Abstract
Although the promiscuous nature of G16 allows it to interact with numerous G protein-coupled receptors, several Gi-linked receptors are incapable of activating phospholipase C via G16. A series of chimeras between Gα16 and Gαz were constructed and assayed for their ability to mediate receptor-induced stimulation of phospholipase C. Two Gα16/z chimeras harboring 25 or 44 Gαz-specific sequences at their C termini (named 16z25 and 16z44) were capable of responding to 14 different Gi-coupled receptors tested, including those that were either unable to associate with Gα16 (melatonin Mel1c) or activate Gα16 weakly (μ-opioid and type 1 somatostatin). Agonist-induced stimulation of phospholipase C was more efficiently mediated (higher maximal and lower EC50 value) by 16z44 than by Gα16. Both 16z25 and 16z44 were also coupled to Gs- and Gq-linked receptors. Incorporation of Gαz sequence at the N terminus of Gα16 did not further enhance the ability of the chimeras to interact with Gi-coupled receptors. Expression of the various chimeras was verified by immunodetection and functional analysis of their constitutively activated mutants. These results show that the incorporation of α4/β6 and α5 regions of Gαz into a Gα16 backbone can improve the recognition of Gi-coupled receptors. Gα16/zchimeras with expanded capability to interact with Gi-linked receptors may be used to link orphan receptors to the stimulation of phospholipase C.
The G proteins are responsible for the efficient transmission of signals from agonist-bound cell surface receptors to different intracellular effectors (Bourne, 1997). Currently, the total number of G protein-coupled receptors (GPCRs) far exceeds the number of known G proteins. Each member of the four different subfamilies of G proteins must therefore be capable of interacting with multiple GPCRs. Depending on their coupling specificity, most GPCRs are often referred as Gq, Gi, or Gs coupled, which reflects their primary signal transduction pathway. Some G proteins are more promiscuous than others by possessing the ability to interact with a large panel of GPCRs. The most notable examples of promiscuous G proteins are the human G16 and its murine homolog, G15. Both G15 and G16 link a variety of Gq-, Gi-, and Gs-coupled receptors to the stimulation of phospholipase C (PLC; Offermanns and Simon, 1995; Lee et al., 1998). Because of their promiscuity, G15 and G16 are ideal candidates for linking orphan receptors (cloned receptors without a known ligand) to PLC and its downstream effectors. Hence, G16 has received considerable attention as a potential tool for drug discovery (Milligan et al., 1996).
Although G15 and G16 are more promiscuous than other G proteins, they cannot be considered as true universal adapters for GPCRs. For example, the CCR2a chemokine receptor (Kuang et al., 1996), the α1A- and α1C-adrenoceptors (Wu et al., 1992) are unable to recognize G16. Indeed, of the 33 different GPCRs examined to date (Wu et al., 1992, 1993; Offermanns and Simon, 1995; Zhu and Birnbaumer, 1996; Kuang et al., 1996; Lee et al., 1998;Parmentier et al., 1998), at least six receptors are incapable of activating G16. Most of the GPCRs that fail to activate G16 belong to the Gi-coupled receptor subfamily. Because approximately 18% of the total number of Gi-coupled receptors examined to date cannot activate G16, the molecular structure of Gα16 may not be optimal for association to GPCRs with a high preference for Gi proteins. This is perhaps unavoidable if Gα16 were to possess the ability to recognize Gq- and Gs-coupled receptors as well. The Gi-coupled receptors constitute an exceedingly large GPCR subfamily that encompasses many newly discovered receptors such as those for chemokines. This poses a serious concern for the use of G16 as an adapter of orphan receptors in drug-screening protocols.
The overall receptor contact regions on the α-subunit of several G proteins have been mapped (reviewed in Conklin and Bourne, 1993; Onrust et al., 1997), and they are in good agreement with the available crystal structure data (Wall et al., 1995; Lambright et al., 1996). Both the N and C termini are sites for receptor contact. Promiscuity for GPCRs can be increased by altering the amino acid sequence of either the N or C terminus. Modification of the last five residues of Gq (Conklin et al., 1993) or Gs (Liu et al., 1995; Conklin et al., 1996) expanded the number of GPCRs that can use the resultant chimeras to regulate the corresponding effectors beyond those of the parental constructs. Likewise, alterations at the N terminus of Gq resulted in additional coupling to several GPCRs that do not normally use Gq for signal transduction (Kostenis et al., 1998). These studies illustrate the possibility of changing the coupling specificity of G proteins by modifying their putative receptor contact sites on the α-subunit. Many Gi-coupled receptors share the ability to inhibit adenylyl cyclase via the pertussis toxin (PTX)-insensitive Gz (Chan et al., 1995, 1998; Lai et al., 1995;Shum et al., 1995; Tsu et al., 1995a,b; Yung et al., 1995). Thus, the incorporation of Gz-specific sequences into a G16 backbone might improve the ability of the resultant chimera to recognize Gi-coupled receptors. The present study describes the construction and characterization of a series of chimeras between the α-subunits of G16 and Gz(Gα16 and Gαz). Indeed, some of the resultant chimeras has increased promiscuity toward Gi-coupled receptors and are more efficient than G16 in their interactions with GPCRs.
