Interaction of the Xanthine Nucleotide Binding Goα Mutant with G Protein-coupled Receptors*

We constructed a double mutant version of the α subunit of Go that was regulated by xanthine nucleotides instead of guanine nucleotides (GoαX). We investigated the interaction between GoαX and G protein-coupled receptors in vitro. First, we found that the activated m2 muscarinic cholinergic receptor (MAChR) could facilitate the exchange of XTPγS for XDP in the GoαXβγ heterotrimer. Second, the GoαXβγ complex was able to induce the high affinity ligand-binding state in the N-formyl peptide receptor (NFPR). These experiments demonstrated that GoαX was able to interact effectively with G protein-coupled receptors. Third, we found that the empty form of GoαX, lacking a bound nucleotide and βγ, formed a stable complex with the m2 muscarinic cholingeric receptor associated with the plasma membrane. Finally, we investigated the interaction of GoαX with receptor in COS-7 cells. The empty form of GoαX bound tightly to the receptor and was not activated because XTP was not available intracellularly. We tested the ability of GoαX to inhibit the activities of several different G protein-coupled receptors in transfected COS-7 cells and found that GoαX specifically inhibited Go-coupled receptors. Thus the modified G proteins may act as dominant-negative mutants to trap and inactivate specific subsets of receptors.

Hundreds of seven-transmembrane receptors activate heterotrimeric G proteins and transduce signals across cell membranes in eukaryotic cells. The stimulated receptors catalyze the exchange of GTP for GDP bound to G protein ␣ subunits. Activated GTP-bound ␣ subunits and free ␤␥ subunits regulate a variety of cellular effectors including enzymes and ion-channels (1)(2)(3). Signaling is normally initiated by the binding of agonist to receptor, which stabilizes the receptor in an active conformation. Receptors function to stimulate the dissociation of GDP bound to the G protein ␣ subunits. The subsequent binding of GTP to the empty ␣ subunit promotes the conformational change of G␣ and dissociation of the ␤␥ subunits. The G protein ␣ subunit in the nucleotide-free state appears to be an important intermediate in the activation. From studies of rhodopsin and transducin, it has been postulated that the empty G protein (nucleotide-free) forms a stable complex with the receptor (4). Both empty forms of Gi and Go ␣ subunits have been purified under harsh conditions (1 M (NH 4 ) 2 SO 4 and 20% glycerol), and they were unstable (5).
We recently reported that a mutant version of Go␣, Go␣X (Go␣D273N/Q205L), was regulated by xanthine nucleotides, not by guanine nucleotides (6). Go␣X bound XDP 1 and XTP instead of GDP or GTP. Go␣X bound G protein ␤␥ subunits only in the presence of XDP, and XTP stimulated dissociation of the Go␣X␤␥ heterotrimer. XTP-bound Go␣X underwent a conformational change similar to the activated wild-type Go␣. In the present study, we investigated the interaction between Go␣X and G protein-coupled receptors. We found that Go␣X mutant proteins retained the receptor binding specificity of the wildtype Go␣ and were able to interact with Go-coupled receptors, such as the m2 muscarinic cholinergic receptor (MAChR), Nformyl peptide receptor (NFPR), and thrombin receptor, but not with m1 MAChR or thyrotropin-releasing hormone (TRH) receptor which do not couple to wild-type Go. Because the concentrations of XDP and XTP are relatively low in vivo (7), Go␣X mutant proteins are essentially nucleotide-free unless exogenous xanthine nucleotides are provided. Thus, Go␣X provides an excellent model to study the receptor interaction of empty G protein ␣ subunits. Consistent with the previously reported studies on the empty form of transducin (4), our data are most readily interpreted as showing that "empty" Go␣X can form a stable complex with appropriate receptors on the membrane. Expression and Purification of His 6 -tagged Go␣-Both wild-type Go␣ and mutant Go␣X were subcloned into the Escherichia coli expression vector pET-15b (Novagen) with His 6 tag at the N terminus. These clones were used to transform the E. coli strain BL21(DE3), and proteins were expressed. Expression and purification of these proteins was described previously (6,8). After harvesting the culture, cell extracts were resuspended in the binding buffer (5 mM imidazole, 0.5 M NaCl, 160 mM Tris-HCl, pH 7.9, and 1 mM ␤Me). The His 6 -tagged proteins were purified from Ni 2ϩ -NTA column according to the protocol provided by Novagen. Purified proteins were stored in TED buffer (20 mM Tris-HCl, pH 8.0, 1 mM EDTA, and 1 mM dithiothreitol) with 0.1 mM MgCl 2 and 0.1 mM nucleotide diphosphate (GDP or XDP as appropriate).

