Modified yeast cells to investigate the coupling of G protein-coupled receptors to specific G proteins

Authors


Summary

G protein-coupled receptors (GPCRs) help to regulate the physiology of all the major organ systems. They respond to a multitude of ligands and activate a range of effector proteins to bring about the appropriate cellular response. The choice of effector is largely determined by the interaction of individual GPCRs with different G proteins. Several factors influence this interaction, and a better understanding of the process may enable a more rational approach to identifying compounds that affect particular signalling pathways. A number of systems have been developed for the analysis of GPCRs. All provide useful information, but the genetic amenability and relative simplicity of yeast makes them a particularly attractive option for ligand identification and pharmaceutical screening. Many, but not all, GPCRs are functional in the budding yeast Saccharomyces cerevisiae, and we have developed reporter strains of the fission yeast Schizosaccharomyces pombe as an alternative host. To provide a more generic system for investigating GPCRs, we created a series of yeast–human Gα-transplants, in which the last five residues at the C-terminus of the yeast Gα-subunit are replaced with the corresponding residues from different human G proteins. These enable GPCRs to be coupled to the Sz. pombe signalling machinery so that stimulation with an appropriate ligand induces the expression of a signal-dependent lacZ reporter gene. We demonstrate the specificity of the system using corticotropin releasing factor (CRF) and CRF-related peptides on two CRF receptors. We find that different combinations of ligand and receptor activate different Gα-transplants, and the specificity of the coupling is similar to that in mammalian systems. Thus, CRF signalled through the Gs- and Gi-transplants, consistent with its regulation of adenylate cyclase, and was more active against the CRF-R1A receptor than against the CRF-R2B receptor. In contrast, urocortin II and urocortin III were selective for the CRF-R2B receptors. Furthermore, urocortin, but not CRF, induced signalling through the CRF-R1A receptor and the Gq-transplant. This is the first time that human GPCRs have been coupled to the signalling pathway in Sz. pombe, and the strains described in this study will complement the other systems available for studying this important family of receptors.

Introduction

G protein-coupled receptors (GPCRs) regulate diverse biological processes by mediating signalling to a wide range of stimuli. In response to stimulation by extracellular ligands, the receptors act through heterotrimeric G proteins to regulate intracellular effector proteins. The G proteins, which are composed of Gα-, Gβ- and Gγ-subunits, play a pivotal role in linking the different GPCRs to the different effectors (Neer, 1995; Hamm and Gilchrist, 1996). In the inactive state, the Gα-subunit is bound to GDP and associated with the Gβγ dimer. Upon ligand binding, the receptor stimulates the release of GDP from Gα, allowing the binding of GTP. Gα-GTP is released from the Gβγ dimer, and the dissociated subunits regulate the activity of effector proteins such as adenylate cyclase, phospholipase C, mitogen-activated protein kinase (MAP kinase) cascades and ion channels. Changes in the activity of these proteins bring about changes in cell behaviour.

G proteins are defined by their Gα-subunits, and at least 20 have been identified in humans (Venter et al., 2001). They exhibit considerable structural and functional diversity but can be assigned to four major families (Gs, Gi/o, Gq and G12) according to sequence homologies (Simon et al., 1991). Individual GPCRs interact with different G proteins, and it is the specificity of this interaction that determines which effectors are targeted and, ultimately, the cellular response. Most receptors interact only with one or a small subset of G proteins, but there are many examples of GPCRs that couple to multiple G proteins (Laugwitz et al., 1993; Law et al., 1993; Chabre et al., 1994). Some, such as the thyrotropin receptor (Laugwitz et al., 1996), couple to members from all four G protein families. There is considerable interest in understanding the specificity of this coupling process at the molecular level (Gudermann et al., 1996; Bourne, 1997; Wess, 1998).

The hypothesis termed ‘agonist-directed trafficking’ complicates the situation further (Kenakin, 1995; 1997). This proposes that different agonists can induce different receptor conformations and that these could differentially favour the coupling to one G protein over another. In functional terms, it means that a GPCR could have different pharmacological profiles depending on which G protein is activated and that the same GPCR could have different roles in different cell types, depending on the local G protein environment. There is now a considerable amount of data to support this hypothesis (Spengler et al., 1993; Robb et al., 1994; Berg et al., 1998; Fitzgerald et al., 1999; Brink et al., 2000; Watson et al., 2000), and it is beginning to impact on several areas. It may, for example, enable a more rational approach to drug design to identify compounds that only affect the coupling of a GPCR to a particular G protein.