Experimental Procedures
Materials.
The Gα16 cDNA was a kind donation from Dr. M. Simon (California Institute of Technology, Pasadena, CA), and all other cDNAs were constructed or obtained as previously described (Wong et al., 1992; Tsu and Wong, 1996; Lee et al., 1998). Simian kidney fibroblast COS-7 cells were obtained from the American Type Culture Collection (ATCC CRL-1651; Rockville, MD). Various receptor agonists were purchased from Research Biochemicals Inc. (Natick, MA), Sigma Chemical Co. (St. Louis, MO), or Tocris Cookson (Bristol, UK). Sequenase Version 2.0 DNA sequencing kit and ECL enhanced chemiluminescence detection kit were purchased from Amersham Pharmacia Biotech. myo-[3H]Inositol was from DuPont-New England Nuclear (Boston, MA). Plasmid DNA purification columns were obtained from Qiagen (Hilden, Germany). Gαz-specific antisera 3A-170 (C-terminal) and 3-18 (N-terminal) were obtained from Gramsch Laboratories (Schwabhausen, Germany) and Calbiochem (La Jolla, CA), respectively.Taq DNA polymerase, restriction endonucleases, custom mutation primers, and cell culture reagents were obtained from Life Technologies Inc. (Grand Island, NY), and all other chemicals were purchased from Sigma Chemical Co. (St. Louis, MO).
Construction of Chimeras and Mutants.
All of the chimeras were constructed by polymerase chain reactions (PCRs) using human Gα16 and rat Gαz cDNAs (subcloned in the XbaI and EcoRI sites of pcDNAI, respectively) as templates with T7 and SP6 promoter sequences as outer flanking priming regions. A pair of chimeric primers covering both the nucleotide sequences of Gα16 and Gαz were designed as appropriate for each chimeric construct. First, two overlapping fragments that corresponded to the portions of Gα16 and Gαz were generated. The 5′ fragment was made with T7 primer and the reversed chimeric primer, whereas the 3′ fragment was made with the forward chimeric primer and a primer annealed to the SP6 promoter on the vector sequence (SP6 primer). The two PCR products were annealed together, and the full-length fragments were made using the T7 and SP6 primers. Specific primers used for the construction of the various chimeras were listed below, with the nucleotide sequences of Gαz underlined: 16z25/S, CAC TAC ACA TGT GCC ACA GAC ACC AGT AAC ATC; 16z25/AS, GAT GTT ACT GGT GTC TGT GGC ACA TGT GTA GTG; 16z44/S, GGC CCC GAG GGC AGC AAC CGA AAC AAG GAG; 16z44/AS,CTC CTT GTT TCG GTT GCT GCC CTC GGG GCC; 16z66/S, GCT ACC TAT TTC CCC GAG TAC AAG GGT CAG; 16z66/AS, CTG ACC CTT GTA CTC GGG GAA ATA GGT AGC; 30z16/S, GAG AGC CAG CGG CAG CGC GGG GAG CTG AAG; and 30z16/AS, CTT CAG CTC CCC GCGCTG CCG CTG GCT CTC. MgCl2 (1.5 mM) was included in the PCR mixture, and the PCR products were amplified with thermal cycling at 94°C for 60 s, 50°C for 90 s, and 72°C for 90 s for 30 cycles using Robocycler 40 from Stratagene (La Jolla, CA). The 30z16z44 and 30z16z66 chimeras were constructed in a similar manner except one of the half-products was amplified by PCR using either the 16z44 or 16z66 chimera as the template. Full-length chimeric α-subunit cDNAs were subcloned into either pcDNA3 or pcDNA3.1Zeo(+) mammalian expression vectors (InVitrogen, San Diego, CA). DNA sequences of the mutants were checked by dideoxynucleotide sequencing method using Sequenase V2.0 kit and restriction mapping.
A GTPase-deficient mutant of Gα16(Gα16QL) with Gln212mutated into Leu was constructed essentially as described previously (Qian et al., 1994). Functional constitutive activity of Gα16QL was confirmed by transiently expressing the construct in COS-7 cells and determining the basal PLC activity in the transfectants. A HindIII/XcmI fragment from Gα16QL was used to construct the GTPase-deficient mutants of 16z25, 16z44, and 16z66. The resultant mutant chimeric constructs were verified by restriction mapping.
Transfection of COS-7 Cells.
Simian kidney fibroblast COS-7 cells were cultured with Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% FCS (v/v), 50 U/ml penicillin, and 50 μg/ml streptomycin at 37°C in humidified air with 5% CO2. Then, 1 × 105cells/well were seeded onto 12-well plates the day before transfection. DEAE-dextran-mediated transfection was performed as described previously (Wong, 1994). Briefly, appropriate amounts of various DNA samples purified by Qiagen column chromatography were mixed with growth medium containing 250 μg/ml DEAE-dextran and 100 μM chloroquine. Cells were incubated with the transfection cocktails for ∼3.5 h and then shocked for 1 min at room temperature in PBS containing 10% dimethyl sulfoxide (v/v). After rinsing with PBS, the cells were returned to growth media for 24 h. Approximately 50% of the cell population will take up the cDNAs as indicated by cotransfecting a plasmid DNA encoding β-galactosidase as a reporter.
