Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The potent platelet-aggregating and vasoconstrictor eicosanoid thromboxane A₂ (TxA₂) has been implicated in diseases affecting the heart, lungs, kidneys, and peripheral vascular system (Fitzgerald et al., 1987; Oates et al., 1988). Its biological effects are mediated by activating specific cell surface receptors and are subject to regulatory controls (Dorn and Davis, 1992; Kinsella et al., 1994; Spurney et al., 1994). In the continuous presence of agonist, TP activation is attenuated by distinct molecular mechanisms. This desensitization is transient and receptor responsiveness returns following removal of agonist (Spurney et al., 1994). The mechanisms for regulating receptor responsiveness have been extensively studied in the cAMP-coupled ₂-adrenergic receptor and the phosphodiesterase-coupled receptor rhodopsin (Dohlman et al., 1991; Lefkowitz, 1993, 1998). In these receptor systems, desensitization is largely caused by direct phosphorylation of receptors at serine and threonine residues by general kinase systems such as protein kinase A (PKA) and protein kinase C as well as by a family of receptor-specific kinases, GPCR kinases or G-protein coupled receptor kinases (for review, see Dohlman et al., 1991). Receptor phosphorylation is followed by binding of a second group of protein cofactors termed arrestins, which desensitize receptor signaling presumably by sterically interfering with receptor-effector coupling (Lefkowitz, 1993).

In contrast to desensitization, the mechanisms involved in return of receptor responsiveness or resensitization are less well characterized but may involve dephosphorylation and recycling of receptor proteins (Lefkowitz, 1998). In this regard, rhodopsin is dephosphorylated by a novel calcium-activated protein phosphatase with a catalytic domain similar to PP1, PP2A, and PP2B (Steele et al., 1992; Kutuzov and Bennett, 1996; Vinos et al., 1997; Huang and Honkanen, 1998). In extraocular receptor systems, GPCR phosphatases appear to conform more closely with the classical scheme for categorizing protein phosphatase activity (Lutz et al., 1993; Pitcher et al., 1995; Shih et al., 1999). Thus, the ₂-adrenergic receptor is dephosphorylated by a PP2A-like enzyme (Pitcher et al., 1995). Cholecystokinin (CCK) receptors are also dephosphorylated by a PP2A-like enzyme, which is activated by pretreatment of the cells with CCK (Lutz et al., 1993). More recently, Malbon and coworkers (Shih et al., 1999) found that 1) both PP2A and PP2B form complexes with the ₂-adrenergic receptor, and 2) inhibition of PP2A or PP2B prevents recovery of ₂-adrenergic receptor responsiveness in the intact cell. These data suggest that serine/threonine protein phosphatases may play an important role in regulating responsiveness of GPCRs.

