Summary
Polarization of early embryos provides a foundation to execute essential patterning and morphogenetic events. In Caenorhabditis elegans, cell contacts polarize early embryos along their radial axis by excluding the cortical polarity protein PAR-6 from sites of cell contact, thereby restricting PAR-6 to contact-free cell surfaces. Radial polarization requires the cortically enriched Rho GTPase CDC-42, which in its active form recruits PAR-6 through direct binding. The Rho GTPase activating protein (RhoGAP) PAC-1, which localizes specifically to cell contacts, triggers radial polarization by inactivating CDC-42 at these sites. The mechanisms responsible for activating CDC-42 at contact-free surfaces are unknown. Here, in an overexpression screen of Rho guanine nucleotide exchange factors (RhoGEFs), which can activate Rho GTPases, we identify CGEF-1 and ECT-2 as RhoGEFs that act through CDC-42 to recruit PAR-6 to the cortex. We show that ECT-2 and CGEF-1 localize to the cell surface and that removing their activity causes a reduction in levels of cortical PAR-6. Through a structure–function analysis, we show that the tandem DH-PH domains of CGEF-1 and ECT-2 are sufficient for GEF activity, but that regions outside of these domains target each protein to the cell surface. Finally, we provide evidence suggesting that the N-terminal region of ECT-2 may direct its in vivo preference for CDC-42 over another known target, the Rho GTPase RHO-1. We propose that radial polarization results from a competition between RhoGEFs, which activate CDC-42 throughout the cortex, and the RhoGAP PAC-1, which inactivates CDC-42 at cell contacts.
Introduction
Embryonic cells must polarize to carry out the essential early developmental events that establish the body plan. For example, cell polarization allows cells to asymmetrically partition fate determinants during division, permits directional cell migration, and enables the morphogenesis of epithelial tissues (Nance and Zallen, 2011; St Johnston and Ahringer, 2010). Many cell types polarize when polarity cues restrict the localization of the PAR proteins PAR-6 (PDZ and CRIB domain protein), aPKC (atypical protein kinase C), and PAR-3 (multi-PDZ domain scaffolding protein) to a subdomain of the cortex, thereby establishing a spatially restricted signaling center (Nance and Zallen, 2011; St Johnston and Ahringer, 2010). Failure to establish or maintain PAR protein asymmetries results in polarity loss and disrupted embryonic development.
Although the cues that trigger polarity can differ among cell types, in many cases polarity cues induce PAR protein asymmetries by altering the activity or localization of Rho family GTPases (Etienne-Manneville, 2004; Iden and Collard, 2008; Macara, 2004). These switch-like guanine nucleotide-binding proteins cycle between active GTP-bound and inactive GDP-bound states. Active Rho GTPases signal by binding to a diverse set of effector proteins that regulate many cellular processes, including cell polarization (Heasman and Ridley, 2008). For example, the Rho GTPase CDC-42 – an ancient polarity regulator first identified for its role in polarizing budding yeast cells (Adams et al., 1990) – is now known to mediate polarization in a multitude of animal cell types (Etienne-Manneville, 2004). CDC-42 polarizes animal cells in part by interfacing directly with PAR proteins: active CDC-42 binds to the CRIB (Cdc42/Rac interactive binding) domain of PAR-6, influencing PAR-6 localization and the activity of its binding kinase aPKC (Joberty et al., 2000; Johansson et al., 2000; Lin et al., 2000; Qiu et al., 2000). Determining how polarity cues regulate the localization of active CDC-42 is therefore central to learning how CDC-42 polarizes cells.
Rho GTPase activity is controlled by regulators that influence whether the Rho protein binds to GDP or GTP (Etienne-Manneville and Hall, 2002). RhoGAPs stimulate the intrinsic GTPase activity of Rho proteins, causing rapid GTP hydrolysis to an inactive GDP-bound state. By contrast, RhoGEFs stimulate the exchange of GDP for GTP, thereby restoring Rho GTPases to their active, GTP-bound state. The specificity and relative localization of RhoGAPs and RhoGEFs within the cell determine when and where a particular Rho GTPase is active. Therefore, one mechanism through which polarity cues can polarize cells is by inducing the asymmetric positioning of a RhoGEF or RhoGAP, which in turn leads to asymmetries in Rho GTPase signaling (Gulli and Peter, 2001; St Johnston and Ahringer, 2010). In the early C. elegans embryo, asymmetrically localized or activated Rho GTPases control two sequential polarization events. Soon after fertilization, the zygote polarizes along its anterior-posterior (A/P) axis in response to a cue provided by the sperm. A/P polarity is necessary for developmental determinants to segregate asymmetrically during the first division. As A/P polarity is established, cortical PAR-6, PKC-3/aPKC, and PAR-3 become restricted to the anterior cortex. The Rho GTPases RHO-1/RhoA and CDC-42 perform distinct roles in A/P polarization. During polarity establishment, the RhoGEF ECT-2 is excluded from the posterior cortex, limiting the activity of its target RHO-1 to the anterior cortex (Jenkins et al., 2006; Motegi and Sugimoto, 2006; Schonegg and Hyman, 2006). In turn, RHO-1 triggers a contraction of cortical actomyosin that moves PAR proteins to the anterior. CDC-42 also becomes restricted to the anterior cortex, where it maintains polarity by preventing PAR proteins from returning to the posterior cortex (Aceto et al., 2006; Gotta et al., 2001; Kay and Hunter, 2001; Motegi and Sugimoto, 2006; Schonegg and Hyman, 2006). The posteriorly localized RhoGAP CHIN-1 and the RhoGEF CGEF-1 contribute to CDC-42 activity, although additional unidentified RhoGEFs are thought to function with CGEF-1 for CDC-42 activation (Kumfer et al., 2010).
A second polarization event regulated by CDC-42 occurs during the four-cell stage, when the embryo polarizes along its radial axis in response to cell contact cues. During radial polarization, cell–cell contacts exclude PAR-6, PKC-3/aPKC, and PAR-3 from the adjacent cortex of each somatic cell, restricting these proteins to the contact-free surface (Nance and Priess, 2002; Nance et al., 2003). Radial polarity is needed for cytoskeletal asymmetries that direct the first cell movements of gastrulation (Nance et al., 2003). An analogous contact-induced and Rho GTPase-dependent radial polarization occurs in early mammalian embryos and is thought to allow the initial segregation of cells into extra-embryonic and embryonic lineages based on their position relative to the embryo surface (Clayton et al., 1999; Cockburn and Rossant, 2010; Johnson, 2009). Radial polarization in mammals is also accompanied by restriction of PAR proteins to contact-free surfaces of blastomeres (Plusa et al., 2005; Vinot et al., 2005). Because of these mechanistic similarities, C. elegans provides an accessible model for understanding this critical event in mammalian development.