Inositol Phosphates (IP) Accumulation Assay.
An aliquot (750 μl) of inositol-free DMEM containing 5% FCS and 2.5 μCi/mlmyo-[3H]inositol was added to each well of transfected COS-7 cells and incubated for 18 to 24 h. The labeling media were subsequently replaced by 1 ml of inositol phosphate assay medium (DMEM buffered with 20 mM HEPES, pH 7.5, containing 20 mM LiCl) for 10 min, and then 1 ml of IP assay medium containing the appropriate agonist was added to the cells for an additional 1 h at 37°C. Reactions were stopped by adding 750 μl of ice-cold 20 mM formic acid and stored at 4°C for 30 min. [3H]IP were separated from other labeled inositol species by sequential ion-exchange chromatography as described previously (Tsu et al., 1995a).
Membrane Protein Preparation and Immunodetection of Recombinant Proteins.
COS-7 cells were grown on 150-mm dishes to 70 to 80% confluence. Transfection was performed as on 12-well plates with proper adjustments to the volumes and amounts of the reagents used. After 48 h in normal growth conditions, cells were washed with Ca2+/Mg2+-free PBS and harvested with 5 ml of Ca2+/Mg2+-free PBS containing 10 mM EDTA. The following procedures were performed at 4°C. Cells were spun down briefly (200g, 5 min); resuspended in hypotonic lysis buffer (50 mM Tris · HCl, 2.5 mM MgCl2, 1 mM EGTA, 0.1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine-HCl, and 1 mM dithiothreitol, pH 7.4); and lysed by one cycle of freeze-thawing followed by 10 passages through a 27-gauge needle. Nuclei were removed by brief spinning, and membranes were collected by spinning the supernatants at 15,000g for 15 min. Membrane pellets were finally resuspended in lysis buffer. Protein concentrations were determined using the Bio-Rad Protein Assay Kit. For immunodetection, 50 μg of each membrane protein sample was resolved on a 10% SDS-polyacrylamide gel and transferred to polyvinylidene difluoride membranes through electroblotting. Protein molecular weight markers were visualized by Coomassie blue staining. Several chimeras were detected by the Gαz-specific antiserum from Calbiochem and with enhanced chemiluminescence (ECL kit; Amersham).
Data Analysis.
[3H]IP was estimated by determining the ratios of [3H]IP to [3H]inositol plus [3H]IP as previously described (Tsu et al., 1995a). Absolute values for IP accumulation varied between experiments, but variability within a given experiment was less than 10% in general. Unless otherwise stated, data shown in the figures and tables represent the mean ± S.E. of three or more independent experiments performed in triplicate. Bonferroni's t test with 95% confidence limit was adopted to verify the significance between different treatment groups within the experiments.
Results
Design of C-Terminal Chimeras.
Although multiple regions in the primary structure of Gα16 are required for receptor coupling (Lee et al., 1995), the molecular determinants for the promiscuous property of Gα16 has not been fully delineated. Numerous studies on other α-subunits have implicated the C-terminal tail of the α-chain as one of the major receptor contact regions (Sullivan et al., 1987; Conklin et al., 1993,1996; Tsu et al., 1997). As a first step toward enhancing the ability of Gα16 to recognize Gi-coupled receptors, we constructed a series of Gα16/z chimeras by incorporating different lengths of Gαz sequences at the C terminus of Gα16. Gαz was selected as a donor for the construction of the chimeras because it recognizes practically all Gi-coupled receptors (Wong et al., 1992; Chan et al., 1995, 1998; Lai et al., 1995; Shum et al., 1995; Tsu et al., 1995a,b; Yung et al., 1995) but is insensitive to PTX. Signals transduced by chimeric Gα16/z can be easily discerned from Gi-mediated signals with the use of PTX. Based on the alignment of the C-terminal sequences of Gα16 and Gαz (Fig.1), we selected three junctional sites for the construction of Gα16/z chimeras.
Because the α5 helix is a known contact region for receptors (Lichtarge et al., 1996), we replaced the entire α5 helix of Gα16 with that of Gαz. Based on the crystal structures of Gαt1(Lambright et al., 1996) and Gαi1 (Wall et al., 1995), the α5 helix of Gα16 is predicted to be composed of the last 25 residues. The resultant chimera was therefore named 16z25; for 16z25 and subsequent chimeras, the number following the letter z indicates the number of Gαz residues present in the C terminus of the construct.
A unique structural feature of Gα16 (and Gα15) is an insertion of 11 residues (amino acids 326–336; based on the alignment of all mammalian Gα-subunits using the Clustal X program), which is absent in all other Gα-subunits, between the α4 helix and the β6 strand. To test whether this insertion is critical for the promiscuity of Gα16, we constructed the 16z44 chimera. In this chimera, half of the α4/β6 insertion was replaced by Gαz residues without shortening the insertion (Fig. 1). Last, we replaced the α4/β6/α5 domains of Gα16 with the cognate regions of Gαz by creating the 16z66 chimera with a junctional site between the αG and α4 helices (Fig. 1). The 16z66 chimera has approximately 20% of the C-terminal residues of Gα16 substituted by those of Gαz and is shorter than Gα16 by 12 amino acids. No epitope tag was engineered into the chimeras in case it disrupts receptor recognition.
