The Proximal Portion of the COOH Terminus of the Oxytocin Receptor Is Required for Coupling to Gq, but Not Gi

As the oxytocin receptor plays a key role in parturition and lactation, there is considerable interest in defining its structure/functional relationships. We previously showed that the rat oxytocin receptor transfected into Chinese hamster ovary cells was coupled to both Gq/11 and Gi/o, and that oxytocin stimulated ERK-2 phosphorylation and prostaglandin E2 synthesis via protein kinase C activity. In this study, we show that deletion of 51 amino acid residues from the carboxyl terminus resulted in reduced affinity for oxytocin and a corresponding rightward shift in the dose-response curve for oxytocin-stimulated [Ca2+] i . However, oxytocin-stimulated ERK-2 phosphorylation and prostaglandin E2 synthesis did not occur in cells expressing the truncated receptor. Oxytocin also failed to increase phospholipase A activity or activate protein kinase C, indicating that the mutant receptor is uncoupled from Gq-mediated pathways. The Δ51 receptor is coupled to Gi, as oxytocin-stimulated Ca2+ transients were inhibited by pertussis toxin, and a Gβγ sequestrant. Preincubation of Δ51 cells with the tyrosine kinase inhibitor, genistein, also blocked the oxytocin effect. A Δ39 mutant had all the activities of the wild type oxytocin receptor. These results show that the portion between 39 and 51 residues from the COOH terminus of the rat oxytocin receptor is required for interaction with Gq/11, but not Gi/o. Furthermore, an increase in intracellular calcium was generated via a Giβγ-tyrosine kinase pathway from intracellular stores that are distinct from Gq-mediated inositol trisphosphate-regulated stores.

Oxytocin (OT) 1 is a nine-amino acid peptide that stimulates uterine smooth muscle and mammary myoepithelial cell contraction, and prostaglandin production by uterine endometrial and amnion cells. Nucleic acid sequencing of cDNA clones of the oxytocin receptor (OTR) indicated that it is a member of the G protein-coupled receptor (GPCR) superfamily (1). As OT plays a pivotal role in parturition and lactation, there is considerable interest in defining the structure of the OTR. It has been shown for a number of GPCRs that several regions in the cytoplasmic domains contribute directly or indirectly to G protein coupling (see Ref. 2 for a review). The juxtamembrane portions of cytoplasmic loop 3 and cytoplasmic loop 2 of several family members have been implicated in receptor-G protein interactions. In addition, the COOH-terminal region of adrenergic receptors (3,4) and other receptor types (5-7) is also required for G protein interactions, but this domain does not appear to be important for receptor function of all GPCR family members (8 -10).
The COOH-terminal domain of some GPCRs plays an important role in G protein isotype selectivity (11,12). At least four isoforms of the prostaglandin EP3 receptor, differing only at their COOH-terminal tails (produced by alternative splicing), couple to different G proteins to activate different second messenger systems (13,14). The COOH terminus of the human parathyroid hormone receptor directs the receptor toward an interaction with G s , whereas a core region composed of the first, second, and third intracellular loops can interact promiscuously with different G proteins (11). A truncated human AT1 receptor mutant lacking the carboxyl-terminal 50 residues is deficient in coupling to G i , but it retains full ability to bind to G q (12).
The COOH terminus of some GPCRs has also been shown to be important for desensitization, which is manifested as a diminution in responsiveness for some period of time following agonist stimulation. COOH-terminal truncation of the ␤ 2 -adrenergic, ␣ 1B -adrenergic, lutropin/choriogonadotropin, platelet-activating factor, and neurokinin-2 receptors has been shown to impair homologous desensitization (15)(16)(17)(18)(19)(20). Work on the ␤-adrenergic receptor, in particular, has indicated that one mechanism of desensitization involves the rapid internalization of membrane-bound receptors following agonist stimulation (15). Removal of the COOH-terminal tail of some GPCRs has been shown to greatly reduce agonist-induced internalization of the receptor while having little or no effect on signal transduction (21)(22)(23)(24)(25).
