The G-Protein ß-Subunit GPB-2 in Caenorhabditis elegans Regulates the G_o–G_q Signaling Network Through Interactions With the Regulator of G-Protein Signaling Proteins EGL-10 and EAT-16

Corresponding author: Ronald H. A. Plasterk, Hubrecht Laboratory, Centre for Biomedical Genetics, Uppsalalaan 8, 3584 CT, Utrecht, The Netherlands., plasterk{at}niob.knaw.nl (E-mail)

Communicating editor: P. ANDERSON

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The genome of Caenorhabditis elegans harbors two genes for G-protein ß-subunits. Here, we describe the characterization of the second G-protein ß-subunit gene gpb-2. In contrast to gpb-1, gpb-2 is not an essential gene even though, like gpb-1, gpb-2 is expressed during development, in the nervous system, and in muscle cells. A loss-of-function mutation in gpb-2 produces a variety of behavioral defects, including delayed egg laying and reduced pharyngeal pumping. Genetic analysis shows that GPB-2 interacts with the GOA-1 (homologue of mammalian G_o) and EGL-30 (homologue of mammalian G_q) signaling pathways. GPB-2 is most similar to the divergent mammalian Gß5 subunit, which has been shown to mediate a specific interaction with a G-subunit-like (GGL) domain of RGS proteins. We show here that GPB-2 physically and genetically interacts with the GGL-containing RGS proteins EGL-10 and EAT-16. Taken together, our results suggest that GPB-2 works in concert with the RGS proteins EGL-10 and EAT-16 to regulate GOA-1 (G_o) and EGL-30 (G_q) signaling.

HETEROTRIMERIC G proteins, consisting of a guanine nucleotide-binding -subunit and a ß dimer, act as signal-transducing molecules that couple serpentine transmembrane receptors to a wide variety of intracellular effectors (KAZIRO et al. 1991 ; SIMON et al. 1991 ). The G-subunit and the Gß dimer can independently interact with and regulate downstream effector molecules (CLAPHAM and NEER 1993 ). Despite the extensive data on the biochemical properties of these components, there are still major questions regarding the mechanisms through which Gß functions in vivo, and the relationship between Gß dimers and their potential G partners. For example, it is unclear how a given Gß dimer interacts with distinct G signaling cascades in a given cell or tissue in vivo. Genetic analysis in the model organism Caenorhabditis elegans may provide insight into the in vivo Gß function and its interaction with previously studied G signaling pathways.

Recent genetic experiments in C. elegans demonstrate that GOA-1 (G_o) antagonizes EGL-30 (G_q) signaling, with the G_o pathway acting upstream of, or parallel to, the G_q pathway (HAJDU-CRONIN et al. 1999 ; MILLER et al. 1999 ). Both the G_o and G_q signaling pathway regulate neurotransmitter secretion in C. elegans (Rand AND NONET 1997 ; LACKNER et al. 1999 ; NURRISH et al. 1999 ). This antagonism between G_o and G_q involves EAT-16, a regulator of G-protein signaling proteins (RGS) and/or DGK-1 (DAG-kinase; HAJDU-CRONIN et al. 1999 ; MILLER et al. 1999 ), both identified as suppressors of activated G_o (HAJDU-CRONIN et al. 1999 ; MILLER et al. 1999 ; NURRISH et al. 1999 ). RGS proteins function as GTPase-activating proteins (GAPs) that accelerate the intrinsic GTPase activity of the G-subunit (N. WATSON et al. 1996 ; KOELLE 1997 ; NEER 1997 ). Genetic analysis shows that EAT-16 (RGS) functions downstream of, or parallel to, GOA-1 (G_o). However, rather than acting as a GAP for GOA-1 (G_o), EAT-16 acts as a GAP for EGL-30 (G_q; HAJDU-CRONIN et al. 1999 ). Thus, G_o could negatively regulate G_q signaling through EAT-16 by functioning as a direct effector of GOA-1 or indirectly via the G-subunit-like (GGL) domain of EAT-16 that may functionally mimic a G-subunit in the heterotrimeric G-protein complex (GUAN and HAN 1999 ). Another GGL-containing RGS protein in C. elegans, EGL-10, was previously identified as an upstream negative regulator of GOA-1 (G_o) signaling (KOELLE and HORVITZ 1996 ). Together, these studies suggest that the RGS proteins EGL-10 and EAT-16 inhibit G_o and G_q signaling in C. elegans, respectively, and that especially EAT-16 has an important role in mediating cross-talk between G_o and G_q. At this point, however, it is not known where and how the Gß-subunits function in the G_o–G_q signaling network.

Two genes encoding G-protein ß-subunits have been identified in the complete C. elegans genome: gpb-1 (VAN DER VOORN et al. 1990 ) and gpb-2. The first G-protein ß-subunit gene, gpb-1, is expressed throughout development in most tissues and in the germline, and loss of gpb-1 results in embryonic lethality (ZWAAL et al. 1996 ). Here, we show that the second G-protein ß-subunit gene, gpb-2, is expressed in most neurons and in muscle cells and that loss of gpb-2 is not lethal, but affects behaviors such as egg laying and pharyngeal pumping. Animals lacking both gpb-2 and goa-1 function have a synthetic larval lethal phenotype, which is rescued when EGL-30 activity is reduced, suggesting that G_q activity causes the gpb-2; goa-1 double mutant synthetic lethality. GPB-2 is most homologous to the divergent mammalian Gß5 subunit that, unlike the other mammalian Gß-subunits (Gß1-4), has the ability to interact with the GGL domain of a subset of RGS proteins in vitro (SNOW et al. 1998 , SNOW et al. 1999 ; LEVAY et al. 1999 ). Our results show that GPB-2 interacts with the two GGL-containing RGS proteins in C. elegans, EGL-10 and EAT-16, and that GPB-2 regulates the functions of these proteins in the G_o–G_q signaling network.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Nematode strains, culturing, and manipulation:
General methods used for culturing, manipulation, and genetics of C. elegans were as described by LEWIS and FLEMING 1995 . DNA transformation assays in C. elegans by microinjection were as described by MELLO and FIRE 1995 . Unless indicated, all experiments were performed at 20°. The following strains were used in this study: Bristol N2, PS4034 [eat-16(sy438)I; HAJDU-CRONIN et al. 1999], DA0823 [egl-30(ad805)I], MT1434 [egl-30(n686)I; BRUNDAGE et al. 1996], NL561 [goa-1(pk62)I; MENDEL et al. 1995], MT363 [goa-1(n363)I; SÉGALAT et al. 1995], DR102 [dpy-5(e61)unc-29(e403)I], CB73 [unc-15(e73)I], CB1282 [dpy-20(e1282)IV], MT8504 [egl-10(md176)V; KOELLE and HORVITZ 1996], MT1083 [egl-8(n488)V; TRENT et al. 1983; MILLER et al. 1999], and KP1097 [dgk-1(nu62)X; NURRISH et al. 1999].

Sequence analysis of GPB-2:
The GENEFINDER prediction of the gpb-2 open reading frame, as annotated in the C. elegans database ACeDB (EECKMAN and DURBIN 1995 ), was confirmed by cloning and sequencing of the cDNA (see Yeast two-hybrid analysis). The coding sequence of GPB-2 has been submitted to the DDBJ/EMBL/GenBank databases under accession no. AF291847. The percentages of identity and similarity given above were derived using the Genetics Computer Group (GCG, version 9.1) GAP program using default parameters. Multiple alignments of GPB-2, mammalian, Dictyostelium, and Drosophila G-protein ß-subunit sequences (Swiss-Prot accession nos. P04901, P54311, P54312, P11016, P29387, AF291846, AF291847, P16520, P36408, O14775, and P29829) were made using the GCG PILEUP program using default parameters.