The receptor for TxA₂ belongs to the large superfamily of GPCRs (Hirata et al., 1991; Namba et al., 1992; Abe et al., 1995). In most cell systems, TP couples to phospholipase C (PLC) through pertussis toxin-insensitive G proteins (Shenker et al., 1991; Offermans et al., 1994). In humans, two isoforms of TP have been identified (Raychowdhury et al., 1994) that, in addition to coupling to PLC, oppositely regulate adenylyl cyclase activity (Hirata et al., 1996). Previous studies suggest that desensitization of TP receptors is associated with phosphorylation of receptor proteins (Habib et al., 1997, 1999; Spurney, 1998). At least some of the phosphorylation sites are found in the distal C terminus (Spurney, 1998). Moreover, deletion of these C-terminal phosphorylation sites inhibits desensitization (Spurney, 1998), suggesting that phosphorylation of TP plays a key role in regulating TP responsiveness. In the present studies, we investigated the serine/threonine protein phosphatases that dephosphorylate TP. We found that recovery of TP responsiveness following desensitization was associated with dephosphorylation of receptor proteins. Both recovery of TP responsiveness and dephosphorylation of TP were inhibited by the cell-permeable PP1 and PP2A inhibitors calyculin and OKA. In vitro, dephosphorylation of TP was prevented by combined PP1/PP2A inhibitors and was partially blocked by selective inhibition of either PP1 or PP2A. TP phosphatase activity did not have an absolute requirement for divalent cations and was found predominantly in cytosolic fractions of the cell. These data suggest that both PP1- and PP2A-like protein phosphatases contribute to recovery of TP responsiveness following agonist-specific desensitization.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. Inhibitor 2 was obtained from Calbiochem (La Jolla, CA). Okadaic acid, calyculin, and microcystin were obtained from Biomol (Plymouth Meeting, PA). The TxA₂ agonist U46619 (Coleman et al., 1981) was obtained from Cayman Chemical (Ann Arbor, MI). All other chemical reagents were obtained from Sigma-Aldrich (St. Louis, MO). The 12CA5 monoclonal antibody was obtained from Boehringer-Mannheim (Indianapolis, IN) and HEK293 cells were obtained from American Type Culture Collection (Rockville, MD). All tissue culture reagents were obtained from Life Technologies (Gaithersburg, MD). Epitope-tagged TP were created using polymerase chain reaction to insert the 12CA5-epitope at both the amino- and C terminus of the mouse TP as previously described (Spurney, 1998). These mutations do not affect TxA₂ binding or agonist-induced PLC activation (Spurney, 1998). 12CA5-tagged TP were subcloned into the mammalian expression vector pcDNA 3.0 (Invitrogen, San Diego, CA) and were used to stably transfect HEK293 as described below.

Culture and Transfection of HEK293 Cells. HEK293 cells were grown and subcultured as previously described (Spurney, 1998). To create cell lines stably expressing 12CA5-tagged TP, our pcDNA 3 expression vector containing 12CA5-tagged TP was transfected into HEK293 cells by the calcium-phosphate method (Sambrook et al., 1989). To isolate permanent transfectants, G418-resistant cells were selected in complete medium containing 800 µg/ml G418. Following G418 selection, individual clones were screened for TxA₂ binding using the stable radiolabeled TxA₂ receptor antagonist [³H]SQ29548 (Ogletree et al., 1985) (New England Nuclear, Boston, MA) as previously described (Spurney, 1998). Three separate clones expressing 200 to 400 fmol of TP/mg of protein were used for the experiments.

Resensitization of Inositol Phosphate Generation. Inositol phosphates were measured as previously described using anion exchange column chromatography (Spurney, 1998). For the resensitization studies, cells were grown in six-well plastic culture dishes (9.5 cm²/well) (Costar, Cambridge, MA) to 80% confluency and then treated for the indicated times with the agents to be tested or their vehicle in 2 ml of Krebs-Ringer buffer at 37°C. After treatment, cells were washed three times with Krebs-Ringer buffer and then allowed to recover for the indicated times in the presence or absence of the cell-permeable phosphatase inhibitors calyculin or OKA. One minute before rechallenge with U46619, 2 M lithium chloride was added to a final concentration of 20 mM. Cells were then stimulated with the indicated concentrations of U46619 or its vehicle. After 2 min, the reaction was stopped and the samples were processed as previously described (Spurney, 1998).

Dephosphorylation of TP in the Intact Cell. For these studies, HEK293 cell clones stably expressing 12CA5-tagged TP were plated in 100-mm dishes and grown to approximately 80% confluency. On the day to the experiment, cells were loaded for 90 min with 0.1 to 0.2 mCi of inorganic phosphate (³²P). After loading, cells were treated in a manner identical to the protocol described above for the resensitization experiments. After the recovery period, TP were isolated by immunoprecipitation as previously described (Spurney, 1998) and the phosphorylation state of TP was assessed by autoradiography after separation on 12% polyacrylamide gels. In previous studies (Spurney, 1998), we found that treatment with the TP agonist U46619 (Coleman et al., 1981) caused phosphorylation of a broad band of 44 kDa in HEK293 cells transfected with 12CA5-tagged TP. The band was not seen in cells transfected with the wild-type receptor lacking the 12CA5 epitope or when the immunoprecipitation was performed in the presence of the hemagglutinin peptide, which is the epitope recognized by the 12CA5 antibody. These data indicate that the 44-kDa band represents the TP receptor.

