Heterotrimeric G proteins in C. elegans*

Heterotrimeric G proteins, composed of alpha, beta, and gamma subunits, are able to transduce signals from membrane receptors to a wide variety of intracellular effectors. In this role, G proteins effectively function as dimers since the signal is communicated either by the G alpha subunit or the stable G betagamma complex. When inactive, G alpha-GDP associates with G betagamma and the cytoplasmic portion of the receptor. Ligand activation of the receptor stimulates an exchange of GTP for GDP resulting in the active signaling molecules G alpha-GTP and free G betagamma, either of which can interact with effectors. Hydrolysis of GTP restores G alpha-GDP, which then reassociates with G betagamma and receptor to terminate signaling. The rate of G protein activation can be enhanced by the guanine-nucleotide exchange factor, RIC-8, while the rate of GTP hydrolysis can be enhanced by RGS proteins such as EGL-10 and EAT-16. Evidence for a receptor-independent G-protein-signaling pathway has been demonstrated in C. elegans early embryogenesis. In this pathway, the G alpha subunits GOA-1 and GPA-16 are apparently activated by the non-transmembrane proteins GPR-1, GPR-2, and RIC-8, and negatively regulated by RGS-7. The C. elegans genome encodes 21 G alpha, 2 G beta and 2 G gamma subunits. The alpha subunits include one ortholog of each mammalian G alpha family: GSA-1 (Gs), GOA-1 (Gi/o), EGL-30 (Gq) and GPA-12 (G12). The remaining C. elegans alpha subunits (GPA-1, GPA-2, GPA-3, GPA-4, GPA-5, GPA-6, GPA-7, GPA-8, GPA-9, GPA-10, GPA-11, GPA-13, GPA-14, GPA-15, GPA-16, GPA-17 and ODR-3) are most similar to the Gi/o family, but do not share sufficient homology to allow classification. The conserved G alpha subunits, with the exception of GPA-12, are expressed broadly while 14 of the new G alpha genes are expressed in subsets of chemosensory neurons. Consistent with their expression patterns, the conserved C. elegans alpha subunits, GSA-1, GOA-1 and EGL-30 are involved in diverse and fundamental aspects of development and behavior. GOA-1 acts redundantly with GPA-16 in positioning of the mitotic spindle in early embryos. EGL-30 and GSA-1 are required for viability starting from the first larval stage. In addition to their roles in development and behaviors such as egg laying and locomotion, the EGL-30, GSA-1 and GOA-1 pathways interact in a network to regulate acetylcholine release by the ventral cord motor neurons. EGL-30 provides the core signals for vesicle release, GOA-1 negatively regulates the EGL-30 pathway, and GSA-1 modulates this pathway, perhaps by providing positional cues. Constitutively activated GPA-12 affects pharyngeal pumping. The G alpha subunits unique to C. elegans are primarily involved in chemosensation. The G beta subunit, GPB-1, as well as the G gamma subunit, GPC-2, appear to function along with the alpha subunits in the classic G protein heterotrimer. The remaining G beta subunit, GPB-2, is thought to regulate the function of certain RGS proteins, while the remaining G gamma subunit, GPC-1, has a restricted role in chemosensation. The functional difference for most G protein pathways in C. elegans, therefore, resides in the alpha subunit. Many cells in C. elegans express multiple G alpha subunits, and multiple G protein pathways are known to function in specific cell types. For example, Go, Gq and Gs-mediated signaling occurs in the ventral cord motor neurons. Similarly, certain amphid neurons use multiple G protein pathways to both positively and negatively regulate chemosensation. C. elegans thus provides a powerful model for the study of interactions between and regulation of G protein signaling.


Introduction 1.G protein structure/G protein cycle
Heterotrimeric G proteins, composed of α, β, and γ subunits, transduce signals from the plasma membrane through a cycle of guanine nucleotide exchange and hydrolysis.When signaling, G proteins effectively function as dimers since the signal is communicated either by the Gα subunit or the stable Gβγ complex.The classical G protein cycle is shown in Figure 1.Gα-GDP associates with Gβγ and with cytoplasmic portions of its typically seven-pass membrane receptor (also known as G protein coupled receptors; GPCRs).Upon ligand activation, the receptor acts as a guanine-nucleotide exchange factor (GEF) stimulating the exchange of GDP for GTP on the α subunit.Gα-GTP dissociates from Gβγ and both entities can interact with effectors.Hydrolysis of GTP restores Gα-GDP, which then reassociates with Gβγ and receptor to terminate signaling.In some pathways, Gα-GDP may interact with an additional GEF protein, RIC-8, which may then re-activate the α subunit (Tall et al., 2003;Reynolds et al., 2005).The rate of GTP hydrolysis can be enhanced by RGS proteins (regulator of G protein signaling; Watson et al. 1996).GEF proteins therefore act as positive regulators of G protein pathways while RGS proteins act as negative regulators (Figure 1).G protein signaling can also be activated through a receptor-independent pathway in which the dissociation of Gα and Gβγ may be facilitated by GPR (G protein regulator) proteins, which compete with Gβγ for binding to the GDP-bound form of Gα and function as guanine nucleotide dissociation inhibitors (GDIs;Manning, 2003).Dissociation of Gα and Gβγ may also be facilitated by GEF proteins (Afshar at al., 2004).Alternatively, GEF proteins may recognize the Gα-GPR complex, and subsequently effect nucleotide exchange to generate an active Gα-GTP signaling molecule (Manning, 2003;Hess et al., 2004).Details of this pathway are discussed further under "Receptor-independent pathway" in the Gαo section.
Figure 1.Depicted is the activation of a heterotrimeric G protein by a G protein coupled receptor (GPCR).In the inactive conformation (1), the Gαβγ heterotrimer is associated with a seven transmembrane-spanning receptor protein, and the Gα subunit is associated with GDP.Upon receptor activation with ligand (*), the receptor undergoes a conformational change and acts as a guanine nucleotide exchange factor (GEF) for the α subunit, stimulating the exchange of GDP for GTP (2).Additional GEFs have been isolated that act on GDP associated Gα subunits not associated with receptor.Upon binding of GTP, the Gα subunit dissociates from the Gβγ dimmer and these proteins are released from the receptor.Dissociated Gα and Gβγ subunits then interact with a variety of effector (E) proteins (3).The signal is attenuated by the intrinsic ability of Gα subunits to hydrolyze GTP, and this activity is accelerated by GTPase-activating proteins (GAPs) (4).Most GAPs contain a regulator of G protein signaling (RGS) domain that often displays higher affinity for the Gα-GTP transition state.
The expression and function of each G protein subunit are described in detail below.