Functional Coupling of Gα16/z Chimeras to δ-Opioid Receptor.
We used a well established transient expression system to examine the ability of the Gα16/z chimeras to interact with Gi-coupled receptors. We have previously shown that coexpression of the δ-opioid receptor (a typical Gi-coupled receptor) and Gα16 in COS-7 cells permits the δ-selective opioid agonist [d-Pen2,d-Pen5]enkephalin (DPDPE) to stimulate the formation of IP (Lee et al., 1998). Here, we adopted the same approach to study the Gα16/zchimeras. In accordance with our earlier report (Lee et al., 1998), 100 nM DPDPE stimulated IP formation by 3-fold in COS-7 cells coexpressing the Gα16 and the δ-opioid receptor (Fig.2). In contrast, agonist treatment failed to evoke formation of IP in COS-7 cells cotransfected with cDNAs encoding the δ-opioid receptor and Gαz (Fig.2). In COS-7 cells coexpressing the δ-opioid receptor with either 16z25 or 16z44, 100 nM DPDPE stimulated the formation of IP by 3- to 4.5-fold (Fig. 2). Interestingly, the DPDPE response was significantly higher in cells transfected with the 16z44 cDNA than those expressing 16z25 or Gα16 (Fig. 2 and Table1).
Unlike 16z25 and 16z44, the 16z66 chimera failed to mediate the DPDPE response under identical experimental conditions (Fig. 2). Hence, we checked the expression of Gα16/z chimeras by Western blot analysis. Because most of the Gα16/z chimeras contained Gαz C-terminal sequences, we could not use commercially available Gα16-specific C-terminal antiserum to verify the expression of these chimeras. Instead, we used a Gαz-specific antiserum for the immunodetection of the 16z25, 16z44, and 16z66 chimeras. As shown in Fig. 3, all three chimeras were detected by the Gαz-specific antiserum 3A-170 in membranes prepared from COS-7 cells transfected with the chimeras.
Next, we tested the possibility that 16z66 has lost the ability to stimulate PLC by using a constitutively active mutant of 16z66. Mutation at codon 212 of Gα16 has been shown to constitutively activate PLC in Swiss 3T3 cells (Qian et al., 1994). Expression of Gα16QL (harboring the Q212L mutation) in COS-7 cells also led to increased basal accumulation of IP, whereas the expression of Gα16 or GαzQL (Wong et al., 1992) did not affect the PLC activity (Fig. 4). The constitutively active mutants of the Gα16/z chimeras were similarly expressed in COS-7 cells. Except for 16z66QL, all Gα16/z chimeras harboring the Q212L mutation constitutively stimulated PLC activity by 3- to 6-fold above those of their corresponding wild-type chimeras (Fig. 4). The lack of constitutive activity of 16z66QL was not due to a deficiency in expression because 16z66QL was expressed to the same level as 16z25QL and 16z44QL (Fig. 3). These results support the notion that 16z66 cannot stimulate PLC.
Promiscuity of 16z25 and 16z44.
The ability of 16z25 and 16z44 to interact productively with the δ-opioid receptor prompted us to further investigate into their capacity to functionally associate with other Gi- and Gs-coupled receptors. COS-7 cells were cotransfected with Gα16, 16z25, or 16z44 and a receptor (0.25 μg/ml per construct) chosen from a panel of Gi- or Gs-coupled receptors that were available in our laboratory. The selected receptors include the adenosine A1, α2- and β2-adrenergic, complement C5a, dopamine D1 and D2, formyl peptide, luteinizing hormone, opioid receptor-like (ORL1), vasopressin V2, somatostatin-1 and -2 (SSTR1 and SSTR2), three subtypes each of melatonin (1a, 1b, and 1c) and opioid (μ, δ, and κ) receptors (Table 1). Transfected cells were assayed for IP formation in the absence or presence of saturating concentrations of the appropriate agonists. All 14 Gi-coupled receptors examined were capable of activating 16z25 and elicit agonist-induced PLC activation (Table 1). The magnitude of PLC stimulation by the various receptors ranged from 1.5- to 4.5-fold. Activation of 16z25 by aminergic receptors (dopamine and melatonin receptors) resulted in up to 3.5-fold stimulation of PLC activity. The receptors for peptide ligands gave slightly higher responses in general. Similar results were obtained with the 16z44 chimera (Table 1). All of the Gi-coupled receptors tested were efficiently coupled to 16z44 and stimulated PLC. The 16z44-mediated PLC responses ranged from 1.7- to 5.5-fold stimulation. It should be noted that none of the Gi-coupled receptors, except SSTR2, was capable of stimulating IP production in the absence of 16z25 or 16z44 (unpublished data; Lee et al., 1998). Collectively, these results indicate that both 16z25 and 16z44 are capable of linking a large variety of Gi-coupled receptors to the stimulation of PLC. In terms of absolute amount of IP formation, 16z25 was generally less efficient than 16z44 in transducing signals from Gi-coupled receptors. When Gi-coupled receptors such as the formyl peptide, melatonin Mel1a, and ORL1 receptors were tested against the 16z66 chimera, no functional coupling could be detected.