In C. elegans, the establishment of radial polarity requires CDC-42 (Anderson et al., 2008). For example, if CDC-42 is removed from early embryonic cells, PAR-6 cannot localize to the cortex; and if constitutively active CDC-42 is expressed in these cells, PAR-6 localizes to all cortical surfaces and polarity is lost. In contrast to the zygote, where cortical CDC-42 is asymmetric (Motegi and Sugimoto, 2006; Schonegg and Hyman, 2006), CDC-42 localizes uniformly to the cortex of early embryonic cells (Anderson et al., 2008). However, cell contacts recruit the RhoGAP PAC-1/ARHGAP21, which inhibits CDC-42 activity at these sites (Anderson et al., 2008). Removing PAC-1 causes PAR proteins to localize pan-cortically, as in embryos expressing constitutively active CDC-42. However, it is not known how CDC-42 is activated at contact-free surfaces, allowing it to polarize cells by recruiting PAR proteins to these sites.
Here, we perform loss-of-function and overexpression screens of C. elegans RhoGEFs to identify those that activate CDC-42 during radial polarization. We show that overexpressing either ECT-2 or CGEF-1 is sufficient to activate CDC-42, causing the ectopic recruitment of PAR-6 to cell contacts. ECT-2 and CGEF-1 are enriched at the cortex, but unlike the RhoGAP PAC-1, both RhoGEFs localize symmetrically. Loss of ECT-2 and CGEF-1 results in a reduction but not elimination of cortical PAR-6, indicating the presence of a partially redundant CDC-42 activation mechanism. Finally, we identify the regions of ECT-2 and CGEF-1 necessary for their function and localization, and define a domain of ECT-2 that appears to control its preference in vivo for CDC-42 over known target RHO-1. We propose that competition between multiple symmetrically localized RhoGEFs and the asymmetrically localized RhoGAP PAC-1 causes the asymmetry in CDC-42 activity that polarizes the embryo radially.
Results
CGEF-1 and ECT-2 are candidate CDC-42 RhoGEFs
RhoGEFs are characterized by the presence of either a Dbl-homology (DH) domain or a DOCK domain (Rossman et al., 2005). C. elegans contains 20 genes whose product is predicted to contain a DH domain and three that contain a DOCK domain (WormBase version WS233; Table 1). We reasoned that if a RhoGEF were required to activate CDC-42, its removal from early embryonic cells would prevent most PAR-6 protein from localizing to the cortex, as occurs in cells depleted of CDC-42 (Anderson et al., 2008). To knock down RhoGEFs, we performed feeding RNAi to target each RhoGEF individually, and monitored cortical PAR-6 levels in living embryos using a transgene expressing functional PAR-6–GFP (Totong et al., 2007). pkc-3 RNAi, which caused PAR-6–GFP to become cytoplasmic, and gfp RNAi, which eliminated visible PAR-6–GFP, were used as controls for RNAi effectiveness. For 22 of the 23 genes, RNAi knockdown produced no discernible defects in PAR-6–GFP localization (Table 1). It was not possible to evaluate the remaining gene, ect-2, because ect-2(RNAi) embryos failed to complete cytokinesis and early embryonic cells did not form, as described previously (Table 1) (Morita et al., 2005; Motegi and Sugimoto, 2006; Schonegg and Hyman, 2006). These results suggest either that RNAi is not completely effective at removing the RhoGEF responsible for activating CDC-42 or that two or more RhoGEFs function redundantly to activate CDC-42.
Gene | Human ortholog | GEF domain | GEF RNAi phenotype† | GEF localization | GEF overexpression phenotype† |
Y95B8A.12 | PLEKHG7 | DH | Polarized | Cytoplasm | Polarized |
unc-73 | TRIO | DH | Polarized | Cytoplasm | Polarized |
cgef-2 | TIAM2 | DH | Polarized | Cytoplasm | Polarized |
unc-89 | TITIN | DH | Polarized | N/A# | N/A |
Y105E8A.24 | ARHGEF2 | DH | Polarized | N/A# | N/A |
tag-218 | ARHGEF16 | DH | Polarized | N/A‡ | N/A |
rhgf-2 | PLEXHG5 | DH | Polarized | Cytoplasm | Absent from contact-free surfaces, clusters on internal membranes¥ |
ect-2 | ECT2 | DH | N/A* | Cortex | Pan-cortical§ |
R02F2.2 | ARHGEF17 | DH | Polarized | Cytoplasm | Polarized |
exc-5 | FGD1 | DH | Polarized | Cytoplasm | Polarized |
tag-77 | FGD6 | DH | Polarized | Cortex | Polarized |
Y37A1b.17 | TUBA | DH | Polarized | Cortex | Polarized |
sos-1 | SOS1 | DH | Polarized | N/A‡ | N/A |
uig-1 | PLEKHG1 | DH | Polarized | Cytoplasm | Polarized |
cgef-1 | MCF2L | DH | Polarized | Cortex, cytoplasm, nucleus | Pan-cortical |
tag-52 | FGD2 | DH | Polarized | Nucleus | Polarized |
vav-1 | VAV1 | DH | Polarized | Cytoplasm | Polarized |
rhgf-1 | ARHGEF12 | DH | Polarized | N/A# | N/A |
F52D10.6 | FERM | DH | Polarized | Cytoplasm | Polarized |
pix-1 | αPIX | DH | Polarized | Cytoplasm | Polarized |
ced-5 | DOCK180 | DOCK | Polarized | Cytoplasm | Polarized |
F22G12.5 | DOCK11 | DOCK | Polarized | N/A# | N/A |
F46H5.4 | DOCK7 | DOCK | Polarized | N/A# | N/A |
Gene | Human ortholog | GEF domain | GEF RNAi phenotype† | GEF localization | GEF overexpression phenotype† |
Y95B8A.12 | PLEKHG7 | DH | Polarized | Cytoplasm | Polarized |
unc-73 | TRIO | DH | Polarized | Cytoplasm | Polarized |
cgef-2 | TIAM2 | DH | Polarized | Cytoplasm | Polarized |
unc-89 | TITIN | DH | Polarized | N/A# | N/A |
Y105E8A.24 | ARHGEF2 | DH | Polarized | N/A# | N/A |
tag-218 | ARHGEF16 | DH | Polarized | N/A‡ | N/A |
rhgf-2 | PLEXHG5 | DH | Polarized | Cytoplasm | Absent from contact-free surfaces, clusters on internal membranes¥ |
ect-2 | ECT2 | DH | N/A* | Cortex | Pan-cortical§ |
R02F2.2 | ARHGEF17 | DH | Polarized | Cytoplasm | Polarized |
exc-5 | FGD1 | DH | Polarized | Cytoplasm | Polarized |
tag-77 | FGD6 | DH | Polarized | Cortex | Polarized |
Y37A1b.17 | TUBA | DH | Polarized | Cortex | Polarized |
sos-1 | SOS1 | DH | Polarized | N/A‡ | N/A |
uig-1 | PLEKHG1 | DH | Polarized | Cytoplasm | Polarized |
cgef-1 | MCF2L | DH | Polarized | Cortex, cytoplasm, nucleus | Pan-cortical |
tag-52 | FGD2 | DH | Polarized | Nucleus | Polarized |
vav-1 | VAV1 | DH | Polarized | Cytoplasm | Polarized |
rhgf-1 | ARHGEF12 | DH | Polarized | N/A# | N/A |
F52D10.6 | FERM | DH | Polarized | Cytoplasm | Polarized |
pix-1 | αPIX | DH | Polarized | Cytoplasm | Polarized |
ced-5 | DOCK180 | DOCK | Polarized | Cytoplasm | Polarized |
F22G12.5 | DOCK11 | DOCK | Polarized | N/A# | N/A |
F46H5.4 | DOCK7 | DOCK | Polarized | N/A# | N/A |
Scored by PAR-6-GFP localization (RNAi screen) or PAR-6 staining (overexpression screen).