β subunits
C. elegans expresses two Gβ subunits encoded by gpb-1 and gpb-2.GPB-1 shares 86% amino acid identity with mammalian Gβ subunits (van der Voorn et al., 1990), and appears to be required to mediate signaling by all Gα subunits.Inactivation of gpb-1 leads to abnormalities early in embryogenesis (Zwaal et al., 1996; see "Receptor-independent pathway" in the Go section).GPB-1 also has roles later in development that were revealed by mosaic rescue of a gpb-1 null mutant (Zwaal et al., 1996).Sterility and abnormalities in the germ line were observed in some adult mosaic animals, suggesting the gpb-1 may also play a role in germ line development.Overexpression of gpb-1 causes lethargic locomotion and delayed egg laying.These later effects are consistent with the expression pattern of gpb-1; expression is seen throughout development in nearly all somatic tissues and in the germ line (Zwaal et al., 1996).Once tissue differentiation occurs, expression is highest in neurons.In larval and adult stages, GPB-1 expression is seen in most or all neurons, the somatic gonad, vulva, and hypodermal seam cells.The intestine, pharynx, and body wall muscles appear to express GPB-1 at lower levels (Zwaal et al., 1996).

γ subunits
Two genes encoding Gγ subunits have been identified in C. elegans, gpc-1 and gpc-2 (Jansen et al., 2002).GPC-1 and GPC-2 show from 22 to 79% amino acid identity to the vertebrate Gγ subunits, but are not clear orthologs of any of the human Gγ subunits (Jansen et al., 2002).GPC-1 is specifically expressed in sensory neurons while GPC-2 shows ubiquitous expression (Jansen et al., 2002).
GPC-1 is expressed in 12 cells in the head and two cells in the tail that are identified as putative chemosensory neurons.Although GPC-1 is not essential for the detection of the water soluble attractants, NaCl, NaAc, and NH4Cl, it is essential for adaptation to these tastants (Jansen et al., 2002).In addition, gpc-1 mutants exhibit defects in the adaptation to soluble chemical repellents mediated by the ASH neurons, although they are not defective in the response to any repellents tested except for quinine (Hilliard et al., 2005).gpc-1 mutants display wild-type adaptation to volatile odorants (Jansen et al., 2002).gpc-1 overexpression may reduce locomotion and egg laying, though a null mutation in gpc-1 has no effect on those phenotypes (Jansen et al., 2002).
Heterotrimeric G proteins in C. elegans

Function and pathways for individual Gα subunits
Many of the Gα subunits have remarkably diverse roles in C. elegans biology, and thus these genes have been identified in many genetic screens.Where not previously identified in forward genetic screens, loss-of-function (lf) mutations have been isolated for all Gα subunit genes (with the exception of gpa-17) using target-selected gene inactivation (Jansen et al., 2002).Gα subunits can be mutated to reduce their intrinsic GTP-ase activity, thus causing constitutive signaling.Such gain-of-function (gf) mutations have been isolated for gsa-1 (Schade et al., 2005) and egl-30 (Bastiani et al., 2003), and have been produced in vitro and expressed as transgenes (QL) for gsa-1, goa-1, egl-30, gpa-1, gpa-2, gpa-3, gpa-12 and odr-3 (see sections on individual Gα subunits for references).In many cases, transgenic overexpression of the wild-type Gα subunit (XS) causes effects similar to the activated transgene (although often not as severe) and reciprocal to the loss-of-function mutation.Thus, loss-of-function phenotypes have been described for each of the C. elegans Gα subunits while gain-of-function phenotypes are known for most.We discuss here the current knowledge of their functions.

Introduction
GSA-1, encoded by gsa-1, shares over 60% identity with its vertebrate ortholog, Gαs.In vertebrates, Gαs is known to activate adenylyl cyclase to produce cyclic AMP, which both activates Protein Kinase A (PKA), and modulates cyclic nucleotide-gated ion channels (reviewed in Sunahara et al., 1996;Walsh and Van Patten, 1994).Vertebrate Gαs can also activate L-type voltage gated calcium channels in skeletal muscle cells and can inhibit cardiac sodium channels (reviewed in Wickman and Clapham, 1995a;Wickman and Clapham, 1995b).In C. elegans, the adenylyl cyclase ACY-1 appears to be the major effector of GSA-1 for growth and locomotion.A receptor with similarity to the vertebrate 5-HT7 family may couple to GSA-1 based on the observation that expression of SER-7b in COS-7 cells results in a dramatic increase in basal cAMP levels over untransfected cells (Hobson et al., 2003).

Phenotypes
In C. elegans GSA-1 is required for viability; loss of GSA-1 function results in L1 arrest (Korswagen et al, 1997).GSA-1 also affects egg laying, locomotion, and necrotic neuronal cell death based on mutational analysis (Korswagen et al., 1997;Berger et al., 1998;Korswagen et al., 1998).Viable animals segregating from mosaically rescued transgenic lines show behavioral defects including sluggish locomotion and delayed egg laying (Korswagen et al., 1997).Gain-of-function mutations in gsa-1, like transgenic overexpression, result in reciprocal phenotypes (Korswagen et al, 1997;Schade et al., 2005).Constitutive activation of transgenic GSA-1 causes necrotic neuronal cell death although similar mutations in chromosomal gsa-1 cause only minimal cell death (Korswagen et al, 1997;Schade et al., 2005).GSA-1 plays an ongoing functional role in regulating movement since gsa-1(QL) transgenes are able to induce hyperactive locomotion even when expressed after adulthood (Schade et al., 2005).Gain-of-function mutation in both gsa-1 and acy-1, which encodes an adenylyl cyclase that is the major effector of GSA-1 for movement (see below), cause hypersensitivity to aldicarb, indicating that the GSA-1 pathway regulates acetycholine release by motor neurons.However, elimination of the neuronal GSA-1 pathway does not affect steady-state levels of neurotransmitter release (Reynolds et al., 2005).The implications for this are discussed further in the "Regulators of EGL-30/G protein signaling network" section.GSA-1 also affects gametogenesis, hermaphrodite genital morphogenesis, growth rate and morphogenesis of the epithelium, as determined in large-scale RNAi assays (Simmer et al., 2003).