Most GPCRs have a conserved cysteine in the COOH-terminal cytoplasmic tail near the seventh transmembrane-spanning region. This cysteine is known to be palmitoylated in rhodopsin (26), the ␤ 2 -adrenergic receptor (27), and the ␣ 2Aadrenergic receptor (28). In addition to the three intracellular loops delimited by transmembrane domains, a putative fourth cytoplasmic loop is formed in many GPCRs by insertion of palmitoylcysteines into the membrane lipid bilayer (29). It has been suggested that the fourth intracellular loop is important for G protein coupling (29), and mutation of Cys-341 in the carboxyl tail of the human ␤ 2 -adrenergic receptor leads to an uncoupled, nonpalmitoylated form of the receptor (27). Studies with m2 receptors indicate that palmitoylation is not an absolute requirement for receptor interaction with G proteins, but it enhances the ability of the receptors to interact with G proteins (30). In other cases, elimination of palmitoylation sites does not affect receptor-G protein interactions (31)(32)(33)(34)(35). However, downregulation of the receptor number after prolonged agonist exposure was completely abolished by this mutation (32). Depalmitoylation also has been shown to increase the rate of the luteinizing hormone/human chorionic gonadotropin receptor internalization (33). In other instances, replacement of Cys results in the specific loss of coupling to one G protein isotype but not another, namely mutation of the human endothelin receptor A resulted in no effect on the ability of endothelin to activate adenylyl cyclase, but inhibited activation of PLC (36).
Previous work from our laboratory, using a CHO cell line that was stably transfected with the rat OTR, showed that OT stimulated rapid increases in intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ), extracellular signal-related kinase-2 (ERK-2) phosphorylation, and PGE 2 synthesis (37,38). Furthermore, the OTR was coupled to both G q/11 and G i in transfected CHO cells (CHO-OTR cells) (38), and in pregnant rat myometrium (39). In the present studies, we have systematically analyzed the importance of the COOH-terminal domain of the OTR on several receptor-associated processes: ligand affinity, G protein coupling via specific signal pathways, receptor desensitization, and selectivity of G-protein coupling. With the exception of a few amino acid residues, the COOH-terminal domain of the OTR is highly conserved between species (Fig. 1). This observation argues in favor of the functional importance of the COOH-terminal domain of the OTR. Previous work with the rat V 1a vasopressin receptor indicated that the COOH-terminal region of the receptor is inaccessible to antibodies directed against a COOH-terminal peptide when the receptor is coupled to G proteins, but is accessible when receptor-G protein complexes are dissociated (40). As the V 1a vasopressin receptor is closely related to the OTR, we have examined the importance of the COOH-terminal domain of the OTR in G protein coupling by creating COOH-terminal deletion mutants. Our approach has been to create COOH-terminal truncations of 22, 39, and 51 residues. The OTR has two adjacent Cys residues at positions 351 and 352, which are potential palmitoylation sites (Fig. 2). To determine the importance of these sites in OT action in the present studies, both these residues were replaced by Ser in one of the mutant OTRs analyzed.
DNA Constructs and Transfections-Rat OTR cDNA was provided by Dr. Stephen J. Lolait. Full-length and carboxyl-terminal mutants, lacking 22, 39, and 51 residues (Fig. 2), were generated from rat OTR cDNA by polymerase chain reaction, using primers sets containing an EcoRI and XhoI end. The amplified DNAs were ligated with the expression vector pcDNA3.1 Myc/His A (Invitrogen, San Diego, CA) in frame with the Myc/His epitopes at the 3Ј-terminus. Mutation of two Cys residues at positions 343 and 344 to Ser was carried out by the method of Higuchi et al. (41). The sequences of all the DNA constructs were verified by DNA sequence analysis. The primer pairs for the full-length cDNA were primer 1 (5Ј-GAATTCCTGAGTCGCGTCGCGTCG-3Ј) and primer 2 (5Ј-CTCGAGTGCTGAAGATGGCTGAGAGC-3Ј). Each truncation mutant was generated with primer 1 and a unique primer 2. The primers 2 for ⌬22, ⌬39, and ⌬51 were 5Ј-CTCGAGGTTGCTCTTCTT-GCTGACAC-3Ј, 5Ј-CTCGAGACGAGCAGAGCAGCAGAAGA-3Ј, and 5Ј-CTCGAGGTGGAAGAGGTGACCTGTGA-3Ј, respectively. For construction of the Cys to Ser replacement mutants, full-length primers 1 and 2 were used along with 5Ј-GTGCAGCGCTTCTTCTCCTCCTCT-GCTCGT-3Ј and 5Ј-ACGAGCAGAGGAGGAGAAGAAGCGCTGCAC-3Ј.
CHO-K1 cells were grown in ␣-minimal essential medium containing 5% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. The expression plasmids were introduced into these cells by calcium phosphate-mediated transfection, and clonal cell lines that stably expressed the cDNAs were obtained by Geneticin (400 g/ml) selection. The cells were maintained under an atmosphere of 5% CO 2 .