Could not score; worms were sterile or embryos produced could not form cells.
Unable to amplify from cDNA.
No visible expression.
Abnormal cell shapes; some anucleate cells.
Small blebs present on cell surfaces in many embryos.
To circumvent potential redundancy among RhoGEFs, we turned to a gain-of-function screen. We reasoned that an overexpressed CDC-42 RhoGEF able to localize to the cortex might override the inhibition of CDC-42 at cell contacts by RhoGAP PAC-1, resulting in pan-cortical CDC-42 activity and therefore pan-cortical PAR-6 localization; this is the phenotype observed in embryos expressing constitutively active CDC-42 (compare Fig. 1C with Fig. 1A) or in pac-1 mutant embryos (Anderson et al., 2008). For 18 of the 23 RhoGEFs, we were able to amplify complete cDNAs from embryonic mRNA (Table 1). We cloned each gene into a heat-shock expression vector that would create an N-terminal hemaglutinin (HA) tag fusion with the RhoGEF, and injected constructs into worms to obtain high-copy extrachromosomal arrays. For 16 of the 18 cloned RhoGEFs, we were able to isolate transgenic lines that, upon heat shock, expressed the HA-tagged RhoGEF in embryos as early as the four-cell stage (supplementary material Table S2).
We immunostained heat-shocked transgenic embryos with α-HA and α-PAR-6 antibodies to evaluate the effect on PAR-6 recruitment; sibling embryos on the same slide that did not inherit the transgenic array were used as heat-shocked controls. While the majority of overexpressed RhoGEFs localized to the cytoplasm and caused no defects in PAR-6 localization (Table 1; supplementary material Fig. S1), three genes in our screen produced defects. Overexpressing RHGF-2 disrupted cell shape and resulted in many anucleate cells (supplementary material Fig. S1D, Fig. S2D,D′). This phenotype is similar to that of embryos expressing constitutively active RHO-1 (supplementary material Fig. S2C′) (Anderson et al., 2008), and is consistent with a recent report showing that RHGF-2 activates RHO-1 (Lin et al., 2012). Overexpressing CGEF-1 or ECT-2 caused PAR-6 to localize pan-cortically (Table 1; Fig. 1D,E). We quantified PAR-6 ectopic recruitment by determining the ratio of PAR-6 staining at contact-free versus contacted surfaces. In comparison to heat-shocked control embryos (Fig. 1A,B), embryos overexpressing CGEF-1 or ECT-2 showed a significant recruitment of PAR-6 to cell contacts, resembling those overexpressing constitutively active CDC-42 (Fig. 1B–E). In addition to these polarity phenotypes, overexpressing ECT-2 also caused small blebs to form on the surface of cells in many embryos – a phenotype similar to but less severe than that seen in embryos overexpressing RHGF-2 (supplementary material Fig. S2A,B, arrows). Overexpressed ECT-2 and CGEF-1 were enriched at the cortex (Fig. 1D′,E′, arrowheads), and were thus positioned at the proper place to activate CDC-42. However, localization of an overexpressed RhoGEF to the cortex was not sufficient to activate CDC-42, since TAG-77 and Y37A1B.17 were enriched cortically but failed to recruit PAR-6 to cell contacts (supplementary material Fig. S1G,H). These findings raise the possibility that CGEF-1, which contributes to CDC-42 activation in the zygote, and ECT-2, previously known only as a RhoGEF for RHO-1, can activate CDC-42 at the cortex of early embryonic cells.
CGEF-1 and ECT-2 affect PAR-6 localization by acting through CDC-42
We next wanted to confirm that CGEF-1 and ECT-2 affected PAR-6 localization by acting through CDC-42 rather than another Rho GTPase. If CGEF-1 and ECT-2 act through CDC-42, overexpressing these RhoGEFs in embryos depleted of CDC-42 should not recruit PAR-6 to cell contacts; PAR-6 should instead remain in the cytoplasm, as in embryos depleted of CDC-42. When we knocked down CDC-42 in embryos using RNAi, PAR-6 levels were markedly reduced at the cortex of early embryonic cells in comparison to control embryos (compare Fig. 2B with Fig. 2A), as observed previously (Anderson et al., 2008). Neither CGEF-1 nor ECT-2 was sufficient to recruit additional PAR-6 to the cortex of cdc-42(RNAi) embryos when overexpressed, whereas both RhoGEFs recruited PAR-6 in control heat-shocked embryos fed bacteria containing empty vector (Fig. 2D–G′). We quantified these results by comparing the enrichment of PAR-6 immunostaining at cell contacts versus the adjacent cytoplasm; overexpressed CGEF-1 and ECT-2 caused a significant enrichment of PAR-6 at cell contacts only when CDC-42 was present at normal levels (Fig. 2C).