GFP reporter constructs:
A translational fusion of gpb-2 to gfp (pRP2016) was constructed by inserting a ±3-kb fragment amplified by PCR on wild-type gpb-2 genomic sequence using the primers green fluorescent protein (GFP)-BTWO1 (5'-ATAGCATGCTTCCTGGTGATCAGGTCATGT) and GFP-BWTO2 (5'-TAGGATCCAATAGCACATGTTGAATCTCC), containing a unique SphI and BamHI site (shown in bold), respectively, into the SphI and BamHI sites of pPD95.77 (A. FIRE, personal communication). The construct pRP2016 was injected at 50 µg/ml with 100 µg/ml pMH86 (HAN and STERNBERG 1991 ) into CB1282 [dpy-20(e1282)IV] to produce NL1562 (dpy-20(e1362)IV; pkEx563[gpb-2::gfp,dpy-20(+)]). A second translational fusion of gpb-2 to gfp (pRP2017) was constructed by inserting a ±3-kb fragment amplified by PCR on wild-type gpb-2 genomic sequence using the primers GFP-BTWO3 (5'-ATAGCATGCGGCTCTAGATAGGTATGTAGA) and GFP-BWTO4 (5'-TAGGATCCCATGCATAATACTTTTCCCAC), containing a unique SphI and BamHI site (shown in bold), respectively, into the SphI and BamHI sites of pPD95.77 (A. FIRE, personal communication). The construct pRP2017 was injected at 50 µg/ml with 150 µg/ml pMH86 (HAN and STERNBERG 1991 ) into CB1282 [dpy-20(e1282)IV]. This transgenic array was integrated by irradiating animals with 40 Gy of gamma radiation from a ¹³⁷Cs source (WAY et al. 1991 ) to produce NL2784 (dpy-20(e1282)IV; pkIs1303[gpb-2::gfp,dpy-20(+)]). NL2784 was backcrossed once to a N2 wild-type background to produce NL2785. Cells were identified in reference to SULSTON and HORVITZ 1977 and WHITE et al. 1986 .

Generation and transgenic rescue of a gpb-2 loss-of-function mutation:
A deletion mutant of gpb-2, pk751, was isolated from a chemical deletion library using primers BTWO5 (5'-ACAATTGGCAAATGAAGCCG) and BTWO6 (5'-TCAACGGAAATTGAGAGATG) and nested primers BTWO7 (5'-CACAAGCTTAATGACATTCC) and BTWO8 (5'-AGAAGCCGTGACGGATGACC) as described by JANSEN et al. 1997 . The PCR product detecting the deletion in pk751 was sequenced and confirmed by Southern analysis (data not shown). The following sequence in capitals is deleted in pk751: (5'—tagtcttcAAAAATTGG—TGTAAAAAaagagtg—). The pk751 allele was backcrossed three times to an N2 wild-type background, resulting in the strain NL2001 [gpb-2(pk751)I].

The behavioral defects of pk751 were complemented with two wild-type gpb-2 genomic constructs, pRP2015 (see below) and pRP2051. The rescuing construct pRP2051, containing a 4.7-kb SpeI-XbaI fragment, including the gpb-2 genomic sequence and 1.3 kb of upstream sequence, was injected at 20 µg/ml with 150 µg/ml pMH86 (HAN and STERNBERG 1991 ) into NL2002 [dpy-20(e1282)IV; gpb-2(pk751)I], resulting in strains NL2771, NL2772, and NL2773 [dpy-20(e1282)IV; pkEx1321, pkEx1322, and pkEx1323, respectively]. Homozygosity for the pk751 allele was confirmed by single worm PCR.

Overexpression of GPB-2:
We generated multiple transgenic lines that overexpress GPB-2 from a transgene carrying a multicopy array, either extrachromosomal or integrated, of the gpb-2 gene under control of its endogenous promoter. The rescuing construct pRP2051 was injected at different concentrations of plasmid DNA with 150 µg/ml pMH86 (HAN and STERNBERG 1991 ) into CB1282 [dpy-20(e1282)IV] at 20 µg/ml to produce NL2757 (dpy-20(e1282)IV; pkEx1318 [gpb-2(+),dpy-20(+)]), at 30 µg/ml to produce NL2758 [dpy-20(e1282)IV; pkEx1319], and at 50 µg/ml to produce NL2759 [dpy-20(e1282)IV; pkEx1320]. All strains obtained showed no defects in egg laying, locomotion, and pharyngeal pumping compared to wild type (data not shown). Another rescuing construct pRP2015, containing a 6.9-kb SpeI fragment including the gpb-2 genomic sequence and 3.5 kb of upstream sequence, was injected at 50 µg/ml with 150 µg/ml pMH86 (HAN and STERNBERG 1991 ) into CB1282 [dpy-20(e1282)IV] to produce NL1557 (dpy-20(e1282)IV; pEx557[gpb-2(+),dpy-20(+)]). This transgenic array was integrated by irradiating animals with 40 Gy of gamma radiation from a ¹³⁷Cs source (WAY et al. 1991 ) to produce NL1559 (dpy-20(e1282)IV; pkIs559[gpb-2(+),dpy-20(+)]). NL1559 was backcrossed two times to an N2 wild-type background to produce NL2004. The strain NL2004 was used in behavioral assays as described below. To confirm that the integrated transgene pkIs559[gpb-2(+)] is functional, we crossed pkIs559[gpb-2(+)] into NL2002 [dpy-20(e1282)IV; gpb-2(pk751)I], resulting in strain NL2006 (dpy-20(e1282)IV;gpb-2(pk751)I; pkIs559[gpb-2(+),dpy-20(+)]).

A construct with gpb-2 expression under the control of a heat-shock promoter (hsp) was generated by cloning a ±800-bp SpeI-AvrII hsp16.2-gpb-2 containing PCR fragment in front of an AvrII-SpeI fragment of gpb-2 from pRP2051. In a first PCR, two fragments were amplified. The first 518-bp fragment was amplified on the hsp16.2 promoter from pPD49.78 (STRINGHAM et al. 1992 ) using the primers HSP1 (5'-GGTCGACACTAGTGGATCAAGAGC) and HSP3 (5'-GAGTTTTCTGGCATGGTACCGTAGACGC). In this way, the PCR fragment becomes flanked on one side by a unique SpeI site (shown in boldface type) and on the other side by the first 14 bp of gpb-2. The second 296-bp fragment was amplified on gpb-2 using the primers HSP2 (5'-CGACGGTACCATGCCAGAAAACTCTCAG) and HSP4 (5'-CTAACCATGGAATTGTAAAG). This PCR fragment is flanked on one side by the last 10 bp of hsp16.2 and on the other side by a unique AvrII site (shown in bold). In a second PCR, both fragments were allowed to anneal the overlapping flanks (underlined) and used as a template to amplify a ±800-bp fragment using the outer primers (HSP1 and HSP4). The SpeI-AvrII PCR fragment was cut and cloned into the XbaI-AvrII sites of the construct pRP2051, resulting in pRP2201. The region of the PCR fragment used in cloning was sequenced completely and found to be free of amplification errors. Injection of pRP2201 was done with varying concentrations of plasmid DNA (10–50 µg/ml) with 150 µg/ml pMH86 (HAN and STERNBERG 1991 ) into CB1282 [dpy-20(e1282)IV] to produce NL2775, NL2776, and NL2780 [dpy-20(e1282)IV; pkEx1324, pkEx1325, and pkEx1326, respectively]. Transgenic animals of different larval stages (L1 to adult) were placed on Escherichia coli OP50-seeded NGM agar plates and heat shocked for 2 hr at 33° to induce promoter activity.

Construction of double mutant strains:
Double mutants used in this study were generated by standard genetic methods. Homozygosity of the alleles was confirmed by either sequencing amplified genomic DNA of strains containing eat-16(sy438), egl-30(ad805), egl-30(n686), and dgk-1(nu62) or by single worm PCR for strains containing gpb-2(pk751), egl-8(n488), and goa-1(pk62). To detect egl-10(md176), we used single worm PCR using primers that fail to amplify on the rearranged egl-10 region (KOELLE and HORVITZ 1996 ). In addition, all double mutant strains were confirmed by crossing with N2 wild-type males to re-isolate both mutations.