Compared with Gα16, the 16z25 and 16z44 chimeras exhibited enhanced capability to interact with Gi-coupled receptors. Most notable of all was their ability to interact productively with the melatonin Mel1c receptor. The Mel1c was unable to activate Gα16, whereas both Mel1a and Mel1b receptors stimulated PLC via Gα16 (Table 1). Other examples of enhanced linkage to Gi-coupled receptors by the 16z25 and 16z44 chimeras included the μ-opioid receptor and SSTR1. Among the three opioid receptor subtypes, activation of Gα16 by the μ-opioid receptor is relatively weak (Lee et al., 1998). In terms of both the absolute activity and the fold of stimulation, μ-opioid receptor-induced responses were more robust with the 16z25 and 16z44 chimeras (Table 1). As for SSTR1, 100 nM somatostatin stimulated IP formation by only 2-fold in COS-7 cells coexpressing Gα16, whereas the same concentration of agonist elicited 3.5-fold of stimulation in cells coexpressing either 16z25 or 16z44 (Table 1).
The 16z44 chimera seemed to produce greater stimulations of PLC activity on receptor activation compared with either Gα16 or 16z25 (Fig. 2 and Table 1). The magnitudes of the agonist-induced responses mediated via 16z44 were generally higher than those obtained with Gα16or 16z25. Hence, we examined the efficiency of coupling between 16z44 and several Gi-coupled receptors. The melatonin Mel1c, SSTR1, and δ-opioid receptors were chosen on the basis of their varying abilities to associate with Gα16. Each receptor was coexpressed with either Gα16or the 16z44 chimera in COS-7 cells and assayed for IP accumulation in response to increasing concentrations of the corresponding agonist. Activation of the melatonin Mel1c receptor by 2-iodomelatonin did not stimulate PLC activity in cells coexpressing the receptor and Gα16 (Fig. 5). In the presence of 16z44, however, 2-iodomelatonin dose-dependently stimulated the formation of IP with an EC50 value of ∼0.4 nM (Fig. 5). Likewise, the SSTR1 was weakly coupled to Gα16 (Fig. 5 and Lee et al., 1998), whereas it efficiently stimulated the PLC activity in the presence of 16z44 with an EC50 value of ∼3 nM somatostatin (Fig. 5); the EC50 of somatostatin for Gα16 cotransfected cells was ∼20 nM. The δ-opioid receptor has been shown to activate Gα16 in a dose-dependent manner (Lee et al., 1998). Replacement of Gα16 by the 16z44 chimera resulted in a more efficient stimulation of PLC by DPDPE (Fig. 5). The 16z44-mediated DPDPE response gave a higher maximal stimulation (twice that of the Gα16-mediated response) and a reduced EC50 value (10 versus 40 nM). Taken together, these studies showed that the 16z44 chimera exhibited enhanced linkage to Gi-coupled receptors compared with its parental Gα16. Preliminary results suggest that such enhanced linkage can be extended to include the CCR1, CCR2b, and CCR5 chemokine receptors.
Coupling of 16z25 and 16z44 to Gs- and Gq-Linked Receptors.
A distinguishing feature of Gα16 is its ability to link a large number of GPCRs to the stimulation of PLC, including those receptors that normally utilize Gs for signal propagation. To confirm that the Gα16/z chimeras can also recognize Gs-coupled receptors, we assessed the ability of 16z25 and 16z44 to interact productively with four different Gs-linked receptors. COS-7 cells were transiently cotransfected with either 16z25 or 16z44 and a Gs-coupled receptor (β2-adrenergic, dopamine D1, luteinizing hormone, or vasopressin V2). Transfected cells were subsequently challenged with the appropriate agonists. In cells coexpressing the β2-adrenergic, dopamine D1, or vasopressin V2receptors with either 16z25 or 16z44, activation of the receptor led to increased production of IP (Table 1). Among these three receptors, only the vasopressin V2 receptor has the ability to utilize endogenous Gq to weakly stimulate IP formation (66.5 ± 17.8% over basal, n = 6). The magnitudes of agonist-induced stimulations mediated via 16z25 or 16z44 were all ∼3-fold and were generally lower than those observed with Gα16 in previous reports (Offermanns and Simon, 1995; Lee et al., 1998). In contrast, activation of the luteinizing hormone receptor did not significantly stimulate IP formation in the presence of either 16z25 or 16z44 (Table 1).
The inability of the luteinizing hormone receptor to interact with 16z25 and 16z44 was unlikely to be due to lack of receptor expression because it was functionally associated to stimulation of adenylyl cyclase in the transfectants (data not shown). Moreover, under identical experimental conditions, coexpression of luteinizing hormone receptor and Gα16 allowed the transfected cells to produce a 3-fold stimulation of PLC in response to 1 μg/ml human choriogonadotropin (Table 1). The lack of coupling of 16z25 and 16z44 to the prostanoid DP and luteinizing hormone receptors suggests that even though these chimeras exhibited enhanced ability to recognize Gi-coupled receptors, their linkage to Gs-coupled receptors seemed to be impaired. We have also examined the ability of 16z66 to interact with the Gs-coupled β2-adrenergic receptor. The β2-adrenergic receptor was coexpressed with one of the three chimeras in COS-7 cells. The 16z66 chimera was not able to stimulate IP formation in response to 10 μM isoproterenol, despite the fact that the same concentration of agonist potently stimulated cAMP accumulation.