Expression
GSA-1 is broadly expressed in neurons and muscles from the embryonic stage of development onward.It may also be expressed in intestinal and some epithelial cells (i.e. of the pharynx and vulva; Korswagen et al., 1997;Park et al., 1997).GSA-1 functions both pre-and post-synaptically in regulating movement since expression of an activated form of its effector, ACY-1, from promoters specific for either neurons or muscles causes hyperactive locomotion (Schade et al., 2005).GSA-1 pathway activity is also required in both neurons and muscles for larval growth, based on studies using heterologous promoter elements fused to acy-1 (Reynolds et al., 2005).

Pathways
Two genetic studies of neurodegeneration have revealed conservation between GSA-1 signaling in C. elegans and the vertebrate Gαs signaling pathway (Korswagen et al., 1997;Berger et al., 1998).One screen employed a transgene with the constitutively active C. elegans gsa-1 gene under the control of the ubiquitously expressed heat-shock promoter (Korswagen et al., 1997) and the other employed a transgene with the constitutively active rat Gαs gene under the control of the glr-1 promoter, expressed in 17 classes of neurons, including interneurons required for locomotion (Berger et al., 1998).In both cases, constitutive activation and overproduction of either C. elegans GSA-1 or rat Gαs induced paralysis and neurodegeneration (Korswagen et al., 1997;Berger et al., 1998).Necrotic cell death was suppressed by mutations in acy-1 (sgs-1) (Korswagen et al., 1997;Berger et al., 1998).Most mutations in acy-1 fail to, or only partially suppress, the body-wall hypercontraction caused by transgenic activated Gαs, and also fail to suppress the hyperactive egg laying caused by overproduction of GSA-1 (Korswagen et al., 1997).acy-1 is expressed in the nervous system and in muscle cells (Berger et al., 1998;Korswagen et al., 1998), and null mutations affect larval viability, larval molting, life span, and pharyngeal pumping while reduction-of-function mutations affect locomotion (Moorman and Plasterk, 2002).
For growth and locomotion rate, ACY-1 appears to be the major effector of GSA-1 since null mutations in acy-1 are epistatic to gsa-1(gf) mutations (Schade et al., 2005).Downstream of ACY-1, the GSA-1 pathway likely converges with the EGL-30 pathway to control synaptic transmission from the neurons that regulate locomotion.RIC-8, a GEF protein, may regulate both GSA-1 and EGL-30 with respect to locomotion (Reynolds et al., 2005; see "Regulators of EGL-30/G protein signaling network" section for further detail).A second adenylyl cyclase gene, acy-2, was found to be essential for larval viability (Korswagen et al., 1998).The terminal phenotype of acy-2(lf) mutants resembles that of gsa-1(lf) mutants and clr-1 mutants (Korswagen et al., 1998), which phenotypically mimic worms after laser ablation of the canal-associated neurons (Kokel et al., 1998).acy-2 is expressed in the CAN cells, and some head ganglia and ventral cord neurons (Korswagen et al., 1998).These results indicate that ACY-2 may be an effector of GSA-1 in the CAN cells.Two other predicted adenylate cyclases, acy-3 and acy-4, have not been characterized genetically, nor have RNAi phenotypes been uncovered.
Gαs-induced cell death occurs via a necrotic rather than an apoptotic pathway, based on both the inability of mutations in ced-3 and ced-4 to suppress the cell death and the observation that the morphology of neuronal degeneration observed is similar to that caused by activating mutations in some ion channel genes, such as deg-1, deg-3, and mec-4.Loss-of-function mutations in unc-36, one of three calcium channel genes potentially regulated by PKA (unc-2, unc-36, and egl-19), were found to confer a slight reduction in cytotoxity in one study and slight suppression of paralysis in another (Korswagen et al., 1997;Berger et al., 1998).Cytotoxic cell death mechanisms conferred by activated Gαs and by mec-4(d), deg-1(d) and deg-3(d) alleles require ER-driven Ca 2+ release (Xu et al., 2001), and converge upon calcium-regulated aspartyl and calpain proteases (Syntichaki et al., 2002).In all cases, degeneration is suppressed in the genetic mutants cad-1, daf-4, and unc-52, or by starvation, which all reduce aspartyl protease activity.Treatment of mec-4(d), deg-1(d), deg-3(d) and gsa-1(gf) mutant animals with Z-Val-Phe-CHO (MDL-28170), a potent calpain inhibitor, significantly reduces degeneration in these mutants (Syntichaki et al., 2002).These studies suggest that an increase in internal calcium triggers the activation of calpain proteases that subsequently engages executioner lysosomal and cytoplasmic aspartyl proteases and ultimately leads to necrotic cell death (Syntichaki et al., 2002).

Phenotypes
A complete loss-of-function mutation of egl-30 is presumably inviable (Brundage et al., 1996).Homozygotes having strong reduction-of-function mutations in egl-30 hatch, but are paralyzed with, at most, feeble contractions of body-wall, pharyngeal, and defecation muscles and arrest throughout larval development (Brundage et al. 1996).Less severe alleles cause a variety of phenotypes including sluggish movement, delayed egg laying, slow pharyngeal pumping (Trent et al., 1983;Brundage et al., 1996), and aldicarb resistance (Miller et al., 1999).Gain-of-function mutations in egl-30 or transgenic overexpression of wild type egl-30 cause hyperactive movement with exaggerated body bends, and hyperactive egg laying (Brundage et al., 1996;Bastiani et al., 2003).Overexpression of a constitutively activated form of EGL-30 results in vacuolization of cells, paralysis, and ultimately death (Bastiani et al., 2003).Both loss and gain-of-function mutations in egl-30 cause misplacement of certain migrating neural cell bodies (Kindt et al., 2002).
Rescuing fusions with GFP have revealed occasional expression in some other muscle types (i.e., body-wall, vulval) and in some epidermal cells (i.e., vulval;Bastiani et al., 2003).Neuronal expression of egl-30 has been shown to be sufficient for the modulation of locomotion, sensitivity to aldicarb, mediation of retrograde signaling from post-synaptic muscle cells to neurons, and for its effect on vulval development (Lackner et al., 1999;Doi and Iwasaki, 2002;Moghal et al., 2003).