Phenotypic analyses of the gpb-2 mutant, double mutant, and transgenic lines:
The rate of egg laying was assayed by two methods as described (BRUNDAGE et al. 1996 ; KOELLE and HORVITZ 1996 ). In short, in the first egg-laying assay the number of unlaid eggs in the uterus was determined. L4 staged hermaphrodites were placed on E. coli OP50-seeded NGM agar plates and were allowed to develop for 36 hr at 20°. Adults were treated with sodium hypochlorite and eggs were counted using a dissection microscope (Leica MZ75). In the second egg-laying assay, the developmental stage of newly laid eggs was determined by placing L4 staged hermaphrodites on OP50-seeded NGM agar plates. After 29 or 53 hr at 20°, the developmental stage of newly laid eggs was determined in a 2.5-hr interval by counting the cells with a high power dissection microscope (Wild M3C). Newly laid eggs were divided into three categories: one to eight-cell, nine-cell to comma stage, and the post-comma stage. The rate of locomotion was assayed by measuring the frequency of body bends of a hermaphrodite. L4 staged hermaphrodites were placed on OP50-seeded NGM agar plates and were allowed to develop for 24 hr at 20°. At this stage, the locomotion of an adult hermaphrodite is less affected by the presence of excessive unlaid eggs. Each animal was transferred to a thin lawn of a freshly OP50-seeded NGM agar plate 15 min before measuring locomotion by counting body bends in 3-min intervals. A body bend was defined as movement of a quarter of a body length in a forward or backward direction (BRUNDAGE et al. 1996 ; KOELLE and HORVITZ 1996 ). Pharyngeal pumping on food was assayed by measuring the number of pumps per minute. L4 staged hermaphrodites were placed on separate NGM agar plates seeded with 10 µl OP50 and were allowed to develop for 29 hr at 20°. Pharyngeal pumps were counted in 2-min intervals in which one count represents three pumps. The number of the counts was multiplied by 1.5 to yield pumps per minute (HAJDU-CRONIN et al. 1999 ). Assays for egg laying in 5 mg/ml serotonin (5-HT) and 0.75 mg/ml imipramine were performed and interpreted as described by TRENT et al. 1983 , with the exception that newly laid eggs were counted after 90 min. Imipramine and 5-HT were dissolved in M9 buffer. Aldicarb sensitivity was measured either by assaying the time course of the onset of paralysis following acute exposure of a population of animals to aldicarb or by quantifying the population growth rates to various concentrations of aldicarb as described by LACKNER et al. 1999 and MILLER et al. 1999 . For statistical analysis among multiple groups, an ANOVA followed by directed Student's t-test was used. All results are given as mean ± standard errors.

Synthetic lethality:
We constructed a gpb-2(pk751)goa-1(pk62)/unc-15(e73) mutant in which all mutations are very closely linked on the right arm of chromosome I. The percentage of survival of the gpb-2 goa-1 double mutant was scored by placing five gpb-2 goa-1 / unc-15 L4 heterozygotes on separate OP50-seeded NGM agar plates. The heterozygous animals were given the opportunity to lay eggs for 1–2 days. All progeny were transferred to separate plates and followed for one generation to determine their genotype. This yielded 131 Unc animals, 635 heterozygotes, and 195 gpb-2 goa-1 double mutant animals of which 95% were arrested at different larval stages. Similarly, we picked all progeny of two gpb-2(pk751)/+; dgk-1(nu62) heterozygotes. This yielded 69 hyperactive dgk-1 animals, 121 heterozygotes that segregated arrested larvae, and 53 gpb-2; dgk-1 double mutant animals.

We constructed transgenic gpb-2 goa-1 double mutant animals containing the gpb-2 wild-type sequence, which rescues the synthetic lethal phenotype. The rescuing construct pRP2051 was injected at 50 µg/ml with 100 µg/ml pRF4 and 50 µg/ml pRP2017 as markers into gpb-2(pk751)goa-1(pk62)/unc-15(e73) heterozygous animals to produce NL2728 gpb-2(pk751)goa-1(pk62);pk1344Ex[gpb-2(+),gpb-2::gfp, rol-6 (su1006)]. Animals that lost this transgenic array in the germline resulted in larval arrest (>96%).

To generate an egl-30(n686)gpb-2(pk751)goa-1(pk62) triple mutant, egl-30 /+ males were mated with gpb-2 goa-1/++ hermaphrodites, and F₁ cross-progeny were transferred to separate plates. From plates that segregated Egl and arrested larvae, 52 Egl F₂ progeny were transferred to separate plates. Two of the 52 Egls were egl-30 gpb-2 goa-1 / egl-30 ++ and egl-30 gpb-2 goa-1 triple mutants were isolated. As a control, we crossed egl-30 gpb-2 goa-1 hermaphrodites with dpy-5 unc-29/++ males and selected 19 Dpy non-Unc and 16 Unc non-Dpy F₂ recombinants. Fourteen of the 19 Dpy non-Unc were dpy-5 gpb-2 goa-1/dpy-5++ unc-29 recombinants and about one-fourth of their progeny arrested as larvae. Five of the 16 Unc non-Dpy were egl-30 gpb-2 goa-1 unc-29/+ dpy-5 ++ unc-29 recombinants and did not produce lethal progeny.

An egl-8(n488); dpy-5(e61)gpb-2(pk751)goa-1(pk62)/+++ mutant was generated. We placed 10 egl-8; dpy-5 gpb-2 goa-1/+++ L4 heterozygotes on separate plates. The heterozygous animals were given the opportunity to lay eggs for 2 days. A total of 119 Dpy F₁ larvae were transferred to separate plates and followed for a generation to determine their genotype. Of the 119 Dpy larvae, 101 arrested and did not produce progeny, and 18 viable Dpy animals were dpy-5 gpb-2 goa-1/dpy-5++ recombinants. Arrested Dpy larvae were seen among the progeny of all recombinants. In all experiments, homozygosity for gpb-2(pk751) and egl-8(n488) and the presence of goa-1(pk62) was confirmed by single worm PCR. Homozygosity of the egl-30 and dgk-1 allele was confirmed by sequencing genomic DNA of animals containing the n686 and the nu62 mutations, respectively.

Yeast two-hybrid analysis:
The coding sequence of GPB-1 (GenBank accession no. AF291846), GPB-2 (AF291847), GPC-1 (AF291848), and GPC-2 (AF291849) and the regions encoding the GGL of EGL-10 (AF291850) and EAT-16 (AF291851) were amplified from total RNA of the C. elegans N2 strain by RT-PCR. First strand synthesis was initiated using a mixture of random hexamers and oligo(dT) oligonucleotides [100 ng/µl total RNA, 200 nM oligo(20)dT, 180 units/µl random hexamers, 0.5 units/µl avian myeloblastosis virus reverse transcriptase (Roche), 1 mM dNTP, 50 mM Tris-HCl pH 8.5, 8 mM MgCl₂, 70 mM KCl, 1 mM dithiothreitol, 10 ng/µl BSA], and cDNAs were amplified by PCR [10 mM Tris-HCl pH 8.3, 50 mM KCl, 1.5 mM MgCl₂, 0.025 units/µl Taq polymerase (GIBCO, Gaithersburg, MD), 0.2 mM dNTP, 200 nM each primer] using specific oligonucleotides (GPB-1: 5'-GGGAATTCATGAGCGAACTTGACCAAC and 5'-CCCTCGAGTTAATTCCAGATCTTGAG; GPB-2: 5'-GGGAATTCATGCCAGAAAACTCTCAGC and 5'-CCCTCGAGTCAAGCCCAAATGCGAATTG; GPC-1: 5'-GGGAATTCATGGAAAACATCAAGGCATC and 5'-CCCTCGAGTTAGAGTACTGAACAGCTT; GPC-2: 5'-GGGAATTCATGGATAAATCTGACATGC and 5'-CCCTCGAGTTAGAGCATGCTGCACTTG; EGL-10-GGL: 5'-CCCCATATGCCTGGATTACGCCGGTGTAC and 5'-CCCCTCGAGACTATCCTCCCAAAGCTTGAG; EAT-16-GGL: 5'-GGCCATATGCGGCAGAATGCACAAGGTTA and 5'-GGCTCGAGTGTGTTTCGAGCACTTGCCGTC) containing recognition sites for restriction enzymes for in-frame cloning in the appropriate two-hybrid bait or prey vector (pGBK-T7 or pGAD-T7, CLONTECH, Palo Alto, CA). Final cDNA constructs were checked by sequencing for potential differences with the spliced gene products predicted by ACeDB (no differences were found) and for the absence of mutations. Combinations of bait and prey plasmids were introduced in yeast strain AH109 by polyethylene glycol/lithium acetate transformation and assayed for interactions by growth on medium lacking histidine and adenine, according to the manufacturer's protocols (CLONTECH).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

