RAB-7 Antagonizes LET-23 EGFR Signaling during Vulva Development in Caenorhabditis elegans

The Rab7 GTPase regulates late endosome trafficking of the Epidermal Growth Factor Receptor (EGFR) to the lysosome for degradation. However, less is known about how Rab7 activity, functioning late in the endocytic pathway, affects EGFR signaling. Here we used Caenorhabditis elegans vulva cell fate induction, a paradigm for genetic analysis of EGFR/Receptor Tyrosine Kinase (RTK) signaling, to assess the genetic requirements for rab-7. Using a rab-7 deletion mutant, we demonstrate that rab-7 antagonizes LET-23 EGFR signaling to a similar extent, but in a distinct manner, as previously described negative regulators such as sli-1 c-Cbl. Epistasis analysis places rab-7 upstream of or in parallel to lin-3 EGF and let-23 EGFR. However, expression of gfp::rab-7 in the Vulva Presursor Cells (VPCs) is sufficient to rescue the rab-7(−) VPC induction phenotypes indicating that RAB-7 functions in the signal receiving cell. We show that components of the Endosomal Sorting Complex Required for Transport (ESCRT)-0, and -I, complexes, hgrs-1 Hrs, and vps-28, also antagonize signaling, suggesting that LET-23 EGFR likely transits through Multivesicular Bodies (MVBs) en route to the lysosome. Consistent with RAB-7 regulating LET-23 EGFR trafficking, rab-7 mutants have increased number of LET-23::GFP-positive endosomes. Our data imply that Rab7, by mediating EGFR trafficking and degradation, plays an important role in downregulation of EGFR signaling. Failure to downregulate EGFR signaling contributes to oncogenesis, and thus Rab7 could possess tumor suppressor activity in humans.


Introduction
The EGFR/Ras GTPase/Mitogen Activated Protein Kinase (MAPK) signal transduction pathway is evolutionarily conserved and is essential for animal development [1,2]. Activating mutations in the EGFR/Ras/MAPK pathway are commonly found in human cancers and mutations in several components of the Ras/ MAPK pathway have been shown to cause Noonan syndrome as well as several related developmental disorders [3][4][5]. The EGFR is activated by ligand binding, which stimulates receptor dimerization, transautophosphorylation of cytoplasmic Tyrosine residues, and recruitment of phospho-Tyrosine binding proteins such as Grb2. The binding of the Grb2 adaptor protein and the associated SOS protein (a Ras Guanine nucleotide Exchange Factor) to the EGFR results in the activation of the membrane associated Ras GTPase, and subsequently activation of the MAPK cascade consisting of Raf, MAPK/ERK Kinase (MEK) and Extracellular Regulated Kinase (ERK) [6]. EGFR and Grb2 also recruit Cbl, an E3-ubiquitin ligase, which ubiquitinates Lysine residues on the EGFR, targeting it for lysosomal degradation [7]. Ubiquitination contributes to EGFR endocytosis and sorting into late endosomes/MVBs. A series of ESCRT complexes (0, I, II, & III) on the MVBs recognize and internalize the ubiquitinated EGFR into intraluminal vesicles (ILVs), sequestering the EGFR away from the cytoplasm [7]. The ultimate fusion of MVBs with the lysosome results in degradation of the EGFR. Since the EGFR can continue to signal from endosomes, regulators of endocytic trafficking are positioned to regulate the duration and strength of signaling.
The Rab5 and Rab7 GTPases are key regulators of early endosome and late endosome trafficking, respectively, and regulate EGFR trafficking to the lysosome [8]. Rab5 regulates EGFR internalization, while EGFR signaling can modulate Rab5 activity by regulating its Guanine nucleotide exchange factor, RIN1, and GTPase Activating Protein, RN-tre [9][10][11][12][13]. In some cases, Rab5 can function as a downstream effector of EGFR signaling [14]. While Rab7 activity promotes EGFR trafficking from late endosomes/MVBs to the lysosome [15][16][17], it is unclear whether inhibiting EGFR degradation late in the endocytic pathway would impact EGFR signaling. EGFR accumulates in the ILVs of MVBs of Rab7 RNAi treated cells, where it would potentially be sequestered away from the cytoplasm and downstream effectors [17].