Materials-Purified
Membrane Preparation from Baculovirus-infected Sf9 Cells-Sf9 cells were grown and infected with recombinant baculoviruses encoding either m2 MAChR or NFPR (9 -11). Infected cells were centrifuged and resuspended at Ͻ10 7 cells/ml in HME/PI buffer (20 mM NaHepes, pH 8.0, 2 mM MgCl 2 , 1 mM EDTA, 0.1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, and 10 g/ml leupeptin). The cell suspension was homogenized by 10 strokes in a glass homogenizer followed by passing through a 27 gauge hypodermic needle several times. The homogenate was briefly centrifuged at 3,000 ϫ g for 10 min, and then the supernatant was collected and centrifuged at 30,000 ϫ g for 30 min. The pellet was washed once with HME/PI, and the final pellet was resuspended in HME/PI at a protein concentration of 5 mg/ml.
Synthesis of XTP␥S-XTP␥S was synthesized from XDP and ATP␥S * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 (6). The binding reaction contained 0.5 g of purified protein in TED buffer, with 0.1 mM MgCl 2 , 1 M ATP, and 0.1 M GTP␥S or XTP␥S (20,000 cpm/pmol). For the time course experiments, 20-l aliquots were withdrawn from a 200-l reaction, diluted 10-fold with ice-cold TED buffer containing 0.1 mM MgCl 2 , filtered through 45-m nitrocellulose, washed, and dried. The amount of bound radioactivity was determined by scintillation counting.
Radioligand Binding of Receptors-The ligand binding assays of membrane-bound receptors were performed as described (9 -11). The total concentration of m2 MAChR and the affinities of NFPR were determined by incubating membranes with 2 nM [ 3 H]QNB or various concentrations of fML[ 3 H]P for 1 h in 20 mM Tris-HCl, pH 7.4, 12.5 mM MgCl 2 , and 1 mM EDTA at 30°C in a final volume of 0.5 ml. Nonspecific binding was defined as binding that was not displaced by 10 M atropine for m2 MAChR or 10 M fMLP for NFPR. Unbound ligands were removed by filtration through Whatman GF/C filters and washing four times using ice-cold binding buffer. The amount of bound radioactivity was determined by scintillation counting.
Binding of Go␣ on Sf9 Cell Membranes-0.2 g of purified Go␣ or Go␣ mutant proteins were incubated with 100 g of Sf9 cell membranes in TED buffer of a final volume of 100 l at room temperature for 1 h. The membranes were centrifuged and subjected to Western blot using antibodies against Go␣.
Cell Culture and Transfection-COS-7 cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum. 1 ϫ 10 5 cells/well were seeded in 12-well plates 1 day before transfection. All transfection assays contained a total amount of 1 g of DNA, and pCIS encoding ␤-galactosidase was used to maintain a constant amount of DNA. To each well, 1 g of DNA was mixed with 5 l of LipofectAMINE (Life Technologies, Inc.) in 0.5 ml of Opti-MEM (Life Technologies, Inc.), and 5 h later, 0.5 ml of 20% fetal calf serum in Dulbecco's modified Eagle's medium was added to the medium. After 48 h, cells were assayed for inositol phosphate levels as described previously (13,14).