GPB-2 is a homologue of the divergent mammalian Gß5 subunit:
The C. elegans genome project identified two G-protein ß-subunit genes. The first G-protein ß-subunit, gpb-1 (located on chromosome II, F13D12.7), is 86% identical to the conserved mammalian Gß-subunits (VAN DER VOORN et al. 1990 ; ZWAAL et al. 1996 ). However, the predicted sequence of gpb-2 (located on chromosome I, F52A8.2) indicates that the second G-protein ß-subunit is much less conserved. We have isolated and sequenced full-length GPB-2 cDNA (GenBank accession no. AF291847) and found no differences with the sequence predicted by GENEFINDER. The deduced GPB-2 protein consists of 356 amino acid residues, encoding the seven WD40 repeats characteristic for the mammalian ß-subunits. However, the amino acid sequence shares only 50% identity with GPB-1 and the mammalian Gß1-4 subunits (Fig 1A). Interestingly, GPB-2 is most similar to the divergent mammalian Gß5 subunit—64% identity and 73% similarity at the amino acid sequence level.

View larger version (18K): In this window In a new window Download PPT slide   Figure 1. An alignment dendogram of various G-protein ß-subunits from different species and the gene structure of gpb-2. (A) Alignment dendogram generated with the PILEUP algorithm using default program parameters. GPB-2 is most related to human Gß5. (B) Genomic organization of the gpb-2 gene. The solid boxes show coding sequence. The region deleted in gpb-2(pk751), two rescuing constructs, pRP2015 and pRP2051, and two gfp reporter constructs, pRP2016 and pRP2017, are indicated. S, SpeI; X, XbaI.

gpb-2 is widely expressed in neurons and muscle cells:
To analyze GPB-2 expression, we used two translational fusions of the gpb-2 upstream control sequence with the gene encoding GFP (CHALFIE et al. 1994 ). One translational fusion contained gfp fused to the third exon and the other to the fifth exon of gpb-2 (Fig 1B). Both gfp reporter constructs showed identical expression patterns throughout development in the nervous system and in muscle cells. Animals with an integrated gpb-2::gfp array showed no expression in embryos until the comma stage when broad expression was seen, excluding the dorsal posterior site of the comma stage embryo (Fig 2A). From the comma stage onward, when tissues differentiate and become organized, until hatching of the embryos, gpb-2 expression was stronger and broadly expressed in the head and tail ganglia. In larvae and adult animals, gpb-2 expression was seen in most or all neurons, including neurons located in the head ganglia (Fig 2B), the ventral nerve cord (Fig 2D), and the tail ganglia (Fig 2C). Additionally, gpb-2 expression was seen in the hermaphrodite-specific neurons (HSNs), which control egg laying (Fig 2F) and the canal cell-associated neurons (CANs), which are thought to function in osmoregulation (Fig 2F). Besides neuronal expression, high expression levels of gpb-2 were also seen in muscle cells of the pharynx (Fig 2B), the vulva muscle cells (Fig 2D), and the body-wall muscle cells (Fig 2E). In males, similar neuronal and muscle expression was observed. Thus, gpb-2 is widely expressed in most neurons and in muscle cells that control several distinct behaviors.

View larger version (36K): In this window In a new window Download PPT slide   Figure 2. Expression pattern of gpb-2 as detected using translational GFP gene fusions. In the embryo, expression of gbp-2 is first detected in the comma stage (A, lower embryo) and the 1.5-fold stage (A, upper embryo). In adult animals, expression is observed in neurons in the head (B; arrow indicates head ganglia; ph, pharyngeal muscle) and tail (C; tg, tail ganglia) regions. (D) Expression of gpb-2 in the vulva muscle cells (vm) is shown, and in the region surrounding the vulva expression of gpb-2 can be seen in the ventral nerve cord neurons (arrows indicate single neurons or groups of neurons). Furthermore, high expression levels of gpb-2 are found in body-wall muscle cells (E) and in the CAN and HSN cells (F).

gpb-2 is not an essential gene:
To examine GPB-2 function in vivo, we identified a loss-of-function gpb-2 mutant using a reverse-genetic approach. An EMS-derived deletion library (JANSEN et al. 1997 ) was screened for deletions in the gpb-2 genomic sequence. We isolated an allele, gpb-2(pk751), in which part of the gpb-2 coding sequence was removed. As is shown in Fig 1B, this 3380-bp deletion of genomic sequence removes the last five exons of gpb-2 together with 1.5 kb of downstream sequence. Consequently, pk751 leaves only one of the seven WD repeats intact. Because seven WD repeats are needed to form a functional ß-subunit, it is likely that pk751 is a null allele. Animals heterozygous for pk751 were wild type in development and behavior. Animals homozygous for pk751 were viable, but showed a variety of behavioral defects (discussed below). The behavioral defects of pk751 were complemented by introducing a transgene carrying a multicopy array, either extrachromosomal or integrated, of the wild-type gpb-2 genomic sequence, demonstrating that the defects are caused by loss of GPB-2 function (Table 1 and Fig 3).

View larger version (17K): In this window In a new window Download PPT slide   Figure 3. A loss-of-function mutation in gpb-2 affects egg laying and pharyngeal pumping; overexpression of GPB-2 does not have a significant effect on the frequency of body bends during locomotion, egg laying, and pharyngeal pumping. Egg laying assayed as the number of eggs contained within the uterus (A), the frequency of body bends during locomotion (B), and the pumping rate (C) of N2 (wild-type) animals, gpb-2 loss-of-function animals (pk751), and animals transgenic for a multicopy array containing the gpb-2 gene (pkIs559[gpb-2(+)]) in a gpb-2 loss-of-function (pk751) background and a wild-type background. ANOVA comparing groups for different variables was conducted. A significant effect of group was noted for the amount of unlaid eggs per adult (F = 43.7, P < 0.001) and for pharyngeal pumping (F = 26.7, P < 0.001). No significant effect of group was noted for frequency of body bends during locomotion (F = 0.85, P = 0.47). Differences between genotypes for the unlaid eggs per adult and pharyngeal pumping were analyzed using Student's t-test. gpb-2(pk751) mutant animals are significantly different with respect to the amount of unlaid eggs per adult and the rate of pharyngeal pumping from N2 (wild-type) animals (P < 0.001), gpb-2; (pkIs559[gpb-2(+)]) animals (P < 0.001), and (pkIs559[gpb-2(+)]) animals (P < 0.001). There was no significant difference between (pkIs559[gpb-2(+)]) and N2 (wild type) for the unlaid eggs per adult (P = 0.051) and the pharyngeal pumping rate (P = 0.22) or between gpb-2(pk751); (pkIs559[gpb-2(+)]) and N2 (wild type) for the unlaid eggs per adult (P = 0.059) and the pharyngeal pumping rate (P = 0.15). About 20 adult animals were assayed for each genotype (see MATERIALS AND METHODS). Values reported are the means ± standard errors.

Mutant gpb-2 animals are viable, either because GPB-2 does not have an essential function or because GPB-1 is redundant. To address this last question genetically is difficult, since gpb-1 embryos arrest at the four-cell stage, whereas maternally rescued animals arrest at the first stage of larval development (ZWAAL et al. 1996 ). However, the fact that gpb-2(pk751) animals are viable, unlike gpb-1 mutants, allows us to study the function of a Gß-subunit in adult behaviors using genetic tools.