In Vitro Dephosphorylation Assays. HEK293 cell clones were plated in 100-mm dishes and grown to approximately 80% confluency. The medium was then replaced with 6 ml of phosphate-free Dulbecco's modified Eagle's medium (Life Technologies) supplemented with 0.2 mCi of ³²P (New England Nuclear). After 90 min, TP were phosphorylated in the intact cell by treatment of with the TxA₂ agonist U46619 (Coleman et al., 1981) (10 µM) for 10 min. Previous studies suggest that this time period results in high levels of TP phosphorylation (Habib et al., 1997, 1999; Spurney, 1998). To provide a sufficient number of phosphorylated receptors for the experiments, one confluent 100-mm dish was used for two different assay conditions. Following agonist treatment, cells were washed twice with 6 ml of ice-cold Dulbecco's phosphate-buffered saline and then were scrapped into 1 ml of ice-cold lysis buffer containing 150 mM NaCl, 50 mM Tris, pH 8.0, 1.0% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 2 mM EDTA, 10 mM sodium fluoride, 10 mM sodium pyrophosphate, 100 nM calyculin, 1 µg/ml leupeptin, 1 µg/ml pepstatin, 1 µg/ml aprotinin, and 100 µg/ml phenylmethylsulfonyl fluoride. Phosphorylated TP were then isolated by immunoprecipitation as previously described (Spurney, 1998). For the dephosphorylation reactions, phosphorylated TP bound to protein A Sepharose was combined and then was washed twice in lysis buffer, three times with phosphatase buffer (50 mM Tris-HCl, 50 mM 2-mercaptoethanol, pH 7.3), and was divided into 200-µl volumes to individual microcentrifuge tubes for the different dephosphorylation conditions. In preliminary studies we found that dividing the phosphorylated TP after isolation under Results in <10% variation in the amount of phosphorylated TP between tubes. Phosphorylated TP were then used as substrate for protein phosphatases prepared from HEK293 cells as described below. The reaction was initiated by adding 1 ml of phosphatase preparation containing the agents to be tested or their vehicle to the phosphorylated TP. Control reactions were initiated by adding 1 ml of phosphatase buffer to the phosphorylated TP. The dephosphorylation reactions were then incubated for 30 min at 37°C with gentle rocking to disperse the protein A Sepharose-bound TP. The reaction were stopped by a brief centrifugation to pellet the protein A Sepharose-bound TP, aspiration of the supernatant, and the addition of lysis buffer. After washing, three times with lysis buffer, TP were eluted from the protein A Sepharose by boiling in Laemmli SDS sample buffer. Phosphorylated proteins were then assessed by autoradiography after separation on 12% polyacrylamide gels with 0.1% SDS as described by Laemmli (1970).

Preparation of HEK293 Cell Homogenates. Before study, HEK293 cells were grown in 100-mm dishes to approximately 80% confluency. Protein phosphatases were then prepared in batch from HEK293 cells. For the dephosphorylation reactions, one 100-mm dish was used for each assay condition. Homogenates were prepared by washing twice with Dulbecco's phosphate-buffered saline and then scraping cells into phosphatase buffer (50 mM Tris-HCl, 50 mM 2-mercaptoethanol) using 2 ml of buffer per 100-mm dish. The preparations were combined and then sonicated with two 10-s bursts (setting 12) from a VirSonic 50 sonicator (Virtis Company, Gardiner, NY). Protein measurements suggest that this strategy results in <15% variation in the amount of protein in the homogenates between dephosphorylation studies. Cytosolic and membrane fractions were prepared by centrifugation of whole-cell homogenates at 100,000g for 10 min at 4°C. The supernatant was removed for the cytosolic fraction studies and the membrane pellet was resuspended in phosphatase buffer by sonication (setting 12) for 10 s. Preliminary studies suggested that approximately three-fourths of the total protein in the HEK293 cell homogenates was found in the cytosolic fraction and approximately one-fourth of total cellular protein was found in the membrane fraction. We therefore diluted the cellular fractions appropriately to obtain similar concentrations of protein. We also saved an aliquot of the cell fractions to measure protein concentration using the method of Bradford (1976).