Pathways
In vertebrates, Gαq activates PLCβ isoforms, which hydrolyze phosphatidylinositol bisphosphate (PIP2) to inositol 1,4,5-trisphosphate (IP3) and diacylglycerol (DAG) (for review, see Sternweis and Smrcka, 1993;Jiang et al., 1994).IP3 interacts with IP3 receptors to stimulate the release of calcium from internal stores while DAG activates PKC.A well-conserved variation of this pathway appears to be employed in C. elegans to mediate acetylcholine release from the ventral cord motor neurons.An increase in the concentration of DAG via activation of EGL-30, and subsequent activation of EGL-8 (PLCβ), is thought to promote presynaptic membrane localization of the DAG receptor protein, UNC-13, a protein that facilitates synaptic vesicle fusion at active zones (Richmond et al., 1999;Richmond et al., 2001).This model is supported by the following observations.Reduction-of-function mutations in egl-30, egl-8, and unc-13 confer reduced sensitivity to aldicarb, an inhibitor of acetylcholinesterase at synapses (Miller et al., 1996;Lackner et al., 1999;Miller et al., 1999).Loss-of-function mutations in dgk-1 (diacylglycerol kinase) can partially bypass the resistance of egl-8 mutants for aldicarb, indicating that acetylcholine release is dependent on DAG levels (Lackner et al., 1999).Both expression of a constitutively active mutation of egl-30 and exposure to phorbol esters cause hyperactive movement and hypersensitivity to aldicarb (Lackner et al., 1999;Reynolds et al., 2005).The effects of increased DAG levels, whether by egl-30(gf) mutations or phorbol ester treatment, are strongly dependent on UNC-13 activity (Lackner et al., 1999;Reynolds et al., 2005).Stimulation of acetylcholine release by phorbol esters, as measured by sensitivity to aldicarb, is blocked by a mutation that eliminates phorbol ester binding to UNC-13 (Lackner et al., 1999).An unc-13(rf) mutation is epistatic to an egl-30(gf) mutation with respect to locomotion.(Reynolds et al., 2005).Expression of an UNC-13 mutant protein that is constitutively membrane-bound restores acetylcholine release to mutants lacking EGL-8 (Lackner et al., 1999).Also, the larval arrest and paralysis of egl-30 null mutants are rescued by phorbol ester treatment (Reynolds et al., 2005).Exogenous treatment of wild-type animals with arecoline, believed to act as an agonist for receptors that are coupled to EGL-30, leads to both hypersensitivity to aldicarb and an increase in the accumulation of UNC-13 at presynaptic membranes (Lackner et al., 1999).Taken together, these data indicate that the pre-synaptic EGL-30 pathway results in synaptic vesicle priming leading to acetylcholine release.This pathway is inhibited by serotonin, via the heterotrimeric Gα protein, GOA-1 (discussed in further detail below; Nurrish et al., 1999).
A number of other processes may also be linked to egl-30-mediated release of acetylcholine.Genetic epistasis with mutations in egl-8 and unc-13 indicate that these genes act downstream of egl-30 with respect to the stimulation of pharyngeal pumping.In this case, itr-1, which is broadly expressed in the pharynx and encodes the C. elegans IP3 receptor, may also act downstream of, or in parallel to, egl-30 (Bastiani et al., 2003).Similarly, an analysis of response to emodepside, a ligand of nematode latrophilin that stimulates neuronal exocytosis and elicits pharyngeal paralysis revealed that latrophilin-dependent neurotransmitter release required egl-30, egl-8 and unc-13 (Willson et al., 2004).Mutations in genes that increase acetylcholine release, such as in unc-43, dgk-1, and egl-8, enhance ectopic axon branching induced by the amino-terminal domain of UNC-6 (Wang and Wadsworth, 2002).
Heterotrimeric G proteins in C. elegans egl-30 gain-of-function, as well as reduction-of-function mutants, display defects in neuronal cell migration.Neuronal cell migration is also affected by mutations in egl-8 and dgk-1, although genetic epistasis analysis has not yet been performed (Kindt et al., 2002).EGL-30 must directly activate other downstream effector molecules in addition to EGL-8, and also mediates and affects other processes in addition to acetylcholine release.The pathway for serotonin-mediated stimulation of egg laying via egl-30 has been shown to be largely independent of EGL-8, thereby specifically implicating another unidentified major effector for EGL-30 (Bastiani et al., 2003).During protraction of the male spicules, egl-30 both enhances a nicotinic pathway and may mediate a muscarinic pathway to affect spicule protraction, and ultimately regulates downstream UNC-68 ryanodine or EGL-19-containing calcium channels (Garcia et al., 2001).EGL-30 activates vulval induction in an EGL-19-dependent manner under specific growth conditions (Moghal et al., 2003).The mechanism by which EGL-30 regulates these channels is not yet known.Further analysis in C. elegans should reveal other direct effectors for EGL-30, and reveal how EGL-30 interacts with other signaling pathways.

