point at which terminal commitment occurs(Wang and Sternberg 1999), we hypothesize that transcrip-tional and potentially post-translational network reprogram-ming, occurring prior to the first cell division, is a criticalcomponent of the high fidelity observed in this developmen-tal decision.

In this study, we use VPC fate patterning to investigate thelong-standing question of the relationship between Ras andRap1. We used CRISPR (clustered regulatory interspacedshort palindromic repeats) to fluorescently tag the endoge-nous RAP-1 and found it to be expressed throughout theanimal and localized to plasma membranes and with in-creased concentration at cell-cell junctions. We find that de-letion of RAP-1 in an otherwise wild-type background resultin mild and low penetrance VPC patterning defects. Use ofsensitized genetic backgrounds indicated a cell-autonomousrole for RAP-1 in promoting 1� fate. CRISPR-mediated muta-tional activation of endogenous RAP-1 revealed that RAP-1 issufficient to induce ectopic 1� cells, and also induce duplica-tion of the excretory duct cell, another cell induction eventdependent on LET-60/Ras-LIN-45/Raf signaling. We foundthat the PXF-1/PDZ-GEF is also required for maximal 1� in-duction, consistent with a role for PXF-1 as the activatingRAP-1 GEF in VPC patterning. Furthermore, we find thatduring the response to EGF signal, transgenic GFP expres-sion from the pxf-1 promoter changes from uniform expres-sion across VPCs increased in presumptive 1� cells butdecreased in presumptive 2� cells. This observation is another

Figure 1 A model of RAP-1 and vulval cell fate patterning. (A) Patterning of the vulva begins with six equipotent vulval precursor cells (VPCs). (B) Initialcell fate specification of the VPCs begins in response to the release of EGF from the nearby anchor cell (AC) through induction of the classic LET-60/Ras-LIN-45/Raf cascade and a subsequent series of lateral signals between the VPCs via Notch signaling. (C) 1� cell fate is then reinforced through thefocused and magnified expression of PXF-1 and resulting RAP-1 signaling. Subsequent patterning results in two fields of proliferative (1, 2�) andnonproliferative (3�) cells. (D) Sequence alignment of LET-60 and RAP-1 with the core effector domain shaded in pink and identical residues, yellow. TheC-terminal hyper-variable and CAAX sequences are outlined in the blue.

RAP-1 Reinforces LET-60 Signaling 1341

confirmation of the reprogramming hypothesis and suggeststhat the activation of RAP-1 is restricted to 1� cells during thephase when initial patterning of VPCs is reinforced. Takentogether, our results support a model in which a PXF-1-RAP-1signal functions as a minor parallel input into the major LET-60-LIN-45 1�-promoting signaling.

Materials and Methods

C. elegans handling and genetics

All strains were derived from the parent N2 wild-type strain.Animals were grown at 20� under standard culturing condi-tions on NGM agar plates seeded with OP50 bacteria unlessstated otherwise (Brenner 1974). Crosses were performed us-ing standard methods. Strain details are shown in Table S1.

The let-60(n1046gf) strain undergoes genetic drift uponcontinuous growth, resulting in increased phenotypestrength (Zand et al. 2011). We therefore established firmguidelines to ensure consistent results. Immediately uponthawing or construction, strains harboring n1046 werescored and then starved and parafilmed as a reference strain.For subsequent experiments, we reestablished growingstrains each week to avoid drift while in continuous culture.Experiments were only reported if the let-60(n1046gf) con-trol was within the well-validated baseline levels of induc-tion. We did not observe any similar drift in rap-1(re180gf) orlet-23(sa62gf) strains. Strains harboring homozygous let-23-(sa62gf)mutations did exhibit delayed growth, necessitatingscoring VPC induction a day later than with other strains.

Plasmids and generation of CRISPR strains

Details of plasmid constructions are available upon request.Plasmids used are shown in Table S4. Single guide RNAsequences and repair templates are listed in Tables S5 andS6, respectively. The rap-1(re180gf) CRISPR strain was gen-erated using the co-CRISPR strategy (Paix et al. 2014) bymicroinjection of pJA58 (50 ng/ml), pNR21 (50 ng/ml),rap-1 (G12V) single-stranded donor oligonucleotide repairtemplate (500 mM), dpy-10(cn64), single-stranded donor ol-igonucleotide repair template (500 mM), and co-injectionmarker pPD118.33 (20 ng/ml) into N2 wild-type animals.Primers and conditions for PCR genotyping are listed in TableS3. PCR detection of rap-1(re180gf) was determined usingan overnight digestion with BamHI (New England Biolabs -NEB), NEB Cutsmart buffer, and water to a final volume of50 ml.

Scoring vulval induction and fate reporters (VPC andexcretory duct)

To score normal and ectopic vulval induction, animals weremounted in M9 on slides with a 3% grooved agar pad con-taining 5 mM sodium azide and examined by DIC/Nomarskioptics (Nikon eclipse Ni). Grooved agar pads were made bypipetting melted agar onto a 33 rpm vinyl record (Traffic –

Low Spark of High-Heeled Boys), then placing a slide on top

(Rivera Gomez and Schvarzstein 2018). For this study, thescale was 0–3 ectopic 1�s, with 0 being wild type. In n1046gfand sa62gf backgrounds, we observed only 3�-to-1� transfor-mations. In the rap-1(re180gf) background, we observedmostly 2�-to-1� but rare 3�-to-1� transformations. Ectopic1� pseudovulvae were scored in late L4 animals as deter-mined by morphology and position relative to the wild-typevulva. Observed ectopic 1� pseudovulvae largely conformedto the previously described “cap-” like structure morphologyin which the entire 1� lineage has delaminated from the cu-ticle (Katz et al. 1996). “Abnormal” vulvae were likewisescored when deviations from the symmetrical wild-typemorphology were observed. To ensure accuracy and repro-ducibility, all presented data are from animals grown concur-rently and are representative nonpooled samples exceptwhere indicated. Signal from 1� fate transcriptional reportersarIs92[Pegl-17::CFP-LacZ], arIs131[Plag-2::2xNLS::YFP], andarIs222[Plag-2::2xNLS-tag-RFP], and LIN-12 signaling re-porter arIs116[Plst-5::2xNLS::YFP] were scored for each ofthe six VPCs with the Nikon Eclipse Ni microscope, with aNikon DS-Fi2 color camera and using the NIS Elements Ad-vanced Research, Version 4.40 (Nikon, Garden City, NY) soft-ware package. L3 animals were scored at the Pn.px (two-cell)stage to ensure that animals were past the point of EGF in-duction of VPCs. For each VPC transcriptional reporter, aconsistent exposure time was used across all samples. ForarIs116[Plst-5::2xNLS::YFP], an extended exposure of 4 secwas used to capture both high and weak VPC expression.Using the same equipment, newly hatched L1 larvae werescored for the presence of a single or duplicated excretoryduct cell(s) using the saIs14[Plin-48∷GFP] transgene (Johnsonet al. 2001).

Duct cell-absent lethality counts

Twenty gravid adults were picked to seeded plates andallowed to lay eggs overnight. Adults were removed and allprogeny were scored 2–3 days later as either live animals ordead “rods” (Ferguson and Horvitz 1985).

RNA interference

Bacterially mediated RNA interference (RNAi) experimentswere conducted at 23� on NGM agar plates supplementedwith 1 mM IPTG and 50 mg/ml carbenicillin as described(Kamath and Ahringer 2003). In our hands, we typically ob-tain more robust RNAi results at 23� than at 20� (Zand et al.2011). All RNAi clones were sequenced to confirm identity.RNAi plates were seeded with 80 ml HT115 bacteria harbor-ing clones of C. elegans genes (Source Bioscience) or thenegative control luciferase (luc; Shin et al. 2018) and allowedto grow overnight at room temperature. Late L4 larvae wereadded to each plate and transferred to a fresh RNAi plate thenext day. Founding parents were then removed the followingday. Late L4 progeny were scored for the formation of theprincipal vulva and/or ectopic pseudovulvae by DIC/Normarski optics 2 days later. However, as noted above, strainshomozygous for the let-23(sa62gf) allele reach maturity at

1342 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

a slower rate, and consequently were scored 4 days afterparents were removed. For each experiment, pop-1(RNAi),which confers 100% embryonic or L1 lethality under optimalconditions, was used as a control for maximal RNAi efficacy;animals were only scored from experiments in which 100% le-thalitywas observed. RNAi strains used are included in Table S2.

VPC-specific RNAi was performed using the genotypelet-23(sa62gf); rde-1(ne219); mfIs70[Plin-31::rde-1(+),Pmyo-2::gfp]. The mfIs70[Plin-31::rde-1(+), Pmyo-2::gfp] inte-grated transgene, which rescues the loss of rde-1 specificallyin the VPCs (Barkoulas et al. 2013), was introduced andtracked in crosses using the dpy-11(e224) unc-76(e911) chro-mosome in trans as a balancer.

