Identification of Gnr1p, a negative regulator of G [alpha] signalling in Schizosaccharomyces pombe, and its complementation by human G [beta] subunits

Fungal Genetics and Biology 43 (2006) 840–851

www.elsevier.com/locate/yfgbi

IdentiWcation of Gnr1p, a negative regulator of G� signalling in Schizosaccharomyces pombe, and its complementation

by human G� subunits

Alan Goddard a,¤, Graham Ladds b, Rachel Forfar a, John Davey a

a Department of Biological Sciences, University of Warwick, Coventry CV4 7AL, UKb Division of Clinical Sciences, Warwick Medical School, University of Warwick, Coventry CV4 7AL, UK

Received 11 May 2006; accepted 5 June 2006Available online 1 August 2006

Abstract

G protein-coupled receptors (GPCRs) are involved in the response of eukaryotic cells to a wide variety of stimuli, traditionally medi-ating their eVects through heterotrimeric G proteins comprised of G�, G� and G� subunits. The Wssion yeast Schizosaccharomyces pombeis an established tool for GPCR research, possessing two G�-dependent signalling cascades. A complete G��� complex has been charac-terised for the glucose-sensing pathway, but only the G� subunit, Gpa1p, has been identiWed in the pheromone-response pathway. Here,we report the use of the yeast two-hybrid system to identify a novel protein, Gnr1p, which interacts with Gpa1p. Gnr1p is predicted tocontain seven WD repeats and to adopt a structure similar to typical G� subunits. Disruption and overexpression studies reveal thatGnr1p negatively regulates the pheromone-response pathway but is not required for signalling. Human G� subunits complement the lossof Gnr1p, functioning as negative regulators of G� signalling in Wssion yeast.© 2006 Elsevier Inc. All rights reserved.

Index descriptors: Schizosaccharomyces pombe; Gnr1p; Gpa1p; G� subunit; WD repeat protein; Pheromone

1. Introduction In addition to G�� heterodimers and GPCRs, G� sub-

G protein-coupled receptors (GPCRs) are a diverse fam-ily of integral membrane proteins involved in the responseof eukaryotic cells to a wide variety of stimuli. The recep-tors transmit their signals via heterotrimeric G proteins,comprised of G�, G� and G� subunits. In an unstimulatedstate, the GPCR is associated with the G protein complex,in which the G� subunit is bound to GDP. Stimulation ofthe receptor results in replacement of GDP with GTP. Thisnucleotide exchange is accompanied by a conformationalchange in the G� subunit, and its dissociation from thereceptor and the G�� dimer. The G� and G�� dimer arethen free to activate a wide variety of eVector molecules,depending on the cellular environment.

* Corresponding author. Fax: +44 2476 523701.E-mail address: [email protected] (A. Goddard).

1087-1845/$ - see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.fgb.2006.06.005

units interact with a variety of proteins which mediate andregulate their function. They transmit signals to appropri-ate intracellular eVectors, and their activity is modulated bydirect interaction with both negative regulators (RGSs –regulators of G protein signalling; Tesmer et al., 1997) andactivators (AGSs – activators of G protein signalling; Blu-mer et al., 2005). Recently, it has been observed that theGPA2 G� subunit in the budding yeast Saccharomycescerevisiae interacts with two kelch-repeat proteins, GPB1and GPB2, which are believed to act as structural G� mim-ics. These proteins, along with a putative G� mimic, GPG1,provide some, but not all, features of a classical G�� dimer(Harashima and Heitman, 2002). It has been suggested thatG� subunits in other systems may interact with kelch-likeproteins (Gettemans et al., 2003).

The Wssion yeast Schizosaccharomyces pombe has beenestablished as a powerful tool for the study of GPCR sig-nalling cascades (Ladds and Davey, 2004). To date, two G�

A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851 841

proteins have been identiWed in this organism, responsiblefor the regulation of two independent pathways (HoVman,2005). The gpa2 ORF encodes a G� subunit which acts withthe Git3p GPCR and the Git5p–Git11p G�� dimer (Wel-ton and HoVman, 2000; Landry and HoVman, 2001) in theglucose-sensing pathway (Isshiki et al., 1992). Gpa2pdirectly activates adenylate cyclase (Ivey and HoVman,2005) and, consequently, disruption of gpa2 in homothallicstrains results in constitutive mating and sporulation, elimi-nates the glucose-induced cAMP response, and reducesbasal cAMP levels (HoVman, 2005). Mutations which dis-rupt either Git3p or the Git5p–Git11p dimer have similareVects to those in Gpa2p.

The second G� subunit, Gpa1p, is involved in thepheromone-response pathway (Obara et al., 1991; Davey,1998). Sz. pombe exists as a haploid organism in one oftwo mating types, P (Plus) or M (Minus). However, underconditions of nutrient starvation, cells undergo a matingresponse and conjugate to form a diploid. This responseis mediated by the reciprocal exchange of small peptidepheromones. M-cells release M-factor and respond to P-factor via the Mam2p receptor (Davey, 1992; Kitamuraand Shimoda, 1991). Conversely, P-cells secrete P-factorand respond to M-factor via the Map3p receptor (Imaiand Yamamoto, 1994; Tanaka et al., 1993). Both of thesereceptors are GPCRs which transduce their signals viathe Gpa1p G� subunit and a MAP kinase cascade. Todate, the direct link between Gpa1p and this cascade hasnot been established, but G� stimulation ultimatelyresults in activation of the transcription factor Ste11p(KjaerulV et al., 2005) and upregulation of genes requiredfor mating.

A central question in this pathway has concerned theexistence, or otherwise, of a G�� dimer. A previous study(Kim et al., 1996) proposed the product of the gpb1 gene asthe G� subunit in the pheromone-response pathway. How-ever, the eVects observed by these researchers have sincebeen demonstrated to be due to Gpb1p (now termed Git5p)acting in conjunction with Gpa2p (Landry et al., 2000). Dis-ruption of Gpa1p results in a sterile phenotype, yet screensfor sterile mutants have failed to reveal the identity of acorresponding G� or G� subunit. It is, however, possiblethat Gpa1p interacts with atypical binding partners in amanner comparable to Sc. cerevisiae GPA2 and kelch-repeat proteins.

The only protein which has been demonstrated to inter-act directly (Chung et al., 2001) and functionally (Laddset al., 2005) with Gpa1p is Mam2p, the P-factor receptor. It

is also likely that the endogenous RGS protein, Rgs1p(Watson et al., 1999), achieves its GTPase activation viadirect interaction with Gpa1p, although this has not beendemonstrated in vivo. Neither a G� subunit nor a directdownstream eVector has been identiWed. In an attempt toidentify G�-binding partners, we performed a yeast two-hybrid screen using Gpa1p as the bait protein. One of theinteractants identiWed encodes a potential WD repeat pro-tein (so named due to repeating tryptophan and aspartateresidues), now termed Gnr1p (G protein negative regulator1). WD repeats are found in a number of classes of proteins,including G� subunits. Disruption of the gnr1 ORF andoverexpression studies demonstrate that Gnr1p is a nega-tive regulator of Gpa1p but is not required for signalling.Data suggest that Gnr1p achieves this modulation by act-ing as a structural mimic of a G� subunit. Further, we dem-onstrate that mutations in gnr1 can be complemented byexpression of human G� subunits, which also act as nega-tive regulators of the pheromone-response pathway in Sz.pombe.