Receptor Binding Assay-Determination of the apparent K d and B max values of each of the cell lines was carried out with cells in six-well plates as described previously (42), using increasing concentrations of 125 I-OTA (0.14 to 100 pM). The cells were incubated for 1 h at 22°C in 1 ml of Tyrode's solution, pH 7.5, containing 0.1% bovine serum albumin. The binding data were examined by nonlinear regression analysis (GraphPad Software Inc., San Diego, CA), and binding constants were determined by assuming a single class of independent binding sites. Oxytocin competition studies were carried out using a fixed concentration of 125 I-OTA and increasing concentrations of OT (0.1-100 nM) in 24-well plates (0.3 ml/plate). The cells were then rinsed (3 ϫ 1 ml) in assay buffer and solubilized in 1 M NaOH; radioactivity was determined with a ␥ counter.
Measurement of Intracellular Free Calcium Concentrations, Inositol Phosphates, and PKC Activation-Real-time recordings of intracellular calcium concentrations ([Ca 2ϩ ] i ) were performed on single cells, as described previously (37). Each point in the figures represents the mean Ϯ S.E. values from 35 cells. The concentration of OT-stimulated synthesis of inositol phosphates was measured as described previously (37). Translocation of PKC was determined by immunoblot analysis of the cytosol and Triton X-100-solubilized cell fractions that were prepared according to Ogiwara et al. (43). PGE 2 Synthesis and MAP Kinase Phosphorylation-PGE 2 synthesis was determined using a PGE 2 enzyme immunoassay system from Amersham Pharmacia Biotech, as described previously (37). The phosphorylation of ERK-2 MAP kinase and the effects of pertussis toxin were determined by immunoblotting, as described previously (38). For analysis of phosphorylated p38 MAP kinase, immunoblots were incubated with antibody directed against dually phosphorylated p38 (New England Biolabs), followed by stripping of the blots and reprobing with an antibody to p38 (Santa Cruz Laboratories).

Expression of Wild Type and Deletion Mutant Constructs-
The COOH-terminal truncation and replacement sites are shown in Fig. 2. All of the cDNAs were expressed in CHO cells, as shown by 125 I-OTA binding to cell surface OTR on intact cells (Table I). The apparent K d values of binding to the mutant receptors were comparable, but the number of receptors sites expressed per cell varied between mutants. The concentration of binding sites for the ⌬51 mutant was about 10% that of the wild type. Efforts to obtain ⌬51 clones with a greater number of binding sites were unsuccessful. We also determined the concentrations of OT reducing 125 I-OTA binding by 50% (IC 50 ). IC 50 values for the ⌬22, ⌬39, and C351S,C352S mutants were comparable to that of the wild type, while the ⌬51 mutant had an IC 50 value that was about 3-7 times greater than the others (Fig. 3).
Effect of Truncation on the Ca 2ϩ Response to OT-We have shown previously that stimulation of CHO-OTR cells with OT results in a rapid, transient increase in intracellular Ca 2ϩ concentrations (37). Both intra and extracellular sources of Ca 2ϩ were involved (37). In the present studies, OT stimulated intracellular Ca 2ϩ in all of the mutant lines (Fig. 4A). However, the ⌬51 mutant required 175 nM OT for stimulation (Fig. 4B), while the other mutants and wild type CHO-OTR cells responded to 10 nM OT or less (data not shown). The reduced sensitivity to OT by the ⌬51 mutant is consistent with the greater IC 50 value of OT displacement of 125 I-OTA (Fig. 3), and suggests that higher concentrations of OT are required for stimulation because of lower affinity for the peptide. Removal of extracellular Ca 2ϩ with EGTA eliminated the sustained [Ca 2ϩ ] i phase in both ⌬51 and wild type cells, indicating that the mutation had no effect on the relative intracellular and extracellular contributions to [Ca 2ϩ ] i (data not shown). These results further indicate that the major source of [Ca 2ϩ ] i arises from intracellular stores. Treatment of CHO cells, lacking OTRs, with up to and including 1 mM OT had no effect on [Ca 2ϩ ] i (data not shown).
Following OT-stimulated elevation in [Ca 2ϩ ] i , using rela-tively high OT concentrations (200 nM), Ca 2ϩ levels returned to near-base-line levels and the cells were refractory to further stimulation by OT in the medium. The cells were rinsed over a 10-min period to dissociate OT from OTRs, and were challenged with 10% of the OT dose to determine whether they became desensitized. In the case of the ⌬51 cells, the second dose was the same as the first because of the lesser initial sensitivity of this mutant to OT (Fig. 4B). Neither the wild type or any of the mutants exhibited a significantly diminished response to OT after the second dose of OT, indicating that the COOH-terminal region of the rat OTR is not involved in desensitization (Fig. 4A). The lack of desensitization was observed over a 10-fold range in OTR concentration (cells expressing wild type OTR versus those expressing the ⌬51 mutation; Table  I). Cells expressing wild type OTR in concentrations that were comparable to that of the ⌬51 mutant also showed no evidence of homologous desensitization (data not shown).