Fluorescence microscopy (imaging)

Animals were mounted live in M9 buffer on slides with a 3%agarpadcontaining5mMsodiumazide.DIC/Nomarski opticsand epifluorescence microscopy were captured using theNikon Eclipse Ni and confocal images using the A1si ConfocalLaser Microscope (Nikon). Captured images were processedusingNISElementsAdvancedResearch, Version 4.40 (Nikon)and Adobe Photoshop CC 2018 software packages (invertedimages).

Data availability

Strains and plasmids are available upon request. Table S1 listsC. elegans strains used in the study. Table S2 lists bacteriallymediated RNAi clones. Table S3 contains primer sequencesused in the study. Table S4 contains plasmids. Table S5 con-tains single guide RNA and PAM sequences for CRISPR. TableS6 repair oligos used for CRISPR and co-CRISPR. Supplemen-tal material available at Figshare: https://doi.org/10.25386/genetics.7130846.

Results

RAP-1 is expressed in VPCs and localized to plasmamembrane and junctions

TheRas superfamily of small GTPases is characterized by lipidmodification of their C-terminal CAAXmotif, resulting in theirsubcellular localization to the lipid bilayers. For Ras itself,this primarily means localization to the plasma membrane(Hancock et al. 1990; Prior and Hancock 2012; Reiner andLundquist 2016). Likewise, Rap1 has been observed to befound on the plasma membrane (Beranger et al. 1991;Pizon et al. 1994). Prior work reported LET-60/Ras to ubiq-uitously expressed in C. elegans (Dent and Han 1998). Al-though the let-60/Ras promoter fusion could not provideinformation regarding subcellular localization, ectopic induc-tion of 1�s conferred by the let-60(n1046gf) activating muta-tion was reversed by treatment of farnesyl transferaseinhibitors, which block C-terminal prenylation and hence ac-tivity (Hara and Han 1995). Consequently, we are reasonablyconfident that C. elegans LET-60 is targeted to the plasmamembrane.

To determine the expression pattern and subcellular local-ization of endogenous RAP-1, we used CRISPR-Cas9-medi-ated genome editing to insert amNeonGreen::3xFlag epitopetag into the 59 end of the endogenous rap-1 gene, generat-ing rap-1(cp151[mNeonGreen^3xFlag::rap-1]) (Figure S3;Dickinson et al. 2015). In keeping with previous observa-tions with let-60, tagged RAP-1 expression was ubiquitousthroughout the animal during development, and localizedto the plasmamembrane (Figure 2). Of particular importanceto this study, RAP-1 expression was observed in the VPCsthroughout the L3 stage, when specification occurs (one-cellstage, Pn.p; Figure 2, A and C), after EGF induction (two-cellstage, Pn.px; Figure 2, B and D), and throughout vulvalmorphogenesis.

We also observed that tagged endogenous RAP-1 is local-ized to cell-cell junctions between hypodermal seam cells(Figure S4A). This observation is consistent with the estab-lished role of RAP-1 in the assembly of cadherin complexes atjunctions during embryogenesis, a process that is redundantwith another small GTPase, RAL-1 (Frische et al. 2007), andour observation that tagged endogenous RAL-1 also localizesto the plasma membrane and cell-cell junctions (Shin et al.2018). This finding also conforms with the established role ofRap1 in junctional biology in Drosophila and mammalian cellculture (Caron et al. 2000; Reedquist et al. 2000; Knox andBrown 2002).

Intriguingly, we observed increased expression of taggedendogenous RAP-1 in the AC (Figure 2, C and D). The AC isknown to undergo polarized invasive behavior, directed to-ward the 1� lineage of the vulva (Sherwood and Sternberg2003; Hagedorn et al. 2009; Ziel et al. 2009).We additionallyobserved increased expression of RAP-1 in the distal tip cell(DTC) of the developing somatic gonad (Figure S4B), a tissuethat also performs invasive migration. However, our laterexperiments indicate that RAP-1 functions cell autonomouslyto regulate VPC fate patterning, so we interpret the elevatedRAP-1 expression in the AC to be unrelated to VPC fatepatterning.

RAP-1 loss results in underinduction of 2� VPCs

The early false lead in mammalian cell culture of Rap1 as acompetitive inhibitor of Ras has long muddied the waters ofour understanding of the role of Rap1 in Ras signaling(Kitayama et al. 1989, 1990; Sakoda et al. 1992; Cook et al.1993); see Introduction). Additionally, investigations of Rap1have been complicated by its multiple isoforms in vertebratesand its essential role in development in Drosophila (Mishraet al. 2005). Previous work showed C. elegans rap-1 putativenull mutant animals to be fertile and viable, includingmNeonGreen^SEC^3xFlag::rap-1 animals retaining theSEC positive/negative selection cassette, which should dis-rupt RAP-1 expression (Frische et al. 2007; Dickinson et al.2015).

Wemore closely examined rap-1 animalsmutant for eitherof two independent, strong loss or null alleles. Both rap-1(tm861) (a deletion that results in a frame shift) and

RAP-1 Reinforces LET-60 Signaling 1343

rap-1(pk2082) (nonsense allele in exon 5 of 6; Frische et al.2007) (Figure S3) conferred low penetrance abnormal vulvalpatterning (Figure 3). Among the rare patterning defects, weobserved defects which we interpret to be missing 2� cells(Figure 3, B and E). In addition, we observed asymmetricalvulvae (Figure 3, C and F) and potentially a migration defectof the 2� lineage, which could be indicative of ambiguous fate(Figure 3D). Because of the rarity of these defects for bothrap-1(tm861) and rap-1(pk2082) (Figure 3, G and H; 5.5%,N = 128; 7.7%, N = 78, respectively), determining quanti-tative breakdowns for each observed phenotype was pro-hibitive. These phenotypes are similar to those observed inanimals harboring the hypomorphic lin-3(e1417), let-60(n2021), and lin-45(n2506)mutations, reducing functionof EGF, Ras, and Raf orthologs, respectively (Wang andSternberg 1999; Zand et al. 2011). Hypomorphic mpk-1(ku1) animals (Wu and Han 1994) showed low-level pat-terning defects consistent with both absent 2� and absent1� cells (Figure S5, A and B, respectively). rap-1 loss in con-junction withmpk-1(ku1) resulted in a significant increase inpatterning defects of the both 2� and 1� lineages (Figure S5,C, D, and G). In the weakly hypomorphic ksr-1(n2509) singlemutant (Kornfeld 1995), we did not observe patterning de-fects (Figure S5E), but did so in the rap-1(tm861); ksr-1(n2509) double mutant (Figure S5, F and H).

As a single mutant, the stronger hypomorphic allele mpk-1(ku1) caused missing 1� cells as well as missing 2� cells,while the weaker hypomorphic allele in the KSR Raf-MEKscaffold, ksr-1(n2509), conferred no observed vulval pheno-type. That rap-1(tm861) enhanced defects in both of these inmutants hypomorphic for components of the 1�-promotingMAP kinase cascade suggests that RAP-1 plays a minor mod-ulatory role in promoting 1� fate. We speculate that theseunderinduced 1� cells express insufficient levels of the triplyredundant DSL ligands, LAG-2, APX-1, and DSL-1 (Chen andGreenwald 2004; Zhang and Greenwald 2011) to reliablyinduce the neighboring VPCs to become 2�.

To further elucidate the effects of rap-1 loss, we utilizedthe transcriptional reporter arIs116[Plst-5::2xNLS::YFP] (Liand Greenwald 2010). In agreement with previous findings,we found that in otherwise wild-type animals this reportershowed strong expression in both P5.px and P7.px. However,we also observed a small portion of wild-type animals exhib-iting very weak expression in neighboring 1 and 3� lineages

(Figure 3, J and L) The strong signal correlates only partiallywith 2� fate; the weak signal does not. Deletion of rap-1resulted in a significant increase in ectopic YFP reporter ex-pression in both 1 and 3� cells (Figure 3, I, K, andM). BecauseYFP is expressed strongly and consistently in 2� lineages, butweakly and sporadically in 1 and 3� lineages, we speculatethat arIs116[Plst-5::2xNLS::YFP] is a sensitive reporter of LIN-12/Notch signaling, and not simply a fate reporter for 2� cells.We hypothesize that RAP-1 reduces inappropriate signalingnoise, but we did not examine sufficient animals to observemissing 2� cells, which occurs only as a fraction of the 5.5%rate of abnormal vulvae.