Several approaches have been used to investigate the specificity of coupling between GPCRs and G proteins. These include the use of G protein subtype selective antisera (Georgoussi et al., 1995), reconstitution studies with purified G proteins (Chan et al., 1995), incorporation of a covalent label only into activated G proteins (Laugwitz et al., 1993; Chakrabarti et al., 1995) and the construction of GPCR–Gα fusion proteins (Seifert et al., 1999; Milligan, 2000). All these approaches have yielded useful information, but they are not easily adapted to provide high-throughput analysis and screens for isolating, for example, conformation-specific drugs. Many researchers have therefore used yeast as a host for studying human GPCRs.

The yeast pheromone response pathway provides a robust and experimentally tractable system for studying GPCRs (Dohlman et al., 1991; Kurjan, 1993; Davey, 1998). Haploid yeast cells exist in one of two mating types, and conjugation occurs between cells of opposite mating type and is controlled by the reciprocal exchange of diffusible pheromones. These peptides are released by cells of one mating type, bind to GPCRs on the surface of a partner of the opposite type and activate intracellular machinery that includes a G protein, a MAP kinase cascade and a transcription factor. Stimulation by mating pheromone induces expression of the genes required to bring about mating-related changes in cell behaviour, including an arrest of the cell cycle, enhanced cell agglutination and cell fusion. Simple genetics can be used to produce strains expressing human GPCRs, and replacing one of the response genes with a reporter construct such as β-galactosidase provides a simple readout for signalling (Pausch, 1997; Reiländer and Weib, 1998). Most studies use the budding yeast Saccharomyces cerevisiae, but not all GPCRs elicit a response after stimulation in this yeast, and the fission yeast Schizosaccharomyces pombe provides an alternative host in which to study human GPCRs.

Several human GPCRs have been expressed in Sz. pombe but failed to couple to the pheromone response pathway (Sander et al., 1994; Arkinstall et al., 1995; Ficca et al., 1995). An improved understanding of how the pheromone response pathway in Sz. pombe is regulated (reviewed by Davey, 1998) allowed the construction of Sz. pombe strains that express pheromone-dependent reporter proteins (Watson et al., 1999; Didmon et al., 2002). In order to develop a more generic system for investigating human GPCRs, we have modified the strains further by introducing a series of yeast–human Gα-transplants in which the last five amino acids at the C-terminus of the yeast Gα-subunit are replaced with the corresponding residues from the different human proteins – the term ‘transplant’ was introduced to discriminate these slightly modified Gα-subunits from Gα-chimeras in which much larger changes are made to the Gα-proteins (Brown et al., 2000). The resulting yeast strains couple GPCRs to the Sz. pombe signalling machinery such that stimulation with an appropriate ligand induces the expression of a signal-dependent lacZ reporter gene. We demonstrate the specificity of the system using corticotropin releasing factor (CRF) and CRF-related peptides to stimulate the CRF-R1A and CRF-R2B receptors, different combinations of ligand and receptor activating different Gα-transplants.

Results

Creating the host yeast reporter strain

P cells, of plus mating type, containing an inte-grated mat1-Pm>lacZ reporter construct express β-galactosidase in response to the M-factor mating pheromone (Watson et al., 1999). These cells were used previously to identify the role of the Rgs1 protein (regulator of G protein signalling) in the pheromone signalling pathway (Watson et al., 1999), but there is significant expression of the reporter construct in the absence of stimulation and this can complicate its use. We therefore generated M cells (minus mating type) containing an sxa2>lacZ reporter (Didmon et al., 2002). Sxa2 is a carboxypeptidase that inactivates extracellular P-factor pheromone by removing the C-terminal leucine residue (Imai and Yamamoto, 1992; Ladds et al., 1996; Ladds and Davey, 2000). The sxa2 gene is only expressed after stimulation with P-factor, and a two-step strategy replaced the open reading frame (ORF) of the chromosomal sxa2 with that from the Escherichia coli lacZ gene to create a strain that expresses β-galactosidase in response to stimulation with P-factor. Integrating the reporter into the yeast chromosome avoids the variability that can complicate the use of plasmid-borne reporters (Aono et al., 1994). Other modifications that we made to the strains included removing the information at the mat2 and mat3 mating loci to ensure that the strains are unable to switch mating type (Klar and Miglio, 1986) and deleting the cyr1 gene (encodes adenylate cyclase) to enable them to respond to stimulation under conditions that support mitotic growth (Davey and Nielsen, 1994; Imai and Yamamoto, 1994; Didmon et al., 2002).

To create a ‘receptor-free’ strain for analysing the interaction of human GPCRs with the pheromone signalling pathway, we deleted the mam2 gene that encodes the P-factor receptor. The only other GPCR in these cells is Git3, but this normally regulates adenylate cyclase in response to environmental glucose, and there does not appear to be any cross-talk between the two pathways (Welton and Hoffman, 2000; Landry and Hoffman, 2001). When transformed with a plasmid containing the Sz. pombe mam2 gene (encoding the P-factor receptor) under the control of the thiamine-repressible nmt1 promoter, this basic reporter strain produces β-galactosidase in response to stimulation with P-factor (Fig. 1). The results are very similar to those obtained when Mam2 is expressed from its chromosomal locus (Didmon et al., 2002), which validates the use of this strain to analyse plasmid-borne GPCRs.