2. Materials and methods

2.1. Strains, reagents and general methods

The yeast strains used in this study are listed in Table 1.General Sz. pombe procedures were performed as describedpreviously (Davey et al., 1995; Ladds et al., 1996), usinglithium acetate for the transformation of yeast. Culturemedia used were yeast extract (YE; for routine cell growth),deWned minimal medium (DMM; for all Sz. pombe assays)and amino acid (AA) medium for auxotrophic selection(Davey et al., 1995). Cell concentrations were determinedusing a Coulter Channelyser (Beckman Coulter, Luton,UK). DNA manipulations were performed by standardmethods. Oligonucleotides were synthesised by InvitrogenLtd. (Paisley, Scotland, UK). AmpliWcation by the polymer-ase chain reaction (PCR) used Pwo DNA polymerase (fromPyrococcus woesei) according to the supplier’s instructions(Boehringer–Mannheim Biochemicals, Lewes, East Sussex,UK). All constructs generated by PCR were conWrmed bysequencing.

2.2. Yeast two-hybrid screening

The AH109 and Y187 Sc. cerevisiae strains, pGADT7-SV40, pGBKT7-p53 and pGBKT7-Lamin-C constructsused in this study were supplied with the Matchmaker yeast

Table 1Schizosaccharomyces pombe strains used in this study. The term sxa2 > lacZ is used to indicate a construct in which the lacZ open reading frame is placedunder the transcriptional control of the sxa2 promoter

Strain Genotype

JY546 mat1-M, �mat2/3::LEU2¡, leu1¡, ura4-D18, cyr1-D51, sxa2 > lacZ (Didmon et al., 2002)JY1314 mat1-M, �mat2/3::LEU2¡, leu1¡, ura4-D18, cyr1-D51, sxa2 > lacZ, gnr1::ura4+

JY1317 mat1-M, �mat2/3::LEU2¡, leu1¡, ura4-D18, cyr1-D51, sxa2 > lacZ, git5::KanR

JY1319 mat1-M, �mat2/3::LEU2¡, leu1¡, ura4-D18, cyr1-D51, sxa2 > mel1, gnr1::ura4+, git5::KanR

842 A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851

two-hybrid kit and all Sc. cerevisiae-based methods wereperformed as detailed by the manufacturer (BD Biosci-ences–Clontech, Oxford, UK; kit version PT13529-1).Screening of the Sz. pombe cDNA library was performedby mating complementing strains as described by the man-ufacturer. All other two-hybrid assays were performed fol-lowing sequential transformations of AH109 withpGBKT7-based plasmids and then pGADT7-based plas-mids. Co-transformants were maintained as single colonieson MM (a deWned minimal medium that lacks leucine andtryptophan) prior to screening for interactions as describedpreviously (Hiskens et al., 2005).

2.3. Preparing the pGADT7 and pGBKT7 expression constructs

The Sz. pombe gpa1 open reading frame (ORF) wasampliWed from genomic DNA using sense oligonucleotideJO2065 (ggatccgcATGGGATGCATGTCGAGT; BamHIsite underlined, ATG initiation codon in bold, non-homol-ogous residues in lower case) and antisense oligonucleotideJO1973 (ggggatccgatCTAAAACATAA; BamHI siteunderlined, stop anticodon in bold). This PCR product wasdigested with BamHI and cloned into BamHI digestedpGBKT7.

The full-length gnr1 ORF was ampliWed from genomicDNA using sense oligonucleotide JO2125 (gaattcATGGATAATTGTGTAAACTCC; EcoRI site underlined,ATG initiation codon in bold) and antisense oligonucleo-tide JO2126 (ggatccTCAATCCATCCAAACCTG; BamHIsite underlined, stop anticodon in bold). This PCR productwas digested with EcoRI and BamHI and cloned intoEcoRI/BamHI digested pGADT7. The git5 ORF wasampliWed from genomic DNA using sense oligonucleotideJO2170 (gaattcATGGATTCTGGGTCAAGAG; EcoRIsite underlined, ATG initiation codon in bold) and anti-sense oligonucleotide JO2171 (gaattcTTACCCTGACGAAGACCA; EcoRI site underlined, stop anticodon in bold).This PCR product was digested with EcoRI and clonedinto the EcoRI site of pGADT7. All human GNB subunitswere cloned into pGADT7 as described in Hiskens et al.(2005).

The git11 ORF was ampliWed from clone pSL25 (Landryand HoVman, 2001) using sense oligonucleotide JO2167(gaattcATGGAAACAGAGGCTTTA; EcoRI site under-lined, ATG initiation codon in bold) and antisense oligonu-cleotide JO2168 (ggggggatccTTAGGAAATAGTACAGCATTTGG; BamHI site underlined, stop anticodon in bold).This PCR product was digested with EcoRI and BamHIand ligated with EcoRI/BamHI digested pGBKT7. TheGNG4 ORF was ampliWed from a cDNA clone from Guth-rie cDNA Resource Centre (www.cdna.org) using sense oli-gonucleotide JO2164 (gaattcATGAAAGAGGGCATGTCT; EcoRI site underlined, ATG initiation codon in bold)and antisense oligonucleotide JO2165 (ggggggatccTTAGAGAATGGTACAAAAG; BamHI site underlined, stopanticodon in bold). This PCR product was digested with

EcoRI and BamHI and ligated with EcoRI/BamHI digestedpGBKT7.

2.4. Preparing the pREP expression constructs

The pREP series of Sz. pombe vectors allows expressionof genes under the control of the thiamine-repressible nmt1promoter (Maundrell, 1993). The gnr1 ORF was ampliWedfrom genomic DNA using sense oligonucleotide JO2100(atcATGGATAATTGTGTAAACTCC; ATG initiationcodon in bold) and antisense oligonucleotide JO2101(ggggatccTCAATCCATCCAAACCTGTAAAG; BamHIsite underlined, stop anticodon in bold). The PCR productwas digested with BamHI and cloned into the EcoRV andBamHI sites of a modiWed pREP3x vector in which aunique EcoRV site has been introduced between the XhoIand BamHI sites. This generated pREP3x-Gnr1.

All human G� subunits were cloned into the modiWedpREP3x vector in an identical manner. The GNB1 ORFwas ampliWed using sense oligonucleotide JO2134 (ATGAGTGAGCTTGACCAGTT; ATG initiation codon inbold) and antisense oligonucleotide JO1733 (ggggatccTTAGTTCCAGATCTTGAG; BamHI site underlined, stopanticodon in bold). The GNB2 ORF was ampliWed usingsense oligonucleotide JO2135 (ATGAGTGAGCTGGAGCAACT; ATG initiation codon in bold) and antisense oli-gonucleotide JO1735 (ggggatccTTAGTTCCAGATCTTGAGGAAG; BamHI site underlined, stop anticodon inbold). The GNB3 ORF was ampliWed using sense oligonu-cleotide JO2136 (ATGGGGGAGATGGAGCAAC; ATGinitiation codon in bold) and antisense oligonucleotideJO1737 (ttggatccTCAGTTCCAGATTTTGAGGAAG;BamHI site underlined, stop anticodon in bold). The GNB4ORF was ampliWed using sense oligonucleotide JO2137(ATGAGCGAACTGGAACAGTTG; ATG initiationcodon in bold) and antisense oligonucleotide JO1739(ggggatccTTAATTCCAGATTCTAAGAAAACTG; BamHI site underlined, stop anticodon in bold). The GNB5ORF was ampliWed using sense oligonucleotide JO2138(ATGGCAACCGAGGGGCTG; ATG initiation codon inbold) and antisense oligonucleotide JO1741 (ggggatccTTAGGCCCAGACTCTGAG; BamHI site underlined, stopanticodon in bold). In each case, the PCR product wasdigested with BamHI and cloned into the EcoRV andBamHI sites of the modiWed pREP3x vector. The resultingconstructs are pREP3x-GNB1, pREP3x-GNB2, pREP3x-GNB3, pREP3x-GNB4 and pREP3x-GNB5.