To examine 1� fate, we used the transcriptional reporter,arIs222[Plag-2::tag-RFP] (Sallee and Greenwald 2015). Tag-RFP was expressed only in 1� lineages, and deletion of rap-1failed to alter this expression (Figure S6, A–D). The arIs222reporter is consistent with a reporter of 1� fate, and not aquantitative readout of 1� signaling strength. Consequently,we cannot assess the role of RAP-1 in regulating inappropri-ate 1� signaling noise.

RAP-1 contributes to excretory duct cell induction

The LET-60/Ras-LIN-45/Raf MAP -MPK-1/ERK MAP kinasecascade in necessary for excretory duct cell development, soloss of cascade function results in rod-like lethality dueto defective excretion of solutes (Beitel et al. 1990; Hanand Sternberg 1990; Han et al. 1990; Yochem et al. 1997;Abdus-Saboor et al. 2011). Neither the rap-1(tm861) dele-tion nor the hypomorphic ksr-1(n2509) conferred rod-likelarval lethality, but the double mutant with rap-1(tm861)resulted in very low levels of rod-like lethality (Figure 4A).The hypomorphic mpk-1(ku1) allele confers rod-like larvallethality at a low level (Wu and Han 1994), and tm861 sig-nificantly enhanced this defect (Figure S5I).

RAP-1 is necessary for maximal induction of 1� VPCs

To determine at which step rap-1 influences VPC fate pattern-ing, we analyzed double mutants between deleted rap-1 andgain-of-function mutations in let-23/egfr, let-60/ras, and lin-45/raf. We used a sensitized genetic background, let-23-(sa62gf) (Katz et al. 1996), which is sufficient to induceectopic 1�s and is sensitive to perturbation of modulatorysignals (Zand et al. 2011). The addition of rap-1(tm861)resulted in significantly decreased ectopic 1�s compared to

Figure 2 Endogenous RAP-1 is expressedthroughout the VPCs. Representative DIC (Aand B) and inverted confocal epifluorescence(C and D) micrographs of rap-1(cp151[mNG^3xFlag::rap-1]) at (A and C) the one-cell(Pn.p) and (B and D) two-cell (Pn.px) stages.mNG, mNeonGreen (Shaner et al. 2013) (rep-resentative of eight confocal and .50 epifluor-escent images). Bar, 10 mm.

1344 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

Figure 3 Loss of rap-1 confers low-penetrance vulval patterning defects. Representative DIC micrographs of vulvae at the late L4 stage. (A) Wild-type,(B–D) rap-1(tm861), and (E and F) rap-1(pk2082) (main vulva, open triangle; vulvae abnormalities, solid triangle). (G and H) Quantification of mutantabnormal vulvae in wild-type vs. (G) rap-1(tm861) and (H) rap-1(pk2082), with percent abnormal vulvae on the y-axis. (I) Quantification of ectopic Plst-5::2xNLS::YFP expression outside of P5.px or P7.px in wild-type vs. rap-1(tm861) animals. Error bars indicate 95% confidence interval based on sample size(G and H) or SEM (I), P values were calculated by Fisher’s exact-test (G and H) or t-test (I). (J and K) Representative inverted epifluorescence micrographsin animals at the two-cell (Pn.px) stage shows increased inappropriate expression of the 2� signaling reporter arIs116[Plst-5::2XNLS::YFP] in (J) wild-typevs. (K) rap-1(tm861) animals. (L and M) Schematic representation of the expression of the arIs116[Plst-5::2XNLS::YFP] signaling reporter across the all sixVPCs for both (L) wild-type and (M) rap-1(tm861). Solid areas indicate strong expression (saturated signal) and shaded areas indicate weak expression(nonsaturated signal). All data are representative nonpooled assays with the mean shown. Numbers in columns indicate N.

RAP-1 Reinforces LET-60 Signaling 1345

let-23(sa62gf) alone (Figure 4B). We also targeted rap-1 lossin a VPC-specific manner using the let-23(sa62gf); rde-1(ne219) background with the mfIs70 integrated transgene(Plin-31 driving VPC-specific rescue with rde-1(+)). In agree-ment with our previous findings, VPC-specific rap-1-directedRNAi depletion resulted in a significant decrease in the for-mation of ectopic 1�s (Figure 4C). These results demonstratethat RAP-1 functions cell-autonomously in the VPCs. To fur-ther evaluate the effects of rap-1 loss, we utilized thelet-60(n1046gf) mutant strain, which harbors the G13Emutation analogous to mutation in mammalian Ras thatlock the protein in the GTP-bound form, thus causing theprotein to be constitutively active (Beitel et al. 1990; Hanet al. 1990). Similar to our findings with let-23(sa62gf),above, rap-1(tm861) reduced ectopic 1� induction in thelet-60(n1046gf) background (Figure S5J). Consequently,we hypothesize that RAP-1 functions in parallel to or down-stream of LET-60 to promote 1� fate. Given the biochemicalrelationships of mammalian Rap1 and Ras to Raf, we favoran interpretation of parallelism, potentially converging onLIN-45/Raf. To begin testing this possibility we utilized acti-vated LIN-45, kuIs57[Pcol-10::lin-45(gf)] as a sensitized back-ground. In keeping with the hypothesis of rap-1 acting inparallel fashion with let-60, rap-1(tm861) failed to alterkuIs57[Pcol-10::lin-45(gf)] induction of ectopic 1� fate (FigureS5K).

We previously noted that let-60(n1046gf), which is pre-dicted to be GAP insensitive, was paradoxically still respon-sive to RNAi-mediated depletion of the negative regulator,

GAP-1 (Zand et al. 2011). We reproduced this result, observ-ing that the let-60(n1046gf) single mutant responded to gap-1(RNAi) by increasing ectopic 1� induction,withRNAi targetingthe LIP-1/ERK phosphatase (Zand et al. 2011) as a positivecontrol and luciferase-directed RNAi as a negative con-trol (Figure 4D). However, we found that rap-1(tm861)let-60(n1046gf) double mutant animals no longer respondedto gap-1(RNAi) while still responding to lip-1(RNAi) (Figure4D). These results are consistent with GAP-1 functioning topromote GTP hydrolysis of both RAP-1 and LET-60/Ras re-lated small GTPases. Accordingly, the mammalian GAP1 sub-family of GAPs, orthologous to C. elegans GAP-1 (Stetak et al.2008; Grewal et al. 2011), has been shown to be bifunctional,targeting mammalian Ras as well as Rap1 and Rap2 as sub-strates (Kupzig et al. 2006).

RAP-1 is sufficient to induce ectopic 1� cells

If RAP-1 functions in parallel to LET-60 to activate LIN-45/Raf and hence induction of 1� cells, we would expect consti-tutively active RAP-1 to promote induction of ectopic 1� cells.We used CRISPR/Cas9 genome editing to introduce the clas-sical G12V activating mutation into the endogenous rap-1locus (Figure 5A). The G1 region in both Ras/Rap1 proteinsis exceptionally well-conserved across species (Figures S1and S2; Wennerberg et al. 2005; Reiner and Lundquist2016), allowing the G12V mutation to be successfully usedfor both Ras and Rap1 in C. elegans (Pellis-van Berkel et al.2005; Zand et al. 2011) and Drosophila (Mishra et al. 2005),as well as mammalian cell culture (Vossler et al. 1997). The

Figure 4 RAP-1 contributes to excretory ductcell and 1� VPC induction. (A) We observedRod-like lethality in rap-1(tm861); ksr-1(n2509)animals but not in the hypomorphic ksr-1(n2509) or rap-1(tm861) single mutant animals.(B) Deletion or (C) VPC-specific RNAi knockdownof rap-1 decreased ectopic 1� induction in thelet-23(sa62gf) (cis-marked with unc-4(e120))background. VPC-specific RNAi: rde-1(ne219);mfIs70[Plin-31::rde-1(+), Pmyo-2::GFP]. Luciferase-directed (luc) RNAi served as a negative control(Shin et al. 2018). (D) let-60(n1046gf) animalsare sensitive to gap-1– and lip-1–directed RNAi(depletion of RasGAP and ERK phosphatase, re-spectively), while rap-1(tm861) let-60(n1046gf)double mutant animals are insensitive. (D) let-60(n1046gf) vs. rap-1(tm861) let-60(n1046gf)animals were scored on separate days. Y-axesindicate number of ectopic 1� cells. N is indicatedby the numbers in the columns. Error bars indi-cate 95% confidence interval based on samplesize (A) or SEM (B–D). P values were calculatedby Fisher’s exact test (A), t-test (B and C), orANOVA (D). All data are representative non-pooled assays, collected on the same day.