OT-stimulated PGE 2 Synthesis-The addition of increasing concentrations of OT to CHO cells expressing wild type, and the ⌬22, ⌬39 mutant OTRs resulted in the release of PGE 2 in a dose-dependent fashion, showing further that these receptors are functionally coupled to signal transduction pathways (Fig.  5). However, cells expressing the ⌬51 mutant did not release PGE 2 in response to OT (Fig. 5). Increasing the concentration of OT to 1 M had no effect (data not shown). It would appear from these results that the region between 39 and 51 residues from the COOH terminus is involved in the ultimate generation of a PGE 2 response. Within this region are two adjacent Cys residues (positions 351 and 352), which have been implicated in G protein interactions via palmitoylation (26 -28). However, mutation of the two Cys residues to Ser had no effect on the ability of OT to stimulate PGE 2 synthesis (Fig. 5). From estimations of EC 50 values (Fig. 5, inset), the potencies of OT in the active mutant lines were generally indistinguishable, in agreement with the IC 50 results. The maximal responses were generally proportional to the B max values, as estimated by 125 I-OTA binding (Table I).
OT-stimulated ERK-2 Phosphorylation-OT stimulation of PGE 2 synthesis in CHO cells transfected with the full-length rat OTR is thought to involve both the phosphorylation of cytosolic phospholipase A 2 by ERK-2/1, and Ca 2ϩ -mediated translocation of cytosolic phospholipase A 2 from the cytosolic to   ] i in ⌬51 cells, we determined whether the inability of OT to stimulate PGE 2 synthesis in these cells was the result of deficient ERK-2 phosphorylation. As shown previously (38), OT (50 nM) causes the rapid (2 min) phosphorylation of ERK-2 in CHO-OTR cells, as evidenced by the electrophoretic mobility shift of a fraction of total ERK-2 on immunoblots (pp42, Fig. 6A). Comparable results were obtained with each of the mutant OTRs after stimulation with 50 nM OT, with the exception of the ⌬51 mutant, which demonstrated only base-line phosphorylation (like OTR-negative CHO cells, Fig. 6A). To take into account the lower affinity of ⌬51 cells for OT, we used increasing concentrations of OT and a more sensitive antibody assay that measures dually phosphorylated (S/T and Y) ERK-2. Concentrations of OT up to 1 M had no effect on ERK-2 phosphorylation in ⌬51 mutant cells after 5 min of treatment (Fig. 6B). In contrast, 100 nM phorbol 12-myristate 13-acetate (PMA) was effective in stimulating ERK-2 phosphorylation in these cells (Fig. 6B). Treatment of the wild type and mutant cell lines and CHO cells with basic fibroblast growth factor (bFGF), 100 ng/ml for 5 min, resulted in ERK-2 phosphorylation in all of the cell lines (Fig. 6C), showing (along with the PMA results in Fig. 7B) that there is no impairment in ERK-2 phosphorylation in ⌬51 cells. bFGF stimulates ERK-2 phosphorylation by a G protein-independent mechanism (44).
Evidence for OT-stimulated Ca 2ϩ Transients That Are G qindependent-G q -mediated PLC activation in ⌬51 cells is not consistent with the ability of OT to elicit increases in [Ca 2ϩ ] i from intracellular stores on the one hand, and the inability of OT to stimulate PGE 2 synthesis and ERK-2 phosphorylation on the other (38). To determine whether OT activates PLC in ⌬51 cells via G q mediation, we measured two sequelae of G q /PLC activity: inositol phosphates production and PKC activation. In these and subsequent experiments, a wild type clonal cell line expressing about the same number of OTRs as the ⌬51 cells was used. Treatment of cells expressing the full-length receptor with 500 nM OT for 30 min resulted in about a 5-fold increase in inositol phosphate (InsP) production (Fig. 7). In contrast, InsP production by ⌬51 cells was unchanged after OT treatment (Fig. 7). However, these cells were capable of being stimulated because sodium fluoroaluminate, a nonspecific G-protein activator, induced a significant increase in InsP production. Treatment of cells expressing full-length OTR with OT for 5 and 15 min resulted in the activation of PKC␣, as measured by an increase in the amount of PKC associated with the membrane fraction (Fig. 8A). There was a barely detectable level of PKC associated with the membrane fraction in ⌬51 cells in the basal state, but no increase following OT treatment (Fig. 8A). PKC was translocated from the cytosol to membrane fractions in ⌬51 cells after treatment with 100 nM PMA, indicating that the lack of effects of OT were not due to impaired PKC activation (Fig. 8B). The lack of OT activation of PKC in the ⌬51 cells is also consistent with the lack of effect of OT on ERK-2 phosphorylation and PGE 2 synthesis, both of which are mediated by PKC (38).