The EGFR/Ras/MAPK pathway is highly conserved in the nematode C. elegans where it is required for specifying cell fates during development ( Figure 1A) [1]. During C. elegans vulva development, LET-23 EGFR and LIN-12 Notch signaling pathways specify three of six VPCs to adopt the 1u and 2u vulval cell fates ( Figure 1B) [18]. The six VPCs are polarized epithelial cells, named P3.p-P8.p, that have LET-23 EGFR localized to the basolateral membrane. The Anchor Cell, in the overlying gonad, secretes the LIN-3 EGF-like ligand, activating LET-23 EGFR signaling cascade most strongly in P6.p, the closest VPC, inducing it to adopt a 1u vulval cell fate. P6.p subsequently activates  Notch signaling in the neighboring cells. LIN-12 Notch signaling, along with the graded LIN-3 signal, induce the P5.p and P7.p cells to adopt the 2u vulval cell fate. The remaining P3.p, P4.p, and P8.p cells adopt a 3u non-vulval cell fate, divide once, and fuse with the surrounding hypodermis (roughly 50% of P3.p cells fuse prior to dividing). The induced P5.p-P7.p cells undergo a stereotypic set of cell divisions to give rise to the 22 cells of the mature vulva, 8 cells from the 1u (P6.p) cell and 7 cells from each of the 2u (P5.p and P7.p) cells ( Figure 1B). Thus, mutations that reduce LET-23 EGFR signaling result in a Vulvaless (Vul) phenotype in which fewer than three VPCs are induced, and mutations that enhance LET-23 EGFR signaling result in a Multivulva (Muv) phenotype in which greater than three VPCs are induced.
Here we present genetic evidence demonstrating a role for RAB-7 as a negative regulator of LET-23 EGFR signaling during C. elegans vulva development. We show that a rab-7 deletion mutant suppresses the Vul and enhances the Muv phenotypes of let-60 ras hypomorphic and hypermorphic alleles, respectively. Similar to previously characterized negative regulators of EGFR signaling, rab-7(ok511) is synthetic Muv in combination with loss of negative regulators unc-101 AP-1 and ark-1 Ack, and suppresses the Vul phenotypes of mutations that disrupt the basolateral localization of LET-23 EGFR. We show that rab-7 functions upstream of or in parallel to lin-3 EGF and let-23 EGFR, and that it functions in the VPCs and regulates LET-23::GFP localization. Importantly, our data suggest that Rab7 could negatively regulate EGFR signaling in humans where increased EGFR signaling can contribute to oncogenesis.

Microscopy and Phenotype Analysis
General methods for Nomarski differential interference contrast (DIC) microscopy of live animals were as previously described [26]. Animals were analyzed on an Axio Zeiss A1 Imager The genes are oriented with the 59 end to the left and 39 to the right with boxes representing the exons and intervening lines as introns and the 39 UTR. The regions coding for the putative switch and nucleotide binding domains in the RAB-7 protein are shown in blue and red, respectively. The ok511 deletion that removes the first three exons of rab-7 as well as the 39 UTR of W03C9.5 is marked with a bracket. The indicated lines represent the genomic clones used for RNAi feeding. doi:10.1371/journal.pone.0036489.g001 compound microscope (Zeiss, Oberkochen, Germany) and images were captured using an Axio Cam MRm camera and AxioVision software (Zeiss, Oberkochen, Germany). Most phenotype analysis was performed comparing rab-7(ok511)/mIn1 heterozygotes and rab-7(ok511) homozygotes from heterozygous mothers in various genetic backgrounds. The mIn1 chromosomal inversion carries an integrated myo-2::GFP (mIs14) [27], and thus, rab-7(ok511)/mIn1 heterozygotes can easily be distinguished from rab-7(ok511) homozygotes by GFP expression in the pharynx. The Muv and Vul phenotypes were scored by counting the numbers of vulval and non-vulval descendants of P (3)(4)(5)(6)(7)(8).p in L4 stage larvae under DIC optics. Animals with fewer than 3 VPCs induced were considered Vul, and animals with greater than 3 VPCs induced were considered Muv. EGL-17::CFP expression was scored in the Pn.p and Pn.px cells of L2-L3 stage larvae and terminal Pn.p lineages of late L4 stage larvae by epifluorescence and DIC optics. For the rescue experiments, animals carrying vhEx1[plin-31::GFP::rab-7+pvha-6::GFP+Cb-unc-119(+)] were selected based on intestinal GFP expression, which due to the mosaic expression pattern, may or may not be present in the VPCs of any given animal. Quantification of LET-23::GFP-positive foci in the hypodermis was determined by capturing epifluorescent images of L3 Pn.p and Pn.px stage animals centered on the vulva, and taken on a focal plane in which the majority of the hypodermal nuclei are in focus. Images were identically cropped and the LET-23::GFP-positive foci were quantified blindly by four individuals. Tallies for each image were averaged and statistical analysis was performed using an unpaired t test (Graphpad Prism 5, San Diego, CA).