GPB-2 modulates behaviors such as egg laying, locomotion, and pharyngeal pumping:
To investigate the function of GPB-2 in adult animals, we quantified the behavioral defects induced by a loss-of-function gpb-2(pk751) mutant and by overexpression of GPB-2 (Table 1 and Fig 3). gpb-2 mutant animals were less active in egg laying and pharyngeal pumping and had a subtle defect in backward movement; gpb-2 mutants tended to move backward with exaggerated flexes in which the tip of the tail touches the head during each body bend (data not shown). No defect in defecation or dauer formation was observed (data not shown). The rate of egg laying was determined by analyzing the stage of newly laid eggs and the number of unlaid eggs in the uterus. We found that gpb-2 adult hermaphrodites laid fewer eggs at the nine-cell to comma stage of development (74%) compared to the wild type (96%) and laid more of their eggs at the post-comma stage (26%) compared to the wild type (0%, Table 1). More strikingly, 24 hr later, the same gpb-2 adult hermaphrodites laid 97% of their eggs at the post-comma stage compared to 6% for wild-type animals. The number of unlaid eggs that accumulated in gpb-2 mutant animals (32.1 ± 1.6) was also higher than in wild-type animals (19 ± 1.2, Fig 3). One possibility is that the serotonergic HSNs or the vulva muscle cells that control egg laying are defective in gpb-2 mutant animals. Egg laying in wild-type animals is stimulated by adding serotonin (5-HT) and imipramine, a reuptake inhibitor of 5-HT. Animals with nonfunctional HSNs respond to exogenous 5-HT, but are resistant to imipramine (TRENT et al. 1983 ). Animals that are resistant to both agents have defects in the response of the egg-laying muscles. Adult gpb-2 mutant animals respond both to 5-HT and imipramine, indicating that the HSNs and the response of the egg-laying muscles are normal (Table 2). Despite the unusual backward motion described, the frequency of body bends during locomotion was not significantly different in gpb-2 mutant animals compared to the wild type (Fig 3). In addition to these behaviors, we noticed that well-fed adult gpb-2 mutant animals had a starved appearance after 3 days. Although an animal can appear starved for a number of reasons, a starved appearance is often associated with a feeding defect (AVERY 1993 ). Because gpb-2 was highly expressed in the pharyngeal muscle cells, we determined whether the starvation phenotype of gpb-2 mutant animals was caused by a reduced pharyngeal pumping. Indeed, the pharynx pumped slowly in gpb-2 mutant animals (129 ± 8 pumps/min) compared to the wild type (230 ± 10 pumps/min, Fig 3).

To analyze the effect of gpb-2 gene dosage on behavior, we made transgenic animals that overexpress GPB-2 from a chromosomal integrated transgene carrying a multicopy array of the gpb-2 gene, pkIs559[gpb-2(+)]. Functionality of the transgene was demonstrated by its ability to complement the gpb-2(pk751) behavioral defects (Table 1 and Fig 3). However, we found that overexpression of gpb-2 under control of its endogenous (pkIs559[gpb-2(+)]) or heat-shock promoter (pkEx1324, pkEx1325, and pkEx1326; data not shown) did not show any obvious phenotypes with respect to egg laying, locomotion, and pharyngeal pumping (Table 1 and Fig 3). Thus, while GPB-2 is necessary for normal egg laying, locomotion, and pharyngeal pumping, overexpression of GPB-2 does not affect these behaviors.

GPB-2 genetically interacts with GOA-1 (G_o) and EGL-30 (G_q):
The behavioral phenotype of gpb-2 mutants affects similar behaviors as goa-1 (G_o) and egl-30 (G_q) mutants. Reduction-of-function egl-30 alleles induce lethargy and delayed egg laying, whereas a putative null allele results in larval lethality (TRENT et al. 1983 ; BRUNDAGE et al. 1996 ). Loss-of-function mutations in goa-1 result in a starvation phenotype, premature egg laying, and hyperactive locomotion (MENDEL et al. 1995 ; SEGALAT et al. 1995 ). To determine whether GPB-2 plays a role in GOA-1 and EGL-30 signaling, we generated double mutants between gpb-2 and goa-1 and gpb-2 and egl-30. When we crossed gpb-2(pk751) into a goa-1(pk62) background, we saw a synthetic phenotype—lethality at the larval stage (>95%). Of the 195 gpb-2 goa-1 double mutant progeny of goa-1 gpb-2/unc-15 heterozygous animals, 10 escaped to adulthood, but they also in turn produced inviable progeny (>95%). There was no phenotypic difference between the progeny of gpb-2 goa-1 mutant animals that escaped to adulthood and gpb-2 goa-1 progeny of heterozygous mothers. It can be argued that this synthetic lethality is a result of an additive feeding defect of both gpb-2 and goa-1 mutants. Therefore, we investigated the pumping rate in gpb-2 goa-1 double mutant animals (155 ± 11 pumps/min, n = 10) from gpb-2 goa-1 heterozygous mothers and found that there was no significant difference (P = 0.18) in the rate of pharyngeal pumping compared to gpb-2 single mutant adult animals (140 ± 8 pumps/min, Table 3). Thus, the synthetic lethality of gpb-2 goa-1 animals is not due to a feeding defect that is a result of a decrease in the pumping rate. To further analyze the synthetic phenotype of gpb-2 goa-1 larvae, we generated gpb-2 goa-1 double mutant animals rescued with the transgene containing the gpb-2 wild-type sequence. Animals that lose the transgene in the germline give broods that lack functional gpb-2 expression (ZWAAL et al. 1996 ). We found that the progeny of gpb-2 goa-1 double mutant animals that lost the gpb-2 rescuing transgene pumped slowly (163 ± 9 pumps/min, n = 10) and showed pharyngeal defects (data not shown); the pharynx was compressed or shrunken in 80% of the animals. This defect may severely impair feeding, causing the lethality that we observed. However, we cannot exclude other explanations for the lethality. Interestingly, there is evidence that behaviors other than feeding are affected by the gpb-2 goa-1 double mutants and not by each single mutant. Before gpb-2 goa-1 double mutant larvae arrested, these animals moved in an uncoordinated fashion and occasionally we observed that gpb-2 goa-1 larvae rotated half or full turns around their longitudinal axes when elongated.

If combining mutations in gpb-2 and goa-1 causes a synthetic lethal phenotype, then double mutants of gpb-2 and dgk-1, which encode a putative downstream effector of GOA-1 signaling, could also result in a similar synthetic phenotype. Like goa-1, dgk-1 mutants are hyperactive in locomotion and egg laying and also have a starvation phenotype, albeit less severe than goa-1 mutants. We constructed double mutants between gpb-2(pk751) and dgk-1(nu62) and found that animals defective in both gpb-2 and dgk-1 also arrest (>98%). One of the 53 gpb-2; dgk-1 double mutant progeny of gpb-2/+; dgk-1 animals survived to adulthood and produced progeny that arrested as larvae (>98%). As for gpb-2 goa-1 double mutants, the phenotype of the progeny of homozygous gpb-2; dgk-1 animals does not differ from the phenotype of the first generation of gpb-2; dgk-1 mutants, excluding any maternal effects. In addition, the phenotype of gpb-2; dgk-1 double mutants resembles the phenotype of gpb-2 goa-1 mutants, e.g., pharyngeal defects (95% of animals) and locomotory behavior. These genetic data suggest that gpb-2 acts in a parallel pathway with goa-1 and dgk-1.

