Trio's Rho-specific GEF domain is the missing Galpha q effector in C. elegans

Abstract

The Galpha(q) pathway is essential for animal life and is a central pathway for driving locomotion, egg laying, and growth in Caenorhabditis elegans, where it exerts its effects through EGL-8 (phospholipase Cbeta [PLCbeta]) and at least one other effector. To find the missing effector, we performed forward genetic screens to suppress the slow growth and hyperactive behaviors of mutants with an overactive Galpha(q) pathway. Four suppressor mutations disrupted the Rho-specific guanine-nucleotide exchange factor (GEF) domain of UNC-73 (Trio). The mutations produce defects in neuronal function, but not neuronal development, that cause sluggish locomotion similar to animals lacking EGL-8 (PLCbeta). Strains containing null mutations in both EGL-8 (PLCbeta) and UNC-73 (Trio RhoGEF) have strong synthetic phenotypes that phenocopy the arrested growth and near-complete paralysis of Galpha(q)-null mutants. Using cell-based and biochemical assays, we show that activated C. elegans Galpha(q) synergizes with Trio RhoGEF to activate RhoA. Activated Galpha(q) and Trio RhoGEF appear to be part of a signaling complex, because they coimmunoprecipitate when expressed together in cells. Our results show that Trio's Rho-specific GEF domain is a major Galpha(q) effector that, together with PLCbeta, mediates the Galpha(q) signaling that drives the locomotion, egg laying, and growth of the animal.

Figure 1.

The missing Gα_q effector pathway and targeted forward genetic screens for finding it. (A) Summary of the Gα_o and Gα_q pathways that drive C. elegans locomotion and egg laying. Solid lines indicate that direct interactions are known or likely, while dashed lines, such as the missing Gα_q effector pathway, indicate probable missing components. The green proteins positively drive locomotion and neurotransmitter release, and reducing their function causes decreased neurotransmitter release, decreased locomotion or paralysis, and decreased egg laying. The red proteins inhibit locomotion and neurotransmitter release, and reducing their function causes increased neurotransmitter release and hyperactive behaviors. Note that the GOA-1 (Gα_o) pathway exerts its inhibitory effects in a Gα_q pathway-dependent manner. Not shown is a third Gα pathway in this network (Gα_s) (Reynolds et al. 2005; Schade et al. 2005; Charlie et al. 2006a, b). The model is based on the following studies: Mendel et al. (1995), Segalat et al. (1995), Brundage et al. (1996), Koelle and Horvitz (1996), Hajdu-Cronin et al. (1999), Lackner et al. (1999), Miller et al. (1999, 2000), Nurrish et al. (1999), Richmond et al. (1999, 2001), Chase et al. (2001), Robatzek et al. (2001), van der Linden et al. (2001), and Bastiani et al. (2003). (B) Two forward genetic screen strategies for identifying the missing Gα_q effector pathway. Both screen strategies use the mutagen ENU to suppress the slow growth and hyperactive phenotypes of mutants with an overactive Gα_q pathway for the purpose of identifying downstream effectors.

Figure 2.

Mutations in unc-73 (Trio) suppress the growth and locomotion phenotypes of mutants with a hyperactivated Gα_q pathway. (A) The unc-73(ce362) mutation causes a striking change in the appearance of a goa-1 (Gα_o)-null mutant. The photographs show how the unc-73(ce362) mutation converts a goa-1-null mutant from an animal with a small, thin, loopy appearance to a larger, fatter looking animal with a more relaxed posture. Compare with wild type in top panel. (B) Mutations in unc-73 (Trio) improve the growth of mutants with a hyperactivated Gα_q pathway. The graph depicts mean growth rates of wild type and the indicated single and double mutants. Growth rate is measured as mean number of progeny produced per animal over a 96-h period starting with newly hatched larvae. Error bars represent the standard errors of three populations of larvae. (C) Mutations in unc-73 (Trio) slow the locomotion of mutants with a hyperactivated Gα_q pathway. The graph compares locomotion rates, expressed as body bends per minute, of wild type and the indicated single and double mutants. Error bars are the standard errors of 10 animals.

Figure 3.