Confocal analysis of LET-23::GFP localization in the VPCs was performed using a Zeiss LSM-510 Meta laser scanning microscope with 63X oil immersion lens in a single or multi-track mode using a single or dual excitation (488 nm for GFP and/or 543 nm for RFP). Images were captured using LSM Image software (Zeiss, Oberkochen, Germany). Whole mount immunostaining of C. elegans was carried out using peroxide tube fixation [28] in 1% paraformaldehyde for 30 minutes at 4uC. Primary antibodies: a goat anti-GFP antibody (Rockland Inc., Gilbertsville, PA) and a mouse MH27 antibody [29] were used at 1:100 and 1:20 dilutions, respectively. Secondary antibodies: a rabbit anti-goat antibody conjugated to AlexaFluor 488 (Invitrogen, Carlsbad, CA) for anti-GFP and a rabbit anti-mouse antibody conjugated to AlexaFluor 568 (Invitrogen, Carlsbad, CA) for MH27 were used at a 1:200 dilution.

RAB-7 antagonizes LET-60 Ras signaling during vulva cell fate specification
To determine if RAB-7 regulates EGFR/Ras signaling during vulva development, we used the rab-7(ok511) deletion allele generated by the C. elegans Knockout Consortium. The rab-7(ok511) allele consists of a 741 base-pair deletion/a 17 base-pair insertion that deletes the first three exons of rab-7 as well as the 39 UTR of the upstream gene, W03C9.5 ( Figure 1C; www.wormbase. org) [30]. ok511 likely represents a null allele of rab-7 as it deletes the sequences encoding the putative switch regions and guanine nucleotide and Mg2+ binding sites essential for GTPase function. Consistent with this prediction, we placed ok511 over the chromosomal deficiency, maDf87, and found that bli-2(e768) rab-7(ok511)/maDf87 animals were indistinguishable from bli-2(e768) rab-7(ok511) homozygotes in their small body size, granular intestinal appearance, enlarged yolk platelets and maternal effect embryonic lethality (data not shown). We believe that the effects of ok511 in this study are specifically due to loss of rab-7 and not due to deletion of the 39 UTR of W03C9.5. As demonstrated later in the paper, rab-7(RNAi), but not W03C9.5(RNAi) suppressed the lin-2(e1309) Vul phenotype, and an extrachromosomal array, expressing rab-7 in the VPCs, reversed the rab-7(ok511); lin-2(e1309) suppressed Vul phenotype.
To determine if LIN-3, EGF-like ligand, is required for the enhanced signaling in rab-7(ok511) animals, we tested if rab-7(ok511) could suppress the Vul phenotype of lin-3(e1417), a strong hypomorphic allele that specifically disrupts expression in the Anchor cell and hence vulva cell fate specification [52]. We found that rab-7(ok511) failed to suppress the Vul phenotype of lin-3(e1417) animals (Table 3), indicating that the LIN-3 ligand is required for the enhanced signaling in rab-7(ok511) mutants. Therefore, RAB-7 acts upstream or in parallel to LIN-3 and LET-23.

Components of the ESCRT complexes antagonize LET-23 EGFR signaling
Components of the ESCRT complexes are required for the sorting of ubiquitinated cargo into the MVBs and have been shown to be negative regulators of RTK signaling in Drosophila and mammalian cells [53][54][55]. RNAi knockdown of Rab7 in Hela cells resulted in the accumulation of EGFR in the ILVs of MVBs [17], where the EGFR would presumably not be able to signal to downstream effectors in the cytoplasm. To determine if LET-23 EGFR might transit through MVBs en route to the lysosome in the VPCs, we tested if components of the ESCRT-0, and -I, complexes could antagonize LET-23 EGFR signaling. Similar to rab-7(RNAi), we found that RNAi of hgrs-1 (ESCRT-0), and vps-28 (ESCRT-I) suppressed the severity of the lin-2(e1309) Vul phenotype, while control RNAi, gfp and W03C9.5, did not suppress lin-2(e1309) ( Table 4). Thus, LET-23 EGFR signaling in the vulva is antagonized by the activity of the ESCRT machinery suggesting that LET-23 EGFR does transit through MVBs in the VPCs.