Statistical Analysis. Data are presented as the mean ± S.E.M. For comparisons between two groups, statistical significance was assessed using an unpaired t test. For comparisons between more than two groups, statistical analysis was performed by analysis of variance followed by Bonferroni's procedure for multiple pairwise comparisons (Wallenstein et al., 1980).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Recovery of TP Responsiveness Is Inhibited by PP1/PP2A Inhibitors and Is Associated with Dephosphorylation of TP. To study recovery of TP responsiveness in the intact cell, HEK293 cells stably expressing 12CA5-tagged TP were treated with the TxA₂ agonist U46619 (1 µM) (Coleman et al., 1981) for 10 min. After washing, the cells were allowed to recover for the indicated times in the presence or absence of the cell-permeable PP1/PP2A inhibitor calyculin (Hardie et al., 1991; Shih et al., 1999). After the recovery period, cells were rechallenged with U46619 and inositol trisphosphate (IP3) generation was measured as described under Experimental Procedures. Data are expressed as a percentage of the baseline response to control for slight differences in responsiveness between assays (average baseline response: 118 ± 23% increase above basal IP3 generation). As shown in Fig. 1A, TP responsiveness recovered over a 20-min time period. Recovery of TP responsiveness was inhibited by 10 nM calyculin (Fig. 1A). The cell-permeable PP1/PP2A inhibitor OKA (Hardie et al., 1991) (1 µM) similarly inhibited recovery of TP responsiveness at this 20-min time point [88 ± 15 (vehicle) versus 40 ± 4 (OKA) percentage of baseline response; P < 0.01]. Calyculin (10 nM) and OKA (1 µM) had no effect on baseline TP responsiveness over a 20-min time period (data not shown). Taken together, these data suggest that PP1- and/or PP2A-like protein phosphatases contribute to recovery of TP responsiveness following agonist-specific desensitization in the intact cell.

View larger version (44K): [in this window] [in a new window]   Fig. 1.   Recovery of TP responsiveness in the intact cell. HEK293 cells stably expressing 12CA5-tagged TP were loaded with either myo-[3H]inositol (A) or 32P (B) as described under Experimental Procedures. After loading, cells were treated with 1 µM U46619 for 10 min. Cells were then washed and allowed to recover for the indicated times in the presence of the cell-permeable PP1/PP2A inhibitor calyculin (10 nM) or its vehicle (dimethyl sulfoxide). After the recovery period, cells were either 1) rechallenged with 1 µM U46619 for 2 min before measurement of IP3 generation measured as described under Experimental Procedures (A), or 2) harvested for isolation of TP by immunoprecipitation (B). As shown in A, TP responsiveness recovered over a 20-min time period. Recovery of TP responsiveness was inhibited by 10 nM calyculin. As shown in B, recovery of TP responsiveness was associated with an apparent decrease in the phosphorylation state of TP. The apparent decrease in TP phosphorylation was completely blocked by calyculin, suggesting that the decrease in phosphorylation was not due to receptor degradation. These data suggest that recovery of TP responsiveness in the intact cell is temporally associated with dephosphorylation of TP and implicate a PP1- and/or PP2A-like protein phosphatase in the dephosphorylation reaction. The IP3 generation studies were preformed in triplicate and data points are the mean of five separate experiments. The phosphorylation experiments were performed in parallel and similar results were obtained in four separate studies. *P < 0.05 versus vehicle.