Figure 1.

Pheromone-dependent expression of the sxa2 > lacZ reporter construct. The Sz. pombe strain JY1169 (contains gpa1 and sxa2>lacZ but lacks mam2 , encodes the P-factor receptor) was transformed with pREP:Mam2 (JD1627, Sz. pombe P-factor receptor) and exposed to P-factor pheromone. Production of β-galactosidase was assayed using ONPG as substrate, and activity is expressed as OD 420 units per 10 6 cells (see Experimental procedures ).

Top. Cells were exposed to P-factor at either 0 or 10−6 M and assayed at hourly intervals.

Bottom. Cells were exposed to various concentrations of P-factor for 16 h before being assayed.

Values are means of triplicate determinations from a typical experiment.

Constructing yeast strains containing Gα-transplants

Some human GPCRs interact with the endogenous Sz. pombe Gpa1, but the majority were not active when expressed in this strain (not shown). To develop a more generic Sz. pombe system, we constructed a series of Gα-transplants in which the C-terminal five amino acids of Gpa1 were replaced with the corresponding residues from mammalian Gα-subunits. These residues play a key role in determining which GPCRs can couple to a particular Gα-subunit (Hirsch et al., 1991; Conklin et al., 1993; Liu et al., 1995; Wess, 1998), and the Gα-transplant approach has been used to couple human GPCRs to the pheromone signalling pathway in S. cerevisiae (Brown et al., 2000; Erlenbach et al., 2001). To avoid disturbing the balance between the Gα-subunits and the rest of the signalling machinery, we integrated the Gα-transplants at the gpa1 chromosomal locus. The Gα-subunit is a positive effector of signalling in Sz. pombe, and overexpression of Gpa1 activates the pheromone signalling pathway even in the absence of agonist (Obara et al., 1991).

Although immunoblotting showed that all the transplants were produced at similar levels to Gpa1 (data not shown), we sought to demonstrate that they were all functional and could each interact with the other signalling components in the reporter strains. The human AGS1 protein (activator of G protein signalling) is a receptor-independent activator of heterotrimeric G proteins (Cismowski et al., 1999; Cismowski et al., 2000). It de-fines a distinct member of the superfamily of ras-related proteins and appears to act in a manner similar to a GPCR in terms of its ability to promote nucleotide exchange on Gα-subunits. Its precise role within the cell and the mechanism by which it activates Gα-subunits remain to be resolved. A construct containing the human AGS1 under the control of the nmt1 promoter was introduced into a series of sxa2>lacZ reporter strains containing the different Gα-transplants. AGS1 induced the expression of the lacZ reporter in all strains, suggesting that each Gα-transplant can activate the intracellular signalling machinery (Fig. 2). As expected, AGS1 had no effect in strains lacking Gα-subunits, and its activity was inhibited by a glycine 31 to valine substitution in a region important for guanine nucleotide binding and hydrolysis (Cismowski et al., 1999).

Figure 2.

Activation of Gα-transplants by AGS1. A series of Sz. pombe reporter cells, each containing a different Gα-transplant but all lacking mam2 , was transformed with pREP:AGS1 (JD1751, human activator of G protein signalling 1) or pREP:AGS1 G31V (JD1752, an inactive form of AGS1 containing a glycine 31 to valine substitution in a region important for guanine nucleotide binding and hydrolysis). Cells were grown in the absence of thiamine to induce expression of the AGS1 proteins from the nmt1 promoter, and production of β- galactosidase from each Gα-transplant strain was assayed using ONPG as substrate (white columns, AGS1 G31V ; filled columns, AGS1). Values are means ± SD of triplicate determinations from a typical experiment.

In mammalian cells, AGS1 is specific for Gi and Go (Cismowski et al., 2000), and there is a similar specificity in S. cerevisiae strains containing either intact mammalian Gα-subunits or yeast/mammalian Gα-chimeras possessing mainly mammalian sequences (Cismowski et al., 1999). The mechanism by which AGS1 recognizes Gα-proteins is under investigation, but its ability to activate all the Gα-transplants and the fact that it is most effective against the endogenous Sz. pombe Gpa1 suggest that the mechanism is likely to involve elements other than the last five residues of the Gα-subunit.