The Presence of an Intracellular, InsP-independent Pathway in OT-stimulated Ca 2ϩ
i Release That Is Mediated by Pertussis Toxin-sensitive G␤␥ and Tyrosine Kinase Activity-Treatment of ⌬51 and wild type OTR expressing cells with thapsigargin (250 nM) depleted intracellular stores of Ca 2ϩ in the absence of  6. A, oxytocin-stimulated ERK-2 phosphorylation, as measured by an electrophoretic mobility shift and immunoblotting of ERK. Antibody reacting to both phosphorylated (pp42) and nonphosphorylated (p42) ERK was used. The antibody also cross-reacted slightly with ERK-1 (p44). B, lack of effect of increasing concentrations of OT on ERK-2 phosphorylation in ⌬51 mutant cells, as measured by immunoblotting. Antibody reacting to dually phosphorylated ERK-2 was used. PMA (100 ng/ml) activation of ERK-2 phosphorylation in ⌬51 mutant cells indicates no impairment in the ability of ERK-2 to be phosphorylated in these cells. C, effect of bFGF (100 ng/ml) on ERK-2 phosphorylation in wild type and mutant cells. The length of time of stimulation was either 2 min (A) or 5 min (B and C).

FIG. 7. Lack of stimulation of InsP production by OT in ⌬51
cells. The nonspecific G protein activator, sodium fluoroaluminate (AlF), was stimulatory in ⌬51 cells. The number of OTR binding sites in the wild type cells was matched approximately to that expressed by ⌬51 cells, by selecting the appropriate clone from a pool of clones with different numbers of binding sites. *, p Ͻ 0.05. extracellular Ca 2ϩ , and inhibited the effects of OT on further increases in [Ca 2ϩ ] i (Fig. 9A). The absence of InsP formation in the ⌬51 cells indicates that InsP 3 -independent stores of Ca 2ϩ must be responsible for the OT-induced rise in [Ca 2ϩ ] i . Cells expressing ⌬51 and wild type OTRs were pretreated with selective inhibitors to determine whether the mutant G i ␤␥-mediated pathways effect an intracellular Ca 2ϩ response to OT stimulation, and whether this process is mediated by tyrosine kinase activation. Pertussis toxin treatment (500 ng/ml for 16 -20 h) completely obliterated the Ca 2ϩ response to OT in ⌬51 expressing cells and reduced the OT-stimulated increase in wild type cells by more than 60% (Fig. 9B). Transfection of the cells with a plasmid expressing the G␤␥ sequestrant ␤ARK1ct (10 g of DNA/6-cm dish, 24 h before addition of OT), completely blocked the effects of OT in ⌬51 cells, as compared with cells transfected with the empty vector (Fig. 9C). However, the G␤␥ sequestrant had no effect on OT-stimulated [Ca 2ϩ ] i in wild type cells (Fig. 9C). Preincubation of cells with increasing concentrations (1, 10, 100 M) of the tyrosine kinase inhibitor, genistein, for 1 h before stimulation with OT resulted in the complete inhibition of the OT-induced [Ca 2ϩ ] i transient in ⌬51 cells, even at the lowest concentration (Fig. 9D). The OT response in wild type cells was reduced by more than 35% by treatment with 1 M genistein). The inhibition was progressively greater with 10 and 100 M genistein (Fig. 9D).
Phosphorylation of p38 MAP Kinase-To further analyze G protein-coupled pathways present in the ⌬51 mutant, we examined p38 MAP kinase phosphorylation. Treatment of both wild type and ⌬51 cells with 250 nM OT for 5 and 10 min resulted in increased phosphorylation of p38, as measured by immunoblotting with an antibody to dually phosphorylated p38 (Fig. 10). Although basal levels of p38 phosphorylation were elevated, the addition of OT further stimulated phosphorylation after 5 min. OT also caused an increase in p38 phosphorylation in cells expressing the wild type OTR after both 5 and 10 min (Fig. 10). Pretreatment of both mutant and wild type cell lines with oxytocin antagonist or pertussis toxin inhibited p38 phosphorylation (Fig. 10). Treatment of either cell type with PMA resulted in increased p38 phosphorylation (Fig. 10), indicating that both G i (pertussis toxin-sensitive) and G q (PKCmediated) pathways are involved in activation of p38. G q is likely dissociated from the ⌬51 OTR, but direct activation of PKC by PMA occurs downstream from OTR-G q coupling.