1346 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

rap-1(re180gf) allele can be detected by PCR and restrictionenzyme digest due to the introduction of the concomitantsilent BamHI site (Figure 5B).

rap-1(re180gf) is sufficient to promote ectopic 1� cell dif-ferentiation (Figure 5C vs. Figure 5D and Figure S7A). Aspredicted based on let-60(n1046gf) (Beitel et al. 1990; Hanet al. 1990), the rap-1(re180gf) mutation acts in a gain-of-function manner, evidenced by semi-dominant induction ofectopic 1� cells (Figure 5E). Depletion of the LIP-1/ERK phos-phatase can increase ectopic 1� induction in sensitized back-grounds (Berset et al. 2001, 2005; Yoo et al. 2004; Zand et al.2011), and we used lip-1(RNAi) as a positive control in thisstudy (Figure 4D). Unexpectedly, lip-1(RNAi) failed to en-hance the ectopic 1� induction in the rap-1(re180gf) back-ground (Figure 5F).

To verify that rap-1(re180gf) was promoting 1� VPC tran-scriptional programs, we utilized two validated reporters of1�-promoting signaling, arIs131[Plag-2::YFP, Pceh-22::GFP]and arIs92[Pegl-17::CFP-LacZ, Pttx-3::GFP] (Yoo et al. 2004;

Zhang and Greenwald 2011). In wild-type animals, expres-sion of Plag-2::YFP was only observed in P6.px (Figure 6, Aand C), compared to rap-1(re180gf) mutants where we alsoobserved YFP in both P5.px and P7.px (Figure 6, B and D). Asimilar increase in ectopic expression of Pegl-17::CFP-LacZ wasobserved with Pegl-17::CFP-LacZ with the addition rap-1-(re180gf) (Figure S7, B–E). For both expression of Plag-2::YFP and Pegl-17::CFP-LacZ, the addition of rap-1(re180gf)resulted in a significant increase in expression outside ofP6.px. Rare expression in, and transformation of 3� VPCswas also observed (Figure S7A).

RAP-1 is sufficient to promote excretory ductcell duplication

Given the synthetic rod-like lethality we observed with rap-1(tm861) and ksr-1(n2509), we considered the potentialthat rap-1(re180gf) would result in duct cell duplication.We observed that some rap-1(re180gf) animals exhibited aventral protrusion near the posterior bulb of the pharynx

Figure 5 RAP-1 is sufficient to promote ectopic 1� cells. (A) Diagram of the strategy for CRISPR/Cas9-dependent knock-in of the activating (G12V)mutation into endogenous rap-1, along with a silent BamHI restriction site for genotyping. (B) PCR genotyping of the rap-1(re180gf) mutant allele,followed by BamHI digestion and agarose gel electrophoresis. (C and D) Representative DIC micrographs at the late L4 stage shows the wild-type vulva(open triangle) in (C) wild-type (WT) animals, and induction of ectopic 1�s (solid triangle) in (D) rap-1(re180gf). Bar, 10 mm. (E) Both heterozygous andhomozygous rap-1(re180gf) animals exhibited increased ectopic 1� induction compared to WT. bal: rap-1 balancer nT1qIs51[Pmyo-2::gfp, Ppes-10::gfp,PF22B7.9::gfp]. (F) RNAi depletion of lip-1 further increased ectopic 1�s (luc = luciferase RNAi negative control). Y-axis shows the mean number of ectopic1� VPCs 6 SEM. N is indicated by the numbers in the columns. P values were calculated by t-test.

RAP-1 Reinforces LET-60 Signaling 1347

(Figure 7B). let-60(n1046gf) confers a similar ventral pro-trusion, which was found to be a duplication of the excretoryduct cell (Beitel et al. 1990; Han et al. 1990; Yochem et al.1997). To determine if RAP-1 activation is sufficient to resultin excretory duct cell duplication, we used the saIs14 (Plin-48::GFP) transgene, a cell fate reporter for the duct cell (Johnsonet al. 2001). Mirroring our results in VPC cell fate patterning,rap-1(re180gf) was sufficient to drive duct-cell duplication,but at a reduced rate in comparison to let-60(n1046gf) (Fig-ure 7). We conclude that RAP-1 contributes to induction ofthe excretory duct cell.

PXF-1/RapGEF is required for maximal inductionof 1� cells

The Ras superfamily of small GTPases are tightly regulated byGEFs, which promote GTP loading and therefore active con-formation, and by GAPs, which promote GTP hydrolysis toGDP and hence inactive conformation (Wennerberg et al.2005). While certain mammalian GAP and GEF families haveoverlapping specificity for both Ras and Rap1, each alsomaintains unique regulators. Our results (Figure 4D, above)are consistent with GAP-1/GAP1 functioning to repress RAP-1 as well as the published LET-60/Ras (Hajnal et al. 1997;Stetak et al. 2008).

Vertebrate C3G, CA-GEF, EPAC, and PDZGEF have beenidentified as GEFs specific for Raps (Raaijmakers andBos 2009; Guo et al. 2016). The C. elegans ortholog ofPDZ-GEF, PXF-1, is essential for development. Loss ofpxf-1 function results in hypodermal defects, inability tomolt, and variable-onset larval lethality. Consistent withredundant roles for RAP-1 and RAP-2 in molting, loss ofeither individually results in live, superficially wild-typeanimals. However, loss of both rap-1 and rap-2 pheno-copies loss of pxf-1, suggesting that PXF-1 functions as ajoint GEF for both RAP-1 and RAP-2, as is seen in mam-mals (Pellis-van Berkel et al. 2005; Raaijmakers and Bos2009).

To determine potential activators of RAP-1 in VPC pattern-ing we assayed putative RapGEFs, pxf-1 and Y34B4A.4,that are expressed in tissues other than neurons (WormBaserelease WS262). Because of the essential role of PXF-1 indevelopment, we again utilized VPC-specific RNAi in a sen-sitized background (let-23(sa62gf); mfIs70[Plin-31::rde-1(+),Pmyo-2::gfp]; rde-1(ne219)), in which the RNAi defective phe-notype of the rde-1 mutation is rescued by VPC-specificexpression of wild-type RDE-1; Barkoulas et al. 2013). VPC-specific pxf-1(RNAi) (Figure 8A) but not Y34B4A.4(RNAi)(Figure S8A) resulted in decreased ectopic 1� induction.These findings are consistent with the PXF-1 RapGEF pro-moting 1� fate. Given this shared phenotype and the previ-ous association of PXF-1 with RAP-1 function in C. elegans,we hypothesize that PXF-1 is the Rap1 GEF in VPC fatepatterning.

Expression from the pxf-1 promoter is dynamicallyrestricted to 1� cells

While both RAP-1 and LET-60 are sufficient to induce trans-formation of ectopic 1� cells, our results suggest such theydiffer both in strength and spatial activity. One interpretationof this observation is that the activity of RAP-1 is more spa-tially restricted in the VPCs than is the activity LET-60, con-sistent with the modulatory role of RAP-1 compared to thecentral role of LET-60. Yet, like LET-60 (Dent and Han 1998),RAP-1 expression is uniform throughout VPCs (Figure 2).Consequently, we examined the expression of the putativeGEF for RAP-1, PXF-1.

A transgenic reporter fusion of the pxf-1 promoter to GFP(PT14G10.2::GFP-I), which includes 22394 to +26 relative tothe translational start codon of pxf-1 exon 1, expressed GFP inthe hypodermis (Figure 8B; Pellis-van Berkel et al. 2005). Weevaluated dynamic PT14G10.2::GFP-I expression in VPCs be-fore and after onset of LIN-3/EGF induction using the firstcell division (Pn.p to Pn.px; a conservative temporal indicatorof induction). Prior to LIN-3/EGF induction, we observed

Figure 6 RAP-1 activation results in ectopicexpression of a 1�-specific VPC transcriptionalreporter. (A and B) Representative inverted epi-fluorescence micrographs in animals at thetwo-cell (Pn.px) stage show the inappropriateexpression of 1� signaling reporter arIs131[P-

lag-2::2XNLS::YFP] in (B) rap-1(re180gf) but not(A) wild type. Bar, 10 mm. (C and D) Schematicrepresentation of the expression of the arIs131[Plag-2::2XNLS::YFP] signaling reporter acrossthe all six VPCs for both (C) control and (D)rap-1(re180gf). N is indicated by the numbersin the columns, with each line representing ananimal.

1348 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

uniform GFP expression in all VPCs (Figure 8, C and E). Afterinduction, we observed GFP expression to be increased in P6.p, the presumptive 1� cell, and reduced in P5.p and P7.p, thepresumptive 2� cells (Figure 8, D and F). We therefore hy-pothesize that pxf-1 expression is dynamically altered in re-sponse to the onset of LIN-3/EGF expression. We speculatethat this regulatory mechanism serves to dynamically restrictRAP-1 activation to the presumptive 1� cell during VPC pat-terning, while abrogating RAP-1 activation in presumptive2� cells (Figure 1).