Productive coupling of Gα16 to Gq-linked receptors usually leads to a potentiation of agonist-induced stimulation of PLC when Gα16 is coexpressed (Offermanns and Simon, 1995). We adopted the same approach to examine the ability of 16z25 and 16z44 to recognize Gq-coupled receptors and selected the bombesin, serotonin1c, and muscarinic m1 receptors for the study. Each of the three Gq-coupled receptors was transiently expressed in COS-7 cells in the absence or presence of 16z44. Agonist-induced activation of these receptors significantly increased IP formation even in the absence of 16z44 (Table 2). Such stimulations were mediated through endogenous Gq/11 proteins. The magnitude of the agonist-induced responses became greater by 35 to 55% in cells coexpressing the 16z44 chimera (Table 2). Similar results were obtained for the 16z25 chimera (data not shown). These studies demonstrated that both 16z25 and 16z44 chimeras retained the ability to interact with Gq-coupled receptors.
Construction and Characterization of N-terminal Gαz/16 Chimeras.
Recent studies by Wess and coworkers (Kostenis et al., 1998) have focused on the extreme N-terminal region of Gαq as a determinant for the selectivity of receptor coupling. Compared with the α-subunits of the Gi subfamily, Gαq and Gα11 are longer by six residues at their N termini. Progressive deletion or substitution with alanine of these “extra” residues produced Gαq mutants that can effectively interact with Gi-coupled receptors (Kostenis et al., 1998). Alignment of the Gα16 and Gαz sequences revealed that their predicted N-terminal α-helices share little homology and that the Gα16 N terminus is longer than that of Gαz by nine residues (Fig. 1). To assess whether the N terminus of Gαz is required for efficient coupling to Gi-linked receptors, we replaced the entire αN helix with that of the first 30 residues of Gαz. The resultant 30z16 chimera is shorter than Gα16 by nine residues (Fig. 1).
As with the Gα16/z chimeras, we tested the ability of the 30z16 chimera to interact with the δ-opioid receptor in COS-7 cells. The δ-agonist DPDPE doubled the IP formation in COS-7 cells coexpressing the δ-opioid receptor and 30z16 (Fig.6). The 30z16-mediated stimulation was lower than that obtained with Gα16. Replacement of the αN helix of Gα16 by the cognate region of Gαz seemed to impair the ability of the chimera to interact with the δ-opioid receptor. Additional experiments showed that 30z16 was also capable of transducing stimulatory signals from the ORL1 and β2-adrenergic receptors to PLC (Fig.7). However, GPCRs such as the melatonin Mel1c and μ-opioid receptors, which are weak or ineffective activators of Gα16, could not stimulate PLC via 30z16 (data not shown). Although the 30z16 chimera can interact with both Gi- and Gs-coupled receptors, its promiscuous property was compromised.
Because the N and C termini of the Gα-subunit are in close proximity, the ability of Gα16/z chimeras to recognize Gi-coupled receptors may be enhanced by having Gαz-specific sequences at both termini of the chimera. Two “z-16-z” chimeras were constructed. We constructed the 30z16z44 and 30z16z66 chimeras to examine whether the inclusion of a Gαz-specific αN helix can enhance or rescue the ability of 16z44 and 16z66 to recognize Gi-coupled receptors. The 30z16z44 chimera behaved like Gα16, whereas 30z16z66 remained unable to respond to agonist-activated δ-opioid receptor (Fig. 6). One interesting observation is that the 30z16z44 chimera was no better than Gα16 in coupling to the δ-opioid receptor but was markedly better than the 30z16 chimera. The magnitude of DPDPE-induced PLC stimulation was actually lower with 30z16z44 than when it was mediated via 16z44 (cf. Fig. 2). Further experiments showed that the 30z16z44 chimera can interact productively with the ORL1 and β2-adrenergic receptors (Fig. 7), as well as the μ-opioid, κ-opioid, and the three melatonin receptors (data not shown). The expression of 30z16, 30z16z44, and 30z16z66 was confirmed by immunoblot analysis using a Gαz-specific N-terminal antiserum (Fig.8). Thus, the inability of 30z16z66 to interact with GPCRs was not due to a lack of expression. Overall, replacement of the αN helix of Gα16 by Gαz-specific sequence did not seem to enhance the specificity of coupling to Gi-linked receptors.
Effects of Inverse Agonists on Receptors Coupled to 16z44.