DISCUSSION
The COOH-terminal portion of GPCRs has been shown to be involved in a number of essential activities. Depending on the particular GPCR, these activities include G protein coupling (3-7), selectivity of G protein isotype (11,(11)(12)(13)(14)45), and ho-mologous desensitization or internalization (15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25). Whereas truncation of some GPCRs has been shown to modify any one or more of its functions, shortening of the COOH terminus of other family members has no apparent effects (8 -10). Thus, no uniform hypothesis regarding the importance of the COOHterminal region can be applied a priori to the OTR, which has not been previously examined in detail (46,47). In view of the high degree of homology of the COOH-terminal domain of OTRs from several species, we reasoned that conservation would be associated with some important function(s). However, FIG. 8. A, lack of PKC activation in ⌬51 cells by OT, as measured by an increase in PKC immunoactivity associated with the membrane fraction. In contrast, OT stimulation caused an increase in PKC concentration in the membrane fraction of cells expressing the wild type OTR. A nonspecific immunoreactive band below the PKC band serves as a convenient indicator of uniform protein loading on the SDS-PAGE gel. The wild type OTR and ⌬51 clones were matched for approximately the same number of OTA binding sites. B, PMA-induced translocation of PKC from the cytosol (CYT) to membrane (MEM) fractions in ⌬51 cells shows that these cells have PKC that is capable of being activated. FIG. 9. A, depletion of intracellular stores by thapsigargin in ⌬51 cells, and elimination of Ca 2ϩ from the medium, results in the lack of any further increases in [Ca 2ϩ ] i after addition of OT, as is also the case with cells expressing the wild type OTR. These findings indicate that OT stimulates the release of Ca 2ϩ from intracellular stores in both ⌬51 and wild type cells. B, OT-stimulated [Ca 2ϩ ] i transients in ⌬51 cells is almost completely inhibited by preincubation with pertussis toxin, 500 ng/ml for 16 -20 h. Pertussis toxin also partially inhibits the response to OT in cells expressing the wild type OTR. C, transfection of ⌬51 cells with the ␤␥ sequestrant, ␤ARK1ct, blocked OT-stimulated increases in [Ca 2ϩ ] i. The empty vector was used as a control. In contrast, ␤ARK1ct had no effect on OT stimulation in the wild type OTR cells.  10. A, oxytocin-stimulated p38 phosphorylation in CHO cells expressing the ⌬51 mutant and wild type OTRs. The immunoblots were first probed with antibody recognizing dually phosphorylated (pT180, pY182) p38, stripped, and then probed with antibody to p38 to show uniformity of protein loading. The cells were treated with OT for 5 or 10 min, or were pretreated with OTA (1 M, 5 min) or pertussis toxin (500 ng/ml, 16 -20 h) before treatment with OT for 5 min. Alternatively, cells were treated with PMA (20 ng/ml) for 5 min. truncation of 22 and 39 residues from the COOH terminus or replacement of the Cys residues, which have been thought to be palmitoylation sites for the formation of a fourth intracellular loop, had no effect on OTR functions. These include ligand affinity, OT stimulation of increases in intracellular Ca 2ϩ , ERK-2 phosphorylation, and PGE 2 synthesis. The mutants were also indistinguishable from the wild type OTR with respect to homologous desensitization.
Treatment of ⌬51 cells with OT caused a Ca 2ϩ i transient, although higher concentrations of OT were required than with the other mutants or wild type receptor. This lower sensitivity of ⌬51 cells is consistent with the lower affinity of the ⌬51 receptor for OT. In contrast, no dose of OT (up to 1 M) stimulated ERK-2 phosphorylation or PGE 2 synthesis in the ⌬51 cells. As we have shown previously, OT-stimulated ERK-2 phosphorylation and PGE 2 synthesis are mediated by PKC (38). PKC activation occurs as a result of increased diacylglycerol synthesis, which is a product of the G q /PLC pathway. Upon activation of cell surface receptors, PKCs translocate from the cytoplasm to membrane surfaces (48). We found that OT failed to activate PKC-␣ translocation in the ⌬51 mutant cells, but not in cells expressing the wild type OTR. The lack of PKC activation in ⌬51 cells would account for the absence of OTstimulated ERK-2 phosphorylation and PGE 2 synthesis. The absence of PLC activation in the ⌬51 cells accounts for the inability of OT to increase total InsP synthesis. Thus, the Ca 2ϩ mobilizing effect of OT in ⌬51 cells occurred independently of OTR-mediated G q /PLC stimulation.