Discussion

Our genetic results indicate that RAP-1 functions cell-autonomously to promote 1� fate. Our results are also consis-tent with RAP-1, in addition to LET-60, being a substrate ofthe bifunctional GAP-1/GAP1. Critically, the constitutivelyactivated endogenous RAP-1 mutant is semidominant and

is sufficient to induce 1� fate. And, like let-60(gf), rap-1(gf)duplicates excretory duct cells. Yet the ectopic 1� inductionand excretory duct cell duplication phenotypes conferred byrap-1(gf) are substantially weaker than those of let-60(gf). Inaddition, the loss of rap-1 resulted in rod-like lethality due topatterning failure of the excretory duct cell, but only in com-bination with hypomorphic mpk-1(ku1) or ksr-1(n2509).These observations indicate that, as expected, LET-60 playsa central role while RAP-1 plays amodulatory role. It is worthnoting that the G12V mutation that we use to activate RAP-1was never isolated for LET-60 in screens, presumably becausefull LET-60 activation is predicted to confer both lethality(Schutzman et al. 2001; Pellis-van Berkel et al. 2005) andsterility (Eisenmann and Kim 1997). G13 mutations are pre-dicted to beweaker in their activation of GTPases (Reiner andLundquist 2016).

We also found that PXF-1/PDZ-GEF, a Rap1GEF in mam-malian cells and in C. elegans, is required to maximally

Figure 7 RAP-1 is sufficient to duplicate excre-tory duct cells. Representative (A–C) DIC and(D–F) inverted epifluorescence micrographs of ex-cretory duct cell fate marker saIs14[Plin-48∷GFP] atthe L1 stage. (A and D) Wild-type animals have asingle duct, but both (B and E) rap-1(re180gf) and(C and F) let-60(n1046gf) are sufficient to dupli-cate the excretory duct. Bar, 10 mm. Black trian-gles ¼ duct cell, outlined triangles ¼duplicateduct cell. (G) Counting animals with two cells pos-itive for Plin-48::GFP expression indicated that bothactivated rap-1 and let-60 resulted in significantlyincreased duct cell duplications compared to thewild type. N is indicated by the numbers in thecolumns. Error bars indicate 95% confidence in-terval based on sample size. P values were calcu-lated by Fisher’s exact test.

RAP-1 Reinforces LET-60 Signaling 1349

promote 1� fate. This function, like that of RAP-1, is also cellautonomous. Thus, PXF-1 may function as the GEF that trig-gers RAP-1 activation in VPCs. Activation upstream of PXF-1remains unknown. Interestingly, like many other signalinggenes in the VPC signaling network (see Introduction), ex-pression from a pxf-1 promoter::GFP reporter changes afterinitial induction. Expression prior to EGF induction is uniformthroughout the VPCs, but after induction GFP signal in-creases in the presumptive 1� cell and is extinguished frompresumptive 2� cells. Accordingly, we hypothesize that as aconsequence of transcriptional reprogramming of pxf-1 ex-pression, the potential for RAP-1 activation is increased inpresumptive 1� cells and decreased in presumptive 2� cells(Figure 1).

RAP-1 may be activated by EGF in the same cells as LET-60, i.e., those VPCs closest to the EGF source, the AC. Yet wepropose that the functions of LET-60 and RAP-1 are qualita-tively different. The quantitatively weaker rap-1(re180gf)confers mostly 2�-to-1� transformations (this study), whilethe stronger let-60(n1046gf) confers mostly 3�-to-1� trans-formations: 2�s are rarely transformed or express markers ofERK activation in the let-60(n1046gf) background (Beitelet al. 1990; Han et al. 1990; Berset et al. 2001; Yoo et al.2004; Zand et al. 2011). For let-60(n1046gf) to transform2� cells to 1� cells, it takes dramatic derepression of theEGFR-Ras-Raf MEK-ERK 1�-promoting cascade, by simulta-neous inactivation of 2�-specific LIP-1/ERK phosphatase andDEP-1/EGFR phosphatase (Berset et al. 2005) or reduction of2�-promoting signals (Yoo et al. 2004; Yoo and Greenwald

2005; Zand et al. 2011). These differences between LET-60and RAP-1 beyond strength of transformation may reflectdifferent regulatory mechanisms between the two. For exam-ple, LET-60 is thought to be activated by the RasGEF SOS-1(Chang et al. 2000; Modzelewska et al. 2007), while our dataare consistent with RAP-1 activation by PXF-1. Conversely,while GAP-1 and GAP-3/RASA1 redundantly inhibit 1� fateinduction, mammalian isoforms of both are potentially bi-functional (Kupzig et al. 2006; Raaijmakers and Bos 2009),and their relative contributions to LET-60 and RAP-1 inacti-vation are unclear.

Mutations at the conserved active site residues 12, 13, and61 of Ras family small GTPases result in their constitutiveactivation. However, the observed isoform and codon speci-ficity in different Ras-driven tumors indicate that these mu-tations are not wholly interchangeable or equivalent, nor isactivation of different human Ras paralogs (Prior et al. 2012).As such, the differences we observed between let-60(n1046gf)and rap-1(re180gf) driving largely 3�-to-1� vs. 2�-to-1� trans-formations may be due to differences in their respective G13Eand G12V activating mutations.

Additionally,we seedifferent consequences of three classesof activating mutation in let-60: n1046gf (G13E) mostly con-fers 3�-to-1� transformations and duct duplications (Beitelet al. 1990; Han et al. 1990), ay75gf,ts (G60R) mostly dereg-ulates fluid homeostasis in the EGL-15/FGF receptor cascade(Schutzman et al. 2001), and ga89gf, ts (L19F) mostly con-fers sterility (Eisenmann and Kim 1997). Yet these pheno-types are dependent on LIN-45/Raf activation of MPK-1/

Figure 8 PXF-1/RapGEF contributes to 1� VPC induction. (A) VPC-specific RNAi depletion of pxf-1 in the let-23(sa62gf) background decreased ectopic1� induction compared to Luciferase (luc) RNAi control (sa62gf was cis-marked with unc-4(e120)). (VPC-specific RNAi: rde-1(ne219);mfIs70[Plin-31::rde-1(+), Pmyo-2::gfp ].). The y-axis indicates number of ectopic 1�s. All data are representative nonpooled assays. Error bars indicate SEM, numbers in columnsindicate N. P values were calculated by t-test. (B) Diagram of the Ppxf-1::GFP reporter construct. DIC micrographs of L3 larvae at the (C) early one-cellstage (Pn.p), prior to EGF signaling (representative of seven images), and the (D) two-cell stage (Pn.px), after EGF signaling (representative of 12 images).(E and F) Representative inverted epifluorescence micrographs of the animals shown in A and B, showing GFP expression from a pxf-1 transcriptionalGFP reporter fusion transgene (bjIs40[Ppxf-1::GFP]). (E) GFP was observed throughout the VPCs (shaded labels) prior to EGF induction at the early one-cellstage. (F) At the two-cell stage, GFP increased in presumptive 1� (solid labels) and decreased in presumptive 2� cells (open labels); expression inpresumptive 3�s did not change (shaded labels). Bar, 10 mm.

1350 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

ERK. So there remain many mysteries about the differentialeffects of Ras family activation, including RAP-1 in this study.

While our data with LET-60 and RAP-1 are consistent withtheir functioning in parallel through lin-45/Raf, neitherGTPase is limited to a sole effector. Our previous work(Zand et al. 2011) showed LET-60 to promote both 1 and2� cell fate through alternate effectors Raf and RalGEF, re-spectively: when RGL-1 is depleted, let-60(n146gf) trans-forms nearly every 3� to 1� fate, as well as some 2�s.Consequently, we hypothesize that it is in part the use ofRGL-1 to promote 2� fate that prevents activated LET-60 fromtransforming 2�s to 1�s, along with the LIN-12/Notch lateralsignal. If RAP-1 utilizes only LIN-45 and not RGL-1, thenactivated RAP-1 could be expected to transform 2�s to 1�s.

Alternatively, RAP-1 may be activated after LET-60. If so,RAP-1 may only be sufficient to transform VPCs that havereceived transient LET-60-LIN-45 activation of MPK-1/ERK;we know that P5,p and P7.p receive a brief pulse of 1�-promotingMPK-1 signal, but this signal is typically quenched by LIN-12lateral signal activating transcription of LIP-1, an ERK phospha-tase (Berset et al. 2001; Yoo et al. 2004).