Stimulation of XTP␥S Binding of Go␣X by Activated M2
Muscarinic Receptor-To test if Go␣X could interact with G protein-coupled receptors, we investigated the receptor-stimulated nucleotide binding of Go␣X. Activated G protein-coupled receptors are known to facilitate the binding of GTP␥S to G protein ␣ subunits. It has been reported that recombinant m2 MAChR from Sf9 cells stimulated the binding of GTP␥S to wild-type Go␣ 2-3-fold in response to muscarinic agonists (9, 10). We infected Sf9 cells with recombinant baculoviruses encoding m2 MAChR and prepared membranes. The concentration of receptor was about 20 pmol/mg of membrane protein, determined from [ 3 H]QNB binding. We have previously shown that Go␣X mutant proteins bind only xanthine nucleotides, but not guanine nucleotides. In this experiment, we reconstituted purified Go␣X with Sf9 cell membranes containing m2 MAChR in the presence of XDP and G protein ␤␥ subunits purified from bovine retina, and followed agonist-dependent stimulation of XTP␥S binding to Go␣X. We found that carbachol accelerated the XTP␥S binding of Go␣X, in a fashion similar to the acceleration of GTP␥S binding observed with wild-type Go␣ (Fig.  1a). In control experiments using wild-type Sf9 cell membranes, both atropine and carbachol had no effect on the XTP␥S binding of Go␣X (Fig. 1a). ␤␥ subunits were required for the carbachol-dependent stimulation of nucleotide binding (Fig.  1b), suggesting that only the trimeric complex of Go␣X with ␤␥ can be activated to exchange XDP for XTP␥S by the ligandbound receptors.
High Affinity Ligand Binding of N-formyl Peptide Receptor Induced by Go␣X-Another well documented indication of receptor-G protein interaction is that GTP or GTP␥S inhibits the high affinity binding of G protein-coupled receptors to their agonists. NFPR receptors expressed in Sf9 cells are known to be in the low ligand affinity state (ϳ60 nM fMLP) (11), presumably because of the lack of mammalian Gi-like G proteins in Sf9 cells, whereas NFPR in neutrophils and NFPR expressed in mouse L cells exhibited high affinity ligand binding (0.5-3 nM fMLP) (15,16). Therefore, we reconstituted Go␣X with NFPR from Sf9 cells and investigated whether Go␣X could induce the high affinity ligand binding state in NFPR receptors. Sf9 cell membranes containing NFPR receptors were prepared as de- scribed under "Experimental Procedures." The expression level of NFPR was about 20 pmol/mg of membrane protein, determined by fML[ 3 H]P binding. We incubated the NFPR with wild-type Go␣ or mutant Go␣X in the presence of ␤␥ and varieties of nucleotides and determined their affinities for the agonist fMLP. As expected, NFPR expressed in Sf9 cells showed low ligand affinity binding of fMLP (ϳ100 nM), and Go␣X alone did not affect ligand binding (Fig. 2a). More interestingly, NFPR exhibited high affinity ligand binding (ϳ10 nM) when Go␣X, ␤␥, and XDP were present (Fig. 2a). Both ␤␥ and XDP were required to induce the high ligand affinity state of NFPR, and XTP inhibited the fMLP binding of the receptors (Fig. 2b). In the control experiments, wild-type Go ␣␤␥ heterotrimer was also found to convert the NFPR to the high affinity ligand binding state, which was inhibited by GTP␥S (Fig. 2b). These experiments demonstrated that the heterotrimeric complex of Go␣X␤␥ can interact efficiently with NFPR.