The git5 ORF was ampliWed from genomic DNA usingsense oligonucleotide JO2098 (atcATGGATTCTGGGTCAAGAG; ATG initiation codon in bold) and antisense oli-gonucleotide JO2099 (aaagatctTTACCCTGACGAAGACCAGA; stop anticodon in bold, BglII site underlined). ThisPCR product was digested with BglII and cloned into theEcoRV and BamHI sites of the modiWed pREP3x vector tocreate pREP3x-Git5.

The Sz. pombe SPAC343.04 ORF (encoding proteinNP_593424, the homologue of the human WDR26 protein)

A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851 843

was ampliWed from genomic DNA using sense oligonucleo-tide JO2071 (atcATGGCTTTGGATGAAAAATTC; ATGinitiation codon in bold) and antisense oligonucleotideJO2072 (ggggatccTTATTGTCGACGTGGATTATC;BamHI site underlined, stop anticodon in bold). This PCRproduct was digested with BamHI and cloned into theEcoRV and BamHI sites of the modiWed pREP3x vector.Introns were removed from the ORF by a series of inversePCRs to create pREP3x-NP_593424.

2.5. Disruption of endogenous Sz. pombe genes

The gnr1 locus was ampliWed from Sz. pombe genomicDNA using sense oligonucleotide JO2121(CTGCTACAATTTCGGAAG; position ¡645 to ¡627relative to gnr1 ATG, half PvuII site underlined) and anti-sense oligonucleotide JO2122 (CTGTCTCGGTTCACTCTC; position 2137 to 2154 relative to gnr1 ATG, half PvuIIsite underlined). This PCR product was cloned into pKS(+)Bluescript (Stratagene) digested with PvuII to createJD2411. The Sz. pombe ura4+ cassette was ampliWed usingsense oligonucleotide JO1049 (CTGGATCCACCATGTAGCTACAAATCC) and antisense oligonucleotideJO1050 (CTGGATCCACCATGTAGTGATATTGAC).This product was cloned into the unique EcoRV site inJD2411 within the gnr1 ORF to create gnr1::ura4+

(JD2538). The strain JY546 (sxa2 > lacZ) was transformedwith the PvuII fragment from JD2538 and integration ofthe ura4 cassette selected for by growth on medium lackinguracil; the resultant strain is JY1314 (sxa2 > lacZ; �gnr1).

The git5 ORF and surrounding regions were ampliWedfrom a clone of git5 in pREP3x using sense oligonucleotideJORep1 (ATCCGATTGTCATTCGGC) and antisense oli-gonucleotide JORep2 (GCAGCTTGAATGGGCTTCC).These oligonucleotides are complementary to the pREP3xvector. This PCR product was cloned into PvuII digestedpKS(+) Bluescript to create JD2506. The kanamycin-resis-tance cassette was liberated from pFA6a-kanMX6 byPvuII/EcoRV digest and cloned into the unique PvuII sitewithin the git5 ORF in JD2506 to generate the git5::KanR

construct (JD2597). The strains JY546 and JY1314 weretransformed with the XhoI/XmnI fragment from JD2597and grown on YE medium for 18 h at 30 °C. Cells were rep-lica plated to YE containing geneticin at 100 �g/ml andgrown as before until colonies had developed. The resultantstrains are JY1317 (gnr1+, �git5) and JY1319 (�gnr1,�git5). All disruptions were conWrmed via PCR screeningand Southern blot analysis.

2.6. Assay of �-galactosidase activity

Assays were performed using a method modiWed fromDohlman et al. (1995) (Didmon et al., 2002; Ladds et al.,2003). Sz. pombe cells were cultured to a density of»5£ 105 cells ml¡1 in DMM and 500 �l aliquots transferredto 2 ml Safe-Lock tubes (Eppendorf, Hamburg, Germany)containing 5 �l of the appropriate ligand (in HPLC-grade

methanol). Tubes were incubated at 29 °C for 16 h on arotating wheel, and 50�l transferred to 750 �l Z-buVer con-taining 2.25 mM o-nitrophenyl-�-D-galactopyranoside(ONPG). Reactions were stopped after 90 min by adding200 �l of 2 M Na2CO3 and �-galactosidase activity calcu-lated as optical density at 420 nm (OD420) per 106 cells(determined using a Coulter Channelyser).

3. Results

3.1. Gpa1p binds to Gnr1p in a two-hybrid screen

In an attempt to identify additional components of theSz. pombe pheromone-response pathway, a yeast two-hybrid screen was performed in Sc. cerevisiae using Gpa1pto screen a Sz. pombe cDNA library. Screening »60,000colonies identiWed a number of interactants, including aclone containing the majority of the ORF SPCC1020.09,which encodes an uncharacterised protein (CAA18997).This protein is predicted to contain WD repeats that arecharacteristic of various proteins involved in macromolecu-lar assemblies, including G� subunits (Clapham and Neer,1997). The ORF has been demonstrated to be non-essential(Decottignies et al., 2003), but no functional studies havebeen reported. In accordance with data presented here, thisORF is termed gnr1 (G protein negative regulator 1) andthe corresponding protein Gnr1p.

To conWrm interaction between full-length Gpa1p andGnr1p, the proteins were expressed from two complement-ing plasmids (pGBKT7-Gpa1 and pGADT7-Gnr1) and co-transformants spotted onto selective plates (Fig. 1). Growthon MM (a deWned minimal medium lacking leucine andtryptophan) conWrmed the presence of the two plasmids.Interaction between Gpa1p and Gnr1p resulted in theexpression of the HIS3, ADE2 and MEL1 reporter genes,and the formation of blue colonies on MM-His-Ade (MMalso lacking histidine and adenine) + X-�-gal (5-bromo-4-chloro-3-indolyl-�-D-galactoside). A positive control wasprovided by strains expressing SV40-T antigen and p53,proteins that interact strongly (Li and Fields, 1993). In con-trast, SV40-T antigen does not interact with Lamin-C andco-transformants were unable to grow on MM-His-Ade + X-�-gal plates. Neither Gpa1p nor Gnr1p were ableto form colonies on MM-His-Ade + X-�-gal when co-expressed with Lamin-C.

A previous study (Kim et al., 1996) proposed a role forGpb1p as a G� subunit for Gpa1p. However, this has sub-sequently been shown to be incorrect (Landry et al., 2000)and it has been demonstrated that this product (nowreferred to as Git5p) forms a heterotrimeric complex withthe G� subunit Git11p and the G� subunit Gpa2p, acting inthe nutrient-sensing pathway (Landry and HoVman, 2001).To investigate if Git5p also interacts with Gpa1p, Git5pwas co-expressed from the pGADT7 vector with pGBKT7-Gpa1p (Fig. 1). Although the resultant strain was capableof growth on MM, conWrming the presence of the two plas-mids, it was unable to form colonies on MM-His-Ade + X-

844 A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851

�-gal, indicating that Gpa1p and Git5p do not interact inthis system.

To provide a more quantitative measurement of the inter-actions, the various co-transformants were analysed in a liq-uid-based �-galactosidase assay for expression of the lacZreporter gene (Fig. 1). Under the conditions of the assay, thestrong interactants SV40-T and p53 produced »10 �-galac-tosidase units, while non-interacting proteins such as SV40-Tand Lamin-C produced »0.5 U. Co-transformants express-ing Gpa1p and Gnr1p produced »7U, indicating a relativelystrong interaction. All other combinations produced no morethan 0.7 U, indicating no interaction.