Surprisingly, when EGL-30 activity was reduced in gpb-2(pk751)goa-1(pk62) double mutant animals by constructing an egl-30(n686)gpb-2(pk751)goa-1(pk62) triple mutant, the lethality of gpb-2 goa-1 double mutants was suppressed. Only 1 of the 169 egl-30 gpb-2 goa-1 triple mutants of egl-30 gpb-2 goa-1/egl-30 ++ heterozygous animals arrested at the larval stage and did not produce offspring. We measured the egg-laying activity, the pharyngeal pumping rate, and the frequency of body bends during locomotion in egl-30(n686)gpb-2(pk751)goa-1(pk62) triple mutants compared to egl-30(n686) single mutants. Like egl-30(n686) animals, the triple mutant laid most eggs at the post-comma stage and became as bloated with retained eggs as the egl-30(n686) single mutant (Table 3). Moreover, the frequency of body bends during locomotion of the triple mutant is not significantly different from egl-30(n686) single mutants (Table 3), although we did observe that egl-30 gpb-2 goa-1 triple mutants have brief periods of rapid forward and backward movement. Because egl-8, which encodes a phospholipase ß (PLCß), is a putative downstream effector of EGL-30, we also generated an egl-8(n488); gpb-2(pk751)goa-1(pk62) triple mutant. Unlike egl-30, egl-8 did not suppress the synthetic lethality of gpb-2(pk751)goa-1(pk62) double mutants, suggesting that EGL-30 activity causes the lethality of gpb-2(pk751)goa-2(pk62) double mutants via an as yet unknown downstream effector. The results of this epistasis analysis are consistent with gpb-2 and goa-1 acting in parallel pathways and upstream of, or parallel to, egl-30.

If egl-30 acts downstream of gpb-2, we would expect egl-30 gpb-2 double mutants to exhibit the phenotypes of egl-30 single mutants. Therefore, we constructed double mutants between gpb-2(pk751) and hypomorphs of egl-30. We found that animals having both mutations in egl-30(ad805), the strongest reduction-of-function allele of egl-30, and gpb-2(pk751) were highly similar to egl-30(ad805) single mutants. egl-30(ad805)gpb-2(pk751) double mutant animals were severely lethargic and laid 100% of their eggs at the post-comma stage (Table 3). However, egg laying was very infrequent, suggesting that gpb-2 may enhance the egl-30 phenotype (Table 3). We also made a double mutant between a loss-of-function gpb-2 mutation and a reduction-of-function egl-30 mutation, n686, and again found that gpb-2(pk751) did not suppress the behavioral defects of egl-30(n686) single mutants; if anything, there was a slight increase in the egl-30 egg-laying defect (Table 3). Again these results suggest that gpb-2 acts upstream of, or parallel to, egl-30 to control egg-laying activity, the locomotion rate, and the pumping rate. Taken together, the genetic data indicate that gpb-2 may regulate both goa-1 (G_o) and egl-30 (G_q) signaling.

GPB-2 interacts with, and is necessary for, both EGL-10 and EAT-16 function:
The mammalian Gß5 subunit interacts with G-subunits (A. J. WATSON et al. 1996 ; ZHANG et al. 1996 ), but Gß5 also has the unique ability to interact with the GGL domain of a subset of RGS proteins (SNOW et al. 1998 , SNOW et al. 1999 ; LEVAY et al. 1999 ). The complete genome of C. elegans contains two GGL-containing RGS proteins, EGL-10 and EAT-16, and two canonical G-protein -subunits, GPC-1 and GPC-2. We examined the possibility that GPB-2, the closest C. elegans homologue of Gß5, could physically interact with the GGL domain of EGL-10 and EAT-16 and the G-subunits, GPC-1 and GPC-2. Yeast two-hybrid analysis showed that GPB-2 bound to the GGL domain of EGL-10 and EAT-16 and the G-protein -subunits, GPC-1 and GPC-2 (Fig 4). We found similar results with GPB-1 (Fig 4). These findings suggest that there is no specificity for GPB-2 and GPB-1 in their ability to bind to EGL-10, EAT-16, GPC-1, or GPC-2 in vitro. However, these results may not reflect the in vivo situation.

View larger version (24K): In this window In a new window Download PPT slide   Figure 4. G-protein ß-subunits interact directly with G-protein -subunits. The cDNA sequences encoding GPB-1 and GPB-2 and GPC-1 and GPC-2, as well as the GGL domains of EGL-10 and EAT-16, were cloned into the two-hybrid prey and bait vector, respectively. Specific two-hybrid interactions were observed for both ß-subunits with all -subunits and -like domains, whereas no interaction was detected with control bait (P53) or prey (SV40 large T) proteins. +, growth on medium lacking histidine and adenine; -, no growth detectable within 10 days on medium lacking histidine and adenine.

Mutations in eat-16 (RGS) result in similar behavioral phenotypes as described for goa-1 (G_o)—starvation, premature egg laying, and hyperactive locomotion (HAJDU-CRONIN et al. 1999 )—but mutations in egl-10 (RGS) result in opposite phenotypes (KOELLE and HORVITZ 1996 ). Genetic analysis showed that EGL-10 is an upstream inhibitor of GOA-1 (G_o) (KOELLE and HORVITZ 1996 ), whereas EAT-16 acts downstream of, or parallel to, GOA-1 (G_o), which in turn inhibits EGL-30 (G_q) activity (HAJDU-CRONIN et al. 1999 ). Given the physical interaction of GPB-2 with EGL-10 and EAT-16 in vitro and the behavioral phenotypes of gpb-2 mutant animals, GPB-2 may be required for the function of either EGL-10 or EAT-16 or both. To test this hypothesis, we analyzed the phenotype of double mutants between gpb-2 and eat-16 or egl-10. We found that the phenotype of double mutants of gpb-2(pk751)eat-16(sy438) and gpb-2(pk751); egl-10(md176) was indistinguishable from the phenotype of gpb-2 mutant animals with respect to the rate of egg laying, pharyngeal pumping, and the frequency of body bends during locomotion (Table 4). Also the phenotype of gpb-2(pk751)eat-16(sy438); egl-10(md176) triple mutant animals was not different from the phenotype of gpb-2 mutants (data not shown). These results suggest that gpb-2 acts downstream of, or parallel to, egl-10 and eat-16 and are also consistent with GPB-2 being needed for both EGL-10 and EAT-16 function.

If it is the case that GPB-2 is required for both EGL-10 and EAT-16 function, then animals defective in both egl-10 and eat-16 should result in a phenotype resembling the gpb-2 mutant phenotype. Indeed, we found that the phenotype of double mutant animals that contain a putative loss-of-function mutation in eat-16 [eat-16(sy438)] and a loss-of-function mutation in egl-10 [egl-10(md176)] resembled the gpb-2 mutant phenotypes with respect to the rate of egg laying, pharyngeal pumping, and the frequency of body bends during locomotion, rather than an eat-16(sy438) or egl-10(md176) phenotype. As shown in Table 4, eat-16(sy438); egl-10(md176) double mutants were wild type with respect to the frequency of body bends during locomotion; however, the egg-laying rate was intermediate to those of gpb-2(pk751) and egl-10(md176) single mutants. Moreover, well-fed adult eat-16(sy438); egl-10(md176) double mutant animals had a similar starved phenotype as gpb-2(pk751) mutants; pharyngeal pumping in eat-16(sy438); egl-10(md176) double mutant animals was reduced (Table 4). Taken together, the genetic and physical interaction of GPB-2 with EGL-10 and EAT-16 is consistent with GPB-2 function being required for both EGL-10 and EAT-16 function and vice versa.

gpb-2 mutant animals are sensitive to the acetylcholinesterase inhibitor aldicarb:
Recent studies showed that the G_o–G_q signaling network regulates acetylcholine release at the C. elegans neuromuscular junction (LACKNER et al. 1999 ; MILLER et al. 1999 ; NURRISH et al. 1999 ). These studies have been facilitated by assaying the response of animals to the drug aldicarb, which is used as a measure of acetylcholine levels at synapses. Aldicarb treatment in wild-type animals causes hypercontraction of body wall muscles and paralysis, as well as inhibition of growth and reproduction. Mutations that decrease acetylcholine release show reduced sensitivity or resistance to aldicarb, whereas mutations that cause elevated levels of acetylcholine at synapses are hypersensitive to aldicarb. For instance, goa-1 (G_o) null mutant animals exhibit hyperactive locomotion and hypersensitivity to aldicarb, while egl-30 (G_q) reduction-of-function mutant animals are lethargic and resistant to aldicarb (MILLER et al. 1999 ).