Binding of Go␣X with M2 Muscarinic Receptor on Sf9 Cell Membranes-The previous two experiments showed that Go␣X␤␥ heterotrimer could interact with the G protein-coupled receptors efficiently and that the interaction was similar to the interaction between wild-type Go and receptors. To investigate receptor interaction of Go␣X more directly, we studied binding of Go␣X to receptor containing Sf9 cell membranes. Purified wild-type Go␣ or Go␣X were incubated with Sf9 cell membranes containing m2 MAChR in the presence of different reagents. The membranes were then pelleted and subjected to Western blotting using antibodies against Go␣ to see if Go␣ remained bound to the membrane. In the control experiments using wild-type Sf9 cell membranes without m2 MAChR, both wild-type Go␣ and Go␣X did not remain associated with the membrane. However, wild-type Go␣ was bound to membrane when it was coincubated with ␤␥. Similarly, Go␣X stayed on the membrane when in complex with ␤␥ in the presence of XDP (data not shown). These experiments using wild-type Sf9 cell membranes showed that Go␣ bound to the membranes only in the ␣␤␥ complex form, presumably because ␤␥ facilitates membrane association. In the experiments using membranes containing m2 MAChR, we found somewhat surprisingly that Go␣X bound to receptor-containing membranes even in the absence of carbachol and without ␤␥ (Fig. 3, lane 12), whereas wild-type Go␣ did not (Fig. 3, lane 1), suggesting Go␣X alone was able to bind to receptor. Interestingly, both XDP and XTP abolished the interaction between Go␣X and m2 MAChR-containing membranes and released Go␣X from the membrane fraction (Fig. 3, lanes 7 and 8), whereas GDP or GTP had no effect (data not shown), suggesting that the nucleotide-free form of Go␣X can recognize and bind to Go-mediated receptor.
As expected, Go␣X stayed on the membrane when both XDP and ␤␥ were present (Fig. 3, lane 10), and XTP promoted dissociation of the Go␣X␤␥ complex (Fig. 3, lane 9). In the case of wild-type Go␣, the binding pattern was the same between membranes with or without the receptors, and XDP or XTP had no effect on binding (Fig. 3, lanes 1-5). In a titration experiment, quantitation of Go␣X revealed that the amount of Go␣X bound to the membrane increased linearly until it reached saturation, and the level of saturation was proportional to the amount of receptor incubated in the reaction (Fig. 4, a-c). Furthermore, similar experiments using Sf9 cell membranes containing NFPR were also performed, and the results were similar (data not shown). These experiments indicated that the empty form of Go␣X, without a bound nucleotide and ␤␥, could form a stable complex with receptor. In summary, these data suggest that Go␣X with XDP bound and ␤␥ bind to membranes, whereas the XTP form is found to be cytoplasmic. The nucleotide-free form is able to bind to Go-mediated receptors.

Dominant-negative Effect of Go␣X on Receptor Activation in COS-7 Cells-Because our experiments suggested that empty
Go␣X was able to bind to the receptor in vitro, we went on to test for this interaction in intact cells. Indeed, we found that Go␣X was able to interact with receptors and inhibit their activities in COS-7 cells consistent with the observation that Go␣X did not dissociate from the receptors without xanthine nucleotides. Thrombin receptors are known to couple with G proteins from both the Gi and Gq families (17). In COS-7 cells transfected with the thrombin receptor, endogenous Gq is activated by the addition of thrombin and stimulates PLC␤ isoforms to elevate cellular IP 3 concentration. Inhibition of receptor activation in transfected cells by wild-type G proteins was observed before (18). Thus if cells are cotransfected with both the thrombin receptor and wild-type Go␣ subunit, the activation of Gq is inhibited because of the competition of Go␣ for the receptor or endogenous G protein ␤␥ subunits (Fig. 5a). We cotransfected thrombin receptor and Go␣X to determine whether Go␣X could compete with endogenous Gq for the receptors. Indeed, we found that Go␣X inhibited Gq activity stimulated by thrombin, and the inhibition was proportional to the amount of Go␣X cDNA used in the transfection (Fig. 5a). Because Go␣X in the absence of XDP does not interact with ␤␥ and does not affect the ␤␥-stimulated PLC␤ activity in COS-7 cells (6), the inhibition by Go␣X of Gq activation stimulated by the thrombin receptor must come from the competitive binding of Go␣X to the receptor. Similar experiments were performed with m1 MAChR and TRH receptor which were known to couple only to the Gq family of G␣ proteins and not to the Go␣ family (10,19,20). Go␣X had no effect on the activation of m1 MAChR or TRH receptor (Fig. 5, b and c). On the other hand, wild-type Go␣, which can compete for endogenous ␤␥, inhibited both m1 MAChR and TRH receptor stimulation as expected. In COS-7 cells, the activation of thrombin receptor, m1 MAChR, and TRH receptor share the same Gq pathway downstream of the receptor. Because Go␣X inhibited only the thrombin receptor activity, but had no effect on m1 MAChR or TRH receptor, we concluded that Go␣X inhibited thrombin receptor stimulation by competitive binding to the receptor.