Because cdc-42(RNAi) embryos also have defects in A/P polarity (Gotta et al., 2001; Kay and Hunter, 2001), we performed two additional experiments to confirm that the inability of overexpressed ECT-2 and CGEF-1 to recruit PAR-6 in cdc-42(RNAi) embryos was not an indirect result of A/P polarity loss. First, we overexpressed ECT-2 and CGEF-1 in par-2(RNAi) embryos, which also fail in A/P polarity but are able to polarize radially (Kemphues et al., 1988; Nance and Priess, 2002) (supplementary material Fig. S3A). Overexpressed ECT-2 (15/20 embryos) and CGEF-1 (20/20 embryos) were both able to recruit PAR-6 ectopically to cell contacts in par-2(RNAi) embryos (supplementary material Fig. S3B,C, arrows). Second, we overexpressed CGEF-1 in ‘cdc-42(ZF1)’ embryos. These embryos have a null mutation in the cdc-42 gene and are rescued by a transgene expressing HA-tagged CDC-42 fused to the ZF1 domain from PIE-1 (Anderson et al., 2008). Because the ZF1 domain forces the fusion protein to degrade rapidly from somatic cells, cdc-42(ZF1) embryos contain CDC-42 in the zygote and develop normal A/P polarity but lack CDC-42 in most early embryonic cells. In live cdc-42(ZF1) embryos, overexpressed GFP-CGEF-1 recruited PAR-6–mCherry to cell contacts effectively in control heat-shocked cdc-42(+) embryos (supplementary material Fig. S4C′, 10/10 embryos), but was unable to recruit PAR-6–mCherry to cell contacts in cdc-42(ZF1) embryos (supplementary material Fig. S4D′, 10/10 embryos). We were unable to test ECT-2 in this assay since gfp-ect-2 transgenes caused synthetic lethality in the cdc-42(ZF1) background. Together, these findings strongly suggest that CGEF-1 and ECT-2 are able to activate CDC-42, which in turn recruits PAR-6 to cell contacts.
CGEF-1 and ECT-2 are cortically enriched
To determine where ECT-2 and CGEF-1 localize in early embryonic cells, we constructed transgenic versions of each endogenous gene tagged with either GFP or mCherry. ECT-2–GFP expressed from its own regulatory elements was markedly enriched at the cortex of blastomeres, and was present at both contacted and contact-free surfaces (Fig. 3A). This pattern is likely to reflect the localization of endogenous ECT-2, since the ect-2-gfp transgene rescued the sterile and embryonic lethal phenotypes of ect-2 mutants and its product enriched at the anterior cortex of one-cell embryos (data not shown), as demonstrated previously for endogenous ECT-2 (Jenkins et al., 2006). Using cgef-1 endogenous regulatory elements, we were not successful in obtaining tagged CGEF-1 transgenic lines that expressed maternally, although zygotic expression of the transgenic protein was evident in later embryos and appeared similar in localization to overexpressed HA–CGEF-1 (data not shown). Since CGEF-1 is known to be expressed maternally (Kumfer et al., 2010), we drove expression of an mCherry-tagged cgef-1 transgene in early embryos using the maternal mex-5 promoter. Like ECT-2–GFP, CGEF-1–mCherry was enriched at the cortex, but was also present at high levels in the cytoplasm and in the nucleoplasm in a cell-cycle-dependent manner (Fig. 3B,C). Thus, tagged ECT-2 and CGEF-1 are found at both contacted and contact-free surfaces of early embryonic cells, where they overlap in expression with CDC-42.
CGEF-1 and ECT-2 function redundantly with another CDC-42 activation mechanism
We examined mutations in cgef-1 and ect-2 to determine if removing these genes singly or in combination caused a loss of PAR-6 at the cortex, as seen in cdc-42(RNAi) and cdc-42(ZF1) embryos. Two viable deletion mutations, gk261 and tm3663, disrupt cgef-1. Both deletions remove part of the DH domain (Fig. 4A), which is essential for the activity of RhoGEFs, and cgef-1(gk261) was previously shown to cause alterations in the extent of PAR-6 asymmetry within the zygote (Kumfer et al., 2010). However, PAR-6 localized normally in early embryonic cells of cgef-1(gk261) and cgef-1(tm3663) mutant embryos (Fig. 4C and data not shown). We quantified these results in live embryos by measuring the ratio of PAR-6–mCherry at cell contacts to the adjacent cytoplasm (Fig. 4B). Whereas cdc-42(RNAi) embryos had a PAR-6–mCherry contact/cytoplasm ratio of 1.0, cgef-1(gk261) mutant embryos had a ratio of 1.4, similar to wild type.
The ect-2(gk44) deletion takes out a small portion of the ect-2 N-terminus, including the putative start codon (Fig. 4A), and is therefore likely a strong loss-of-function or null mutation. ect-2(gk44) homozygotes grow to be sterile adults, as described previously (Zonies et al., 2010). To circumvent the essential function of maternal ect-2 in the germline and in polarizing the zygote (Jenkins et al., 2006; Motegi and Sugimoto, 2006; Schonegg and Hyman, 2006), we used the ZF1 tagging approach to remove ECT-2 specifically from early embryonic cells. ECT-2–GFP–ZF1 rescued the sterility of ect-2(gk44) mutants, localized to the cortex of early embryonic cells until the four-cell stage, and then degraded rapidly from somatic cells (Fig. 4D′). In these ‘ect-2(ZF1)’ embryos, A/P polarity was normal (as assessed by asymmetries in cell size and P granule localization, data not shown), but cytokinesis failed in cells where ECT-2–GFP–ZF1 was degraded (‘m’ in Fig. 4D′). Consequently, ect-2(ZF1) embryos died with multiple, multinucleate cells. PAR-6 in fixed embryos (Fig. 4D) and PAR-6–mCherry in live embryos (quantified in Fig. 4B) was still enriched at contact-free surfaces of ect-2(ZF1) cells, when compared to cdc-42(RNAi) embryos, but at levels reduced significantly when compared to wild-type embryos (Fig. 4B). The levels of cortical PAR-6 and PAR-6–mCherry in ect-2(ZF1); cgef-1(gk261) embryos were similar to that of ect-2(ZF1) embryos, if not slightly reduced (Fig. 4B,E). Thus, while ECT-2 and CGEF-1 are sufficient to activate CDC-42, removal of these genes together reduces but does not eliminate CDC-42 function. ECT-2 appears to have a more important role in CDC-42 activation since ect-2(ZF1) but not cgef-1 embryos showed a significant reduction in cortical PAR-6–mCherry levels. These findings also indicate the presence of an alternative CDC-42 activation pathway that is independent of ECT-2 and CGEF-1. RNAi of the remaining RhoGEFs individually in the ect-2(ZF1); cgef-1(gk261) background did not prevent PAR-6 cortical localization (data not shown), suggesting that a more complex redundant RhoGEF network controls CDC-42 activity at the cortex of early embryonic cells.