Our previous work, showing that wild type OTRs transfected into CHO cells are coupled to G i as well as G q (38), led us to consider whether the OT-stimulated increase in [Ca 2ϩ ] i in ⌬51 cells is mediated by G i instead of G q . In Rat-1 and COS-7 cells, GPCRs coupled to pertussis toxin-sensitive G proteins mediate ERK-1/2 activation via a G␤␥ subunit complex signal pathway that is dependent upon tyrosine phosphorylation and p21 ras activation (49,50). Cellular expression of a specific G␤␥ subunit sequestrant peptide derived from the carboxyl-terminal G␤␥ subunit binding domain of the ␤-adrenergic receptor kinase 1, ␤ARK1ct, was shown to inhibit LPA and ␣ 2A adrenergic receptor-mediated Shc phosphorylation in COS-7 cells (49). ⌬51 cells were pretreated with either a G i inhibitor (pertussis toxin), the G ␤␥ sequestrant (␤ARK1ct), or a tyrosine kinase inhibitor (genistein). All three agents inhibited OT-stimulated increases in [Ca 2ϩ ] i in ⌬51 expressing cells; pertussis toxin and genistein also reduced the effects of OT in wild type cells. Therefore, based on our findings, it would appear that the ⌬51 mutant OTR is functionally coupled to G i , but lacks the coupling to G q . Both G i and G q coupling to the OTR occurs in cells expressing the wild type OTR. We would extrapolate from the results that truncation of the COOH-terminal tail by 22 or 39 residues, and mutating residues 351 and 352 (Cys 3 Ser) do not affect G q coupling.
Because OT-stimulated ERK-2 phosphorylation in CHO-OTR cells is largely mediated by G q , we examined pertussis toxin-sensitive p38 MAP kinase phosphorylation as an indicator of a G i -mediated process. This pathway has been shown to be activated upon stimulation of both G q/11 -coupled m1 and G i -coupled m2 muscarinic-acetylcholine receptors (51). Overexpression of G␤␥ or a constitutively activated mutant of G␣ 11 , but not G␣ i , also stimulated p38 kinase activity (51). p38 kinase phosphorylation was stimulated by OT in ⌬51-expressing cells by a process that was completely inhibited by pertussis toxin. Because PMA also stimulated p38 phosphorylation in these cells, it is apparent that both G q and G i mediate p38 kinase activity. Nagao et al. (52) showed that G␣ q/11 stimulates p38 MAP kinase activity through PKC and Src family kinase-dependent pathways. These findings suggest that PMA stimulation of p38 phosphorylation in ⌬51 cells occurs through PKC activation of tyrosine kinase-regulated steps. Parenthetically, the present findings are the first observations of an effect of OT on p38 phosphorylation.
Pathways connecting G i activation and intracellular Ca 2ϩ transients in ⌬51 cells are not currently known. Depletion of intracellular Ca 2ϩ stores with thapsigargin, an inhibitor of endoplasmic reticulum Ca 2ϩ -ATPase activity, resulted in the loss of the Ca 2ϩ i transient following OT stimulation of ⌬51 cells. ␣ 1 -Adenoreceptors utilize two different G␣ subunits to increase [Ca 2ϩ ] i in rat myocytes (53). G␣ q appears to activate InsP production and induce the release of Ca 2ϩ from intracellular stores, while G␣ 11 may enhance the Ca 2ϩ -activated Ca 2ϩ influx that replenishes intracellular Ca 2ϩ stores (53). This mechanism does not appear to occur with the OTR, as both the pertussin toxin-sensitive and insensitive pathways involved intracellular Ca 2ϩ stores. If G␤␥-stimulated tyrosine kinases resulted in activation of PLC-␥, we would have expected to find a rise in inositol phosphates in the ⌬51 cells after OT treatment. G␤␥ activation of PLC-␤ isoforms to stimulate PtdInsP2 hydrolysis independent of protein tyrosine kinases (54) also appears to be absent in ⌬51 cells. Previous work from this laboratory demonstrated that OT inhibited (Ca 2ϩ ϩ Mg 2ϩ ) ATPase activity in sarcolemmal membranes from the rat uterine myometrium (55). It was postulated that inhibition of the calcium pump would hinder the extrusion of intracellular Ca 2ϩ , thus allowing OT-stimulated elevations in [Ca 2ϩ ] i to be maintained. Although mechanisms coupling the OTR and (Ca 2ϩ ϩ Mg 2ϩ )ATPase have not been described, it is unlikely that this pathway accounts for [Ca 2ϩ ] i increases in ⌬51 cells because the source of Ca 2ϩ in ⌬51 cells was thapsigargin-sensitive.
