Distinct cell guidance pathways controlled by the Rac and Rho GEF domains of UNC-73/TRIO in Caenorhabditis elegans

Abstract

The cytoskeleton regulator UNC-53/NAV2 is required for both the anterior and posterior outgrowth of several neurons as well as that of the excretory cell while the kinesin-like motor VAB-8 is essential for most posteriorly directed migrations in Caenorhabditis elegans. Null mutations in either unc-53 or vab-8 result in reduced posterior excretory canal outgrowth, while double null mutants display an enhanced canal extension defect, suggesting the genes act in separate pathways to control this posteriorly directed outgrowth. Genetic analysis of putative interactors of UNC-53 or VAB-8, and cell-specific rescue experiments suggest that VAB-8, SAX-3/ROBO, SLT-1/Slit, and EVA-1 are functioning together in the outgrowth of the excretory canals, while UNC-53 appears to function in a parallel pathway with UNC-71/ADAM. The known VAB-8 interactor, the Rac/Rho GEF UNC-73/TRIO operates in both pathways, as isoform specific alleles exhibit enhancement of the phenotype in double-mutant combination with either unc-53 or vab-8. On the basis of these results, we propose a bipartite model for UNC-73/TRIO activity in excretory canal extension: a cell autonomous function that is mediated by the Rho-specific GEF domain of the UNC-73E isoform in conjunction with UNC-53 and UNC-71 and a cell nonautonomous function that is mediated by the Rac-specific GEF domain of the UNC-73B isoform, through partnering with VAB-8 and the receptors SAX-3 and EVA-1.

Figure 1

Excretory canal morphology in unc-53, vab-8, and unc-53; vab-8 animals and genomic characterization of unc-53 and vab-8. (A–D) Fluorescence micrographs of animals carrying the ppgp-12::gfp transgene, allowing for the visualization of the excretory cell body and canals (anterior and posterior termini are marked by long thin arrows, vulva is marked by an arrowhead, terminal bulb of the pharynx is marked by a short arrow, C–F.) Bar, 100 µm. Anterior is to the left and animals are displayed laterally. (A) Morphology of the wild-type excretory cell body and processes. The excretory cell body is positioned on the ventral side of the posterior pharyngeal bulb and extends two canals toward the anterior of the animal to the tip of the head and two canals posteriorly to the tail. (B) Excretory canal outgrowth phenotype of unc-53 (n166). The posterior canals terminate midway at the vulva. (C) In vab-8(e1017) mutants, the posterior canals terminate beyond the vulva within the posterior gonad arm. (D) The unc-53(n166); vab-8(e1017) double mutants exhibit enhanced canal defects where termination occurs before reaching the vulva, and often before reaching the anterior gonad arm. (E) Structure of the unc-53 gene. The start of the various UNC-53L and UNC-53S isoforms are indicated by arrows. The promoter for UNC-53SA is between exons 5 and 8, and the promoter for UNC-53SB is located between exons 8 and 13 (Stringham et al. 2002). Alternatively spliced exons are shown in pink. unc-53(n152) is a 319-bp deletion removing parts of exons 18 and 19, producing a stop codon in exon 20 (Stringham et al. 2002), and n166 is a single nucleotide C-to-T transition in exon 19 that introduces a premature stop codon (Schmidt et al. 2009). The longest polypeptide, UNC-53LA, is 1654 amino acids and contains a calponin homology domain (CH, red; amino acids 11–109), two LKK motifs (LKK, purple; 114–133 and 1097–1116), two proline-rich SH3-binding motifs (SH3b, green; 487–495 and 537–545), two coiled-coil regions (CC, blue; 890–923 and 1078–1113), and an AAA domain (yellow; 1292–1425). n166 introduces a premature stop codon at amino acid 949. Both n152 and n166 remove the coiled-coil, LKK, and AAA domains from all isoforms. (F) Structures of the six characterized vab-8 transcripts (Wolf et al. 1998). Exons are numbered above. The first six exons encode the kinesin-like motor domain. The position of the vab-8(e1017) null allele is indicated, where a C-to-T transition at bp 10,647 results in premature stop codon (Wolf et al. 1998). The VAB-8L protein contains a kinesin-like motor domain at its N terminus, and a domain predicted to form a coiled-coil is shared with all isoforms of VAB-8. (G) Quantification of posterior excretory canal outgrowth defects. The outgrowth of the posterior canals was divided into five regions (1–5) before the anterior gonad arm to the tail as shown. The stop point of canals was determined by fluorescence microscopy. Chi-squared analysis was used to establish statistical significance between mutants using GraphPad Prism 5 (Sigma Stat). For this comparison, phenotypes were grouped into two categories and the mutant compared to a baseline of either wild type, or the most severe single allele in the case of double mutants, as indicated. *P value is not statistically significant.