GEF regulation
ric-8 encodes a GEF expressed in neurons that affects egg laying and locomotion.RIC-8 acts genetically upstream of the EGL-30 signaling pathway (Miller and Rand, 2000).Like egl-30(rf) mutations, ric-8(rf) mutations confer aldicarb resistance, suggesting that RIC-8 may act as a positive regulator of EGL-30 for acetylcholine release (Miller et al., 1996;Miller and Rand, 2000).However, although gain-of-function mutations in egl-30 suppress the paralysis of ric-8(rf) mutants, even a strong gain-of-function mutation in egl-30 is unable to suppress a ric-8 null mutation (Reynolds et al., 2005;Schade et al., 2005).An in vitro biochemical analysis of vertebrate Ric-8 indicates that this protein can interact with, and act as a GEF for, multiple Gα subunits (Tall et al., 2003).Similarly, C. elegans RIC-8 can interact with or stimulate the GEF activity in vitro of at least two Gα subunits, GOA-1 and GPA-16 (Afshar et al., 2004;Couwenbergs et al., 2004, Hess et al., 2004).Further studies are needed to address the in vivo specificity of this GEF and its role in EGL-30 signaling.
The observation that goa-1(lf) mutants are resistant to serotonin-induced slowing of locomotion led to the proposal that serotonin and GOA-1 mediate a neuromodulatory effect on acetylcholine release from the ventral cord motorneurons.Indeed, serotonin induces resistance to aldicarb and requires goa-1 to mediate this effect (Nurrish et al., 1999).GOA-1 also negatively regulates the localization of UNC-13 (Nurrish et al., 1999).Mutations in egl-30, and mutations in genes that function in the EGL-30 pathway, are epistatic to mutations in goa-1 with respect to aldicarb sensitivity, suggesting that activation of GOA-1 by serotonin negatively regulates egl-30-mediated acetylcholine release via UNC-13 (Lackner et al., 1999;Miller et al., 1999;Nurrish et al., 1999; for further detail, see the "Pathways" subheading in the EGL-30 section).These pathways appear to intersect at the level of the RGS protein EAT-16, shown to negatively regulate EGL-30, and/or at the level of diacylglycerol kinase, DGK-1, which affects accumulation of DAG (see Figure 2).DGK-1 or EAT-16 might be effectors of GOA-1 since loss-of-function mutations in eat-16 and dgk-1 were isolated in a screen for suppressors of the lethargic and egg-laying defective phenotypes conferred by a goa-1(QL) transgene (Hajdu-Cronin et al., 1999;Miller et al., 1999;Nurrish et al., 1999).

Heterotrimeric G proteins in C. elegans
Genetic analysis suggests that this signaling network regulates both locomotion and egg laying (Hajdu-Cronin et al., 1999;Miller et al., 1999;Nurrish et al., 1999).It is also likely that unc-43 may act upstream of this signaling network to regulate goa-1 (Robatzek and Thomas, 2000).GPB-2, the C. elegans ortholog of vertebrate Gβ5, has been shown to regulate the EGL-30/GOA-1 signaling network (Chase et al., 2001;Robatzek et al., 2001;van der Linden et al., 2001).Mutations in gpb-2 were isolated together with mutations in eat-16, goa-1, and dgk-1 as suppressors of unc-43(gf) (Robatzek and Thomas, 2000).Genetic analysis using a targeted deletion of gpb-2 also indicates that GPB-2 modulates the EGL-30/GOA-1 signaling network (see Figure 2; Chase et al., 2001;van der Linden et al., 2001).Like vertebrate Gβ5, GPB-2 has been shown to interact with both C. elegans GGL-containing RGS proteins in vitro (van der Linden et al., 2001).These RGS proteins, EGL-10 and EAT-16, regulate GOA-1 and EGL-30, respectively.(Koelle and Horvitz, 1996;Hajdu-Cronin et al., 1999).Since single amino acid substitutions within the gpb-2 coding region can confer either hyperactive or lethargic egg laying, it is possible that the mutations preferentially affect interactions of GPB-2 with either EGL-10 or EAT-16 (Robatzek et al., 2001).Putative null alleles of gpb-2 are viable, affect egg laying, pharyngeal pumping, and locomotion, and phenotypically resemble animals lacking both EAT-16 and EGL-10 (Chase et al., 2001;van der Linden et al., 2001).gpb-2 is broadly expressed in neurons and in muscle (Chase et al., 2001;van der Linden et al., 2001), so it could be a general regulator of the GOA-1 and EGL-30 signaling pathways.Signaling mechanisms that drive interaction of GPB-2 with either RGS protein remain to be elucidated.et al., 2005).Proteins highlighted in yellow activate neurotransmitter release and proteins highlighted in blue inhibit neurotransmitter release.The putative regulatory "switch" protein, GPB-2, is highlighted in green.Solid lines indicate strong support for direct interactions, and dotted lines indicate a genetic pathway for which there may be intermediate components.The major stimulatory pathway is the Gαq pathway, indicated by the red arrows.The Gαo pathway inhibits the Gαq pathway by inhibiting presynaptic localization of UNC-13, required for synaptic vesicle priming.DGK-1 and EAT-16 are proposed effectors of GOA-1, but there are no biochemical data to support these models yet.GPB-2 has been shown to interact with both EAT-16 and EGL-10 in vitro, and GPB-2 is believed to stimulate the GAP function of these RGS proteins.Functional Gαq and Gαs pathways are required for locomotion, and the Gαs pathway acts downstream of DAG.Support for this model is based on the following studies: (Brundage et al., 1996;Hajdu-Cronin et al., 1999;Lackner et al., 1999;Miller et al., 1999;Nurrish et al., 1999;Richmond et al., 1999;Robatzek and Thomas, 2000;Chase et al., 2001;Richmond et al., 2001;Robatzek et al., 2001;van der Linden et al., 2001;Reynolds et al., 2005;Schade et al., 2005).

Regulation of the EGL-30 pathway by GSA-1
In a screen for suppressors of paralysis caused by ric-8(rf) mutations, Schade et al. (2005) recovered gain-of-function alleles of egl-30, gsa-1, and acy-1 as well as loss-of-function alleles of kin-2, which encodes the regulatory subunit of PKA.Epistasis results using a ric-8 null mutation suggest that RIC-8 is a regulator of both the EGL-30 and GSA-1 pathways in neurons (Reynolds et al., 2005).It has not yet been demonstrated whether RIC-8 can function as a GEF for either of these Gα subunits.EGL-30 is implicated as a critical component of the core synaptic vesicle priming pathway.On the other hand, GSA-1 does not appear to affect steady-state neurotransmitter release, based on the essentially normal aldicarb sensitivity of an acy-1 null mutant.(Reynolds et al., 2005).Although these proteins do not appear to affect the same process, their pathways appear to merge downstream of DAG (see Figure 2).Either hyperactivation of the EGL-30 pathway or treatment with phorbol esters can rescue acy-1 null mutants with respect to paralysis.Hyperactivation of the GSA-1 pathway in ric-8 null mutants, in combination with phorbol ester treatment, restores locomotion.The authors propose that neuronal GSA-1 is required downstream of DAG for the EGL-30-dependent priming pathway to drive locomotion (Reynolds et al., 2005).
In summary, EGL-30 appears to be negatively regulated by the RGS protein EAT-16 for movement and egg laying (Hajdu-Cronin et al., 1999).In motor neurons, the EGL-30 pathway acts to generate the core signals for synaptic vesicle priming.This pathway is positively regulated by the GEF protein RIC-8 and is negatively regulated by the GOA-1 pathway (Lackner et al., 1999;Miller et al., 1999;Nurrish et al., 1999;Miller and Rand, 2000).The GSA-1 pathway may modify the EGL-30 pathway in motor neurons, perhaps by providing positional information to stabilize vesicle priming at selected synapses optimal for coordinated locomotion (see Figure 2; Reynolds et al., 2005).