Since gpb-2 interacts with the goa-1 (G_o) and egl-30 (G_q) signaling pathways, we measured the sensitivity of gpb-2 animals to the acetylcholinesterase inhibitor aldicarb. We observed that gpb-2 mutant animals exhibit an altered response to aldicarb-induced paralysis compared to wild-type animals; gpb-2(pk751) animals are more sensitive to aldicarb (Fig 5A). Similar aldicarb sensitivities of gpb-2(pk751) mutant animals were obtained by measuring the population growth rates (Fig 5B). These results suggest a presynaptic locus of action of gpb-2. Although gpb-2 mutant animals did not exhibit hyperactive locomotion, in the absence of aldicarb they did tend to move backward with exaggerated body flexion. To further investigate whether GPB-2 is required for both EGL-10 and EAT-16 function, we analyzed the response of gpb-2 eat-16 and gpb-2; egl-10 double mutant animals to aldicarb. A putative loss-of-function mutation in eat-16 [eat-16(sy438)] and a loss-of-function mutation in egl-10 [egl-10(md176)] alone showed opposite effects to aldicarb—hypersensitivity and resistance, respectively (Fig 5B). When we crossed gpb-2 into an eat-16 or egl-10 background, we found that the aldicarb sensitivity of gpb-2(pk751)eat-16(sy438) and gpb-2(pk751); egl-10(md176) double mutants was indistinguishable from the gpb-2(pk751) single mutant (data not shown), confirming again that GPB-2 is needed for both EGL-10 and EAT-16 function.

View larger version (26K): In this window In a new window Download PPT slide   Figure 5. Loss-of-function gpb-2 mutant animals cause aldicarb sensitivity. (A) Aldicarb sensitivity in gpb-2(pk751), goa-1(n363), and egl-30(ad805) single mutants is quantified by assaying the time course of the onset of paralysis following acute exposure to aldicarb. In each experiment, 25 animals were placed on aldicarb plates and prodded for 10 min over a 2-hr period to determine if they were able to move. The goa-1(n363) allele is a deletion allele that removes a region containing the goa-1 gene (SEGALAT et al. 1995 ). egl-30(ad805) is the strongest reduction-of-function allele (BRUNDAGE et al. 1996 ). Reducing the function of gpb-2 leads to a significantly increased rate of paralysis compared to N2 (wild type), whereas reduction of goa-1 and egl-30 function results in animals that show an increased and decreased rate of paralysis compared to N2 (wild type), respectively. A two-factor repeated-measure ANOVA comparing groups over time was conducted. A significant effect between gpb-2(pk751) and N2 (wild type) was noted (F = 19, P < 0.001), over a time course of 10 min to 90 min (F = 71.6, P < 0.001) with no significant interaction. (B) Aldicarb sensitivity is quantified by measuring the population growth rates of N2 (wild-type) animals, gpb-2(pk751), eat-16(sy438), or egl-10(md176) single mutants, and eat-16(sy438); egl-10(md176) double mutants on various concentrations of aldicarb. One hundred percent represents the number of progeny produced from a starting population of L1 larvae over a 96-hr period in the absence of aldicarb. Animals that lost both eat-16 and egl-10 function result in aldicarb sensitivities not significantly different from wild type (N2), whereas gpb-2 animals are significantly sensitive to aldicarb. Significant effects between groups over increasing aldicarb concentration were analyzed using a two-factor repeated-measure ANOVA. The following significant effects were obtained: N2 (wild type) from gpb-2(pk751) (F = 12.6, P = 0.005), over an aldicarb concentration range of 10 to 200 µM (F = 50.8, P < 0.001) with no significant interaction. No significant effect between N2 (wild type) and eat-16(sy438); egl-10(md176) was observed (F = 0.006, P = 0.94), with a significant effect over an aldicarb concentration range of 10 to 200 µM (F = 66.8, P < 0.001). Error bars represent the standard error of the mean of triplicate experiments.

Despite the fact that eat-16; egl-10 double mutant animals resemble the phenotype of gpb-2 mutant animals with respect to the rate of egg laying, pharyngeal pumping, and the frequency of body bends during locomotion, we found that eat-16; egl-10 double mutant animals had aldicarb sensitivities that were not significantly different from those of wild type, whereas gpb-2 mutant animals were sensitive (Fig 5B). Interestingly, eat-16; egl-10 double mutants did not exhibit the exaggerated backward flexion seen in gpb-2 mutant animals, suggesting that the hyperflexion and aldicarb sensitivity phenotypes are related. While it is not completely clear whether eat-16 and egl-10 are complete loss-of-function mutations, our data suggest that GPB-2 has functions that are independent of EAT-16 and EGL-10 function.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Two G-protein ß-subunit genes, gpb-1 and gpb-2, are present in the complete C. elegans genome. We show here that, like GPB-1, GPB-2 is widely expressed, even though both have different and specific functions. GPB-2 is most related to mammalian Gß5, which specifically interacts with the GGL domain of a subset of RGS proteins (SNOW et al. 1998 ; LEVAY et al. 1999 ). Our data illustrate a genetic and physical interaction of GPB-2 with the C. elegans GGL-containing RGS proteins, EGL-10 and EAT-16, which functionally regulate the G_o–G_q signaling network.

What is the in vivo function of G-protein ß-subunits?
The first G-protein ß-subunit in C. elegans, GPB-1, is highly similar to the mammalian Gß1-4 subunits (VAN DER VOORN et al. 1990 ; ZWAAL et al. 1996 ), whereas GPB-2 is most related to the mammalian Gß5 subunit. Both GPB-2 and Gß5 differ in amino acid sequence from the other four mammalian Gß-subunits in conserved regions, but particularly in their amino-terminal regions. The Gß1-4 subunits contain glutamine residues in the conserved amino-terminal region that are absent in GPB-2 and Gß5 (WATSON et al. 1994 ). The amino-terminal region of the Gß-subunit is involved in the coiled-coil interaction with the amino-terminal region of the G-subunit (LUPAS et al. 1991 ; SONDEK et al. 1996 ).

Like Gß5, GPB-2 is expressed in neuronal tissues. In addition, GPB-2 is widely expressed in muscle cells. Both GPB-1 and GPB-2 have similar expression patterns and therefore may have redundant function in certain cells. However, on the basis of the phenotypes of gpb-1 and gpb-2 mutants, they seem to have distinct functions. gpb-1 null mutants are embryonic lethal and progeny of maternally provided gpb-1 animals arrest at the first stage of larval development (ZWAAL et al. 1996 ); gpb-2 null mutant animals develop normally to adulthood even though GPB-2 is expressed from the embryo to the adult. These results suggest that GPB-2 cannot compensate for GPB-1, but does not exclude the possibility that GPB-1 can compensate for GPB-2. Mosaic gpb-1 animals exhibit hyperactive locomotion and premature egg laying (ZWAAL et al. 1996 ), whereas loss of gpb-2 results in animals that have delayed egg laying and reduced pharyngeal pumping. Thus, loss of gpb-1 results in different adult behavioral phenotypes compared to loss of gpb-2, suggesting that GPB-1 and GPB-2 have distinct and specific functions. In agreement with this hypothesis, overexpression of GPB-1 results in animals that are lethargic and bloated with eggs (ZWAAL et al. 1996 ), whereas overexpression of GPB-2 has no effect on egg laying, pharyngeal pumping, or the frequency of body bends during locomotion.