Figure 2

Anterior excretory canal morphology in wild-type, unc-53(n166), vab-8(e1017), and unc-53(n166); vab-8(e1017) animals. (A–D) Fluorescence micrographs showing a lateral view of hermaphrodites carrying the ppgp-12::gfp transgene. The stop point of the anterior canals was scored with respect to the wild-type position near the head (arrows mark the final positions of the anterior excretory canals). Anterior is to the right. (A) Wild-type canals extend to the anterior end of the animal. (B) The anterior canals terminate prematurely in the strong allele unc-53(n166). (C) The anterior canals were considered wild type in vab-8(e1017) mutants, though the excretory cell body was displaced posteriorly with respect to the static terminal bulb of the pharynx (arrowhead). (D) unc-53(n166); vab-8(e1017) animal showing the anterior canals are severely truncated and often absent. (E) Quantification of anterior longitudinal extension defects. unc-53(n166) (n = 72), vab-8(e1017) (n = 111), and unc-53(n166); vab-8(e1017) (n = 95).

Figure 3

Loss of molecules involved in the gonad-independent pathway for SM migration disrupt excretory canal extension. The outgrowth of the posterior canals was divided into five regions (1–5) before the anterior gonad arm to the tail as shown. The stop point of canals was determined by fluorescence microscopy. Loss of UNC-53, VAB-8, UNC-71, and UNC-73 perturb posterior canal outgrowth. unc-71 and unc-73(ev802) mutants do not enhance the migration phenotype seen in unc-53(n166). By contrast, vab-8(e1017) mutants were enhanced in double-mutant combination with unc-71 or unc-73(ev802). Chi-squared analysis was used to establish statistical significance between mutants using GraphPad Prism 5 (Sigma Stat). For this comparison, phenotypes were grouped into two categories and the mutant compared to a baseline of either wild type or the most severe single allele in the case of double mutants, as indicated. *P value is not statistically significant.

Figure 4

Loss of molecules known to function in dorsoventral guidance disrupt excretory canal migration. (A–C) The outgrowth of the posterior canals was divided into five regions (1–5) before the anterior gonad arm to the tail as shown. The stop point of canals was determined by fluorescence microscopy. Loss of UNC-53, VAB-8, EVA-1, and SLT-1 perturb posterior canal outgrowth. (A) sax-3, (B) slt-1, and (C) eva-1 mutants do not enhance the extension phenotype seen in vab-8(e1017). By contrast, unc-53(n166) mutants were enhanced in double-mutant combination with sax-3, slt-1, or eva-1. (A) sax-3(ky123); quEx168 [psax-3::sax-3::gfp; odr-1::RFP] rescue strain was able to partially rescue defects seen in unc-53(n166); sax-3(ky123) double mutants, and rescued animals resembled unc-53 mutants, which is consistent with unc-53 and sax-3 functioning in parallel pathways. Chi-squared analysis was used to establish statistical significance between mutants using GraphPad Prism 5 (Sigma Stat). For this comparison, phenotypes were grouped into two categories and the mutant compared to a baseline of either wild type or the most severe single allele in the case of double mutants, as indicated. *P value is not statistically significant. ***psax-3::sax-3::gfp; unc-53(n166); sax-3(ky123) is the sax-3(ky123); quEx168 [psax-3::sax-3::gfp; odr-1::RFP] rescue strain in unc-53(n166); sax-3(ky123) double mutants.