Introduction
In C. elegans, goa-1 encodes the only clear member of the mammalian Gi/o class of Gα subunits.In mammals, Gi/o subunits transduce signals from several hormones and neurotransmitters including acetylcholine, dopamine and serotonin (Neves et al., 2002).The predicted amino acid sequence of C. elegans GOA-1 is over 80% identical to that of mammalian Gαo (Lochrie et al., 1991).Characterization of goa-1 mutants and intensive RNAi analysis has revealed that Gαo signaling in C. elegans regulates several fundamental processes including cell division, neuronal migration, and synaptic transmission.Large-scale RNAi screens have also revealed roles for GOA-1 in egg laying, vulval development, embryonic viability, locomotion, fertility, and the regulation of spindle orientation (Fraser et al., 2000;Simmer et al., 2003).

Expression
In larval and adult C. elegans, goa-1-reporter gene fusions are expressed throughout the nervous system and in a few non-neuronal cell types including pharyngeal, enteric, and sex-specific muscles, and the hermaphrodite distal tip cells (Mendel et al., 1995;Ségalat et al., 1995).The neuronal expression of GOA-1 has been confirmed by immunolocalization (Miller and Rand, 2000).

Heterotrimeric G proteins in C. elegans
In embryos, GOA-1 is present at the cell membrane in all cells (up to at least the 20-cell stage).In the early embryo, GOA-1 is also localized to the cell cortex and to regions surrounding the centrosomal asters (Miller and Rand, 2000;Gotta and Ahringer, 2001).

Pathways
In neuronal migration and synaptic transmission GOA-1 signaling is mediated by its GTP-bound form, is positively regulated by ligand-activated receptors, and is negatively regulated primarily by the RGS protein EGL-10 (see "Receptor-mediated pathway" below).In early embryonic cleavages, and perhaps in many mitotic cell divisions, GOA-1 function is partially redundant with that of a structurally similar Gα subunit, GPA-16.In this pathway, Gα signaling is receptor independent, negatively regulated by RGS-7, and positively regulated by two nearly identical GPR proteins, GPR-1 and/or GPR-2, and by the GEF protein RIC-8 (see "Receptor-independent pathway" below).

Receptor-mediated pathway
GOA-1 appears to negatively regulate synaptic transmission because goa-1(lf) mutants exhibit hyperactive locomotion and egg laying, and are hypersensitive to aldicarb (Mendel et al., 1995;Ségalat et al., 1995;Miller et al., 1999;Nurrish et al., 1999).Although immediate downstream effectors for GOA-1 have not been definitively identified, GOA-1 likely affects locomotion rate by negatively regulating the acetylcholine release pathway mediated by EGL-30 (discussed under "Negative regulation of the EGL-30 pathway by GOA-1").The role of GOA-1 in egg laying is likely complex since it acts both pre-and post-synaptically (Mendel et al., 1995;Nurrish et al., 1999;Shyn et al., 2003).Presynaptically, GOA-1 may function in egg laying as it does in the ventral cord motor neurons; inhibiting neurotransmitter release by negatively regulating the EGL-30 pathway (Bany et al., 2003).Postsynaptically, GOA-1 appears to negatively regulate vulva muscle activity independent of input from the egg-laying motor neurons (Shyn et al., 2003).
For locomotion and egg laying, GOA-1 is negatively regulated by the RGS protein EGL-10.Mutations in egl-10 cause phenotypes opposite to those seen in goa-1, and goa-1(lf) mutations are epistatic to those in egl-10 ( Koelle and Horvitz, 1996).Two other RGS proteins, encoded by rgs-1 and rgs-2, can also act as GTPase activators for GOA-1 (Dong et al., 2000).Loss-of-function mutations in rgs-1 or rgs-2 are not opposite to those of goa-1.Instead, RGS-1 and RGS-2 are redundantly required to rapidly induce egg laying when starved animals are re-fed (Dong et al., 2000).
The signaling molecules that directly activate a GOA-1-coupled receptor have not yet been defined.However, several neurotransmitters and neuromodulators appear to act upstream of GOA-1.goa-1(lf) mutants are partially resistant to the effects of serotonin on locomotion, egg-laying and defecation, suggesting that GOA-1 may act as a serotonin effector (Mendel et al., 1995;Ségalat et al., 1995).Exogenous serotonin reduces locomotion rate by reducing acetylcholine release at neuromuscular junctions, resulting in resistance to aldicarb.This effect requires GOA-1 activity (Nurrish et al., 1999).Although expression of goa-1 in ventral cord motor neurons is sufficient to alter aldicarb responsiveness, the site of action of serotonin's effect on movement has not been established (Nurrish et al., 1999).Serotonin also acts upstream of GOA-1 in a pathway that mediates the enhanced slowing response seen when starved worms are reintroduced to food (Sawin et al., 2000;Ranganathan et al., 2001).Additionally, spontaneous activity of the egg-laying motorneurons is silenced by serotonin through a pathway that requires GOA-1 activity (Shyn et al., 2003).GOA-1 also acts downstream of serotonin in the migrating neurons ALM L/R, BDU L/R, SDQR and AVM, perhaps transducing the serotonin signal to the UNC-2 calcium channel (Kindt et al., 2002).In addition to serotonin, other neurotransmitters act upstream of GOA-1, including FRMFamides (Nelson et al., 1998;Rogers et al., 2001), acetylcholine (Bany et al., 2003), octopamine (Rogers et al., 2001) and dopamine (Sawin et al., 2000).GOA-1 may also transduce a signal present in a dauer pheromone preparation that confers resistance to volatile anesthetics (van Swinderen et al., 2002).