Dephosphorylation of TP in Vitro Is Prevented by Inhibiting PP1 and PP2A Activity. To further study the TP phosphatase, TP phosphorylated in the intact cell were isolated by immunoprecipitation and were used as substrate for protein phosphatases prepared from HEK293 cells as described under Experimental Procedures. As shown in Fig. 2, there was an apparent reduction in the phosphorylation state of TP during incubation with whole-cell homogenates prepared from HEK293 cells. The PP1 and PP2A inhibitors calyculin (100 nM) and microcystin (100 nM) (Hardie et al., 1991) completely inhibited this apparent decrease in TP phosphorylation, suggesting that TP were dephosphorylated rather than degraded. Figure 3 shows the time course and protein dependence of the in vitro dephosphorylation assay. As shown in Fig. 3A, TP were dephosphorylated over a 30-min time period. Figure 3B shows that TP phosphatase activity was proportional to the protein concentration. These data are consistent with results in the intact cell (Fig. 1) and suggest that TP are dephosphorylated by a PP1- and/or a PP2A-like protein phosphatase.

View larger version (65K): [in this window] [in a new window]   Fig. 2.   Dephosphorylation of TP by PP1- and PP2A-like protein phosphatases. TP phosphorylated in the intact cell were isolated by immunoprecipitation and was used as substrate for phosphatases prepared from HEK293 cells. Dephosphorylation reactions were performed in the presence of the combined PP1 and PP2A inhibitors calyculin (100 nM) or microcystin-LR (100 nM) or their vehicles (dimethyl sulfoxide). The phosphorylation state of TP was assessed by autoradiography after separation on 12% polyacrylamide gels. Incubation of TP with HEK293 cell homogenates caused a decrease in the apparent phosphorylation state of TP. This apparent decrease in TP phosphorylation was completely inhibited by calyculin and microcystin-LR, suggesting that the decrease in phosphorylation was not due to receptor degradation. These data suggest that a PP1- and/or PP2-like phosphatase dephosphorylate TP. Similar results were obtained in three separate experiments. The positions of molecular weight markers are indicated in kilodaltons.

View larger version (56K): [in this window] [in a new window]   Fig. 3.   Time course and protein dependence of TP dephosphorylation. TP phosphorylated in the intact cell were isolated by immunoprecipitation and was used as substrate for phosphatases prepared from HEK293 cells. The TP substrate was incubated with HEK293 cell homogenates for 0, 10, 20, or 30 min (A) or was incubated for 30 min with HEK293 cell homogenates diluted 1:1 (not diluted), 1:2, or 1:4 with phosphatase buffer (50 mM Tris-HCl, 50 mM 2-mercaptoethanol). As shown in A, TP were dephosphorylated over a 30-min time period. As shown in B, the rate of dephosphorylation was proportional to the protein concentration. Similar results were obtained in three separate experiments.

View larger version (29K): [in this window] [in a new window]   Fig. 4.   Inhibition of TP phosphatase activity by PP1 or PP2A inhibitors. In vitro dephosphorylation reactions were performed as described under Experimental Procedures in the presence or absence of I2, a specific protein inhibitor of PP1 (100 nM), or OKA at concentrations that specifically inhibited PP2A (1 nM). The phosphorylation state of TP was assessed by autoradiography after separation on 12% polyacrylamide gels. Both I2 and OKA partially blocked dephosphorylation of TP. These data suggest that a PP1- and PP2-like phosphatase dephosphorylate TP. Similar results were obtained in four separate experiments. *P < 0.05 versus basal, **P < 0.025 versus basal.