Figure 5

Genomic organization of the predicted unc-73 transcripts and corresponding UNC-73 isoforms. (A) The locations of mutations of the unc-73 gene are indicated above the locus (modified from Steven et al. 2005). The mutation ok936 (in pink) affects the SH3 domain and results in an Unc phenotype, the mutations rh40 and e936 (in red) affect the RacGEF domain, and ev802 (in green) eliminates the RhoGEF domain and is associated with L1 lethality (Steven et al. 2005). Exons of the predicted unc-73 transcripts (A, B, C1, C2, D1, D2, E, and F) are shown. Only isoform E shows expression within the excretory cell (Steven et al. 2005). (B) The predicted UNC-73 isoforms are shown. unc-73 encodes proteins with several domains, including the two tandem RhoGEF and PH domain (gray) combinations (RacGEF (red) and RhoGEF (green), a Sec14p motif (blue), eight spectrin-like repeats (yellow), a variant SH3 domain (pink), an immunoglobulin domain (Ig, orange), and a fibronectin type III (FnIII) domain (brown).

Figure 6

UNC-73 may function in both the UNC-53 and the VAB-8 pathways for excretory canal outgrowth. Alleles affecting the RacGEF domain of UNC-73 [unc-73(rh40) and unc-73(e936)] showed an enhancement in excretory canal truncation in double-mutant combination with unc-53(n166) but not with vab-8(e1017), suggesting that the RacGEF domain of UNC-73 may be required solely for mediating the VAB-8 pathway. In contrast, unc-73(ev802); unc-53(n166) double mutants are no more severely affected than the unc-53 single mutant alone, while unc-73(ev802); vab-8(e1017) animals exhibited severe enhancement of defects when compared to vab-8 null mutants. This suggests that one or more of the UNC-73 isoforms other than isoform B may be unc-53 specific. Supporting this hypothesis, transgenic analysis showed that the punc-73E::unc-73E::gfp transgenic strain was sufficient to partially rescue the posterior canal defects seen in unc-73(ev802), suggesting that UNC-73E, which contains the RhoGEF domain, is required cell autonomously together with UNC-53 for the proper migration of the excretory canals. RNAi data showing unc-73 RNAi in the background of unc-53; vab-8 double mutants or unc-53 RNAi in the background of vab-8; unc-73 double mutants exhibits no more severe canal defects than either double mutant alone, suggesting that unc-73 is likely not functioning in a third, separate pathway. The outgrowth of the posterior canals was divided into five regions (1–5) before the anterior gonad arm to the tail as shown. Also, the rho-1(ok2418)/nT1 heterogyzous strain was able to enhance defects seen in vab-8 null mutants but not in unc-53 null mutants. Chi-squared analysis was used to establish statistical significance between mutants using GraphPad Prism 5 (Sigma Stat). For this comparison, phenotypes were grouped into two categories, and the mutant compared to a baseline of either wild type or the most severe single allele in the case of double mutants, as indicated. *P value is not statistically significant. **pmEx289 is: dpy-5(e907)/dpy-5(e907); pmEx289 [rCes pmyo-2::GFP + pCeh361 + WRM0628aD12].

Figure 7

A model for two parallel pathways controlling posterior extension of the excretory canals. Triangles represent a proposed guidance cue. (A) Genetic analysis suggests that VAB-8, SAX-3/ROBO, SLT-1/Slit, and EVA-1 are functioning together in the migration of the canals. VAB-8L may regulate the SAX-3 receptor via the RacGEF activity of UNC-73B by promoting localization of SAX-3/ROBO to the cell surface or inhibiting its removal by endocytosis (Watari-Goshima et al. 2007). (B) UNC-53 functions in a cell autonomous pathway with UNC-73E and UNC-71/ADAM to promote migration of the excretory canals. In this pathway the RhoGEF domain of UNC-73E activates UNC-53 (either through direct binding or indirectly) to promote formin-mediated assembly of actin filaments. In addition, UNC-53 binds ABI that forms part of the WAVE complex, which has been shown to mediate the actin nucleation activity of ARP2/3. UNC-53 therefore may be a scaffold that coordinates multiple signals to the actin cytoskeleton machinery.