Analysing GPCRs in yeast strains containing Gα-transplants

The Gα-transplants were tested for their ability to interact with the Sz. pombe Mam2 P-factor receptor and to allow signalling in response to stimulation by P-factor mating pheromone (Fig. 3). Although sxa2 is not normally expressed in the absence of pheromone stimulation (Imai and Yamamoto, 1992; Ladds et al., 1996), the reporter strains lack a functional cyr1 gene, and the consequent derepression of the pheromone response pathway causes low-level expression of sxa2 (and other pheromone-dependent genes) even in the absence of signalling (Yabana and Yamamoto, 1996). However, the level of β-galactosidase produced in the absence of P-factor was lower in strains expressing Mam2 compared with the equivalent strains containing a vector-only control (Fig. 3A). A similar reduction in ligand-independent signalling in strains containing a GPCR has been observed in S. cerevisiae (for examples, see Brown et al., 2000; Sommers et al., 2000) and, under quite different conditions, in Sz. pombe (Aono et al., 1994). One explanation could be that the receptor inhibits low-level spontaneous activation of the G protein. It is possible, for example, that GPCRs interact with G proteins in the absence of an extracellular agonist and that this interaction maintains the G protein in an inactive state. When the GPCRs are absent, the G proteins are no longer held in their inactive state and are able to activate their target proteins, albeit it to a relatively low level.

Figure 3.

Characterization of the Mam2 P-factor receptor in Gα- transplants. A series of Sz. pombe reporter cells, each containing a different Gα-transplant but all lacking mam2 , was transformed with pREP:Mam2 (JD1627, Sz. pombe P-factor receptor) or pREP (control vector).

A. Cells were grown in the absence of thiamine to induce expression from the nmt1 promoter, and production of β-galactosidase from each Gα-transplant strain was assayed using ONPG as substrate. The β-galactosidase activity for each strain expressing Mam2 is shown as a percentage of the equivalent strain transformed with the control vector (‘no GPCR’). Values are means ± SD of triplicate determinations from a typical experiment.

B. Cells expressing Mam2 were exposed to P-factor at either 0 or 10 −6  M for 16 h, and β-galactosidase production was assayed using ONPG (white columns, no P-factor, filled columns, P-factor at 10 −6  M). Values are means ± SD of triplicate determinations from a typical experiment.

To investigate ligand-dependent activation of the signalling pathway, strains containing the different Gα-transplants and expressing the Sz. pombe Mam2 GPCR were exposed to P-factor and assayed for β-galactosidase activity (Fig. 3B). Although Mam2 appeared to interact with all the Gα-transplants, as judged by its ability to reduce ligand-independent signalling (Fig. 3A), activation of the signalling pathway was only possible with the G16-, Gi3-, Gq- and Gi2-transplants (Fig. 3B). Signalling of Mam2 through the endogenous G protein, Gpa1, is included for comparison.

Specific interaction of GPCRs and Gα-transplants

To investigate the specificity of coupling between GPCRs and the different G proteins in our system, we generated yeast strains expressing receptors for corticotropin releasing factor (CRF). CRF and CRF-related peptides me-diate complex endocrine and behavioural functions (Dautzenberg et al., 2001; Dautzenberg and Hauger, 2002). They regulate the stress response via activation of the pituitary adrenal axis and are involved in the immune response, cardiovascular function, cognitive function, ingestive behaviour and reproductive function. These multiple actions in various tissues are mediated by a series of CRF receptors generated by alternative processing from two distinct genes. By convention, the receptors are numerically ordered with splice variants designated by capital letters: CRF-R1A-D and CRF-R2A-C. Four related peptide agonists have been described in mammals; CRF, urocortin (UCN), urocortin II (UCN II, also called stresscopin related peptide) and urocortin III (UCN III, also called stresscopin) (reviewed by Dautzenberg and Hauger, 2002).

CRF receptors activate a number of different signalling pathways, and regulation of adenylate cyclase (Bilezikjian and Vale, 1983; Chen et al., 1986), phospholipase C (Nabhan et al., 1995; Kiang, 1997) and MAPK cascades (Rossant et al., 1999; Brar et al., 2000) have all been reported (reviewed by Dautzenberg et al., 2001). Pivotal to this flexibility is the ability of receptors to couple to multiple G proteins, the specificity of coupling depending upon the precise combination of receptor, ligand and cell type (Grammatopoulos et al., 1999; 2000; 2001; Karteris et al., 2000).