3.2. Sequence analysis of Gnr1p

Sequence analysis and electronic annotation suggestedthat the product of the SPCC1020.09 ORF contains WD

Fig. 1. Gpa1p interacts with Gnr1p in the yeast two-hybrid system.Screening a Sz. pombe cDNA library in the Sc. cerevisiae yeast two-hybridsystem using Gpa1p as a bait protein identiWed a number of interactants,including the product of the SPCC1020.09 ORF (now termed Gnr1p).Yeast transformants containing pGBKT7-Gpa1 and pGADT7-Gnr1grew on MM (a deWned minimal medium lacking leucine and tryptophan)conWrming the presence of the two plasmids, and on MM-His-Ade (MMalso lacking histidine and adenine) + X-�-gal (5-bromo-4-chloro-3-indo-lyl-�-D-galactoside) conWrming an interaction between the two proteins,resulting in the expression of the HIS3, ADE2 and MEL1 reporter genesand the formation of blue colonies. Co-transformants expressing Gpa1pand Git5p failed to grow on MM-His-Ade + X-�-gal, suggesting that thetwo proteins do not interact. Likewise, Gpa1p and Gnr1p failed to formcolonies on MM-His-Ade + X-�-gal when co-expressed with Lamin-C.Control strains expressing the strongly interacting SV40-T antigen andp53 or the non-interacting SV40-T and Lamin-C are included for compar-ison. The various co-transformants were also analysed in a liquid-based�-galactosidase assay for expression of the lacZ reporter gene, anotherindication of the interaction between the two proteins of interest. Underthe conditions used, a strong interaction between two proteins (such asSV40-T antigen and p53) generates »10 U of �-galactosidase activity. Theresults are presented as mean of triplicate determinations § SD.

SV

40 &

p53

SV

40 &

Lam

in-C

Gnr

1p &

Lam

in-C

Gpa

1p &

Git5

p

Gpa

1p &

Gnr

1p

MM

-X+edA-siH-MM α lag-

01

8

6

4

2

0

β-ga

lact

osid

ase

units

Gpa

1p &

Lam

in-C

repeats that are characteristic of a number of protein fami-lies, including G� subunits. To more closely examine thepredicted structure of this protein, THREADER (Joneset al., 1999) was utilised to determine the closest structuralalignments. This identiWed GNB1 from Bos taurus, a com-ponent of the G�� dimer that has been demonstrated tocontain seven WD repeats forming a characteristic �-pro-peller structure (Sondek et al., 1996). It was therefore possi-ble that Gnr1p represented a G�-like subunit in the Gpa1ppathway. Gnr1p shows »17% identity and »24% similarity(either identical or conserved amino acid changes) withGit5p and »12% identity and »18% similarity with B. tau-rus GNB1 (Git5p and GNB1 share »41% identity and»46% similarity; Fig. 2). The seven predicted WD repeatsin these proteins are indicated in Fig. 2.

Gnr1p is predicted to contain a »30 amino acid coiled-coilN-terminal region prior to the Wrst WD repeat, consistentwith members of the G� family. This region is important inthe G�–G� interaction (Wall et al., 1995; Lambright et al.,1996; Pellegrino et al., 1997). It has been observed thatGit5p lacks this N-terminal extension (Landry et al., 2000).Unusually, Gnr1p is also predicted to possess a C-terminalextension of »70 amino acids after the Wnal WD repeat,containing a highly charged region (residue 371 to residue385). This extension is not found in either GNB1 or Git5pand appears to be unique to Gnr1p.

3.3. Disruption of gnr1 confers an increased response to pheromone stimulation

To investigate if Gnr1p is a regulator of Gpa1p, a previ-ously described Sz. pombe strain was utilised in which sig-nalling through the pheromone-response pathway can bequantitated via the reporter enzyme �-galactosidase (Did-mon et al., 2002). BrieXy, transcription of the Escherichiacoli lacZ gene is linked to the promoter of the pheromone-responsive gene sxa2. Sxa2p is a carboxypeptidase which isonly expressed after pheromone stimulation (Imai andYamamoto, 1992; Ladds et al., 1996). Replacing the sxa2ORF with lacZ results in a reporter strain that produces�-galactosidase in response to pheromone (Didmon et al.,2002).

The gnr1 ORF was disrupted in the sxa2 > lacZ reporterstrain, and the gnr1+ (JY546) and �gnr1 (JY1314) strainsassayed for �-galactosidase activity at various levels ofpheromone stimulation. The responses seen for strainseither containing or lacking the pREP3x plasmid wereidentical, hence only those in the presence of pREP3x areillustrated (Fig. 3). A low level of �-galactosidase produc-tion in the absence of pheromone was observed for eachstrain, probably due to spontaneous Gpa1p activation andpheromone-independent transcription from the sxa2 pro-moter. The gnr1+ strain exhibited the characteristicresponse to pheromone (Didmon et al., 2002); an increasein �-galactosidase activity was Wrst detectable at 10¡7 Mpheromone and increased to »21 U at 10¡6 M pheromone.The strain in which gnr1 had been disrupted showed an

A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851 845

increased response to pheromone. �-galactosidase activitywas Wrst elevated at 3£10¡8 M pheromone and increasedto »35 U at 10¡6 M pheromone. A concomitant increasewas observed at intermediate pheromone concentrations.Given the structural and interaction data, this suggests thatGnr1p is a negative regulator of Gpa1p signalling, acting asa structural mimic of a G� subunit and directly associatingwith Gpa1p. Other than the eVect on signalling, there wereno observable phenotypic diVerences between gnr1+ and�gnr1 strains. The reporter strains described lack sxa2 andare sterile. However, disruption of gnr1 in mating-compe-tent cells had no eVect on the mating response.

3.4. Expression of gnr1 reduces the pheromone response

We next investigated the eVects of plasmid-borne Gnr1pexpression. To enable such studies, the gnr1 ORF wascloned into the pREP3x vector to create pREP3x-Gnr1,allowing expression under the control of the thiamine-repressible nmt1 promoter (Maundrell, 1993). pREP3x andpREP3x-Gnr1 were introduced into JY546 (gnr1+) andJY1314 (�gnr1), and strains assayed for �-galactosidaseproduction in response to pheromone (Fig. 3).

In a strain containing a wild-type copy (JY546),expression of gnr1 from pREP3x resulted in decreased �-

galactosidase production at all levels of pheromone stim-ulation. The level of activity was reduced from »21 to»5 U at 10¡6 M pheromone. Expression of Gnr1p inJY1314 led to a similar decrease in the level of �-galacto-sidase activity to that observed in JY546, from »35 to»10 U at 10¡6 M pheromone, indicating complementa-tion of the �gnr1 phenotype. The reduction at high levelsof pheromone stimulation was less than observed in thegnr1+ strain, possibly due to the lack of endogenousGnr1p.

3.5. EVects of Git5p on the pheromone response

As the predicted structure of Gnr1p was similar to thatof Git5p (Fig. 2), and both proteins interact with G� sub-units, it was investigated whether Git5p also functions as anegative regulator of Gpa1p in vivo. The git5 ORF was dis-rupted in gnr1+ (JY546) and �gnr1 (JY1314) strains. Astrain lacking git5 (JY1317; gnr1+, �git5+) responded topheromone in a manner comparable to the gnr1+ wild-typestrain (JY546). Additionally, a strain lacking both git5 andgnr1 (JY1319) responded similarly to the �gnr1 strain(JY1314), with no further increase in responsiveness. Theseresults suggest that Git5p does not normally inXuence thepheromone-response pathway.