View larger version (47K): [in this window] [in a new window]   Fig. 5.   Dose-dependent inhibition of TP phosphatase activity by OKA. In vitro dephosphorylation reactions were performed in the presence of either vehicle or 0.1 to 1000 nM concentrations of OKA. The phosphorylation state of TP was assessed by autoradiography after separation on 12% polyacrylamide gels. A, densitometric analysis of the autoradiographs. Little inhibition of TP phosphatase activity was detected at 0.1 nM concentrations of OKA. At 1 nM concentrations of OKA, approximately 40% of the TP phosphatase activity was inhibited, whereas 80 to 90% of the TP phosphatase activity was inhibited at OKA concentrations of 10 to 1000 nM. B, representative autoradiograph of the dose dependence of TP phosphatase inhibition by OKA. Data points are the mean ± S.E. of five experiments.

TP Phosphatase Activity Does Not Have an Absolute Requirement for Divalent Cations. We first determined the effect of adding divalent cations to the dephosphorylation reactions. As shown in Fig. 6A, additions of calcium or magnesium did not enhance dephosphorylation of TP. To further investigate the role of divalent cations in TP dephosphorylation, we determined the effect of the divalent cation chelator EDTA on dephosphorylation of TP. As shown in Fig. 6B, addition of EDTA to the dephosphorylation reaction did not prevent dephosphorylation of TP but caused a modest and reproducible inhibition of TP dephosphorylation. EDTA did not affect binding of 12CA5-tagged TP by the 12CA5 antibody (data not shown). These data suggest that TP phosphatase activity does not have an absolute requirement for divalent cations, although the presence of divalent cations may be necessary for full activity as has been reported for other PP1 and PP2A substrates (Cohen, 1991).

View larger version (52K): [in this window] [in a new window]   Fig. 6.   Effect of divalent cations on TP phosphatase activity. TP phosphorylated in the intact cell were used as substrate for phosphatases prepared from HEK293 cells. Dephosphorylation reactions were performed in the presence or absence of either 0.1 mM CaCl2 or 5 mM MgCl2 (A) or 5 mM EDTA (B). As shown in A, TP phosphatase activity was not affected by additions of CaCl2 or MgCl2. In contrast, EDTA caused a small but consistent inhibition of TP phosphatase activity (B). Similar results were obtained in three to four separate experiments per compound.

TP Phosphatase Activity Is Found Predominantly in Cytosolic Fractions of HEK293 Cells. To determine the subcellular distribution of the TP phosphatase activity, we prepared cytosolic and membrane fractions from HEK293 cells as described under Experimental Procedures. Whole-cell, cytosolic, and membrane fractions were prepared so that the homogenates contained comparable amounts of protein (under Experimental Procedures). Dephosphorylation reactions were then performed using these subcellular fractions. As shown in Fig. 7, TP were dephosphorylated by whole-cell homogenates and cytosolic fractions of the cell. In contrast, little TP phosphatase activity was detected in membrane fractions of HEK293 cells. These data indicate that TP phosphatase activity is primarily cytosolic.

View larger version (66K): [in this window] [in a new window]   Fig. 7.   TP phosphatase activity in subcellular fractions. In vitro dephosphorylation reactions were performed using whole-cell homogenates, cytosolic cell fractions, or membrane fractions prepared as described under Experimental Procedures. The HEK293 cell fractions were prepared so that the homogenates contained comparable amounts of protein (under Experimental Procedures). For this experiment, protein concentration (mg/ml) for whole-cell, cytosolic, and membrane fractions were 0.219, 0.203, and 0.185, respectively. TP phosphatase activity was found primarily in cytosolic fractions of the cell. Similar results were obtained in two separate experiments.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