Strains expressing either the CRF-R1A or the CRF-R2B receptor had lower ligand-independent β-galactosidase activity than the corresponding strains containing a vector-only control (not shown), consistent with the suggestion that both receptors were able to interact with the intracellular signalling machinery. The strains were then exposed to a variety of CRF-related peptides and assayed for β-galactosidase activity (Fig. 4). All ligands were used at 10−6 M as this gave the maximum response for each combination of ligand and receptor, and all β-galactosidase assays were performed 16 h after exposure to the ligand. As expected, the CRF receptors activated multiple G proteins, and the specificity of signalling in the yeast cells was broadly similar to that observed in mammalian systems. CRF signalled through the Gs- and Gi-transplants, consistent with its regulation of adenylate cyclase (reviewed by Dautzenberg and Hauger, 2002), and was more active against the CRF-R1A receptor than against the CRF-R2B receptor (Perrin et al., 1999; Dautzenberg et al., 2001). In contrast, UCN II and UCN III were selective for the CRF-R2B receptors (Hsu and Hsueh, 2001; Lewis et al., 2001; Reyes et al., 2001). UCN, but not CRF, induced signalling through the CRF-R1A receptor and the Gq-transplant. This is consistent with the ability of UCN to use Gq to activate MAP kinase cascades in HEK293 cells, cultured human pregnant myometrial cells (Grammatopoulos et al., 2000) and cardiac myocytes (Brar et al., 2000). Finally, the G16-transplant couples to several ligand–receptor combinations, reminiscent of its ‘promiscuous’ behaviour in mammalian cells (Offermanns and Simon, 1995; Milligan et al., 1996).

Figure 4.

CRF and the specificity of signalling in Gα-transplant strains. A series of Sz. pombe reporter cells, each containing a different Gα-transplant but all lacking mam2 , was transformed with pREP:CRF-R 1A (JD1742, human corticotropin releasing factor receptor, type 1A) or pREP:CRF-R 2B (JD1743, human CRF receptor, type 2B). Cells were exposed to a series of CRF-related ligands at either 0 or 10 −6  M for 16 h; corticotropin releasing factor (CRF), urocortin (UCN), urocortin II (UCN II, also called stresscopin related peptide) and urocortin III (UCN III, also called stresscopin). Production of β-galactosidase was assayed using ONPG (white columns, no ligand, filled columns, ligand at 10 −6  M). Values are means ± SD of triplicate determinations from typical experiments.

Discussion

We have constructed a series of Sz. pombe strains containing Gα-transplants that enable exogenous GPCRs to couple to the yeast pheromone response pathway and lead to the expression of a signal-dependent lacZ reporter gene. The specificity of the system has been demonstrated using two subtypes of the human CRF receptor and a series of CRF-related peptides; different ligand–receptor combinations interact with different Gα-transplants, and the specificity of signalling in the yeast cells is broadly similar to that observed in mammalian systems.

The simplicity of the yeast system makes it possible to investigate the interaction of a receptor with an individual Gα-subunit, without the complicating presence of other Gα-subunits. Such studies are possible in mammalian cells, but they require strategies that label only activated G proteins and the subsequent use of antisera selective for the different Gα-proteins. Not only are the yeast a more convenient alternative, but they offer the unique opportunity of identifying compounds that affect parti-cular signalling pathways. As the function of a particular GPCR depends upon which G protein it activates (Spengler et al., 1993; Robb et al., 1994; Berg et al., 1998; Fitzgerald et al., 1999; Brink et al., 2000; Watson et al., 2000), there is clear value in identifying compounds that differentially affect the coupling of the receptor to the different G proteins. For example, a compound that mimics the effects of UCN on CRF-R1A and activates Gq and the MAP kinase pathway may prevent the death of myocytes in response to hypoxia/ischaemia (Brar et al., 2000; Latchman, 2001). The process of selective activation of G proteins presumably involves the receptor adopting different conformations, but the details are unknown, and the drug discovery process will involve large-scale screening of compound libraries. The yeast would be an ideal host for such screens.

The Gα-transplants could be useful predictors of the physiological coupling specificity of uncharacterized GPCRs (including orphan receptors) and may provide information about potential signalling pathways activated by these receptors. However, it is important to realize the potential limitations of the yeast system. It is clear, for example, that determinants beyond the C-terminus of the Gα-subunit can influence coupling to the receptor (Lee et al., 1995), and the specificity of the Gα-transplant may not be exactly that of the corresponding Gα-subunit. Furthermore, a particular GPCR–Gα-transplant combination may be functional in yeast, but it is possible that the receptor and the G protein may not normally be expressed in the same cell. This is particularly relevant for the ‘promiscuous’ G16-transplant, as the corresponding G16-subunit has a limited tissue distribution.