A possible explanation for these distinct phenotypic patterns is that overexpression of GPB-1, but not GPB-2, may sequester one or more specific G-subunits. For example, reduction-of-function mutation in egl-30 (G_q) has a phenotype similar to GPB-1 overexpression—lethargic and delayed egg laying (BRUNDAGE et al. 1996 ). The GPB-1 overexpression phenotype may therefore be due to sequestration of EGL-30. Consistent with this hypothesis, we identified a putative gain-of-function allele of egl-30 as a suppressor of the lethargy and egg-laying defective phenotype that results from overexpression of GPB-1 (A. M. VAN DER LINDEN, H. C. KORSWAGEN and R. H. A. PLASTERK, unpublished results). Furthermore, gpb-2—or components that interact with GPB-2—may already be expressed at high levels in a wild-type background, in such a way that extra GPB-2 will not have an effect. Finally, it is possible that differential interactions with G-subunits or G-like subunits account for the activation of different downstream signaling pathways, resulting in different effects on behavior.

Interaction of GPB-2 with EGL-10 and EAT-16:
The genome of C. elegans harbors two canonical G-protein -subunits, GPC-1 and GPC-2 (JANSEN et al. 1999 ), and two GGL-containing RGS proteins, EGL-10 and EAT-16 (KOELLE and HORVITZ 1996 ; HAJDU-CRONIN et al. 1999 ). Like the specific interaction of Gß5 with the -like (GGL) domains of RGS6, 7, 9, and 11 (SNOW et al. 1998 , SNOW et al. 1999 ; LEVAY et al. 1999 ), we show that GPB-2, the closest C. elegans homologue of Gß5, interacts with EGL-10 and EAT-16. Using the yeast two-hybrid interaction assay in vitro, we demonstrate specific interactions between Gß-subunits and G-like subunits, but we do not see any differences between GPB-1 and GPB-2 in this regard. In vivo, we found genetic interactions of GPB-2 with EGL-10 and EAT-16, consistent with GPB-2 function being required for both EGL-10 and EAT-16 function. In gpb-2 eat-16 and gpb-2; egl-10 double mutants, functional GPB-1 is present in the cells that control the observed phenotypes, e.g., egg laying and muscle activity, suggesting that GPB-1 is not functionally redundant with GPB-2 for the phenotypes described. After completion of this study we became aware of two recent results that support these conclusions: first, interactions between GPB-2 and EGL-10 and EAT-16 have been independently observed (CHASE et al. 2001 ), and second, consistent with the observed feeding defects, eat-11 is allelic to gpb-2 and appears to interact with eat-16 and egl-10 (ROBATZEK et al. 2001 ).

Role of GPB-2 in the GOA-1 (G_o)–EGL-30 (G_q) signaling network:
In mammals, it is clear that Gß5-RGS complexes can act as GAPs for G-subunits. GAPs inhibit G activity by catalyzing the exchange of GTP bound to G to GDP. Genetic and biochemical data argue that EAT-16 (RGS) functions as a GAP for EGL-30 (G_q; HAJDU-CRONIN et al. 1999 ), whereas genetic data suggest that EGL-10 (RGS) inhibits GOA-1 (G_o) activity (KOELLE and HORVITZ 1996 ). The results show that GPB-2 interacts with both EGL-10 and EAT-16 to regulate GOA-1 and EGL-30 activity, respectively. This conclusion is emphasized by the finding that loss of gpb-2 does not result in phenotypes similar to either eat-16 or egl-10. Only eat-16; egl-10 double mutant animals resemble a gpb-2 mutant phenotype, e.g., the rate of egg laying, pharyngeal pumping, or the frequency of body bends during locomotion. In line with data from studies on Gß5, a possible role of GPB-2 may thus be to modulate the specificity and/or activity of EGL-10 and EAT-16 for GOA-1 and EGL-30, respectively (WITHEROW et al. 2000 ). It is, however, also likely that GPB-2 function is not completely dependent on EGL-10 and EAT-16. For example, animals defective in gpb-2 exhibit exaggerated backward motion and aldicarb sensitivity phenotypes, whereas eat-16; egl-10 double mutant animals do not.

Previous genetic data indicated that GOA-1 (G_o) antagonize EGL-30 (G_q) signaling directly or indirectly through EAT-16 and/or DGK-1 (DAG-kinase; HAJDU-CRONIN et al. 1999 ; MILLER et al. 1999 ). In this model both EAT-16 and DGK-1 function downstream of, or parallel to, GOA-1 (G_o) and act synergistically to negatively regulate EGL-30 (G_q) signaling (Fig 6). This synergy was proposed because animals defective in both EAT-16 and DGK-1 function are synthetic lethal, and this lethality could be suppressed by reducing EGL-30 activity (HAJDU-CRONIN et al. 1999 ). Thus, downstream signaling components of EGL-30 may cause the lethality of the eat-16; dgk-1 double mutant. In support of this model, we found similar genetic interactions with GPB-2, GOA-1, and DGK-1. gpb-2; dgk-1 and gpb-2 goa-1 double mutants resulted in a synthetic lethal phenotype. Moreover, the synthetic lethal phenotype of gpb-2 goa-1 double mutants was suppressed when EGL-30 activity was reduced. No suppression was observed when the activity of EGL-8—a potential downstream effector of EGL-30—was reduced. This result is in agreement with the model that EGL-8 does not mediate all the effects of EGL-30 (MILLER et al. 1999 ), but that there should be other, so far unknown, downstream effectors. An interesting finding is that, unlike gpb-2 goa-1 double mutants, goa-1 eat-16 double mutants are not lethal (HAJDU-CRONIN et al. 1999 ), suggesting that gpb-2 does not interact with only eat-16 but also with another unknown pathway. A potential pathway is the dgk-1 pathway, since dgk-1; gpb-2 double mutants also result in a synthetic lethal phenotype.

View larger version (31K): In this window In a new window Download PPT slide   Figure 6. A model for GPB-2 to regulate behavior (e.g., locomotion rate, egg-laying activity) by interacting with both EGL-10 and EAT-16, which in turn inhibit GOA-1 (Go) and EGL-30 (Gq) activity, respectively. This model is based on data presented here and on data from KOELLE and HORVITZ 1996 , HAJDU-CRONIN et al. (1999), NURRISH et al. 1999 , LACKNER et al. 1999 , and MILLER et al. 1999 . GPB-2—most closely related to mammalian Gß5—is required for the function of the GGL-containing RGS proteins EGL-10 and EAT-16. EGL-30 activates EGL-8—encoding a phospholipase Cß (PLCß)—which cleaves phosphatidylinositol 4,5-biphosphate (PIP2) into diacylglycerol (DAG) and inositol-1,4,5-triphosphate (BERRIDGE 1984 ). DGK-1—encoding a diacylglycerol kinase (DAG-kinase)—may be activated by GOA-1 to reduce DAG levels. The locomotion rate and egg-laying activity is indirectly regulated by levels of DAG. GOA-1 may inhibit EGL-30 (Gq) activity directly through downstream effectors such as EAT-16 and/or DGK-1, indirectly via the GGL domain of EAT-16 as proposed by GUAN and HALL (1999), or act in parallel to EGL-30.

Recent literature has designed a number of interactions in the GOA-1 (G_o)–EGL-30 (G_q) network that functions in locomotion and egg laying in C. elegans (HAJDU-CRONIN et al. 1999 ; LACKNER et al. 1999 ; MILLER et al. 1999 ; NURRISH et al. 1999 ). In this signaling network, GOA-1 (G_o) and EGL-30 (G_q) are assumed to couple to G-protein-coupled receptors and are shown to be negatively regulated by the GGL-containing RGS proteins EGL-10 and EAT-16, respectively (Fig 6). The rate of locomotion and the activity of egg laying are stimulated by activation of EGL-30 via EGL-8 (phospholipase Cß), whereas activation of GOA-1 inhibits both behaviors, probably via DGK-1 (DAG-kinase). The data presented here support a model in which GPB-2 plays an important role in the GOA-1–EGL-30 signaling network by interacting with both GGL-containing RGS proteins EAT-16 and EGL-10 (Fig 6). In view of this model, two questions can be asked. First, does GPB-2 inhibit both EGL-30 and GOA-1 and, second, do our data distinguish between GOA-1 acting in parallel to EGL-30 or GOA-1 acting upstream of EGL-30? With respect to the first question, our epistatic analysis of gpb-2 with goa-1 and egl-30 appears consistent with two hypotheses: either GPB-2