To test if Go␣X could bind to other Go-coupled receptors in cells, we looked into the interaction between Go␣X and m2 MAChR. Because m2 MAChR couples only to the Gi family of G␣ proteins and not to the Gq family (9, 21), we could not assay their interaction in the same way as the thrombin receptor by monitoring PLC activities in COS-7 cells transfected with the receptor and Go␣X. Therefore, we constructed an artificial pathway by cotransfecting both m2 MAChR and G15␣ into COS-7 cells. G15␣ is known as a promiscuous G protein that can be activated by all kinds of G protein-coupled receptors, and G15␣ also activates PLC␤ isoforms (21). In cells cotransfected with both m2 MAChR and G15␣, we were able to activate endogenous PLC␤ isoforms by the addition of the muscarinic agonist carbachol. We found that this m2 MAChR stimulation pathway could also be inhibited by Go␣X (Fig. 5d). All these experiments suggested that Go␣X was able to interact with G protein-coupled receptors in cells and retained the receptor specificity of wild-type Go␣; it coupled with thrombin receptor and m2 MAChR, but not with m1 MAChR or TRH receptor. Furthermore, Go␣X exhibited dominant-negative inhibitory effects against these receptors in cells.

DISCUSSION
Go␣X (Go␣D273N/Q205L) was the first reported mutant of heterotrimeric G protein ␣ subunits that bound xanthine nucleotides, not guanine nucleotides (6). It bound ␤␥ only in the presence of XDP and could be activated by XTP. We continued to study the interaction of Go␣X with G protein-coupled receptors in this report. The interaction of G proteins and their receptors is best demonstrated in two experiments: agoniststimulated GTP␥S binding of G protein ␣ subunits and inhibition of high affinity ligand binding of the receptors by GTP␥S. To test if Go␣X can interact with G protein-coupled receptors and be activated by their agonists, we reconstituted purified Go␣X with Sf9 cell membranes containing m2 MAChR or NFPR. First, we found that binding of XTP␥S to Go␣X was stimulated by the muscarinic agonist carbachol in the presence of m2 MAChR. In similar experiments using wild-type Go␣, GTP␥S binding was also stimulated by carbachol. In both cases, ␤␥ was required for the carbachol-dependent nucleotide binding, suggesting that only Go␣X␤␥ heterotrimer could interact with the receptors effectively. Second, we tested Go␣X to determine whether it could induce the high affinity state in NFPR receptors expressed in Sf9 cells. The NFPR expressed in these cells is known to be in the low affinity state probably because of lack of mammalian Gi-like proteins in Sf9 cells (11). In our experiments, we found that Go␣X could convert NFPR into the high affinity state in the presence of ␤␥ and XDP, and this effect was inhibited by XTP. These two experiments demonstrated that Go␣X, when in complex with ␤␥ and XDP, could interact with G protein-coupled receptors effectively and be activated by the agonists.