One of the more interesting observations in this study is the lower level of ligand-independent signalling in strains containing a GPCR that interacts with the signalling machinery compared with strains that lack a receptor (Fig. 3A). Possible explanations are under investigation, but it could be that a GPCR normally maintains the G protein in an inactive conformation and that there is an increased level of spontaneous activation when the receptor is absent. Similar observations have been reported in equivalent S. cerevisiae strains (Brown et al., 2000; Sommers et al., 2000) and previous Sz. pombe reporter strains (Aono et al., 1994). Some GPCRs fail to reduce the ligand-independent signalling in our Sz. pombe strains (not shown) and, without exception, these receptors fail to induce lacZ expression when exposed to their appropriate ligands. Negative results are difficult to interpret, but these GPCRs may not be interacting with the signalling machinery. Such observations may help to solve a conundrum encountered when screening for ligands that affect the activity of newly identified GPCRs: how to be confident that the screen will identify ligands when the lack of a ligand means that it is not possible to demonstrate that the screen works and that the GPCR is functional. A reduction in ligand-independent signalling after expression of a particular GPCR may provide confidence that the receptor is coupled to the signalling machinery. This would encourage the use of the yeast strains in screens to identify ligands that affect the activity of the receptor.

This is the first report of the successful coupling of a human GPCR to the pheromone response pathway in Sz. pombe. The human dopamine D2S (Sander et al., 1994), neurokinin NK2 (Arkinstall et al., 1995) and β2-adrenergic receptors (Ficca et al., 1995) were expressed but failed to couple to the signalling machinery, probably because of nutritional repression of the expression of components of the signalling machinery (Davey and Nielsen, 1994). In contrast, many human GPCRs have been functionally expressed in S. cerevisiae (Pausch, 1997; Reiländer and Weib, 1998). However, not all GPCRs elicit a response after stimulation by appropriate ligands. This is often because the receptor fails to interact with the S. cerevisiae signalling machinery and, although modified Gα-subunits extend the range of GPCRs that can be studied (Brown et al., 2000), it is still not possible to investigate all receptors in this host. Poor expression of the receptor, inappropriate glycosylation or mislocalization within the cell can all prevent a response, as can the inability of some ligands to penetrate the yeast cell wall. Indeed, it will be interesting to compare the relative abilities of the two yeast to functionalize different GPCRs.

In many respects, the two yeast differ from each other as much as each does from human cells. Sz. pombe is phylogenetically distant from S. cerevisiae (Sipiczki, 2000), and several aspects of its cell and molecular biology seem to resemble a higher eukaryotic cell more closely (Forsburg, 1999). Some features may be of particular relevance to the study of GPCRs. The Gα-subunit, for example, is a positive effector of signalling in Sz. pombe (Obara et al., 1991) rather than the negative effector in S. cerevisiae (Nakayama et al., 1988). There is also a more highly developed intracellular membrane system that contains galactosyl transferase (Chappell et al., 1994), which may explain why, in a comparative study, production of the human D2S dopamine receptor was considerably higher in Sz. pombe than in S. cerevisiae (Sander et al., 1994). Ligand accessibility also appears to be less of a problem. Diphtheria toxin is a protein of ≈ 60 kDa that can readily cross the Sz. pombe cell wall and gain access to the cytoplasm (Davey, 1992). S. cerevisiae cells are resistant to the toxin unless the cell wall is removed to form spheroplasts (Murakami et al., 1982). Even relatively small peptides, such as the ligand for the human chemoattractant C5a receptor, can fail to cross the S. cerevisiae cell wall (Baranski et al., 1999). The Sz. pombe strains described in this study should be a useful complement to the other systems currently available for studying human GPCRs.

Experimental procedures

Strains, reagents and general methods

The yeast strains used in this study are listed in Table 1. General yeast procedures were performed as described previously (Davey et al., 1995; Ladds et al., 1996) using lithium acetate for the transformation of yeast. Culture media used were YE (yeast extract; for routine cell growth) and DMM (a defined minimal medium for selective growth and all assays) (Davey et al., 1995). Cell concentrations were determined using a Coulter Channelyser (Beckman Coulter). DNA manipulations were performed by standard methods. Oligonucleotides were synthesized by Alta Bioscience (University of Birmingham, Birmingham, UK). Amplification by the polymerase chain reaction (PCR) used Pwo DNA polymerase (from Pyrococcus woesei) according to the supplier's instructions (Boehringer Mannheim). This polymerase has a 3′−5′ exonuclease (proofreading) activity that reduces the introduction of errors during amplification. All constructs generated by PCR were sequenced by the dideoxynucleotide method using double-stranded DNA as template and a series of oligonucleotide primers designed to generate overlapping sequence data.