Receptor-independent pathway
Loss-of-function mutations in goa-1 cause a low percentage of embryonic lethality (Miller and Rand, 2000).Strong synthetic embryonic lethality is seen when both GOA-1 and a structurally similar Gα subunit, GPA-16, are depleted, either by RNA interference (RNAi) or mutation (Gotta and Ahringer, 2001;Bergmann et al., 2003).Depletion of both Gα subunits causes multiple defects in cell cleavage in the early embryo.During the first mitotic cleavage, nuclear rotation does not occur, spindle movement to the posterior is reduced or absent, and the posterior Heterotrimeric G proteins in C. elegans spindle aster remains spherical and fails to oscillate (Gotta and Ahringer, 2001;Tsou et al., 2003).GOA-1 and GPA-16 are also required to generate most or all of the pulling force on spindle poles during mitosis (Colombo et al., 2003).Although first cleavage is nearly symmetric in embryos depleted for GOA-1 and GPA-16, other asymmetries in the zygote (P0) appear normal, suggesting that the Gα subunits act downstream of or in parallel to the PAR proteins (Gotta and Ahringer, 2001).Following the first cleavage in GOA-1/GPA-16 depleted embryos, nuclei are mispositioned and nuclear rotation fails to occur in the blastomere P1 (Gotta and Ahringer, 2001;Tsou et al., 2003).During the next rounds of cell division, centrosomes fail to separate and spindle orientations are incorrect (Gotta and Ahringer, 2001).Embryos arrest with polyploid nuclei and few cells (Srinivasan et al., 2003).
In the early embryo, GOA-1 and GPA-16 partner with the Gβ subunit GPB-1 and the Gγ subunit GPC-2 (Gotta and Ahringer, 2001).It had previously been proposed that some of the aberrations in early cleavage are due to Gβγ signaling (Gotta and Ahringer, 2001).However, potent depletion of GOA-1/GPA-16 is epistatic to depletion of GPB-1 or GPC-2, indicating that regulation of spindle positioning and nuclear rotation in the early embryo acts via Gα signaling (Tsou et al., 2003).The G proteins, as well as their regulators described below, are required maternally to promote normal embryonic development (Zwaal et al., 1996;Miller and Rand, 2000;Srinivasan et al., 2003).
Additional proteins have been identified that function along with GOA-1 and GPA-16 in the embryo.ric-8(lf) mutations result in the same embryonic phenotypes seen when both GOA-1 and GPA-16 are depleted (Miller and Rand, 2000;Afshar et al., 2004;Couwenbergs et al., 2004).RIC-8 can physically interact with both GOA-1 and GPA-16 and has been shown to act as a GEF for GOA-1 (Afshar et al., 2004;Couwenbergs et al., 2004).
Several models describing the receptor-independent G protein cycle involved in spindle force have been proposed.Since both GOA-1 and GPR-1/2 act positively to regulate cell cleavage, it has been proposed that the signaling molecule is a Gα-GDP/GPR complex (Colombo et al., 2003;Gotta et al., 2003;Afshar et al., 2004).Alternatively, Gα-GTP produced by the GEF activity of RIC-8 could be the signaling molecule (Gotta et al., 2003;Srinivasan et al., 2003).Many models suggest that the role of GPR-1/2 is to compete with Gβγ for binding to Gα with the resulting Gα-GDP/GPR complex acting as a substrate for RIC-8 activity (Gotta et al., 2003;Srinivasan et al., 2003;Couwenbergs et al., 2004;Hess et al., 2004).However, RIC-8 appears to act prior to GPR-1/2 since RIC-8 is required for the association of GPR-1/2 with GOA-1, but GPR-1/2 is not required for the association of RIC-8 with GOA-1 (Afshar et al., 2004).At present none of the proposed models encompass all the available data.Moreover, much of the in vitro analysis has not included GPA-16, whose role might not be identical to that of GOA-1.Determination of the active signaling molecule or molecules will be critical to predicting and interpreting epistatic interactions among components of this pathway.

Phenotypes
Analysis of a gpa-12 null mutation as well as a large-scale RNAi analysis, indicate that gpa-12 is not required for viability and that reducing or eliminating its expression does not elicit obvious defects in C. elegans (Kamath et al., 2003;van der Linden et al., 2003).However, another RNAi analysis revealed effects on viability as well as on locomotion, oogenesis, and touch sensitivity (Yau et al., 2003).The reason for these differences is not clear.It is possible that in the latter study, the effects of RNAi may be due to partial silencing of other Gα subunit genes.
Animals that overexpress constitutively activated GPA-12 exhibit developmental growth arrest due to a dramatic reduction in pharyngeal pumping.

Expression
gpa-12 is predominantly expressed in the pharynx and in hypodermal cells, and has shown variable expression in some other muscle and intestinal cells, and in a few neurons in the head, ventral cord and tail of the animal (Jansen et al., 1999;van der Linden et al., 2003;Yau et al., 2003).

Pathways
A genetic screen to recover mutations that suppress the growth arrest phenotype caused by activated GPA-12 revealed one potential downstream target, TPA-1 (van der Linden et al., 2003).TPA-1 is a protein kinase C homolog most similar to the vertebrate novel calcium-independent PKCθ/δ (van der Linden et al., 2003).The site-of-action for growth arrest mediated by GPA-12, as well as suppression mediated by TPA-1, was shown to be in the pharyngeal muscle, based on heterologous promoter fusions with myo-2 (van der Linden et al., 2003), suggesting that these proteins may directly interact with one another.
Activation of Rho by Gα12/13 has been shown, in some cases, to occur specifically through interaction of these Gα proteins with RhoGEFs (Kozasa et al., 1998;Suzuki et al., 2003;Dutt et al., 2004).A CeRhoGEF homolog has been shown to interact with GPA-12 when these proteins are coexpressed in COS-7 cells (Yau et al., 2003).The CeRhoGEF, in turn, has been shown to stimulate serum response factor (SRF) activation, and this activation is inhibited by C3 botulinus toxin, which specifically inactivates Rho by ADP ribosylation.While CeRhoGEF expression appears primarily neuronal, at most weak expression of GPA-12 is observed in neurons, so this pathway may be conserved only in specific tissues in C. elegans.RNAi analysis of CeRhoGEF indicates that it plays a role in embryonic development (Yau et al., 2003).