Desensitization of GPCRs is largely mediated by mechanisms involving direct phosphorylation of receptor proteins (Dohlman et al., 1991; Lefkowitz, 1993, 1998). This desensitization is transient, and receptor responsiveness returns promptly following agonist removal (Spurney et al., 1994; Lefkowitz, 1998). The rapidity of recovery of GPCR responsiveness suggests that resensitization does not require new protein synthesis, but may involve dephosphorylation and recycling of receptor proteins (Lefkowitz, 1998). Thus, the phosphorylation state of GPCRs is likely to be dependent not only on the activity of protein kinases but also on the activity of protein phosphatases. In the present studies, we investigated the serine/threonine protein phosphatases that dephosphorylate mouse TP. We found that TP responsiveness recovered rapidly following agonist-specific desensitization. Recovery of TP responsiveness was associated with dephosphorylation of receptor proteins. Both dephosphorylation of TP and recovery of TP responsiveness could be blocked by cell-permeable PP1/PP2A inhibitors in the intact cell. To further identify the TP phosphatase we developed an in vitro dephosphorylation assay. Using this in vitro system, we found that TP were dephosphorylated by whole-cell homogenates prepared from HEK293 cells. Dephosphorylation of TP was completely inhibited by the PP1 and PP2A inhibitors calyculin and microcystin, and was partially inhibited by the specific PP1 inhibitor I2 as well as OKA at concentrations that specifically inhibit PP2A. TP phosphatase activity did not have an absolute requirement for divalent cations and was found primarily in cytosolic fractions of the cell. Although further studies will be necessary to directly link PP1- and PP2A-like protein phosphatases to dephosphorylation of TP, these pharmacological data provide evidence that the phosphorylation state of TP is modulated by protein phosphatases with PP1- and PP2A-like properties. By regulating the phosphorylation state of TP, protein phosphatases may modulate tissue responsiveness to TxA₂.

Although much has been learned about the kinases that phosphorylate GPCRs (Hausdorff et al., 1990; Lefkowitz, 1993, 1998), less is known about dephosphorylation of GPCRs by protein phosphatases. The most progress has been made with rhodopsin due to the availability of mutations in lower organisms that cause light-dependent retinal degeneration (Vinos et al., 1997). For example, the Drosophila retinal degeneration C gene encodes an unusual protein phosphatase that has a catalytic domain similar to PP1, PP2A, and PP2B and a C-terminal domain with EF-hand calcium binding motifs (Steele et al., 1992; Vinos et al., 1997). Dephosphorylation of rhodopsin in mammals is also mediated by a calcium-activated phosphatase (Kutuzov and Bennett, 1996); and recently, a novel protein phosphatase has been identified in human retina that encodes a protein homologous to the retinal degeneration C gene product (Huang and Honkanen, 1998). In extraocular GPCR systems, PP2A-like protein phosphatases appear to play dominant roles in dephosphorylating GPCR proteins. In this regard, both ₂-adrenergic receptors and CCK receptors are dephosphorylated by PP2A-like enzymes (Lutz et al., 1993; Pitcher et al., 1995). We also found that a PP2A-like protein phosphatase contributed to dephosphorylation of TP in our in vitro system. However, a portion of TP phosphatase activity in vitro was modulated by inhibitors of PP1, suggesting that a PP1-like phosphatase contributes to dephosphorylation of TP. In support of this hypothesis, relatively large concentrations of OKA were required to inhibit both recovery of TP responsiveness and dephosphorylation of TP in the intact cell, suggesting that a least a portion of TP phosphatase activity was relatively resistant to OKA. Since OKA inhibits PP2A 10- to 100-fold more potently than PP1 (Cohen, 1991; Hardie et al., 1991), this finding is consistent with a role for PP1 in dephosphorylating TP in the intact cell. To our knowledge, a role for PP1-like protein phosphatases in dephosphorylating GPCRs has not been previously reported. This observation may of importance with regard to the TP because activation of TP generates second messengers that have the potential for modulating PP1 activity as discussed below.