Fig. 2. Sequence comparison of Gnr1p, Git5p and GNB1. Sequences of Sz. pombe Gnr1p, Git5p and B. taurus GNB1 were aligned using MultAlin version5.4.1. (Corpet, 1988), with some manual adjustment. Symbol comparison table blosum62, gap weight 12 and gap length weight 2 were used. Gaps intro-duced to maximise the alignment are indicated by ¡. Residues shaded in black are identical and those shaded in grey are conservative changes. The sevenpredicted WD repeats are underlined.

846 A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851

Although disruption of git5 had no eVect upon sensitiv-ity to pheromone stimulation, it was possible that Git5pwould be able to act as a structural mimic of Gnr1p. Toexamine if Gnr1p could be functionally replaced by Git5p,the corresponding ORF was cloned into the pREP3xexpression vector. pREP3x and pREP3x-Git5 were intro-duced into both �gnr1 (JY1314; Fig. 4A) and gnr1+

reporter strains (JY546; Fig. 4B) and �-galactosidase activ-ity assayed at various levels of pheromone stimulation. Sig-nalling in JY1314 was reduced from »35 to »10 U at10¡6 M pheromone upon overexpression of Git5p. Overex-pression in JY546 also led to a decrease in signalling, from»22 to »6 U in the presence of 10¡6 M pheromone. Thissuggests that Git5p is capable of functionally replacingGnr1p and attenuating the pheromone-response pathwaywhen overexpressed.

3.6. Human G� subunits function in Sz. pombe

Fission yeast is traditionally a good system for studyingGPCR signalling cascades and a variety of proteins havebeen successfully replaced with human counterparts,including GPCRs and AGS proteins (Ladds et al., 2003),and G� subunits and RGS proteins (G. Ladds, unpub-lished). To determine if human G� subunits could functionin a similar manner to Gnr1p, all Wve subunits were clonedinto the pREP3x expression vector. These plasmids wereintroduced into �gnr1 (JY1314; Fig. 4A) and gnr1+ (JY546;Fig. 4B) strains, and �-galactosidase activity determined atvarious levels of pheromone stimulation.

Expression of the human G� subunits decreasedreporter activity at all levels of pheromone stimulation.GNB1, GNB2, GNB3 and GNB4 all functioned to similarlevels, reducing reporter activity to »10 U at 10¡6 M phero-mone in JY1314 (�gnr1) and to »7 U at the same phero-

mone concentration in JY546 (gnr1+). GNB5 functionedslightly less eVectively, resulting in a decrease to »15 U inJY1314 and to »10 U in JY544 when stimulated with10¡6 M pheromone. The lower level of activity observed inJY546 compared to JY1314 is possibly due to the presenceof endogenous Gnr1p.

To determine if other WD repeat proteins could act in acomparable manner to Gnr1p, Git5p and the human G�subunits, the product of the Sz. pombe NP_593424 ORFwas cloned and expressed in both gnr1+ (JY546) and �gnr1(JY1314) strains. This protein is the Sz. pombe homologueof the human WDR26 protein (Zhu et al., 2004). WDR26 isreported to be a G�-like subunit which acts to suppress theactivity of MAP kinase cascades (Zhu et al., 2004). Overex-pression of NP_593424 did not aVect signalling through thepheromone-response pathway in either JY1314 (Fig. 4A) orJY546 (Fig. 4B). This suggests that the eVects observed arespeciWc for the Sz. pombe Gnr1p and Git5p subunits andthe classical human G� subunits.

3.7. Interaction of G�, G� and G� subunits

To establish whether the human G� subunits coulddirectly interact with Gpa1p, a two-hybrid assay was per-formed similar to that described earlier. BrieXy, all G� sub-units were cloned into pGADT7 and each co-transformedwith either Gpa1p or the human G� subunit GNG4. GNG4has previously been demonstrated to interact with all humanG� subunits in this system (Yan et al., 1996), providing apositive control for interaction. Colonies were observed forall co-transformants on MM but only for the GNB-GNG4and Git5p–GNG4 pairings on MM-His-Ade +X-�-galplates (Fig. 5). This indicates that human G� subunits andGit5p are incapable of direct interaction with Gpa1p in thissystem (Fig. 5A), but are able to bind GNG4 (Fig. 5B).

Fig. 3. EVect of Gnr1p and Git5p disruptions on pheromone signaling. The sxa2 > lacZ reporter strain (JY546) and the same strain containing a disruptionof the gnr1 ORF (JY1314), both containing pREP3x or pREP3x-Gnr1, were exposed to various concentrations of pheromone and �-galactosidase activityassayed after 16 h. Reporter strains in which the git5 ORF (JY1317), or the gnr1 and git5 ORFs (JY1319) had been disrupted and containing pREP3x werealso assayed. Activity is expressed as OD420 per 106 cells. The results are presented as mean of triplicate determinations § SD.

0

10

20

30

40β-

gala

ctos

idas

e un

its

)M(]enomorehP[

0

01 8-

01x3 8-

01 7-

645YJ( 1rng +, 5tig +)

+x3PERp

01x3 7-

01 6-

4131YJ(Δ 1rng , 5tig +)

+x3PERp

7131YJ( 1rng +, Δ 5tig )

+x3PERp

9131YJΔ( 1rng , Δ 5tig )

+x3PERp

645YJ( 1rng +, 5tig +)

+1rnG-x3PERp

4131YJ(Δ 1rng , 5tig +)

+1rnG-x3PERp

A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851 847

It is therefore possible that the human G� subunits andGit5p require a G� subunit to modulate Gpa1p activity inSz. pombe. A single G� subunit, Git11p, has been identiWedin Sz. pombe, which has been demonstrated to interact withGit5p in the yeast two-hybrid system and form a functionalpartner in vivo (Landry and HoVman, 2001). We thereforeinvestigated if Git11p could interact with the human G�subunits. Git11p was cloned into the pGBKT7 expressionvector and co-transformed with the human G� subunits,Gnr1p and Git5p. As earlier, strains were plated onto MMand MM-His-Ade + X-�-gal medium (Fig. 5C). Gnr1p andGit11p do not interact in this system but formation of bluecolonies on MM-His-Ade + X-�-gal plates indicated thatGit5p and all human GNB subunits were capable of inter-acting with Git11p.

To allow quantitation of the relative strengths of theseinteractions, liquid �-galactosidase assays were conducted(Fig. 5). This demonstrated, as previously (Fig. 1), a fairlystrong interaction between Gnr1p and Gpa1p, producing»7 U of activity. Additionally, all human G� subunitsinteracted with GNG4, producing »4 U. The interaction

between Git5p and GNG4 was slightly weaker. Similarly,all human G� subunits interacted with Git11p, producing»4 U, but the interaction between Git5p and Git11p wasslightly stronger, producing »5 U. All other protein pairsfailed to generate more than 1 U, indicating very weak, ifany, interaction.

4. Discussion

4.1. Gnr1p is a negative regulator of the pheromone-response pathway

This study details the identiWcation and characterisationof a novel Sz. pombe protein which interacts with Gpa1p inthe yeast two-hybrid system. The data presented demon-strate that the product of the SPCC1020.09 ORF (Gnr1p)negatively regulates the activity of Gpa1p. It appears thatthe level of gnr1 expression is key in determining theresponsiveness of Sz. pombe to pheromone stimulation.Strains which lack a functional Gnr1p become hypersensi-tive to pheromone, whereas those in which Gnr1p is

Fig. 4. EVect of G� subunit expression on pheromone signaling. (A) A strain containing the sxa2 > lacZ reporter but lacking gnr1 (JY1314) or (B) thesxa2 > lacZ reporter strain (JY546) was transformed with pREP3x, pREP3x-Git5, pREP3x-GNB1, pREP3x-GNB2, pREP3x-GNB3, pREP3x-GNB4,pREP3x-GNB5 or pREP3x-NP_593424. Cells were grown in deWned minimal medium in the absence of thiamine to induce expression from the nmt1 pro-moter. Each strain was exposed to various concentrations of pheromone and �-galactosidase production assayed 16 h after stimulation. Activity isexpressed as OD420 per 106 cells. The results are presented as mean of triplicate determinations § SD.