Discussion
Here we show that RAB-7 antagonizes LET-23 EGFRmediated vulval cell induction in C. elegans. Our genetic analysis using the rab-7(ok511) deletion mutant is consistent with RAB-7 acting as a negative regulator of LET-23 EGFR signaling to a similar extent as previously described negative regulators [32,33,38,59,60]. Similar to previously described negative regulators, rab-7(ok511) mutant animals display no defects in vulval cell fate specification. We found that rab-7(ok511) enhanced the Muv phenotype of a let-60 ras gain-of-function mutation and suppressed the Vul phenotype of a let-60 ras weak loss-of-function mutant. rab-7(ok511) strongly suppressed the Vul phenotype of mutations in lin-2, lin-7, and lin-10, as well as the let-23(sy1) allele. Furthermore, rab-7(ok511) was synthetic Muv with mutations in the negative regulators ark-1 Ack, and unc-101 AP-1 m1. Our genetic data indicate that rab-7 is a potent negative regulator of LET-23 EGFR signaling.
Genetic epistasis suggests that RAB-7 antagonizes LET-23 EGFR signaling in a manner that is distinct from previously described negative regulators. We show that rab-7(ok511) cannot suppress the Vul phenotypes of strong loss-of-function mutations in let-60 ras, let-23 EGFR, and lin-3 EGF suggesting that rab-7 is required upstream or in parallel to these genes. The fact that expression of GFP::RAB-7 in the VPCs can rescue the suppression of the lin-2(e1309) Vul phenotype by rab-7(ok511) indicates that RAB-7 functions in parallel in the VPCs to regulate signaling. The requirement for LIN-3 and LET-23 would be consistent with a known role of mammalian Rab7 in trafficking activated EGFR to the lysosome for degradation [16,17]. Also consistent with RAB-7 functioning at the level of LET-23 EGFR, rab-7(ok511) has stronger genetic interactions with negative regulators, ark-1 and unc-101, that act at the level of LET-23 EGFR, than with downstream negative regulators gap-1 and lip-1. While the weak interaction between rab-7(ok511) and sli-1 Cbl would be consistent with both acting in the same pathway to target LET-23 EGFR for lysosomal degradation, sli-1 mutations do suppress let-23(sy97) [38] while rab-7(ok511) does not, suggesting that SLI-1 and RAB-7 are antagonizing LET-23 EGFR signaling via different mechanisms. In fact, rab-7 is distinct from other negative regulators in that loss of rab-7 fails to suppress lin-3(e1417) Vul phenotype, while mutations in unc-101, ark-1, sli-1, gap-1, and sli-3 can suppress the Vul phenotype of lin-3 alleles e1417 and/or n378 [32,38,39,43,60].
The synthetic Muv phenotypes seen in unc-101; rab-7 and rab-7; ark-1 double mutants suggests that RAB-7 functions in parallel to UNC-101 and ARK-1 to antagonize LET-23 EGFR signaling at a common point of the pathway. ARK-1 is a non-receptor tyrosine kinase related to Ack that can interact with the SEM-5 Grb2 adaptor [32]. Although neither the kinase substrate(s) for ARK-1 nor the mechanism by which it inhibits signaling are known, the genetic data point to LET-23 as the likely target. ARK-1 might function to inactivate LET-23 through phosphorylation, while RAB-7 mediates LET-23 trafficking to the lysosome, thus functioning in parallel to reduce the amount of active LET-23. The adaptor complex AP-1 mediates trafficking between the Trans-Golgi, endosomes, and the plasma membrane, where it functions to sort and cluster cargo into clathrin coated vesicles [44]. Although the mechanism is not understood, two partially redundant AP-1 m1 subunits, UNC-101 and APM-1, negatively regulate LET-23 EGFR signaling [43,61]. Despite the fact that RAB-7 and AP-1 regulate distinct steps in the vesicular trafficking network, the strong Muv phenotype of unc-101(sy108); rab-7(ok511) double mutants might represent compounded defects in the sorting and trafficking of LET-23 and/or other transmembrane regulators of the LET-23 signaling, such as the DEP-1 protein tyrosine phosphatase [33,62]. The fact that the induced cells in unc-101(sy108); rab-7(ok511) failed to invaginate or express the egl-17::CFP marker suggested that they did not fully adopt a 1u or 2u vulval cell fate implies that AP-1 and RAB-7 might regulate additional inductive pathways such as LIN-12 Notch signaling.