RAP-1 mutant phenotypes and reprogramming of expres-sion from the pxf-1 promoter after EGF induction are consis-tent with other transcriptional reprogramming events invulva signaling. We hypothesize that many levels of suchregulation impose higher developmental fidelity on the sys-tem: a collective reconfiguration of signaling protein expres-sion could refine the action of signals, thus providing thetransition from a ligand gradient to discrete binary outputs.In other words, reconfiguration of signals could decrease thepotential for developmental noise inherent in a gradient, orcontradictory signals within the same cell. We potentially seethis noise in our data with lst-5::YFP promoter fusion havingweak expression outside of P5.px and P7.px that was signif-icantly increased with the loss of rap-1. However, interpreta-tion of these data are complicated, in that we also see loss ofrap-1 resulting in both underinduction of 2� VPCs and poten-tial mixed fates. The underinduced 2� events represent only aportion of the rare abnormal phenotypes we observed withrap-1 loss may explain why we did not observe any loss ofPlst-5::YFP expression in either P5.px or P7.px.

Our results are also consistent with the role of Rap1 indevelopmental signaling as a “shadow Ras” that echoes andreinforces the main function of Ras signaling through Raf. Itis difficult, however, to prove that RAP-1 signals through LIN-45/Raf. Yet our results are consistent with this model, asare the molecular characteristics of RAP-1 as a close rela-tive of LET-60. However, an alternative model is that RAP-1acts to reinforce other aspects of the 1�-promoting signal.For example, RAP-1 could promote or stabilize EGFR sig-naling, as Drosophila Rap1 has been shown to regulateEGFR membrane localization and polarity through its junc-tional effector Canoe/afadin/AF6 (O’Keefe et al. 2009;Baril et al. 2014). Such a function, piggybacking on thestrong induction of 1� cells by LET-60, could induce occa-sional ectopic 1� cells.

Our studyprovidesmechanistic insights into the functionofRAP-1, the C. elegans ortholog of mammalian Rap1 proteinsthat function as putative human oncogenes. We additionallyestablish a role for the RapGEF PXF-1 as the putative up-stream GEF for RAP-1 in VPC fate patterning, and argue thattranscriptional control of the pxf-1 promoter over time pro-vides an example of how the VPC signaling network is dy-namically reprogrammed during development to preventpotentially conflicting signals. We hypothesize that PXF-1-RAP-1 functions as a modulatory parallel input to rein-force and spatially restrict LET-60-LIN-45–dependent 1� cellinduction.

Acknowledgments

Some strains were provided by the Caenorhabditis GeneticsCenter (CGC), which is funded by National Institutes ofHealth Office of Research Infrastructure Programs (P40OD010440) and I. Greenwald. WormBase was used regu-larly. This work was supported by National Institutes ofHealth grants R01-GM121625 and R21-HD090707 to D.J.R., an American Cancer Society PF-16-083-01 post-doctoralfellowship to N.R.R., and a Howard Hughes PostdoctoralFellowship from the Helen Hay Whitney Foundation toD.J.D.

Literature Cited

Abdus-Saboor, I., V. P. Mancuso, J. I. Murray, K. Palozola, C. Norriset al., 2011 Notch and Ras promote sequential steps of excre-tory tube development in C. elegans. Development 138: 3545–3555. https://doi.org/10.1242/dev.068148

Altschuler, D. L., and F. Ribeiro-Neto, 1998 Mitogenic and oncogenicproperties of the small G protein Rap1b. Proc. Natl. Acad. Sci. USA95: 7475–7479. https://doi.org/10.1073/pnas.95.13.7475

Aoki, Y., T. Niihori, S. Inoue, and Y. Matsubara, 2016 Recent ad-vances in RASopathies. J. Hum. Genet. 61: 33–39. https://doi.org/10.1038/jhg.2015.114

Baril, C., M. Lefrancois, M. Sahmi, H. Knaevelsrud, and M. Therrien,2014 Apical accumulation of the Sevenless receptor tyrosinekinase during Drosophila eye development is promoted by thesmall GTPase Rap1. Genetics 197: 1237–1250. https://doi.org/10.1534/genetics.114.166272

Barkoulas, M., J. S. van Zon, J. Milloz, A. van Oudenaarden, and M.A. Felix, 2013 Robustness and epistasis in the C. elegans vulvalsignaling network revealed by pathway dosage modulation. Dev.Cell 24: 64–75. https://doi.org/10.1016/j.devcel.2012.12.001

Beitel, G. J., S. G. Clark, and H. R. Horvitz, 1990 Caenorhabditiselegans ras gene let-60 acts as a switch in the pathway of vul-val induction. Nature 348: 503–509. https://doi.org/10.1038/348503a0

Beranger, F., B. Goud, A. Tavitian, and J. de Gunzburg, 1991 Associationof the Ras-antagonistic Rap1/Krev-1 proteins with the Golgi complex.Proc. Natl. Acad. Sci. USA 88: 1606–1610. https://doi.org/10.1073/pnas.88.5.1606

Berset, T., E. F. Hoier, G. Battu, S. Canevascini, and A. Hajnal,2001 Notch inhibition of RAS signaling through MAP kinasephosphatase LIP-1 during C. elegans vulval development.Science 291: 1055–1058. https://doi.org/10.1126/science.1055642

RAP-1 Reinforces LET-60 Signaling 1351

Berset, T. A., E. F. Hoier, and A. Hajnal, 2005 The C. eleganshomolog of the mammalian tumor suppressor Dep-1/Scc1 in-hibits EGFR signaling to regulate binary cell fate decisions.Genes Dev. 19: 1328–1340. https://doi.org/10.1101/gad.333505

Bögershausen, N., I. C. Tsai, E. Pohl, P. O. Kiper, F. Beleggia et al.,2015 RAP1-mediated MEK/ERK pathway defects in Kabukisyndrome. J. Clin. Invest. 125: 3585–3599. https://doi.org/10.1172/JCI80102

Bos, J. L., 2005 Linking Rap to cell adhesion. Curr. Opin. Cell Biol.17: 123–128. https://doi.org/10.1016/j.ceb.2005.02.009

Braendle, C., and M. A. Felix, 2008 Plasticity and errors of arobust developmental system in different environments. Dev.Cell 15: 714–724. https://doi.org/10.1016/j.devcel.2008.09.011

Brenner, S., 1974 The genetics of Caenorhabditis elegans. Genet-ics 77: 71–94.

Caron, E., A. J. Self, and A. Hall, 2000 The GTPase Rap1 controlsfunctional activation of macrophage integrin alphaMbeta2 byLPS and other inflammatory mediators. Curr. Biol. 10: 974–978. https://doi.org/10.1016/S0960-9822(00)00641-2

Chang, C., N. A. Hopper, and P. W. Sternberg, 2000 Caenorhabditiselegans SOS-1 is necessary for multiple RAS-mediated develop-mental signals. EMBO J. 19: 3283–3294. https://doi.org/10.1093/emboj/19.13.3283

Chen, N., and I. Greenwald, 2004 The lateral signal for LIN-12/Notch in C. elegans vulval development comprises redundantsecreted and transmembrane DSL proteins. Dev. Cell 6: 183–192. https://doi.org/10.1016/S1534-5807(04)00021-8

Cook, S. J., B. Rubinfeld, I. Albert, and F. McCormick,1993 RapV12 antagonizes Ras-dependent activation of ERK1and ERK2 by LPA and EGF in Rat-1 fibroblasts. EMBO J. 12:3475–3485. https://doi.org/10.1002/j.1460-2075.1993.tb06022.x

Dent, J. A., and M. Han, 1998 Post-embryonic expression patternof C. elegans let-60 ras reporter constructs. Mech. Dev. 72: 179–182. https://doi.org/10.1016/S0925-4773(98)00026-4

Dickinson, D. J., A. M. Pani, J. K. Heppert, C. D. Higgins, and B.Goldstein, 2015 Streamlined genome engineering with a self-excising drug selection cassette. Genetics 200: 1035–1049.https://doi.org/10.1534/genetics.115.178335

Eisenmann, D. M., and S. K. Kim, 1997 Mechanism of activationof the Caenorhabditis elegans ras homologue let-60 by a novel,temperature-sensitive, gain-of-function mutation. Genetics 146:553–565.

Ferguson, E. L., and H. R. Horvitz, 1985 Identification and char-acterization of 22 genes that affect the vulval cell lineages of thenematode Caenorhabditis elegans. Genetics 110: 17–72.