)M( ]enomorehP[

0

01 8-

01 7-

01 6-

A

B

β-ga

lact

osid

ase

units

-x3PERpx3PERp1BNG

-x3PERp2BNG

-x3PERp3BNG

-x3PERp4BNG

-x3PERp5BNG

-x3PERp5tiG

-x3PERp

424395_PN

0

10

20

30

40

0

10

20

30

40

β-ga

lact

osid

ase

units

-x3PERpx3PERp1BNG

-x3PERp2BNG

-x3PERp3BNG

-x3PERp4BNG

-x3PERp5BNG

-x3PERp5tiG

-x3PERp

424395_PN

848 A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851

expressed to a higher level become hyposensitive. It is possi-ble, given that it is predicted to contain seven WD repeats,that Gnr1p achieves this modulation by acting as a struc-tural mimic of a G� subunit. The G�� dimer of a G proteintraditionally acts to increase the stability of the G�-GDP(inactive) form by altering the conformation of the switchII region of the G� (Coleman et al., 1994; Lambright et al.,1994). Concomitantly, association of the G�� and G�results in an increase in the rate of association with GDPand a decrease in the rate of dissociation (Brandt and Ross,

1985). This is consistent with the results observed in whichGnr1p appears to act as a negative regulator of Gpa1p.

Interestingly, disruption of a G� subunit often results ina decrease in response through the associated pathway. Forexample, disruption of Git5p results in a similar phenotypeto a disruption of Gpa2p, i.e., a defect in adenylate cyclaseactivation (Landry et al., 2000). This may be as lipid-modi-Wcation of the G�� can be essential for correct membrane-localisation of the heterodimer (Simonds et al., 1991) andconsequently the G� protein. For example, the Git5p–

Fig. 5. Yeast two-hybrid analysis of G�–G�–G� interactions. The Sc. cerevisiae strain AH109 was transformed with pGBKT7-Gpa1, pGBKT7-GNG4and pGBKT7-Git11. Each of these strains was subsequently transformed with pGADT7-GNB1, pGADT7-GNB2, pGADT7-GNB3, pGADT7-GNB4,pGADT7-GNB5, pGADT7-Gnr1 or pGADT7-Git5. Each pairing was plated onto MM to select for the presence of the two plasmids and onto MM-His-Ade + X-�-gal to determine if the two proteins interact. Interacting proteins induce expression of the HIS3, ADE2 and MEL1 reporters, allowing growthof blue colonies on this selective medium. All strains grew on MM indicating the presence of the two relevant plasmids. Yeast strains expressing Gpa1pand Gnr1p grew on MM-His-Ade + X-�-gal indicating an interaction between these proteins. Likewise, GNB1, GNB2, GNB3, GNB4, GNB5 and Git5pwere capable of growth on MM-His-Ade + X-�-gal when co-expressed with GNG4 or Git11p. The absence of growth for the other protein combinationsindicates that they interact very weakly. The various co-transformants were also analysed in a liquid-based �-galactosidase assay for expression of thelacZ reporter gene, another indication of the interaction between the two proteins of interest. Under the conditions used, a strong interaction between twoproteins (such as SV40-T antigen and p53) generates »10 U of �-galactosidase activity. The results are presented as mean of triplicatedeterminations § SD.

1B

NG

2B

NG

3B

NG

4B

NG

5B

NG

p1rn

G

p5tiG

MM

MM-His-Ade+ X-α-galGpa1p(Gα)

A

MM

MM-His-Ade+ X-α-gal

GNG4(Gγ)

B

MM

MM-His-Ade+ X-α-gal

Git11p(Gγ)

C

10

8

6

4

2

0

8

6

4

2

0

8

6

4

2

0

βstinu

esadisotcalag-β

stinuesadi sotcalag-

βstinu

esadisotcalag-

10

10

A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851 849

Git11p dimer may be required to target Gpa2p to theplasma membrane. In contrast, disruption of gnr1enhanced signalling via Gpa1p. It is therefore likely thatGnr1p functions solely as a negative regulator of Gpa1pand plays no positive role in G� activation. This could oVeran explanation as to why gnr1 has not been identiWed inprevious screens. Additionally, the increase in pheromonesignalling upon disruption was probably not suYcient toallow detection in a previous growth-based screen for nega-tive regulators of pheromone signalling (Didmon et al.,2002).

Gnr1p probably acts with Rgs1p and Mam2p to main-tain Gpa1p in an inactive GDP-bound state, although viasubtly diVerent mechanisms. Mam2p may sequester Gpa1pinto a “preactivation complex” (Ladds et al., 2003), Rgs1pseems to accelerate the rate of return from the GTP- toGDP-bound state (Watson et al., 1999), and it wouldappear that Gnr1p acts to stabilise the Gpa1p–GDP state.

4.2. Overexpression of Git5p attenuates the pheromone response

Although disruption of Git5p had no eVect upon thepheromone-response pathway, overexpression was able tocomplement a mutation of Gnr1p and reduce signalling ina wild-type strain. Data presented in this study (Fig. 3) andelsewhere (Landry et al., 2000; Landry and HoVman, 2001)demonstrate that Git5p functions only in the nutrient-sens-ing pathway. It is likely that at endogenous levels, Git5pand Gpa1p are not associated. The Git5p–Git11p dimermay possess a higher aYnity for Gpa2p than for Gpa1pand the reverse may be true for Gnr1p. It is also possiblethat Gpa2p is in excess of Git5p and Git11p, and thereforeacts to sequester the G�� dimer away from Gpa1p. Suchstoichiometry would be consistent with that observed in Sc.cerevisiae where the ratio is 5 G�:3 G�:1 G� (Ghaemmagh-ami et al., 2003). Overexpression of Git5p in our strainscould perturb this stoichiometry, leading to association ofthe Git5p–Git11p dimer with Gpa1p. These data suggestthat the relative expression levels of all components withinthe signalling cascades are crucial in regulating both GPCRpathways and may act to prevent crosstalk between them.

4.3. Human G� subunits can complement a �gnr1 mutation

It was demonstrated that all Wve human G� subunitswere capable of complementing a �gnr1 mutation andcould reduce signalling in a strain containing endogenousGnr1p (Fig. 4). GNB5 was slightly less eVective at comple-menting the mutation at high pheromone concentrationsthan the other subunits. This could be due to lower expres-sion levels in our strains or due to a weaker associationwith Gpa1p; GNB5 is more divergent than the other fourG� subunits. GNB5 has also been demonstrated to comple-ment a �STE4 (G�) mutation in Sc. cerevisiae (Ajit andYoung, 2004). This is, to our knowledge, the Wrst demon-stration of functionality of the remaining human G� sub-

units in yeast and also that all Wve are capable of acting asnegative regulators of G� signalling in such a system. TheeVects observed upon the expression of these subunitsappears to be speciWc as the Sz. pombe homologue of thehuman G�-like WDR26 (NP_593424) protein had no eVectupon the pheromone-response pathway.

Additionally, it was demonstrated that all human G�subunits could interact not only with the human GNG4subunit but also with the Sz. pombe G� (Git11p). Git5p wasalso capable of interacting with both G� subunits. This isnot unprecedented as human G� subunits have been shownto interact with GNG4 (Yan et al., 1996), Git11p was iden-tiWed in a yeast two-hybrid screen using Git5p as bait, andthe two G� subunits share a high degree of similarity (Lan-dry and HoVman, 2001). Gnr1p bound Gpa1p in the two-hybrid system, although neither the human G� subunitsnor Git5p were capable of direct interaction. Therefore, it ispossible that the eVects observed in vivo are mediated bythe presence of Git11p, which allows association of the G�subunit and Gpa1p. Such a subunit may not be required tofacilitate the Gnr1p–Gpa1p interaction.