Table 1. .Schizosaccharomyces pombe strains.
StrainGenotype
JY546 mat1-M , Δmat2/3::LEU2 , leu1 , ura4-D18 , cyr1-D51 , sxa2 > lacZ ( Didmon et al., 2002 )
JY1168 mat1-M , Δmat2/3::LEU2 , leu1 , ura4-D18 , cyr1-D51 , sxa2 > lacZ , mam2::ura4 +
JY1169 mat1-M , Δmat2/3::LEU2 , leu1 , ura4-D18 , cyr1-D51 , sxa2 > lacZ , mam2-D10
JY1170 mat1-M , Δmat2/3::LEU2 , leu1 , ura4-D18 , cyr1-D51 , sxa2 > lacZ , mam2-D10 , gpa1::ura4 +
JY1157JY1170 with Gs-transplant
JY1158JY1170 with Go-transplant
JY1159JY1170 with Gi2-transplant
JY1160JY1170 with Gi3-transplant
JY1161JY1170 with Gz-transplant
JY1162JY1170 with G12-transplant
JY1163JY1170 with G13-transplant
JY1164JY1170 with G14-transplant
JY1165JY1170 with Gq-transplant
JY1166JY1170 with Gpa1
JY1167JY1170 with G16-transplant

Assay of β-galactosidase activity

Assays were performed using a method modified from Dohlman et al. (1995) by Didmon et al. (2002). Sz. pombe cells were cultured to a density of ≈ 5 × 105 cells ml−1 in DMM, and 500 µl aliquots were transferred to 2 ml Safe-Lock tubes (Eppendorf) containing 5 µl of the appropriate ligand (in HPLC-grade methanol). Tubes were incubated at 29°C for 16 h on a rotating wheel, and 50 µl was transferred to 750 µl of Z-buffer containing 2.25 mM ONPG. Reactions were stopped after 90 min by adding 200 µl of 2 M Na2CO3, and β-galactosidase activity was calculated as optical density at 420 nm (OD420) per 106 cells (determined using a Coulter Channelyser).

Removing the endogenous Mam2 GPCR from the sxa2>lacZ reporter strain

The mam2 gene was a gift from C. Shimoda (Kitamura and Shimoda, 1991) and was cloned as a BglII (position −346 relative to ATG) to HindIII (position 1383, the TAA stop codon is at position 1045) into pSP71. Inverse PCR using the antisense primer JO215 (ggatccTCAGAGGGAGCAAGAAC; introduces a BamHI site immediately upstream of the ATG) and the sense primer JO216 (ggatcCTTACGCCTGAATG TATC; introduces a BamHI site immediately downstream of the TAA stop codon) created JD430 (pSP71:Mam2-D10). A BamHI-digested ura4+ cassette was introduced into the BamHI site of JD430 to generate JD431 (pSP71:Mam2:: ura4+). The mam2::ura4+ disruption fragment was released from JD431 by digestion with BglII and HindIII and used to transform the sxa2>lacZ reporter strain JY546 (Didmon et al., 2002). Southern blot analysis of the Ura+ transformants confirmed disruption of mam2, and the resulting strain (JY1168) was then transformed with the BglII–HindIII fragment from JD430 (pSP71:Mam2-D10) to remove the ura4+ cassette and create JY1169 (M-cells containing the sxa2>lacZ reporter but lacking mam2).

Constructing the Gα-transplant strains

The gpa1 locus of the Sz. pombe chromosome from the EcoRI site (position −676 relative to the ATG for gpa1) to the BglII site (position 1938, the TAG stop codon is at position 1221) was amplified by PCR using the sense primer JO1271 (CTTACGAATTCTAATAGCTCG) and the antisense primer JO1272 (TTTAATAGATCTTCCATACCG). The product was cloned into the EcoRI–BglII sites of pSP71 to generate JD1645 (pSP71:Gpa1). This was used as template for inverse PCR using the antisense primer JO1273 (ggatccGGTGAAAAGGCAGCTACAGATTCC; upper case letters are complementary to positions −1 to −24 of gpa1, the primer introduces a BamHI site immediately upstream of the ATG) and the sense primer JO1274 (ggatCCTTAACTATCTTCAT AATCT; positions 1235–1255, the primer introduces a BamHI site immediately downstream of the TAG stop codon). The PCR product was ligated to itself to generate JD1632 (pSP71:Gpa1-D12). A BamHI-digested ura4+ cassette was introduced into the BamHI site of JD1632 to generate JD1634 (pSP71:Gpa1::ura4+). The gpa1::ura4+ disruption fragment was released from JD1634 by digestion with EcoRI and BglII and used to transform JY1169 (sxa2>lacZ reporter strain lacking the mam2 gene). Southern blot analysis of the Ura+ transformants confirmed disruption of gpa1, and the resulting strain (JY1170) was used as the starting point for generating the Gα-transplants.