The CGEF-1 DH-PH domain mediates CGEF-1 activity but not localization
CGEF-1, like many RhoGEFs, contains tandem DH and PH domains. DH domains are structurally conserved regions of about 200 amino acids, and it is thought that specific interactions between regions in the DH domain and the Rho GTPase confer GEF-GTPase specificity. PH domains in RhoGEFs can serve as a protein-targeting domain or can enhance GTP exchange activity in vitro (Rossman et al., 2005). To determine the domains of CGEF-1 that are needed for function and cortical localization, we first tested the sufficiency of its tandem DH-PH domains to recruit PAR-6. The CGEF-1 DH and PH domains reside within a central region of the protein (Fig. 5A). To determine whether the tandem DH-PH domains are sufficient for CGEF-1 function, we asked whether overexpressing the DH-PH domains fused to GFP (DH-PHCGEF-1–GFP) could recruit PAR-6 to cell contacts. DH-PHCGEF-1–GFP expressed from the heat-shock promoter localized to the cytoplasm and had no effect on PAR-6 localization in early embryonic cells (Fig. 5C,C′,E), suggesting that other regions of CGEF-1 are required for its cortical localization. Because DH-PHCGEF-1–GFP failed to localize, we could not assess whether the DH-PH domains are sufficient for RhoGEF activity in vivo. Therefore, we asked whether forcibly targeting DH-PHCGEF-1–GFP to the plasma membrane was sufficient to recruit PAR-6 to cell contacts. To target DH-PHCGEF-1–GFP to the plasma membrane, we fused it to the PH domain from rat PLC1δ1, which was shown previously to promote plasma membrane association in early C. elegans embryos (Audhya et al., 2005). DH-PHCGEF-1–GFP–PHPLC1∂1 localized to the cell surface and recruited PAR-6 to cell contacts as efficiently as full-length CGEF-1 (compare Fig. 5D,D′,E with Fig. 1D), whereas PAR-6 was not recruited to cell contacts in control embryos expressing GFP–PHPLC1∂1 alone (Fig. 5B′,E). These findings indicate that the tandem DH-PH domains confer GEF activity, while sequences outside of the DH-PH domains are needed to target CGEF-1 to the cell cortex.
N-terminal regulation of ECT-2
ECT-2 also has tandem DH and PH domains, and in addition contains two N-terminal BRCA1 C terminus (BRCT) domains; the BRCT domains have been shown to inhibit the GEF activity of mammalian Ect2 toward known target RhoA (Somers and Saint, 2003; Yüce et al., 2005; Zhao and Fang, 2005). The presence of inhibitory BRCT domains in overexpressed full-length ECT-2 might explain why we observed no major disruption of cytokinesis, which requires RHO-1 activity (Jantsch-Plunger et al., 2000). We first tested this hypothesis by overexpressing the ECT-2 DH-PH domains alone (DH-PHECT-2–GFP). However, the DH-PHECT-2–GFP fusion did not localize to the cortex and had no effect on PAR-6 localization or on cytokinesis (Fig. 6B,B′). We then targeted the DH-PH domains to the membrane by fusing with PHPLC1∂1. Overexpressed DH-PHECT-2–GFP–PHPLC1∂1 localized to the cortex and caused significant defects in cellular architecture: cell membranes were rounded, many anucleate cells could be found throughout the embryo, and PAR-6 accumulated in patches in the center of the embryo (Fig. 6C,C′). These defects are very similar to those of rho-1(CA) and rhgf-2(OE) embryos, as described above (supplementary material Fig. S2). This finding is consistent with a role for the BRCT domains in inhibiting the GEF activity for ECT-2 towards RHO-1 in vivo. To determine if the C-terminus has a role in regulating the localization or specificity of ECT-2, we overexpressed the DH-PHECT-2 together with the C terminus (DH-PH-CtermECT-2–GFP). In contrast to the DH-PH domain alone, including the C-terminus of ECT-2 enabled the fusion protein to localize to the cortex (compare Fig. 6D with Fig. 6B,). Embryos expressing DH-PH-CtermECT-2–GFP exhibited defects in cellular architecture, including the formation of anucleate cells (see Fig. 6D). Together, these ECT-2 structure–function experiments indicate that the DH-PHECT-2 domain is sufficient for ECT-2 activity, the C-terminus helps localize ECT-2 to the cell cortex, and the N-terminus containing the BRCT domains represses ECT-2 activity towards RHO-1 in vivo. Given that full-length ECT-2 was able to recruit PAR-6 to cell contacts, the BRCT domains may not inhibit the activity of ECT-2 towards CDC-42 in vivo, and could be involved in determining which Rho GTPase ECT-2 targets (see Discussion).
Discussion
Establishing asymmetric CDC-42 activity during contact-induced polarization
We showed previously that the contact-induced radial polarization of C. elegans early embryonic cells occurs when RhoGAP PAC-1 inhibits CDC-42 activity at cell contacts (Anderson et al., 2008). However, it was unknown how CDC-42 was activated at contact-free surfaces, allowing it to recruit PAR-6 and associated PAR proteins to these sites. Here, we have identified two RhoGEFs, CGEF-1 and ECT-2, that function redundantly to activate CDC-42 during radial polarization. Although we did not observe active CDC-42 directly upon manipulating CGEF-1 or ECT-2 levels in the embryo, our genetic experiments using active CDC-42-binding protein PAR-6 as a readout for CDC-42 activity strongly suggest that CDC-42 is their target in vivo. First, both CGEF-1 and ECT-2 are enriched at the cell cortex in early embryonic cells, mirroring the localization of CDC-42. Second, we showed that overexpressed CGEF-1 or ECT-2 are able to recruit PAR-6 ectopically to cell contacts in a CDC-42-dependent manner; this phenotype is identical to that of embryos expressing constitutively active CDC-42 (Anderson et al., 2008) (this study). Although it is formally possible that CGEF-1 and ECT-2 activate another Rho GTPase that in turn activates CDC-42, the phenotypes observed upon expressing constitutively active versions of other Rho GTPases in the early embryo do not support such a model. For example, expressing constitutively active CED-10/Rac at this stage has no effect on PAR-6 localization (Anderson et al., 2008), and expressing constitutively active RHO-1 causes severe defects in cell shape and cleavage that we do not observe when overexpressing CGEF-1 or full-length ECT-2 (Anderson et al., 2008 and this study).