Because cells lack xanthine nucleotides, Go␣X provides an excellent model to study empty G protein ␣ subunits. The empty form of G␣ is an important intermediate in receptor activation and has long been proposed to form a stable complex with activated receptors. However stable interaction between empty G proteins and their receptors was only reported in the transducin-rhodopsin system. Empty transducin apparently formed a stable complex with light-activated rhodopsin and stayed on the rod outer segment membrane. Interestingly, deactivation of the rhodopsin did not lead to the dissociation of transducin from the complex (4). In this report, we showed that empty Go␣X was able to bind to the receptor on the membrane in the absence of ␤␥ subunits and without agonists, and the interaction could be abolished by either XDP or XTP. The amount of Go␣X associated on the membranes with m2 MAChR was proportional to the amount of receptor at saturation. Interestingly, binding of Go␣X alone did not convert the receptor to the high ligand affinity conformation, which required the ␣␤␥ complex. Therefore, the binding of Go␣X alone to the receptor is not functional in contrast with the binding in the presence of ␤␥ and XDP.
Because Go␣X appears to form a stable complex with the receptor, we tested whether Go␣X could inhibit receptor activation in cells. In transfected COS-7 cells, we showed that Go␣X was able to inhibit thrombin receptor or m2 MAChR stimulated PLC␤ activities via the Gq or G15 pathway, but had no effect on m1 MAChR or TRH receptor stimulation. Because both thrombin receptor and m2 MAChR are known to couple with wild-type Go, and m1 MAChR and TRH receptor only couple with Gq, we interpret the data to mean that Go␣X retained the receptor specificity of wild-type Go and was able to interact with Go-coupled receptors in cells. The inhibitory binding of Go␣X enables us to specifically block Go-coupled receptors in certain systems. This could be a useful means to analyze FIG. 5. The negative inhibitory effects of Go␣X on receptor-stimulated PLC␤ activation in COS-7 cells. 1 ϫ 10 5 cells/well were seeded in a 12-well plate and then were transfected with cDNAs encoding the indicated G proteins (Go␣DN designates Go␣D273N) and thrombin receptor (a), m1 MAChR (b), TRH receptor (c), or m2 MAChR (d). In panels a-c, the amount of receptor cDNA used in each well was 0.25 g, and the amount of Go␣ cDNA was 0.75 g/well unless otherwise indicated. In panel d, the amount of both m2 MAChR and G15␣ cDNA was 0.2 g/well and that of Go␣ was 0.6 g/well. The total amount of cDNA for each well was adjusted to 1.0 g by the addition of CMV-LacZ cDNA. After cells were labeled with [ 3 H]inositol overnight, they were incubated in the medium containing 0.1 unit/ml thrombin (a), 1 M carbachol (b and d), or 1 M TRH (c) before levels of inositol phosphates were determined. different receptor-stimulated signal transduction pathways, and could perhaps be useful in drug screening associated with G protein-coupled receptors.
In the previous report (6), we showed that the single Go␣ mutant, Go␣D273N, lost the ability to bind either guanine nucleotides or xanthine nucleotides and could not bind ␤␥ under any conditions. Surprisingly, Go␣D273N can still bind to receptors. In transfected COS-7 cells, we found that Go␣D273N inhibited thrombin receptor and m2 MAChR activation, in a fashion similar to Go␣X (Fig. 5, a and d). Go␣D273N also retained the same receptor specificity as wild-type Go␣; i.e. it had no effect on m1 MAChR or TRH receptor stimulated pathways (Fig. 5, b and c). In the Sf9 cell membrane binding assay, it only bound to the m2 MAChR membranes, not to the control wild-type Sf9 cell membranes. However in contrast to Go␣X, Go␣D273N was not released from the m2 MAChR membranes by XDP or XTP, consistent with its inability to bind nucleotides (data not shown). The reason that Go␣D273N mutant proteins do not bind xanthine nucleotides is not clear. Apparently it must have a structure similar to that of the empty Go␣ which enables it to bind receptors, but the structure is probably not stable locally around the nucleotide binding pocket. Nevertheless, Go␣D273N may also be useful as a dominant-negative inhibitor of receptor functions.