In Hela cells, the overexpression of a dominant negative-Rab7(N125I) or Rab7 RNAi inhibited EGF-induced EGFR degradation, and the EGFR accumulated in ILVs of MVBs [16,17], where the EGFR would be unable to engage downstream signaling molecules. However, our findings suggest that LET-23 EGFR is still competent to signal in the VPCs when rab-7 activity is inhibited, leading us to question whether LET-23 also transits through MVBs. Four ESCRT complexes (0, I, II, and III) are required for MVB formation and sort ubiquitinated EGFR (and other cargos) into ILVs [63]. Components of the ESCRT-0 and ESCRT-I complexes antagonized RTK signaling in mammalian cells and in Drosophila [53][54][55]. While the ESCRT components have been shown to regulate LET-23::GFP localization in the embryonic hypodermis, they have not been shown to modulate LET-23 EGFR signaling [20]. Here we demonstrated that RNAi of hrgs-1, and vps-28, components of the ESCRT-0, and ESCRT-I complexes, respectively, suppressed the severity of the lin-2(e1309) Vul phenotype suggesting that the ESCRT complexes negatively regulate LET-23 EGFR signaling and that in the VPCs, LET-23 EGFR is targeted to the lysosome via MVBs. Therefore, the enhanced LET-23 EGFR signaling in rab-7(ok511) mutants is not for a lack of trafficking through MVBs.
Consistent with RAB-7 regulating LET-23 EGFR trafficking, we demonstrated that RAB-7 acts within the VPCs and influences LET-23::GFP localization. LET-23::GFP localizes to foci in the VPCs of rab-7(ok511) animals, that are more apparent in the lin-2(e1309); rab-7(ok511) background. We further explored this using a LET-23::GFP that is expressed in the hypodermal syncytium where is can be seen in endosomal foci [20]. We found that there is nearly a two-fold increase in the number LET-23::GFP foci in rab-7(ok511) animals consistent with LET-23::GFP accumulating in an endosomal compartment. Since Rab7 regulates early to late endosome maturation [64][65][66][67], in addition to transition of cargo between late endosomes/MVBs and lysosomes [17,[68][69][70], some LET-23 EGFR could become trapped on early endosomes prior to entry into MVBs. However, we failed to detect any significant colocalization between LET-23::GFP and the early endosome marker, EEA-1, in either wild-type or rab-7(2) animals. Alternatively, LET-23 EGFR could enter into MVBs, but in the absence of lysosomal degradation might exit the MVB via backfusion. Viral nucleocapsids can escape the MVB via back-fusion of ILVs [71], however it is not known whether the EGFR or other cell surface receptors can exit MVBs by this manner. In either case, the LET-23 EGFR could conceivably signal from these internal membranes or be recycled back to the plasma membrane to reengage LIN-3 EGF. Consistent with LET-23 EGFR recycling, several regulators of endosome recycling can promote LET-23 EGFR signaling during vulva development (A. Holmes and G. Michaux, personal communication).
We previously identified rab-7 in an RNAi screen for regulators of LET-23 EGFR signaling during embryogenesis for specification of the excretory duct cell [72]. In that process, rab-7 appears to promote LET-23 EGFR signaling. However, many of the genes identified in the screen appear to act indirectly to promote LET-23 EGFR signaling, and we cannot rule out that rab-7 might also indirectly promote excretory duct cell fate specification. Alternatively, it might be possible that RAB-7 has cell type specific effects on LET-23 EGFR signaling. In contrast to what is seen in Hela cells, Rab7 has recently been suggested to promote EGFR stability in A431 and MCF7 cancer cells by protecting EGFR from proteosomal degradation [73]. We have not found a role for rab-7 in modulating LET-23 EGFR signaling during the specification of the P12.pa hypodermal cell (data not shown). However, the P12.pa cell fate is specified at an earlier developmental stage than the vulval cells and could be maternally rescued in rab-7 homozygous progeny of a heterozygous parent.
In summary, we show that RAB-7 antagonizes LET-23 EGFR signaling during C. elegans vulva development. The requirements for rab-7 in LET-23 EGFR signaling are similar to, but distinct from those of previously described negative regulators. Because the EGFR, as well as many of its downstream effectors, can have oncogenic properties in humans [5], our findings that RAB-7 antagonizes LET-23 EGFR signaling, suggest the possibility of Rab7 having tumor suppressor activities in humans like that of c-Cbl through promoting downregulation of activated RTKs.