A notable feature of cells coexpressing a GPCR and 16z25 or 16z44 was their elevated basal IP production. The increased basal IP accumulation can be seen with the vast majority of the receptors tested (Table 1), and it resembles the constitutive activity of GPCRs. This interpretation is supported by the fact that no elevation of basal activities could be observed in COS-7 cells expressing 16z44 alone; coexpression with a Gi- or Gs-coupled receptor is required (Y. H. Wong, unpublished results). If the enhanced linkage of the chimeras to GPCRs promotes the formation of constitutively active receptors, it may offer an opportunity to identify inverse agonists that act at various GPCRs. To test this hypothesis, COS-7 cells were cotransfected with cDNAs encoding 16z44 and either the δ-opioid or β2-adrenergic receptor, and the ability of known inverse agonists to suppress the elevated basal IP levels of the transfectants was examined. Compared with cells coexpressing the δ-opioid receptor and Gα16, 16z44 transfectants exhibited increased basal IP production, but neither ICI-174,864 (an inverse agonist) nor naloxone (a neutral antagonist) affected the elevated basal levels (Fig.9). The expression of δ-opioid receptors in the transfectants was confirmed by the DPDPE-induced stimulation of PLC (Fig. 9). Similar results were obtained with cells coexpressing the β2-adrenergic receptor and 16z44. Two inverse agonists of β2-adrenergic receptor, ICI-118,551 and timolol, were incapable of reducing the elevated basal level associated with the coexpression of 16z44 (Fig.9). These results suggest that although 16z44 may provide enhanced linkage to GPCRs, it does not necessarily promote the formation of constitutively active GPCRs.
Discussion
As one of the largest protein families found in nature, it is estimated that several thousand different GPCRs may exist in the human genome. The recent discovery of more than 1000 genes encoding known and orphan GPCRs in the Caenorhabditis elegans genome (Bargmann, 1998) lends further credibility to this estimation. Human Gα16 possesses the rare ability to recognize a wide spectrum of GPCRs, which can facilitate the characterization of orphan GPCRs. Numerous biochemical, structural, and molecular genetic studies have revealed that the docking site for receptors is composed of multiple regions on the Gα-subunit (reviewed by Bourne, 1997). The five regions of the Gα-subunit involved in receptor recognition are the α2 helix, the β6-α5 loop, the α5 helix, and the two extreme termini. By replacing one or more of these regions in Gα16 with sequences from Gαz, we have successfully created chimeric Gα16/z proteins that exhibit enhanced linkage to Gi-coupled receptors. The 16z44 chimera appeared to be particularly effective in this respect.
Compared with Gα16, both 16z25 and 16z44 can additionally couple to the melatonin Mel1c receptor and produce higher magnitudes of PLC stimulation with some of the Gi-coupled receptors (e.g., μ-opioid and SSTR1). In the 16z25 chimera, the α5 helix of Gα16 was replaced by that of Gαz. Numerous studies have attested to the importance of the α5 helix in receptor coupling. Amino acids within the α5 helix of Gαs (Conklin et al., 1996;Sullivan et al., 1987), Gαi (Tsu et al., 1997), Gαt1 (Martin et al., 1996), and Gαq (Conklin et al., 1993, 1996) have been shown to alter the specificity of receptor coupling. The α5 helix of Gαz is quite different from that of Gα16 with less than 35% homology (Fig. 1). Alignment of the α5 helices of Gα16 and Gαz shows that every third or fourth amino acid from position −1 are different between the two Gα-subunits. Using the crystal structures of Gαt1 and Gαi1 as a basis, we generated a molecular model to highlight the structural differences in the α5 helices of Gα16 and Gαz (Fig.10A). When the Gαz-specific residues are superimposed on the α5 helix of Gα16, the receptor-facing plane are predominantly composed of Gαz-specific amino acids.
Not only did 16z44 recognize more Gi-coupled receptors than Gα16, the magnitudes of the stimulations were also higher. Comparison of the EC50 values obtained with 16z44 and Gα16 transfectants reflected that 16z44 was more efficient in linking the Gi-coupled receptors to the activation of PLC. Another notable feature of 16z44 is its elevated basal activities. Cells coexpressing 16z44 generally exhibited basal PLC activities that were much higher than those coexpressing Gα16 (Table 1). This elevation in basal IP accumulation resembles the constitutive activity of GPCRs, but we were unable to detect inverse agonist effects on the δ-opioid and β2-adrenergic receptors. As a result of elevated basal activities and in terms of percentage stimulation, 16z44-mediated responses were not significantly better than those mediated by Gα16. However, the absolute levels of IP accumulation transduced by 16z44 were often greater than those of Gα16. Basal IP accumulation was also higher in cells coexpressing the 16z25 chimera, albeit to a lesser extent.
Structural differences between 16z25 and 16z44 are primarily located in the α4/β6 loop (residues 318–335 of Gα16). The α4/β6 loop is one of several secondary structures forming the contact surface for receptors (part of the A1 cluster as described byLichtarge et al., 1996). The α4 helix and α4/β6 loop region of Gαi1 are important for specific recognition of receptors (Bae et al., 1997). The α4/β6 loop of 16z25 is identical with that of Gα16 because there is no substitution of Gαz-specific residues in this region of 16z25 (Fig. 1). Based on the crystal structure coordinates of trimeric Gt1 (Lambright et al., 1996), the α4/β6 loop of 16z44 is predicted to be a more flexible structure with an energy level higher than those of 16z25 and Gα16 (Fig. 10B). This increased flexibility may partially account for the enhanced potency of 16z44 to transduce signals from Gi-coupled receptors. The model generated for 16z66 (Fig. 10B) fitted well to the known structures of Gαt1 and Gαi1. Compared with Gα16, 16z66 has a smaller α4/β6 loop and a tighter α4 helix. One would presume that the close resemblance of 16z66 to Gαt1 and Gαi1 implies enhanced capability of the chimera to interact with Gi-coupled receptors, yet our results were contrary to such a prediction.