Three observations suggest that CGEF-1 and ECT-2 activate CDC-42 constitutively, throughout the cortex. First, CGEF-1 and ECT-2 are enriched pan-cortically, at both contacted and contact-free surfaces; their localization contrasts with that of RhoGAP PAC-1, which is found exclusively at cell contacts. Second, we showed that when CGEF-1 or ECT-2 is overexpressed, each is able to recruit PAR-6 to all cell surfaces. Finally, in pac-1 mutant embryos, PAR-6 is found at all cell surfaces, indicating that CDC-42 is activated pan-cortically (Anderson et al., 2008). Collectively, these observations suggest a competition model for CDC-42 activation during radial polarization. We propose that the contact/contact-free asymmetry in CDC-42 activity results from a competition between pan-cortical activation by CGEF-1, ECT-2, and other RhoGEFs (see below), together with targeted inactivation at cell contacts by RhoGAP PAC-1. In wild-type embryos, PAC-1 inhibition of CDC-42 at cell contacts would outcompete activation by the RhoGEFs, restricting CDC-42 activity to contact-free surfaces. This binary system for regulating asymmetric CDC-42 activity is ideally suited for the dynamic regulation of polarity that occurs as cell contact patterns change due to cell division and ingression (Anderson et al., 2008); repolarization of cells would require only that PAC-1 quickly recognize and relocalize to nascent cell contacts. Indeed, we showed previously that PAC-1 is rapidly recruited to new cell contacts generated by combining cells in culture (Anderson et al., 2008). In future studies, it will be important to determine how PAC-1 recognizes and localizes to cell contacts, and how it outcompetes the RhoGEFs present at these sites.
Our findings, coupled with observations in other systems, illustrate the diversity in ways that cortical cues can trigger an asymmetry in Cdc42 activity that polarizes cells. The mechanisms of Cdc42 activation during cell polarization have been studied extensively in the budding yeast S. cerevisiae. In yeast, active Cdc42p promotes formation of the daughter cell bud at a discrete cortical site. In contrast to C. elegans early embryonic cells, Cdc42p is not evenly distributed, but instead concentrates at the site of bud formation in response to cortical cues from the previous budding site (Johnson et al., 2011; Park and Bi, 2007; Perez and Rincón, 2010). Cdc42p is activated by its RhoGEF Cdc24p, which shows a similar asymmetric localization. Both proteins are coupled through a positive feedback loop mediated by the Bem1p scaffolding protein, which helps cluster active Cdc42p and Cdc24p in the same cortical subdomain. Thus while in C. elegans early embryonic cells CDC-42 localizes pan-cortically, is activated constitutively, and is inhibited by a RhoGAP at cortical sites determined by the polarity cue (cell contacts), Cdc42p in S. cerevisiae localizes asymmetrically and is activated asymmetrically in response to cortical polarity cues (the previous bud site). It remains to be determined whether in C. elegans there is an equivalent physical coupling between CDC-42 and its RhoGEFs (analogous to yeast Bem1p), or whether this feature of Cdc42 activation is needed only when Cdc42 and its RhoGEFs must localize in an asymmetric manner. An interesting possibility is that PAR-6 could serve this role, since PAR-6 binds directly to active CDC-42 (Aceto et al., 2006; Gotta et al., 2001) and mammalian Par6 binds to Ect2 (Liu et al., 2004). Despite these differences in mechanism, the output in yeast and C. elegans is the same – active CDC-42 is confined to a limited cortical subdomain.
A novel role for ECT-2 in CDC-42-mediated polarization
We report for the first time an in vivo role for ECT-2 in activating CDC-42 during cell polarization. ECT-2 is a highly conserved RhoGEF that is known largely for its role in activating RHO-1/RhoA during cytokinesis (Barr and Gruneberg, 2007; Piekny et al., 2005), and as a proto-oncogene in mammals (Fields and Justilien, 2010). Loss of ECT-2 in C. elegans and other animals blocks cytokinesis and results in multinucleate cells (Morita et al., 2005) – a phenotype we observe in early embryonic cells within ect-2(ZF1) embryos. Our overexpression and loss-of-function experiments indicate that ECT-2 also has a previously unrecognized role in activating CDC-42 during cell polarization in vivo. This finding is consistent with in vitro experiments demonstrating that Ect2 has GEF activity towards Cdc42, in addition to RhoA and Rac1 (Solski et al., 2004; Tatsumoto et al., 1999). Moreover, expression of dominant-negative Ect2 in Xenopus oocytes (Zhang et al., 2008a), or Ect2 knockdown in malignant glioblastoma cells or astrocytomas (Fortin et al., 2012; Weeks et al., 2012), leads to a decrease in Cdc42 activity. It is not clear in these examples whether Ect2 impacts Cdc42 activity directly. Our in vivo experiments suggest that ECT-2 regulates CDC-42 directly in C. elegans early embryonic cells, since acute overexpression of full-length ECT-2 results in a phenotype identical to that of acute overexpression of constitutively active CDC-42 (and differing from embryos overexpressing constitutively active RHO-1; see below). The identification of ECT-2 as an activator of CDC-42 in polarizing C. elegans cells raises the possibility that Cdc42-mediated cell polarity defects contribute to the transforming activity of oncogenic Ect2 in mammalian tumor cells. Consistent with a role for mammalian Ect2 in cell polarization, altering Ect2 activity in 3D cysts of cultured MDCK cells disrupts polarization and lumen formation (Liu et al., 2006).
ECT-2 has also been shown to regulate cell polarity in the one-cell C. elegans embryo, where it controls RHO-1 activity during polarity establishment (Jenkins et al., 2006; Motegi and Sugimoto, 2006; Schonegg and Hyman, 2006). Prior to polarization, RHO-1 and ECT-2 are enriched throughout the cortex. A cue from the sperm centrosomes causes ECT-2 to clear from the posterior cortex, locally reducing RHO-1 activity, and triggering a RHO-1-dependent asymmetric contraction of cortical actomyosin to the anterior. These events polarize the cell by inducing the anterior displacement of cortical PAR-6, PAR-3, PKC-3, and CDC-42. CDC-42 does not appear to be required for polarity establishment in the one-cell embryo, but rather functions in polarity maintenance; in cdc-42(RNAi) embryos, PAR-6 disappears from the anterior cortex and polarity is lost. Intriguingly, the phenotype of ect-2(RNAi) embryos is similar to embryos lacking both RHO-1 and CDC-42 (Motegi and Sugimoto, 2006). Our findings raise the possibility that ECT-2, in addition to activating RHO-1 during polarity establishment, may also activate CDC-42 during polarity maintenance and therefore couple these two phases of polarization.