Four mutations that suppress mutants with an overactive Gα_q pathway disrupt UNC-73’s Rho-specific GEF domain. Shown are scale drawings comparing the domain structures of human Trio, human p63RhoGEF, and the eight alternative gene products of the C. elegans unc-73 (Trio) locus as described by Steven et al. (2005). Red arrows indicate the sites of the four suppressor mutations that disrupt the Rho-specific GEF domain (allele names ce362, ox317, ce367, and ox322). A red line denotes the boundaries of the ev802 deletion mutation (Steven et al. 2005). The graph compares the mean locomotion rates of wild type and the various Rho-specific GEF domain mutants, including the ev802-null mutant. ev802 [D1] is a strain containing the deletion mutation that has been rescued for its larval arrest phenotype with the D1 transcript (Steven et al. 2005). Since D1 is mainly expressed in the pharynx and egg-laying muscles (Steven et al. 2005), this strain is close to a null for UNC-73 in most of the nervous system; however, small amounts of expression in neurons driving locomotion cannot be ruled out, and there appears to be a slight improvement in locomotion rate over the ev802 null that lacks that transgene. Error bars are the standard errors of 10 animals.

Figure 4.

The sluggishness of unc-73 RhoGEF mutants is caused by defects in neuronal function, not neuronal development. (A) UNC-73’s RhoGEF domain functions in adults, after neuronal development is complete. Locomotion rates of control mutants compared with an unc-73(ce362) mutant that expresses the unc-73 cDNA under control of an inducible heat-shock promoter. Dark-blue and light-blue bars depict mean locomotion rates of eight adults of each strain before and after a 50-min heat shock followed by a 3-h recovery period. Error bars are standard errors of the means. See also the Supplementary Movies. (B) UNC-73’s RhoGEF domain functions in the nervous system. Mean locomotion rates, in body bends per minute, of the unc-73(ce362) mutant plus or minus the ceIs39 [rab-3∷unc-73e cDNA] transgene that expresses the unc-73e cDNA throughout the nervous system. Error bars are standard errors of the means of 10 animals.

Figure 5.

unc-73 RhoGEF mutants exhibit drug sensitivities consistent with a mild deficit in ACh release. (A) unc-73(ce362) mutants exhibit delayed aldicarb-induced paralysis at later time points relative to wild type and neuronally rescued unc-73(ce362). Shown are the percentage of animals that are paralyzed at various time points during incubation on a 2 mM aldicarb plate. The ceIs39 [rab-3∷unc-73e cDNA] transgene expresses the unc-73e cDNA pan-neuronally. The aldicarb resistance of egl-8(md1971), a probable null mutant of PLCβ (Miller et al. 1999), is significantly stronger than unc-73(ce362) during the first hour of exposure. Data are the means and standard errors of three trials and are representative of two independent experiments. (B) unc-73(ce362) mutants are hypersensitive to the ACh receptor agonist levamisole relative to wild type and neuronally rescued unc-73(ce362). Shown are the percentage of animals that are paralyzed at various time points during incubation on an 800 μM levamisole plate. The ceIs39 [rab-3∷unc-73e cDNA] transgene restores wild-type levamisole sensitivity to unc-73(ce362). Also shown is unc-29(x29), which lacks functional levamisole receptors (Gally et al. 2004) and is completely resistant to levamisole-induced paralysis. Data are the means and standard errors of three trials.

Figure 6.

Double mutants containing null or near-null mutations for both EGL-8 (PLCβ) and UNC-73 RhoGEF domain phenocopy Gα_q-null mutants. (A) egl-8 (PLCβ) and unc-73 RhoGEF mutants show a synthetic interaction for locomotion. The graph shows mean locomotion rates of the indicated strains. Error bars are the standard errors of 10–12 animals. The egl-30 and egl-8 mutations used here are early stop codons, except for egl-30(ad805), which is a splice site mutation (Brundage et al. 1996; Lackner et al. 1999; Miller et al. 1999). See also the Supplementary Movies. (B) Gα_q and PLCβ/Trio RhoGEF-null mutants share the folded-hatchling phenotype. (Top panel) Photographs of a newly hatched wild-type larva next to two unhatched eggs containing late-stage folded embryos. (Bottom panels) Due to lack of locomotion, newly hatched mutant larvae remain folded as they were in the egg for hours after hatching. Genotypes of mutants are egl-30(ad810) (bottom left) and unc-73(ev802); egl-8(md1971) (bottom right). For scale, C. elegans eggs are ∼50 μm in the long axis. See also the Supplementary Movies. (C) The synthetic interaction involves growth as well as locomotion. Shown are the mean growth rates of the indicated strains. Growth rate is measured as the mean number of progeny produced per animal over a 96-h period. Note that the growth rates of PLCβ/UNC-73 RhoGEF double mutants resemble the Gα_q null and are much more severe than either single mutant. Error bars are the standard errors of three populations of larvae. (D) Phorbol esters rescue the paralysis of PLCβ/Trio Rho GEF double null mutants. Locomotion rates of wild type and the indicated mutants after 2 h of exposure to carrier (ethanol) or phorbol ester in ethanol. Data are means and standard errors of eight animals using 6-min locomotion assays. See also the Supplementary Movies.