Mobilization of Ca 2ϩ from intracellular stores is mediated by three major receptors on the endoplasmic reticulum, the InsP 3 , ryanodine, and sphingosine-1-phosphate systems. Although we measured total InsPs to ensure that transient increases in InsP 3 would not go undetected; we were unable to detect any increase after OT treatment of ⌬51 cells. These results indicate that the increase in Ca 2ϩ i in ⌬51 cells caused by OT is not mediated by the InsP3 receptor. We did not study the effects of ryanodine on [Ca 2ϩ ] i release, as the ryanodine receptor/Ca 2ϩ release channel, which is an essential component of excitationcontraction coupling in striated muscle cells, is not found in any significant concentration in CHO cells (56,57). Sphingosine 1-phosphate has been shown to release Ca 2ϩ from the endoplasmic reticulum in several cell types in conjunction with occupancy of surface IgG receptors (58 -61). Sphingosine-1phosphate biosynthesis is catalyzed by sphingosine kinase (62,63), a ubiquitous enzyme found in the cytosol (64,65) and the endoplasmic reticulum (58). However, addition of the sphingosine analogue, DL-threo-dihydrosphingosine, which is a competitive inhibitor of sphingosine kinase activity (60), had no effect on OT-induced Ca 2ϩ i transients (data not shown). Thus, the signals mediating the release of Ca 2ϩ from intracellular stores are not known at the present time.
Residues in the N-terminal part of the COOH terminus of the human V 2 vasopressin receptor have been shown to be necessary for correct folding; the COOH-terminal residues are also important for efficient cell surface expression (66). Although the reduced affinity of the ⌬51 mutant for OT could be a result of its dissociation from G q , it might be due instead to modified protein folding as has been indicated for the V 2 vasopressin receptor. The distinction is important because if G qcoupled OTR has a higher affinity for OT than the G i -coupled form, the preferred pathway for OT-stimulated increases in [Ca 2ϩ ] i with low concentrations of OT would be through PLC activation and InsP 3 -mediated stimulation of Ca 2ϩ release from intracellular stores. However, because pertussis toxin and genistein substantially inhibited the OT-stimulated increase in [Ca 2ϩ ] i in wild type cells, the OTR-associated G␤␥/protein tyrosine kinase pathway appears to be as important as the G q / PLC pathway for increasing [Ca 2ϩ ] i . Of about 30 stable clones of ⌬51 examined, only the one used in these studies had 125 I-OTA binding activity. These results suggest that processing of the ⌬51 mutant protein is impaired.
The results of our studies are summarized in Fig. 11. The model postulates that, in addition to elements of the third intracellular loop, the proximal portion of the COOH-terminal domain of the OTR is required for coupling to G q . In contrast, coupling to G i occurs without participation of the COOH-terminal region. The loss of association of the ⌬51 mutant with G q leads to a loss of OT/PLC-mediated events, which include InsP production, PKC activation, ERK-2 phosphorylation, and PGE 2 synthesis (38). Phosphorylation of p38 MAP kinase occurred through a pertussis toxin-sensitive pathway both in the ⌬51 and wild type cells. Although p38 phosphorylation is also G q (PKC)-dependent through a process that likely involves PKC activation of p21 ras , pertussis toxin-sensitive phosphorylation of p38 serves as a useful indicator of G i coupling to ⌬51. We have also shown that the G i -tyrosine kinase-mediated pathway is intact in ⌬51 cells, as indicated by the pertussis toxin and genistein sensitivity of OT-stimulated increases in [Ca 2ϩ ] i . Inositol trisphosphate arising from the G q -initiated pathway mediates the release of Ca 2ϩ from intracellular stores via InsP 3 receptors. Endoplasmic reticulum receptors mediating intracellular Ca 2ϩ release from G i /protein tyrosine kinase pathways are presently unknown. Understanding why G i and G q contribute about equally to OT-stimulated rises in [Ca 2ϩ ] i remains to be established.