Regulation of ECT-2 and CGEF-1 localization and activity
Our structure–function experiments identified regions of CGEF-1 and ECT-2 important for GEF activity and localization to the plasma membrane. For both CGEF-1 and ECT-2, we showed that the tandem DH-PH domains have GEF activity on their own, although this activity was apparent only when we targeted the DH-PH domains to the membrane using the heterologous PHPLC1∂1 domain. Therefore, the DH-PH domains are sufficient for GEF activity, while sequences outside of the DH-PH domain target the proteins to the plasma membrane. For ECT-2, we defined this region as the C-terminus. A role for the C-terminus in targeting mammalian Ect2 and Drosophila Pebble/Ect2 to the plasma membrane has been described previously (Solski et al., 2004; van Impel et al., 2009; Su et al., 2011), suggesting that this region engages a molecularly conserved mechanism that directs the protein to the cell surface.
In addition, we showed that the N-terminal region of ECT-2 is important for determining which Rho GTPase ECT-2 targets. Previous structure–function studies of Ect2 have shown that the N-terminal BRCT domains inhibit its GEF activity towards RhoA, and that these domains are regulated during cytokinesis to control RhoA activity in the cleavage furrow (Kim et al., 2005; Somers and Saint, 2003; Yüce et al., 2005; Zhao and Fang, 2005). Consistent with these findings, we showed that removing the N-terminal BRCT domains caused a dramatic enhancement of phenotypes that are identical to those produced by overexpressing constitutively active RHO-1, including altered cell shapes, anucleate cells, and cell surface blebs. However, overexpressing full-length ECT-2 caused a CDC-42-dependent recruitment of PAR-6 to cell surfaces, identical to phenotypes observed when overexpressing constitutively active CDC-42. These observations suggest that when the N-terminus is present, ECT-2 constitutively activates CDC-42, and when the N-terminus is absent, ECT-2 constitutively activates RHO-1. We propose two models that could explain how the N-terminus might control ECT-2 target specificity. First, it is possible that the BRCT domains (or other sequences in the N-terminus) inhibit ECT-2 activity towards RHO-1 but have no effect on ECT-2 activity towards CDC-42. Second, it is possible that the BRCT domains dampen GEF activity towards both RHO-1 and CDC-42, and that upon overexpressing N-terminally truncated ECT-2, we are unable to detect PAR-6 cortical recruitment because of the dramatic cellular defects that accompany RHO-1 activation.
Redundant mechanisms of CDC-42 activation
Our genetic experiments argue that CDC-42 can be activated by at least two RhoGEFs, although additional unidentified mechanisms contribute to CDC-42 activation or PAR-6 recruitment. In the one-cell embryo, a small amount of cortical PAR-6 is recruited independently of CDC-42, by binding protein PAR-3 (Beers and Kemphues, 2006). Therefore it is possible that in early embryonic cells lacking both CGEF-1 and ECT-2, some of the remaining cortical PAR-6 we observe is recruited by PAR-3. However, even if this were the case, the observation that cortical PAR-6 levels are significantly lower in embryos lacking CDC-42 than in embryos lacking both ECT-2 and CGEF-1 indicates that some CDC-42 remains active in the RhoGEF-depleted embryos. Using RNAi, we were unable to identify a single additional RhoGEF that eliminated cortical PAR-6 in embryos lacking ECT-2 and CGEF-1. Therefore, it is possible that a complex network of multiple RhoGEFs contributes to CDC-42 activation. Our overexpression screen may not have picked up these genes for various reasons, including interference with GEF activity or localization by the HA tag, incorrect annotation of the predicted gene, or our inability to amplify cDNA or obtain expressing transgenic lines (7 genes of 23). A similar complex regulation of CDC-42 may occur in the one-cell embryo, where cgef-1 mutants show lower but not depleted levels of a cortical CDC-42 activity probe (Kumfer et al., 2010), and only minor defects in the positioning of the anterior PAR-6 domain. ECT-2 is also present in the one-cell embryo (Motegi and Sugimoto, 2006), and as we note above, its removal causes phenotypes consistent with a role in activating CDC-42. Therefore, as we show in early embryonic cells, ECT-2 and CGEF-1 may also function redundantly to activate CDC-42 during polarity maintenance in the one-cell embryo. Indeed, given the large number of RhoGEFs present in C. elegans and other animals in comparison to the number of Rho GTPases, complex and redundant activation of a given Rho GTPase by multiple RhoGEFs may be a common theme.
Materials and Methods
Strains and transgenes
Strains and mutations used in this study are described in supplementary material Table S1. Deletion mutant strains ect-2(gk44), cgef-1(gk261) (both isolated by the C. elegans Gene Knockout Consortium) and cgef-1(tm3663) (isolated by the National BioResource Project and a kind gift of S. Mitani, Tokyo Women's Medical University) were out-crossed four times prior to analysis. Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440).
Transgenes created for the RhoGEF screen are described in supplementary material Table S2. Transgenes created for cgef-1 and ect-2 structure–function studies are described in supplementary material Table S3.
Transgene construction
RNAi feeding constructs
RNAi feeding constructs made in this study were constructed by amplifying pieces of the targeted gene coding sequence (supplementary material Table S4) and cloning into the multiple cloning site of vector pPD129.36 (Timmons and Fire, 1998). The par-2 RNAi feeding construct was made using forward and reverse oligos CAATCACCCCACACCGCCA and CCGGCTCCAGAGTGTCC to amplify par-2 cDNA; PCR product was cloned into pPD129.36.
RhoGEF constructs for overexpression screen
RhoGEF heat shock overexpression constructs were assembled using Gateway™ cloning (Invitrogen). Sequences encoding tandem HA domains were inserted immediately prior to the 5′ att recombination site in pCD6.09AP (Hao et al., 2006), a destination vector containing the hsp16-41 promoter (‘Phsp’). RhoGEF coding sequences were amplified from C. elegans cDNA, as described in supplementary material Table S2, and recombined into entry vector pDONR221 (Invitrogen). Entry clones were sequenced and recombined into the modified pCD6.09AP destination vector.
Phsp::gfp-cgef-1
Assembled using the strategy for HA-tagged RhoGEFs described above, except the tandem ha sequences in pCD6.09AP destination vector were replaced with gfp, amplified by PCR from the plasmid pPD95.75 (1995 Fire Lab vector kit).
cgef-1 and ect-2 constructs for structure–function analysis
gfp, vector pCD6.09AP, and coding regions to be assayed were fused using Gibson end-joining (Gibson et al., 2009) (supplementary material Table S3). The rat PLC1δ1 PH domain was amplified as described (Audhya et al., 2005).