Frische, E. W., W. Pellis-van Berkel, G. van Haaften, E. Cuppen, R.H. Plasterk et al., 2007 RAP-1 and the RAL-1/exocyst pathwaycoordinate hypodermal cell organization in Caenorhabditis ele-gans. EMBO J. 26: 5083–5092. https://doi.org/10.1038/sj.emboj.7601922

Gloerich, M., and J. L. Bos, 2011 Regulating Rap small G-proteinsin time and space. Trends Cell Biol. 21: 615–623. https://doi.org/10.1016/j.tcb.2011.07.001

Greenwald, I., and R. Kovall, 2013 Notch signaling: genetics andstructure (January 17, 2013), WormBook, ed. The C. elegansResearch Community, WormBook, doi/10.1895/wormbook.1.10.2,http://www.wormbook.org. https://doi.org/10.1895/wormbook.1.10.2

Greenwald, I. S., P. W. Sternberg, and H. R. Horvitz, 1983 Thelin-12 locus specifies cell fates in Caenorhabditis elegans. Cell34: 435–444. https://doi.org/10.1016/0092-8674(83)90377-X

Grewal, T., M. Koese, F. Tebar, and C. Enrich, 2011 Differentialregulation of RasGAPs in cancer. Genes Cancer 2: 288–297.https://doi.org/10.1177/1947601911407330

Guo, X. X., S. An, Y. Yang, Y. Liu, Q. Hao et al., 2016 Rap-inter-acting proteins are key players in the Rap symphony orchestra.Cell. Physiol. Biochem. 39: 137–156. https://doi.org/10.1159/000445612

Gyan, E., M. Frew, D. Bowen, C. Beldjord, C. Preudhomme et al.,2005 Mutation in RAP1 is a rare event in myelodysplastic syn-dromes. Leukemia 19: 1678–1680. https://doi.org/10.1038/sj.leu.2403882

Hagedorn, E. J., H. Yashiro, J. W. Ziel, S. Ihara, Z. Wang et al.,2009 Integrin acts upstream of netrin signaling to regulateformation of the anchor cell’s invasive membrane in C. elegans.Dev. Cell 17: 187–198. https://doi.org/10.1016/j.devcel.2009.06.006

Hajnal, A., C. W. Whitfield, and S. K. Kim, 1997 Inhibition ofCaenorhabditis elegans vulval induction by gap-1 and by let-23 receptor tyrosine kinase. Genes Dev. 11: 2715–2728. https://doi.org/10.1101/gad.11.20.2715

Han, M., and P. W. Sternberg, 1990 let-60, a gene that spec-ifies cell fates during C. elegans vulval induction, encodes aras protein. Cell 63: 921–931. https://doi.org/10.1016/0092-8674(90)90495-Z

Han, M., R. V. Aroian, and P. W. Sternberg, 1990 The let-60 locuscontrols the switch between vulval and nonvulval cell fates inCaenorhabditis elegans. Genetics 126: 899–913.

Hancock, J. F., H. Paterson, and C. J. Marshall, 1990 A polybasicdomain or palmitoylation is required in addition to the CAAXmotif to localize p21ras to the plasma membrane. Cell 63: 133–139. https://doi.org/10.1016/0092-8674(90)90294-O

Hara, M., and M. Han, 1995 Ras farnesyltransferase inhibitorssuppress the phenotype resulting from an activated ras mutationin Caenorhabditis elegans. Proc. Natl. Acad. Sci. USA 92: 3333–3337. https://doi.org/10.1073/pnas.92.8.3333

Hariharan, I. K., R. W. Carthew, and G. M. Rubin, 1991 The Dro-sophila roughened mutation: activation of a rap homologdisrupts eye development and interferes with cell determina-tion. Cell 67: 717–722. https://doi.org/10.1016/0092-8674(91)90066-8

Johnson, A. D., D. Fitzsimmons, J. Hagman, and H. M. Chamberlin,2001 EGL-38 Pax regulates the ovo-related gene lin-48 duringCaenorhabditis elegans organ development. Development 128:2857–2865.

Kamath, R. S., and J. Ahringer, 2003 Genome-wide RNAi screen-ing in Caenorhabditis elegans. Methods 30: 313–321. https://doi.org/10.1016/S1046-2023(03)00050-1

Katz, W. S., G. M. Lesa, D. Yannoukakos, T. R. Clandinin, J. Schles-singer et al., 1996 A point mutation in the extracellular do-main activates LET-23, the Caenorhabditis elegans epidermalgrowth factor receptor homolog. Mol. Cell. Biol. 16: 529–537.https://doi.org/10.1128/MCB.16.2.529

Kitayama, H., Y. Sugimoto, T. Matsuzaki, Y. Ikawa, and M.Noda, 1989 A ras-related gene with transformation suppres-sor activity. Cell 56: 77–84. https://doi.org/10.1016/0092-8674(89)90985-9

Kitayama, H., T. Matsuzaki, Y. Ikawa, and M. Noda, 1990 A do-main responsible for the transformation suppressor activity inKrev-1 protein. Jpn. J. Cancer Res. 81: 445–448. https://doi.org/10.1111/j.1349-7006.1990.tb02589.x

Knox, A. L., and N. H. Brown, 2002 Rap1 GTPase regulation ofadherens junction positioning and cell adhesion. Science 295:1285–1288. https://doi.org/10.1126/science.1067549

Kornfeld, K., D. B. Hom, and H. R. Horvitz, 1995 The ksr-1 geneencodes a novel protein kinase involved in ras-mediated signal-ing in C. elegans. Cell 83: 903–913.

Krauthammer, M., Y. Kong, B. H. Ha, P. Evans, A. Bacchiocchi et al.,2012 Exome sequencing identifies recurrent somatic RAC1mutations in melanoma. Nat. Genet. 44: 1006–1014. https://doi.org/10.1038/ng.2359

1352 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner

Kupzig, S., D. Deaconescu, D. Bouyoucef, S. A. Walker, Q. Liu et al.,2006 GAP1 family members constitute bifunctional Ras andRap GTPase-activating proteins. J. Biol. Chem. 281: 9891–9900. https://doi.org/10.1074/jbc.M512802200

Li, J., and I. Greenwald, 2010 LIN-14 inhibition of LIN-12 con-tributes to precision and timing of C. elegans vulval fate pattern-ing. Curr. Biol. 20: 1875–1879.

Lindsay, C. R., S. Lawn, A. D. Campbell, W. J. Faller, F. Rambowet al., 2011 P-Rex1 is required for efficient melanoblast migra-tion and melanoma metastasis. Nat. Commun. 2: 555. https://doi.org/10.1038/ncomms1560

Maertens, O., and K. Cichowski, 2014 An expanding role forRAS GTPase activating proteins (RAS GAPs) in cancer. Adv.Biol. Regul. 55: 1–14. https://doi.org/10.1016/j.jbior.2014.04.002

Mavromatakis, Y. E., and A. Tomlinson, 2012 The role of thesmall GTPase Rap in Drosophila R7 photoreceptor specification.Proc. Natl. Acad. Sci. USA 109: 3844–3849. https://doi.org/10.1073/pnas.1115108109

McLaughlin, S. K., S. N. Olsen, B. Dake, T. De Raedt, E. Lim et al.,2013 The RasGAP gene, RASAL2, is a tumor and metastasissuppressor. Cancer Cell 24: 365–378. https://doi.org/10.1016/j.ccr.2013.08.004

Mishra, S., S. M. Smolik, M. A. Forte, and P. J. Stork, 2005 Ras-independent activation of ERK signaling via the torso receptortyrosine kinase is mediated by Rap1. Curr. Biol. 15: 366–370.https://doi.org/10.1016/j.cub.2005.02.022

Modzelewska, K., M. G. Elgort, J. Huang, G. Jongeward, A. Laur-itzen et al., 2007 An activating mutation in sos-1 identifies itsDbl domain as a critical inhibitor of the epidermal growth factorreceptor pathway during Caenorhabditis elegans vulval devel-opment. Mol. Cell. Biol. 27: 3695–3707. https://doi.org/10.1128/MCB.01630-06

O’Keefe, D. D., E. Gonzalez-Nino, M. Burnett, L. Dylla, S. M. Lam-beth et al., 2009 Rap1 maintains adhesion between cells toaffect Egfr signaling and planar cell polarity in Drosophila.Dev. Biol. 333: 143–160. https://doi.org/10.1016/j.ydbio.2009.06.032

Paix, A., Y. Wang, H. E. Smith, C. Y. Lee, D. Calidas et al.,2014 Scalable and versatile genome editing using linear DNAswith microhomology to Cas9 Sites in Caenorhabditis elegans.Genetics 198: 1347–1356. https://doi.org/10.1534/genet-ics.114.170423

Papke, B., and C. J. Der, 2017 Drugging RAS: know the enemy.Science 355: 1158–1163. https://doi.org/10.1126/science.aam7622

Pellis-van Berkel, W., M. H. Verheijen, E. Cuppen, M. Asahina, J. deRooij et al., 2005 Requirement of the Caenorhabditis elegansRapGEF pxf-1 and rap-1 for epithelial integrity. Mol. Biol. Cell16: 106–116. https://doi.org/10.1091/mbc.e04-06-0492

Pizon, V., M. Desjardins, C. Bucci, R. G. Parton, and M. Zerial,1994 Association of Rap1a and Rap1b proteins with late en-docytic/phagocytic compartments and Rap2a with the Golgicomplex. J. Cell Sci. 107: 1661–1670.