4.4. Gnr1p may function as a monomer

It is possible that the Gpa1p–Gnr1p complex interacts inthe absence of a G�-like protein. It is not unprecedented foryeast G proteins to be atypical. For example, Git5p in Sz.pombe lacks the N-terminal coiled-coil which is generallynecessary for G�–G� interactions, yet interacts as stronglyas human G� subunits with Git11p and GNG4. Addition-ally, kelch-repeat proteins mimic the action of the G� sub-unit in the Sc. cerevisiae GPA2 nutrient-sensing pathway,even in the absence of the G�-like protein GPG1 (Hara-shima and Heitman, 2002). Finally, GPA1 (G�) and STE4(G�) from this organism were demonstrated to interact inthe yeast two-hybrid system in a strain apparently lackingSTE18 (G�) (Clark et al., 1993).

Sequence analysis of Gpa1p and Gnr1p revealedextensions which are not present in the majority of mam-malian G� and G� subunits. For example, Gpa1p has anN-terminal extension of »30 residues when compared tomammalian G� proteins. It has been demonstrated that theN-terminus of G� proteins plays a key role in the interac-tion with the G�� dimer (Clapham and Neer, 1997). Thisextension could strengthen the Gpa1p–Gnr1p interactionand remove the need for a G�-like subunit. Alternatively, oradditionally, the C-terminal extension of Gnr1p could playa role in the association. This extension is not found inhuman G� subunits, nor in Git5p, but is »70 amino acidsin length, a similar size to most G� subunits. Most aminoacids within the highly charged region have a negativecharge (glutamate or aspartate). It is therefore possible thatthis region may interact with residues of opposite charge,either in other proteins or within Gnr1p itself. Hypotheti-cally, this extension could substitute for a G� subunit andstabilise the Gnr1p–Gpa1p interaction. Obviously, thesehypotheses do not preclude the presence of a G� or G�-like

850 A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851

subunit. Yeast two-hybrid screens for proteins which inter-act with Gnr1p may reveal the presence of further bindingpartners.

Despite similarities in predicted structure, Gnr1p doesnot seem to function as a typical G� subunit as it plays noapparent role in activation of Gpa1p. Instead, it appears toserve solely as a negative regulator of G� signalling. A simi-lar role as a “pseudostructural inhibitor” (Ivey and HoV-man, 2002) has been proposed for the kelch-repeat proteinsfound in Sc. cerevisiae (Harashima and Heitman, 2002).The presence of typical and atypical G�-binding partners inSz. pombe may prevent cross-talk between the two G pro-tein-mediated signalling cascades, and it is possible thatproteins with similar functions exist in higher eukaryotes.Model organisms may play key roles in determining theprecise action of such proteins. To date, human GPCRs,G� subunits, RGS and AGS proteins have been shown tofunction in Sz. pombe (Ladds et al., 2003, G. Ladds, unpub-lished). The addition of human G� subunits to this reper-toire should enable analysis of a complete human GPCRsignalling cascade in this model system.

Acknowledgments

This work was supported by Ph.D. studentships from theBiotechnology and Biological Sciences Research CouncilUK (AG and RF) and a Research Fellowship from theUniversity Hospitals of Coventry and Warwickshire NHSTrust (GL). We thank Charlie HoVman (Boston College)for provision of constructs, Vilmos Fülöp (University ofWarwick) for expert assistance with THREADER, GregHannon (Cold Spring Harbour Laboratory) for provisionof the Sz. pombe cDNA library and Claire Hill (Universityof Warwick) for critical reading of the manuscript.

References

Ajit, S.K., Young, K.H., 2004. Enhancement of pheromone response byRGS9 and G�5 in yeast. Biochem. Biophys. Res. Commun. 324, 686–691.

Blumer, J.B., Cismowski, M.J., Sato, M., Lanier, S.M., 2005. AGS proteins:receptor-independent activators of G-protein signaling. Trends Phar-macol. Sci. 26, 470–476.

Brandt, D.R., Ross, E.M., 1985. GTPase activity of the stimulatory GTP-binding regulatory protein of adenylate cyclase, Gs. Accumulationand turnover of enzyme–nucleotide intermediates. J. Biol. Chem. 260,266–272.

Chung, K.S., Won, M., Lee, S.B., Jang, Y.J., Hoe, K.L., Kim, D.U., Lee,J.W., Kim, K.W., Yoo, H.S., 2001. Isolation of a novel gene fromSchizosaccharomyces pombe: stm1+ encoding a seven-transmembraneloop protein that may couple with the heterotrimeric Galpha 2 protein,Gpa2. J. Biol. Chem. 276, 40190–40201.

Clapham, D.E., Neer, E.J., 1997. G protein �� subunits. Annu. Rev. Phar-macol. Toxicol. 37, 167–203.

Clark, K.L., Dignard, D., Thomas, D.Y., Whiteway, M., 1993. Interactionsamong the subunits of the G protein involved in Saccharomyces cerevi-siae mating. Mol. Cell. Biol. 13, 1–8.

Coleman, D.E., Berghuis, A.M., Lee, E., Linder, M.E., Gilman, A.G.,Sprang, S.R., 1994. Structures of active conformations of Gi alpha 1and the mechanism of GTP hydrolysis. Science 265, 1405–1412.

Corpet, F., 1988. Multiple sequence alignment with hierarchical clustering.Nucleic Acids Res. 16, 10881–10890.

Davey, J., 1992. Mating pheromones of the Wssion yeast Schizosaccharo-myces pombe: puriWcation and structural characterization of M-factorand isolation and analysis of two genes encoding the pheromone.EMBO J. 11, 951–960.

Davey, J., 1998. Fusion of a Wssion yeast. Yeast 14, 1529–1566.Davey, J., Egel, R., Nielsen, O., 1995. Pheromone procedures in Wssion

yeast. In: Adolph, K.W. (Ed.), Methods in Molecular Genetics 6B. Aca-demic Press, San Diego, pp. 247–263.

Decottignies, A., Sanchez-Perez, I., Nurse, P., 2003. Schizosaccharomycespombe essential genes: a pilot study. Genome Res. 13, 399–406.

Didmon, M., Davis, K., Watson, P., Ladds, G., Broad, P., Davey, J., 2002.Identifying regulators of pheromone signalling in the Wssion yeastSchizosaccharomyces pombe. Curr. Genet. 41, 241–253.

Dohlman, H.G., Apaniesk, D., Chen, Y., Song, J., Nusskern, D., 1995. Inhi-bition of G protein signalling by dominant gain-of-function mutationsin Sst2p, a pheromone desensitisation factor in Saccharomyces cerevi-siae. Mol. Cell. Biol. 5, 3635–3643.

Gettemans, J., Meerschaert, K., Vandekerckhove, J., De Corte, V., 2003. Akelch beta propeller featuring as a G beta structural mimic: reinventingthe wheel? Science 191, PE27.

Ghaemmaghami, S., Huh, W.K., Bower, K., Howson, R.W., Belle, A.,Dephoure, N., O’Shea, E.K., Weissman, J.S., 2003. Global analysis ofprotein expression in yeast. Nature 425, 737–741.

Harashima, T., Heitman, J., 2002. The Galpha protein Gpa2 controls yeastdiVerentiation by interacting with kelch repeat proteins that mimicGbeta subunits. Mol. Cell 10, 163–173.