Expression
For most of the gpa genes, expression is confined to small subsets of mostly sensory neurons.The exceptions are: gpa-7, expressed in most neurons and in all muscle cells, gpa-14, expressed in sensory and a few other neurons and in the vulva muscles, gpa-15, expressed in the distal tip cell and anchor cell as well as sensory neurons, gpa-16, expressed in adults in several neurons, the pharynx and in body wall and vulva muscles, and gpa-17, expressed in the intestine (Jansen et al., 1999;J. Burghoorn and G. Jansen, personal communication).gpa-1, gpa-7 and gpa-15 are also expressed in male-specific neurons (Jansen et al., 1999).See Table 1.
Heterotrimeric G proteins in C. elegans

Phenotypes
The function of each gpa gene has been investigated by examining both loss-of-function (lf) mutants, most obtained by target-selected gene inactivation, and gain-of-function animals overexpressing either a wild type (XS) or constitutively activated (QL) transgene (Zwaal et al., 1997;Roayaie et al., 1998;Jansen et al., 1999).odr-3(lf) mutants were isolated in screens for mutants defective in osmotic avoidance and in chemotaxis to benzaldehyde (Roayaie et al., 1998).Phenotypes caused by mutations in each gpa gene, where known, are described briefly below.Negative results are also included to indicate which phenotypes have been examined.Since many of the GPA proteins function redundantly, interactions between them are described under 'Pathways' below.
gpa-1(lf) animals are wild type for response to both soluble and volatile attractants and repellents.gpa-1(XS) transgenics are defective for response to soluble attractants and repellents, and males have reduced potency (Jansen et al., 1999).gpa-1(QL) animals have no visible phenotype and males have apparently normal mating behavior (Jane Mendel, unpublished observations).
gpa-2(lf) mutants are wild type for response to both soluble and volatile attractants and repellents (Roayaie et al., 1998;Jansen et al.,1999).Roayaie et al. (1998), reported that gpa-2(lf) mutations enhance the mild defect in chemotaxis to butanone seen in odr-3(lf) mutants, indicating that GPA-2 and ODR-3 act redundantly in response to this odorant.However, Lans et al., (2004), found improved odorant responses in gpa-2(lf)odr-3(lf) double mutants for all odorants except isoamylalcohol and butanone, indicating a negative function for GPA-2 in detection of most odorants.In their study, butanone response was too low in odr-3 mutants to detect a redundant stimulatory function for GPA-2 (Lans et al., 2004).gpa-2(lf) mutants are partially resistant to dauer pheromone with respect to dauer formation and recovery from the dauer stage (Zwaal et al., 1997).gpa-2(XS) transgenics are completely resistant to dauer pheromone, whereas expression of the (QL) transgene causes both constitutive dauer formation and resistance to exogenous pheromone (Zwaal et al., 1997).gpa-2(QL) animals are wild type for response to water-soluble attractant and repellents, but are defective for attraction and aversion to certain volatile odorants.Male potency is not affected by the (QL) transgene (Jansen et al., 1999).
gpa-3(lf) mutants are defective in response to water-soluble attractants and repellents, and are partially resistant to dauer induction by pheromone, but are sensitive to pheromone inhibition of dauer recovery (Zwaal et al., 1997;Jansen et al., 1999;Hilliard et al., 2004).gpa-3(lf) mutants are also defective in response to the volatile attractant pyrazine (Jansen et al., 1999).GPA-3 is sufficient for detection of all odorants except butanone (Lans et al., 2004).Loss of gpa-3 activity enhances the chemotaxis defect of odr-3(lf) mutants to all odorants indicating that GPA-3 is redundant with ODR-3, and that it plays a stimulatory role in odorant detection (Lans et al., 2004).gpa-3(QL) transgenics are defective in response to water-soluble attractants and to several volatile odorants, and like gpa-3(XS) transgenics are resistant to dauer pheromone (Zwaal et al., 1997;Jansen et al., 1999).The underlying basis for the chemosensory and dauer defects in gpa-3(QL) transgenics may be due to defects in morphological development of the amphid sensory neurons since these animals are defective for dye filling of amphid neurons (Zwaal et al., 1997).The gpa-3(QL) transgene causes a Daf-c phenotype (Zwaal et al., 1997).
gpa-6(lf) and gpa-6(XS) mutants are wild type for chemotaxis to soluble and volatile attractants and repellents.gpa-6(lf) mutants show and increased preference for NaCl when presented with a choice between NaCl and NaAc (Jansen et al., 1999).gpa-7(lf) mutants are wild type for chemotaxis, except for a defect in aversion to low concentrations of SDS.gpa-7(XS) transgenics are defective for chemotaxis to both water soluble and volatile attractants and repellents (Jansen et al., 1999).gpa-7(lf) mutants are egg-laying defective, while gpa-7(XS) transgenics are hyperactive for this behavior, suggesting that gpa-7 plays a stimulatory role in egg-laying (Jansen et al., 1999).However, mutations have no significant effect on vulva muscle activity (Shyn et al., 2003).

Figure 2 .
Figure2.Depicted is the G protein signaling network that controls the release of acetylcholine from the motorneurons that regulate locomotion and egg laying (adapted fromReynolds et al., 2005).Proteins highlighted in yellow activate neurotransmitter release and proteins highlighted in blue inhibit neurotransmitter release.The putative regulatory "switch" protein, GPB-2, is highlighted in green.Solid lines indicate strong support for direct interactions, and dotted lines indicate a genetic pathway for which there may be intermediate components.The major stimulatory pathway is the Gαq pathway, indicated by the red arrows.The Gαo pathway inhibits the Gαq pathway by inhibiting presynaptic localization of UNC-13, required for synaptic vesicle priming.DGK-1 and EAT-16 are proposed effectors of GOA-1, but there are no biochemical data to support these models yet.GPB-2 has been shown to interact with both EAT-16 and EGL-10 in vitro, and GPB-2 is believed to stimulate the GAP function of these RGS proteins.Functional Gαq and Gαs pathways are required for locomotion, and the Gαs pathway acts downstream of DAG.Support for this model is based on the following studies:(Brundage et al., 1996;Hajdu-Cronin et al., 1999; Lackner et al., 1999; Miller et al., 1999;Nurrish et al., 1999; Richmond et al., 1999; Robatzek and Thomas, 2000;Chase et al., 2001; Richmond et al., 2001; Robatzek et al., 2001; van der Linden et al., 2001; Reynolds et al., 2005;Schade et al., 2005).