To reduce the region to be amplified by PCR, an SpeI–PstI fragment from JD1645 (pSP71:Gpa1, the fragment corresponds to bases 1147–1475 and includes the last 24 codons of gpa1) was cloned into pKS-Bluescript to generate JD1647 (pKS:Gpa1-Ctermin). JD1647 was used as template for inverse PCR using the antisense primer JO1354 (TAGAT TGTTGGACATAATCGTATCTTGAACGG, positions 1206–1175 of gpa1) and a series of sense primers designed to change the last five amino acids of Gpa1 (normally QSLMF): JO1344 for Gq (EYNLV) (GAATATAATCTTGTTTAGATGA ATTTTTCCTTAAC); JO1345 for Gs (QYELL) (CAATAT GAACTTCTTTAGATGAATTTTTCCTTAAC); JO1346 for Go (GCGLY) (GGATGCGGACTTTATTAGATGAATTTTTCCTTA AC); JO1347 for Gi2 (DCGLF) (GATTGCGGACTTTTTTA GATGAATTTTTCCTTAAC); JO1348 for Gi3 (ECGLY) (GAAT GCGGACTTTATTAGATGAATTTTTCCTTAAC); JO1349 for Gz (YIGLC) (TATATTGGACTTTGCTAGATGAATTTTTCCTT AAC); JO1350 for G12 (ENVRF) (GATATTATGCTTCAATAG ATGAATTTTTCCTTAAC); JO1351 for G13 (RLVFR) (CAA CTTATGCTTCAATAGATGAATTTTTCCTTAAC); JO1352 for G14 (EFNLV) (GAATTTAATCTTGTTTAGATGAATTTTTCCT TAAC); and JO1353 for G16 (FKDVR) (GAAATTAATCTT CTTTAGATGAATTTTTCCTTAAC). The changes incorporated the most commonly used codons in gpa1. This selection of changes spans the full diversity of known mammalian Gα-subunits (G11 and Gq are identical in this region, as are Gs and Golf, and Gi1, Gi2 and Gt). The altered SpeI–PstI fragments were used to replace the corresponding region from JD1645 to create a series of full-length Gα-transplant constructs (pSP71:Gα-transplant). Each construct was digested with EcoRI and BglII, and the released fragments were used to transform JY1170 (sxa2>lacZ, mam2-D10, gpa1::ura4+). Southern blot analysis of the Ura transformants confirmed the integration of the Gα-transplants at the gpa1 locus and created the series of Gα-transplant reporter strains.

Preparing the pREP expression constructs

The pREP series of Sz. pombe vectors allows the expression of genes to be under the control of the thiamine-repressible nmt1 promoter (Maundrell, 1993). The Sz. pombe mam2 ORF was amplified from Sz. pombe genomic DNA using the sense primer JO1508 (ATGAGACAACCATGGTGG, position 1–18) and the antisense primer JO1509 (TTACGTCCA CTTTTTAGTTTCAG, position 1047–1025). The PCR product was cloned into the unique EcoRV site of a modified pREP vector to generate JD1627 (pREP:Mam2). The human CRF-R1A ORF was amplified from placenta cDNA using the sense primer JO1283 (ATGGGAGGGCACCCGCAGC, position 1–19) and the antisense primer JO1284 (TCAGACTGCTGTG GACTGC, position 1335–1317). The PCR product was cloned into pREP vector to generate JD1742 (pREP:CRF-R1A). The human CRF-R2B ORF was amplified from placenta cDNA using the sense primer JO1415 (ATGAGGGGTCCCT CAGGGCCCCCAGGCCTCCTC, position 1–33) and the antisense primer JO1416 (TCACACAGCGGCCGTCTGCTTG ATGCTGTGG, position 1317–1287). The PCR product was cloned into pREP vector to generate JD1743 (pREP:CRF-R2B). The human AGS1 ORF was amplified from myometrium cDNA using the sense primer JO1610 (ggggatccAT GAAACTGGCCGCGATGATC, position 1–21, the primer introduces a BamHI site immediately upstream of the ATG) and the antisense primer JO1611 (ggggatcCTAGCTGAT GACGCAGCGCTCCT, position 846–824, the primer introduces a BamHI site immediately downstream of the TAG stop codon). The PCR product was cloned into the BamHI site of pKS to generate JD1750 (pKS:AGS1) and of pREP to generate JD1751 (pREP:AGS1). AGS1G31V, an inactive form containing a glycine 31 to valine substitution in a region important for guanine nucleotide binding and hydrolysis (Cismowski et al., 1999), was prepared by inverse PCR using JD1750 as template, the sense primer JO1612 (Gtc TCGTCCAAGGTGGGCAAG, position 91–111, the primer converts the Gly-31 codon to Val) and the antisense primer JO1613 (GAGGATGACCATGCGATAG, position 90–72). The modified ORF was then cloned as a BamHI fragment into pREP to generate JD1752 (pREP:AGS1G31V).

Acknowledgements

This work was supported by grants from the University Hospitals of Coventry and Warwickshire NHS Trust, and the Cancer Research Campaign (SP1972, K.D.). We thank Olaf Nielsen (University of Copenhagen), Masayuki Yamamoto (University of Tokyo) and Chikashi Shimoda (University of Osaka) for plasmids and strains.

Ancillary