Pect-2::ect-2-gfp
gfp was amplified and inserted immediately prior to the stop codon of ect-2 within fosmid WRM0614aC02, using recombineering and galK replacement as described (Zhang et al., 2008b). The ect-2-gfp fusion, plus 2.2 kb of upstream sequence and 3.8 kb of downstream sequence, was subcloned into PCR-amplified unc-119 plasmid pPUB using recombineering (Sarov et al., 2006) and the following PCR homology arms: 5′ of ect-2: TACTTTTTTTGGTTATCTTGAAGCTCTTTTTTCTAATTTATCATTTTTCC and 3′ of ect-2: AATGCACACAATTCTACTGAAACTTCTTGGAAGACGAACCTTCCATCCATC.
Pect-2::ect-2-gfp-zf1
Constructed as Pect-2::ect-2-gfp, except that a fusion between gfp and zf1 was inserted into ect-2, as described (Nance et al., 2003).
Pmex-5::cgef-1-mCherry::tbb-23′UTR
Assembled using Multisite Gateway™ (Invitrogen) and the following constructs: pCFJ150 destination vector (Frøkjaer-Jensen et al., 2008), 5′ entry clone pJA281 (Pmex5::mCherry) (Zeiser et al., 2011), middle entry clone cgef-1a in pDONR221 (see above), and 3′ entry clone pCM1.36 (tbb-23′UTR) (Merritt et al., 2008).
Ppar-6::par-6-mCherry-ha-ha
gfp was removed from the PstI site of par-6-gfp (Nance et al., 2003) and replaced with mCherry fused to tandem ha sequences. unc-119(+) from plasmid pJN254 was inserted into the vector NotI site (Nance et al., 2003).
Worm transformation
Integrated transgenes of Pmex-5::cgef-1-mCherry::tbb-23′UTR and Ppar-6::par-6-mCherry were created by microparticle bombardment into unc-119(ed3), as described (Praitis et al., 2001). Integrated transgenes of Pect-2::ect-2-gfp, Pect-2::ect-2-gfp-zf1 were bombarded into strain FT359. The unc-119(+) gene was present in the vector of each construct. All other transgenes were injected into N2 worms to create extrachromosomal arrays. Injection mixes contained 1–10 ng/µl of the transgene and 80 ng/µl rol-6 transformation marker pRF4 (Mello et al., 1991).
Immunofluorescence
Embryos were freeze-fractured, fixed in methanol alone, or in methanol followed by paraformaldehyde, as described previously (Anderson et al., 2008). The following antibodies were used: rabbit α-GFP 1∶2000 (Abcam), chicken α-GFP 1∶1000 (Chemicon), rabbit α-PAR-6 1∶20,000 (Schonegg and Hyman, 2006), mouse α-HA 1∶1000 (Covance), mouse α-ERM-1 1∶50 (Hadwiger et al., 2010), mouse ‘K76’ P granules 1∶1000 (Strome and Wood, 1983). Images shown were captured using a Zeiss AxioImager, 40×1.3NA objective, and Axiocam MRM camera. Images were deconvolved using AxioVision software, and are shown as maximum intensity projections of 3–5 adjacent planes spaced 0.3 µm apart. Unless otherwise indicated, a minimum of 20 embryos per genotype was scored in each immunostaining experiment.
RhoGEF RNAi screen
The RNAi screen for RhoGEFs was performed twice, first by feeding L4 xnIs3 [Ppar-6::par-6-gfp] larvae for 36 hours at 25°C, then by feeding for 60 hours at 20°C, using established methods (Timmons and Fire, 1998), except 0.2% β-lactose was substituted for IPTG. All feeding constructs were in bacterial strain HT115 and empty vector pPD129.36 was used as a control. Feeding RNAi constructs were obtained from the Ahringer (Kamath and Ahringer, 2003) and Vidal (Rual et al., 2004) libraries (supplementary material Table S4) and were sequence-verified. For genes with no available clone or an incorrect clone, new feeding clones were constructed by cloning genomic DNA or cDNA into pPD129.36 (supplementary material Table S4). The control gfp feeding construct was described previously (Timmons and Fire, 1998), and pkc-3 feeding plasmid was a gift from B. Leung.
RhoGEF heat shock overexpression screen
Gravid transgenic worms (Rollers) were heat-shocked at 34°C for 1 hour. Hermaphrodites were chopped to liberate embryos, which were immediately fixed, co-stained for HA and PAR-6, and imaged as described above. Levels of recruitment of PAR-6 to cell contacts were compared in 12- to 24-cell stage immunostained embryos in which the contact between two AB lineage cells was in focus. PAR-6 levels were measured by averaging the cortical maxima at the contact-free surfaces of each cell (10 transects per cell per embryo) and at the contacted surface between the two cells (10 transects). Ratios were generated between contact-free and contact values and are presented as the arithmetic mean.
ect-2 and cgef-1 overexpression in cdc-42(RNAi) embryos
L4 worms were fed empty vector (see above), cdc-42 RNAi (Anderson et al., 2008), or par-2 RNAi (this study, see above) for 48 hours at room temperature. Under cdc-42 RNAi conditions, a subset of embryos became osmotically sensitive (Osm), as described (Schonegg and Hyman, 2006); only non-Osm embryos were analyzed. To quantify levels of PAR-6 at cell contacts, a single 5 µm transect was drawn perpendicular to the contact of AB lineage cells, centered at the cell contact, and the intensity profile was measured along the line. Intensity was normalized relative to the lowest pixel intensity. Five embryos were averaged in each experiment.
Analysis of PAR-6 in cgef-1 and ect-2 mutants
Ppie-1::par-6-mCherry was crossed into cgef-1(gk261) mutants, ect-2(ZF1) mutants, and cgef-1(gk261); ect-2(ZF1) double mutants. cdc-42 RNAi in the Ppie-1::par-6-mCherry background was carried out as described above. Cortical levels of PAR-6-mCherry were measured by averaging the cortical maxima at the contact-free surfaces of AB lineage cells (10 transects per cell per embryo). Cytoplasmic PAR-6 was quantified by averaging pixels along a 5 µm line drawn parallel to the cortex but not through the nucleus. Ratios were generated between cortical and cytoplasmic profiles.
Acknowledgements
We thank Julie Ahringer, Tony Hyman, Benjamin Leung, Shohei Mitani, and Geraldine Seydoux for their generous gifts of strains and reagents; and Steve Armenti, Ryan Cinalli, Ann Wehman, and Kenneth Ho for comments on the manuscript. Some antibodies and strains were obtained from the NIH-funded Developmental Studies Hybridoma Bank and Caenorhabditis Genetics Center.
Author contributions
E.C. and J.N. designed the experiments. E.C. performed all experiments and analysis. E.C. and J.N. wrote the manuscript.
Funding
This work was funded by the National Institutes of Health (NIH) [grant number R01 GM078341 to J.N.]. Deposited in PMC for release after 12 months.