Hiskens, R., Vatish, M., Hill, C., Davey, J., Ladds, G., 2005. SpeciWc in vivobinding of the activator of G protein signalling 1 to the G�1 subunit.Biochem. Biophys. Res. Commun. 337, 1038–1046.

HoVman, C.S., 2005. Except in every detail: comparing and contrasting G-protein signaling in Saccharomyces cerevisiae and Schizosaccharomycespombe. Eukaryot. Cell 4, 495–503.

Imai, Y., Yamamoto, M., 1992. Schizosaccharomyces pombe sxa1+ andsxa2+ encode putative proteases involved in the mating response. Mol.Cell. Biol. 12, 1827–1834.

Imai, Y., Yamamoto, M., 1994. The Wssion yeast mating pheromone P-fac-tor: its molecular structure, gene structure, and ability to induce geneexpression and G1 arrest in the mating partner. Genes Dev. 8, 328–338.

Isshiki, T., Mochizuki, N., Maeda, T., Yamamoto, M., 1992. Characteriza-tion of a Wssion yeast gene, gpa2, that encodes a G alpha subunitinvolved in the monitoring of nutrition. Genes Dev. 6, 2455–2462.

Ivey, F.D., HoVman, C.S., 2002. Pseudostructural inhibitors of G proteinsignaling during development. Dev. Cell 3, 154–155.

Ivey, F.D., HoVman, C.S., 2005. Direct activation of Wssion yeast adenylatecyclase by the Gpa2 Galpha of the glucose signaling pathway. Proc.Natl. Acad. Sci. USA 102, 6108–6113.

Jones, D.T., Tress, M., Bryson, K., Hadley, C., 1999. Successful recognitionof protein folds using threading methods biased by sequence similarityand predicted secondary structure. Proteins 3, 104–111.

Kim, D.U., Park, S.K., Chung, K.S., Choi, M.U., Yoo, H.S., 1996. The Gprotein beta subunit Gpb1 of Schizosaccharomyces pombe is a negativeregulator of sexual development. Mol. Gen. Genet. 252, 20–32.

Kitamura, K., Shimoda, C., 1991. The Schizosaccharomyces pombe mam2gene encodes a putative pheromone receptor which has a signiWcanthomology with the Saccharomyces cerevisiae Ste2 protein. EMBO J.10, 315–321.

KjaerulV, S., Lautrup-Larsen, I., Truelsen, S., Pedersen, M., Nielsen, O.,2005. Constitutive activation of the Wssion yeast pheromone-responsivepathway induces ectopic meiosis and reveals Ste11 as a mitogen-acti-vated protein kinase target. Mol. Cell. Biol. 25, 2045–2059.

Ladds, G., Davey, J., 2004. Analysis of human GPCRs in Wssion yeast.Curr. Opin. Drug Discov. Devel. 7, 683–691.

Ladds, G., Rasmussen, E.M., Young, T., Nielsen, O., Davey, J., 1996. Thesxa2-dependent inactivation of the P-factor mating pheromone in theWssion yeast Schizosaccharomyces pombe. Mol. Microbiol. 20, 35–42.

Ladds, G., Davis, K., Hillhouse, E.W., Davey, J., 2003. ModiWed yeast cellsto investigate the coupling of G protein-coupled receptors to speciWc Gproteins. Mol. Microbiol. 47, 781–792.

A. Goddard et al. / Fungal Genetics and Biology 43 (2006) 840–851 851

Ladds, G., Davis, K., Das, A., Davey, J., 2005. A constitutively activeGPCR retains its G protein speciWcity and the ability to form dimers.Mol. Microbiol. 55, 482–497.

Lambright, D.G., Noel, J.P., Hamm, H.E., Sigler, P.B., 1994. Structuraldeterminants for activation of the alpha-subunit of a heterotrimeric Gprotein. Nature 369, 621–628.

Lambright, D.G., Sondek, J., Bohm, A., Skiba, N.P., Hamm, H.E., Sigler,P.B., 1996. The 2.0 A crystal structure of a heterotrimeric G protein.Nature 379, 311–319.

Landry, S., HoVman, C.S., 2001. The git5 Gbeta and git11 Ggamma forman atypical Gbetagamma dimer acting in the Wssion yeast glucose/cAMP pathway. Genetics 157, 1159–1168.

Landry, S., Pettit, M.T., Apolinario, E., HoVman, C.S., 2000. The Wssionyeast git5 gene encodes a Gbeta subunit required for glucose-triggeredadenylate cyclase activation. Genetics 154, 1463–1471.

Li, B., Fields, S., 1993. IdentiWcation of mutations in p53 that aVect itsbinding to SV40 large T antigen by using the yeast two-hybrid system.FASEB J. 7, 957–963.

Maundrell, K., 1993. Thiamine-repressible expression vectors pREP andpRIP for Wssion yeast. Gene 123, 127–130.

Obara, T., Nakafuku, M., Yamamoto, M., Kaziro, Y., 1991. Isolation andcharacterization of a gene encoding a G-protein alpha subunit fromSchizosaccharomyces pombe: involvement in mating and sporulationpathways. Proc. Natl. Acad. Sci. USA 88, 5877–5881.

Pellegrino, S., Zhang, S., Garritsen, A., Simonds, W.F., 1997. Thecoiled-coil region of the G protein beta subunit. Mutational analy-sis of Ggamma and eVector interactions. J. Biol. Chem. 272, 25360–25366.

Simonds, W.F., Butrynski, J.E., Gautam, N., Unson, C.G., Spiegel, A.M.,1991. G-protein beta gamma dimers. Membrane targeting requires sub-unit coexpression and intact gamma C–A–A–X domain. J. Biol. Chem.266, 5363–5366.

Sondek, J., Bohm, A., Lambright, D.G., Hamm, H.E., Sigler, P.B., 1996.Crystal structure of a G-protein beta gamma dimer at 2.1A resolution.Nature 379, 369–374.

Tanaka, K., Davey, J., Imai, Y., Yamamoto, M., 1993. Schizosaccharomy-ces pombe map3+ encodes the putative M-factor receptor. Mol. Cell.Biol. 13, 80–88.

Tesmer, J.J., Berman, D.M., Gilman, A.G., Sprang, S.R., 1997. Structure ofRGS4 bound to AlF4¡-activated G(i alpha1): stabilization of the tran-sition state for GTP hydrolysis. Cell 89, 251–261.

Wall, M.A., Coleman, D.E., Lee, E., Iniguez-Lluhi, J.A., Posner, B.A., Gil-man, A.G., Sprang, S.R., 1995. The structure of the G protein heterotri-mer Gi alpha 1 beta 1 gamma 2. Cell 83, 1047–1058.

Watson, P., Davis, K., Didmon, M., Broad, P., Davey, J., 1999. An RGSprotein regulates the pheromone response in the Wssion yeast Schizo-saccharomyces pombe. Mol. Microbiol. 33, 623–634.

Welton, R.M., HoVman, C.S., 2000. Glucose monitoring in Wssion yeast viathe Gpa2 galpha, the git5 Gbeta and the git3 putative glucose receptor.Genetics 156, 513–521.

Yan, K., Kalyanaraman, V., Gautam, N., 1996. DiVerential ability to formthe G protein betagamma complex among members of the beta andgamma subunit families. J. Biol. Chem. 271, 7141–7146.

Zhu, Y., Wang, Y., Xia, C., Li, D., Li, Y., Zeng, W., Yuan, W., Liu, H., Zhu,C., Wu, X., Liu, M., 2004. WDR26: a novel Gbeta-like protein, sup-presses MAPK signaling pathway. J. Cell. Biochem. 93, 579–587.