Figure 7.

Activated EGL-30 (Gα_q) coimmunoprecipitates in a stable complex with UNC-73E (Trio RhoGEF) when coexpressed in HEK293 cells. HEK293 cells were transfected with 1 μg of unc-73E cDNA, 1 μg of activated egl-30 (Gα_q Q205L), and up to 2 μg of empty control vector. N-terminally Flag-tagged UNC-73E was precipitated with 2 μg of monoclonal Flag-M2 antibody and was analyzed by probing a Western blot with a Flag antibody or a specific Gα_q/11 antibody. The bottom panel shows expression levels of UNC-73E and EGL-30 (Gα_q Q205L) in 20 μg of cell lysates.

Figure 8.

Activated EGL-30 (Gα_q), but not activated GOA-1 (Gα_o), synergizes with Trio and UNC-73E to activate RhoA in HEK293 cells. (A) Shown are the domain arrangements of the full-length and truncated Trio and unc-73 cDNAs used for the cell transfection experiments. (B) Trio and UNC-73E enhance EGL-30 (Gα_q) Q205L-mediated SRE-driven transcription. The ability of C. elegans Gα_q (Q205L) and Gα_o (Q205L) to activate RhoA was tested in the presence or absence of Trio and UNC-73E expression using an assay that monitors Rho-dependent transcription of an SRE-controlled luciferase reporter gene. Data are the mean ratios of firefly/Renilla luciferase units ± standard errors (n = 6 transfections). (C) UNC-73E enhances RhoA activation in a Rho effector pull-down assay. RhoA activation was analyzed by a Rhotekin–Rho-binding domain pull-down assay. The amount of total RhoA as well as the expression levels of UNC-73E, EGL-30 (Gα_q Q205L), and GOA-1 (Gα_o Q205L) were analyzed in parallel using anti-RhoA-, anti-Flag-M2-, anti-Gα_q/11-, and anti-GOA-1 (Gα_o)-specific antibodies, respectively. Data are representative of duplicate independent experiments.

Figure 9.

An UNC-73 RhoGEF mutation strongly impairs Gα_q-stimulated RhoA activation in cells and purified proteins. (A) The unc-73(ce362) mutation severely impairs Gα_q-stimulated RhoA activation in intact cells. The ability of C. elegans Gα_q (Q205L) to activate RhoA was tested in the presence of wild-type or mutant UNC-73E 234–577 (ce362 allele) using an assay that monitors Rho-dependent transcription of an SRE-controlled luciferase reporter gene. The inset at the top of the graph shows the expression levels of wild-type and mutant UNC-73 proteins, with lanes corresponding to the graph bars. Note that the mutant protein stimulates luciferase production only slightly more than Gα_q alone, which exerts its effects through an endogenous Trio RhoGEF expressed in these cells. Data are the mean ratios of firefly/Renilla luciferase units ± standard errors (n = 4 transfections). (B) p63RhoGEF-I205N (the ce362 mutation reproduced on p63RhoGEF) binds Gα_q with threefold lower affinity. One representative competition experiment is shown, wherein increasing concentrations of unlabeled wild-type p63RhoGEF (residues 149–502) and its I205N mutant were mixed with 100 nM of fluor-labeled p63RhoGEF, and then incubated with Gα_q immobilized on microspheres. Bead-associated fluorescence is reported as the median fluorescence intensity (MFI) for each concentration of protein, measured in duplicate. Calculated K_is (assuming K_d = 50 nM for fluor-labeled p63RhoGEF) were 30 ± 4 nM and 100 ± 9 nM for wild type and I205N, respectively. The Gα_q used in this experiment is the Gα_i/q chimera reported in Tesmer et al. (2005). (C) p63RhoGEF-I205N is impaired in its activation by Gα_q. Shown is a representative experiment. RhoA exchange activity was measured by fluorescence polarization. The calculated K_d (which assumes that the increase in rate is entirely due to the association of p63RhoGEF with Gα_q) is 90 ± 25 nM and 70 ± 35 nM for wild type and the I205N mutant, respectively. The Gα_q used in this experiment is the Gα_i/q chimera reported in Tesmer et al. (2005).