Prior, I. A., and J. F. Hancock, 2012 Ras trafficking, localizationand compartmentalized signalling. Semin. Cell Dev. Biol. 23:145–153. https://doi.org/10.1016/j.semcdb.2011.09.002

Prior, I. A., P. D. Lewis, and C. Mattos, 2012 A comprehensivesurvey of Ras mutations in cancer. Cancer Res. 72: 2457–2467. https://doi.org/10.1158/0008-5472.CAN-11-2612

Raaijmakers, J. H., and J. L. Bos, 2009 Specificity in Ras and Rapsignaling. J. Biol. Chem. 284: 10995–10999. https://doi.org/10.1074/jbc.R800061200

Reedquist, K. A., E. Ross, E. A. Koop, R. M. Wolthuis, F. J. Zwartk-ruis et al., 2000 The small GTPase, Rap1, mediates CD31-in-duced integrin adhesion. J. Cell Biol. 148: 1151–1158. https://doi.org/10.1083/jcb.148.6.1151

Reiner, D. J., 2011 Ras effector switching as a developmentalstrategy. Small GTPases 2: 109–112. https://doi.org/10.4161/sgtp.2.2.15775

Reiner, D. J., and E. A. Lundquist, 2016 Small GTPases (January17, 2006), WormBook, ed. The C. elegans Research Commu-nity, WormBook, doi/10.1895/wormbook.1.67.1, http://www.wormbook.org.

Rivera Gomez, K. A., and M. Schvarzstein, 2018 Immobilizationnematodes for live imaging using an agarose pad produced witha Vinyl Record. microPublication Biology. https://doi.org/10.17912/QG0J-VT85

Sakoda, T., K. Kaibuchi, K. Kishi, S. Kishida, K. Doi et al., 1992 smg/rap1/Krev-1 p21s inhibit the signal pathway to the c-fos promoter/enhancer from c-Ki-ras p21 but not from c-raf-1 kinase in NIH3T3cells. Oncogene 7: 1705–1711.

Sallee, M.D., and I. Greenwald, 2015 Dimerization-driven degra-dation of C. elegans and human E proteins. Genes Dev 29(13):1356–1361.

Schutzman, J. L., C. Z. Borland, J. C. Newman, M. K. Robinson, M.Kokel et al., 2001 The Caenorhabditis elegans EGL-15 signal-ing pathway implicates a DOS-like multisubstrate adaptor pro-tein in fibroblast growth factor signal transduction. Mol. Cell.Biol. 21: 8104–8116. https://doi.org/10.1128/MCB.21.23.8104-8116.2001

Sergent, J., 1990 Furtive incursions into bicameral minds. Inte-grative and coordinating role of subcortical structures. Brain113: 537–568. https://doi.org/10.1093/brain/113.2.537

Shaner, N. C., G. G. Lambert, A. Chammas, Y. Ni, P. J. Cranfillet al., 2013 A bright monomeric green fluorescent proteinderived from Branchiostoma lanceolatum. Nat. Methods 10:407–409.

Shaye, D. D., and I. Greenwald, 2002 Endocytosis-mediateddownregulation of LIN-12/Notch upon Ras activation in Caeno-rhabditis elegans. Nature 420: 686–690. https://doi.org/10.1038/nature01234

Shaye, D. D., and I. Greenwald, 2005 LIN-12/Notch traffickingand regulation of DSL ligand activity during vulval inductionin Caenorhabditis elegans. Development 132: 5081–5092. https://doi.org/10.1242/dev.02076

Sherwood, D. R., and P. W. Sternberg, 2003 Anchor cell invasioninto the vulval epithelium in C. elegans. Dev. Cell 5: 21–31.https://doi.org/10.1016/S1534-5807(03)00168-0

Shin, H., R. E. W. Kaplan, T. Duong, R. Fakieh, and D. J. Reiner,2018 Ral signals through a MAP4 kinase-p38 MAP kinase cas-cade in C. elegans cell fate patterning. Cell Rep. 24: 2669–2681e2665.

Srijakotre, N., J. Man, L. M. Ooms, C. M. Lucato, A. M. Ellisdonet al., 2017 P-Rex1 and P-Rex2 RacGEFs and cancer. Bio-chem. Soc. Trans. 45: 963–977. https://doi.org/10.1042/BST20160269

Sternberg, P. W., 2005 Vulval development (June 25, 2005),WormBook, ed. The C. elegans Research Community, Worm-Book, doi/10.1895/wormbook.1.6.1, http://www.wormbook.org.

Stetak, A., P. Gutierrez, and A. Hajnal, 2008 Tissue-specific func-tions of the Caenorhabditis elegans p120 Ras GTPase activatingprotein GAP-3. Dev. Biol. 323: 166–176. https://doi.org/10.1016/j.ydbio.2008.08.026

Sundaram, M. V., 2013 Canonical RTK-Ras-ERK signaling and re-lated alternative pathways (July 1, 2013), WormBook, ed. TheC. elegans Research Community, WormBook, doi/10.1895/wormbook.1.80.2, http://www.wormbook.org. https://doi.org/10.1895/wormbook.1.80.2

Vossler, M. R., H. Yao, R. D. York, M. G. Pan, C. S. Rim et al.,1997 cAMP activates MAP kinase and Elk-1 through a B-Raf-and Rap1-dependent pathway. Cell 89: 73–82. https://doi.org/10.1016/S0092-8674(00)80184-1

RAP-1 Reinforces LET-60 Signaling 1353

Wang, M., and P. W. Sternberg, 1999 Competence and commit-ment of Caenorhabditis elegans vulval precursor cells. Dev. Biol.212: 12–24. https://doi.org/10.1006/dbio.1999.9357

Wennerberg, K., K. L. Rossman, and C. J. Der, 2005 The Rassuperfamily at a glance. J. Cell Sci. 118: 843–846. https://doi.org/10.1242/jcs.01660

Wu, Y., and M. Han, 1994 Suppression of activated Let-60 rasprotein defines a role of Caenorhabditis elegans Sur-1 MAP ki-nase in vulval differentiation. Genes Dev. 8: 147–159. https://doi.org/10.1101/gad.8.2.147

Yochem, J., M. Sundaram, and M. Han, 1997 Ras is required for alimited number of cell fates and not for general proliferation inCaenorhabditis elegans. Mol. Cell. Biol. 17: 2716–2722. https://doi.org/10.1128/MCB.17.5.2716

Yoo, A. S., and I. Greenwald, 2005 LIN-12/Notch activation leadsto microRNA-mediated down-regulation of Vav in C. elegans.Science 310: 1330–1333. https://doi.org/10.1126/science.1119481

Yoo, A. S., C. Bais, and I. Greenwald, 2004 Crosstalk between theEGFR and LIN-12/Notch pathways in C. elegans vulval devel-opment. Science 303: 663–666. https://doi.org/10.1126/sci-ence.1091639

York, R. D., H. Yao, T. Dillon, C. L. Ellig, S. P. Eckert et al.,1998 Rap1 mediates sustained MAP kinase activation induced

by nerve growth factor. Nature 392: 622–626. https://doi.org/10.1038/33451

Zand, T. P., D. J. Reiner, and C. J. Der, 2011 Ras effector switch-ing promotes divergent cell fates in C. elegans vulval patterning.Dev. Cell 20: 84–96. https://doi.org/10.1016/j.devcel.2010.12.004

Zhang, X., and I. Greenwald, 2011 Spatial regulation of lag-2transcription during vulval precursor cell fate patterning in Cae-norhabditis elegans. Genetics 188: 847–858. https://doi.org/10.1534/genetics.111.128389

Zhao, Z., C. C. Chen, C. D. Rillahan, R. Shen, T. Kitzing et al.,2015 Cooperative loss of RAS feedback regulation drives my-eloid leukemogenesis. Nat. Genet. 47: 539–543. https://doi.org/10.1038/ng.3251

Ziel, J. W., E. J. Hagedorn, A. Audhya, and D. R. Sherwood,2009 UNC-6 (netrin) orients the invasive membrane of theanchor cell in C. elegans. Nat. Cell Biol. 11: 183–189. https://doi.org/10.1038/ncb1825

Zwartkruis, F. J., R. M. Wolthuis, N. M. Nabben, B. Franke, and J. L.Bos, 1998 Extracellular signal-regulated activation of Rap1fails to interfere in Ras effector signalling. EMBO J. 17: 5905–5912. https://doi.org/10.1093/emboj/17.20.5905

Communicating editor: M. Sundaram

1354 N. R. Rasmussen, D. J. Dickinson, and D. J. Reiner