Because 16z66 was expressed to the same extent as the other chimeras, its inability to mediate receptor-induced stimulation of PLC might be due to the disruption of effector recognition domains like the α4/β6 region. The α4 and the α4/β6 loops of Gαq are known to be involved in the activation of PLC (Arkinstall et al., 1995; Medina et al., 1996). When the putative PLC regulatory domains of Gαq are mapped onto a molecular model of Gα16 (shown in blue, Fig. 10C), these regions were not substituted by Gαz-specific residues in 16z66. Several residues around the switch III region of Gαqhave also been shown to be critical for the regulation of PLC (Venkatakrishnan and Exton, 1996), and they are conserved in Gα16: residues 241 to 243 and 254 to 255 (Fig.10C). We used molecular modeling of Gα16 to identify amino acids distal to or in the α4 helix that may interact with these PLC-activating residues. Residue Leu254 is predicted to interact with several amino acids in the α4 helix (Ile312, Met315, Tyr316, and Thr317) and the α4/β6 loop (Asp325). Four of these five potential sites (except Ile312) were actually replaced by Gαz-specific residues in 16z66 (Fig. 1). Like mutating Leu254 itself, the disruption of intramolecular interactions may severely curtail the ability of 16z66 to stimulate PLC, resulting in a lack of constitutive activity of 16z66QL. Another possibility is that 16z66 cannot adopt the GTP-bound active form. Further experiments are needed to distinguish the molecular basis for the lack of function of 16z66.
Although we did not embark on an extensive study to determine the importance of N-terminal sequences of Gαz in receptor recognition, the present study shows that the αN helix of Gαz alone could not confer specificity for Gi-coupled receptors. This is in stark contrast to results obtained in similar studies where the N terminus of Gαz alone was sufficient to allow a Gαz/t1 chimera to respond to the δ-opioid receptor (Tsu et al., 1997). Close proximity of the two termini in the crystal structures of Gαt1 (Lambright et al., 1996) and Gαi1 (Wall et al., 1995) supports the involvement of the N terminus in receptor recognition. Biochemical evidence are also available to substantiate this notion for the coupling of receptors to Gαo (Denker et al., 1995) and Gαt1 (Dratz et al., 1993), two members of the Gi subfamily. Hence, an intact N terminus may be required for Gα16 to efficiently associate with GPCRs. In this respect, our results are in accordance with those reported by Lee et al. (1995), where the N-terminal 209 residues of Gα16 were found to be essential for activation by the Gi-coupled C5a receptor.
Despite their enhanced linkage to Gi-coupled receptors, neither 16z25 nor 16z44 can be considered as a universal adapter for GPCRs. Compared with Gα16, the ability of 16z25 and 16z44 to recognize the luteinizing hormone receptor was actually diminished. Naturally, one would expect that when the specificity for Gi-coupled receptors increases in a Gα-subunit, its specificity for Gs-linked receptors will be inversely affected because the two sets of GPCRs are designed to produce opposite effects on adenylyl cyclase. Perhaps it is impractical to engineer a universal G protein adapter for GPCRs even though such a construct is highly desirable for the characterization of orphan receptors. Nevertheless, our study has successfully demonstrated the feasibility of improving the recognition of specific subsets of GPCRs by altering receptor contact regions on Gα16. With the recent explosion of the number of orphan receptors being cloned, the chimeras described herein may become invaluable tools for their characterization. For example, the 16z44 chimera can be incorporated into a variety of cell-based assays for the rapid detection of receptor activation. Conceivably, the ability of Gα16 to recognize Gs-coupled receptors can be similarly enhanced by incorporating Gαs-specific regions on a Gα16 backbone. Several chimeric Gα16 with expanded capability of receptor recognition may collectively serve as a true “universal adapter” for orphan GPCRs.
Acknowledgments
We are extremely grateful to those who made the various cDNAs available for this study: Drs. G. Bell, H. R. Bourne, C. Evans, Y. Kaziro, S. Reppert, M. I. Simon, and D. Segaloff.
Footnotes
-
Send reprint requests to: Dr. Yung H. Wong, Department of Biology, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China. E-mailboyung{at}ust.hk
-
↵1 Present address: Center for Biotechnology, NorthWestern University, 1801 Maple Avenue, Evanston, IL 60201.
-
This work was supported by the Research Grants Council of Hong Kong (HKUST 6096/98M).
- Abbreviations:
- GPCR
- G protein-coupled receptor
- DPDPE
- [d-Pen2,d-Pen5]enkephalin
- PLC
- phospholipase C
- PCR
- polymerase chain reaction
- PTX
- pertussis toxin
- IP
- inositol phosphates
- DMEM
- Dulbecco's modified Eagle's medium
- Received April 22, 1999.
- Accepted September 3, 1999.
- The American Society for Pharmacology and Experimental Therapeutics