Recent studies suggest that serine/threonine protein phosphatase activity is regulated in a sophisticated manner by a combination of targeting and regulatory subunits (38, 39). In this regard, PP1 activity is modulated by a heat-stable protein inhibitor of PP1 termed inhibitor 1 (I1) (Cohen, 1989; Hunter, 1995). When phosphorylated by PKA, I1 forms a complex with PP1 and inhibits PP1 activity (Cohen, 1989; Hunter, 1995). I1 is readily dephosphorylated by both PP2A and PP2B (Cohen, 1989). The ability of the calcium-dependent phosphatase PP2B to modulate the phosphorylation state of I1 provides a potential mechanism for regulation of TP phosphatase activity by TxA₂. Stimulation of TP by TxA₂ activates PLC (Spurney et al., 1993). Activation of PLC generates the second messengers diacyl glycerol and inositol phosphates, which increase PKC activity and cause release of calcium from intracellular stores, respectively. The initial increase in calcium from intracellular stores is short-lived, but may be accompanied by a more sustained increase in intracellular calcium as a result of influx of calcium outside the cell (Spurney et al., 1993). Increased intracellular calcium levels have the potential to activate PP2B (Cohen, 1989). PP2B, in turn, may regulate the activity of PP1 by modulating the phosphorylation state of I1 (Cohen, 1989). An additional level of complexity may occur with human TP isoforms with oppositely regulate adenylyl cyclase (Hirata et al., 1996) and, therefore, might modulate PKA-induced phosphorylation of I1. Thus, activation of TP has the potential to stimulate both kinase and phosphatase activity. The coordinated stimulation of these signaling cascades could promote not only desensitization of TP signaling in the continuous presence of agonist but also return of TP responsiveness following agonist removal.

Like PP1, the activity of PP2A can also be modulated (Chen et al., 1992). In this regard, studies by Brautigan and coworkers (Chen et al., 1992) found that phosphorylation of PP2A on tyrosine 307 (Tyr³⁰⁷) inhibits PP2A activity. This Tyr³⁰⁷ phosphorylation is stimulated by serum, v-scr, and either insulin or epidermal growth factor (EGF) receptor activation (Chen et al., 1992). Although the details of the signaling pathways involved remain to be delineated, the ability of receptor tyrosine kinases such as the EGF receptor to modulate PP2A activity provides a potential mechanism for cross talk between GPCR and receptor tyrosine kinase systems. For example, EGF-induced Tyr³⁰⁷ phosphorylation of PP2A might inhibit the activity of PP2A toward GPCR substrates. Such a reduction in PP2A activity could potentially inhibit recycling and reactivation of GPCR systems in which PP2A-like protein phosphatases play a role in dephosphorylating receptor proteins.

TP phosphatase activity was not affected by additions of the divalent cations calcium or magnesium. In contrast, the divalent cation chelator EDTA partially inhibited dephosphorylation of TP by HEK293 cell homogenates, suggesting that divalent cations were necessary for full TP phosphatase activity. A similar divalent cation requirement for full activity has been reported for PP1 and PP2A toward other protein substrates (Cohen, 1991). Moreover, inhibition of TP phosphatase activity by EDTA was incomplete, indicating that the TP phosphatase does not have an absolute requirement for divalent cations. Taken together with the inhibitor data, these data are consistent with the notion TP is dephosphorylated by PP1- and PP2A-like protein phosphatases.

TP phosphatase activity was found predominantly in cytosolic fractions of the cell. A similar subcellular location has been reported for protein phosphatases that dephosphorylate CCK receptors (Lutz et al., 1993). Presumably, this cytosolic phosphatase activity becomes closely associated with the GPCR during the dephosphorylation reaction. If this association is constitutive, it is possible that our in vitro dephosphorylation reaction was unable to detect the small amount of TP phosphatase activity constitutively associated with the membrane-bound receptor. Alternatively, the association of TP and its phosphatase may be a dynamic process as has been reported for phosphatases associated with the ₂-adrenergic receptor (Shih et al., 1999). Further studies will be necessary to determine whether either of these possibilities contributes to dephosphorylation of TP by serine/threonine protein phosphatases.

In summary, we found that TP were rapidly dephosphorylated by protein phosphatases prepared from HEK2932 cells. Dephosphorylation of TP was completely prevented by combined PP1 and PP2A inhibitors and was partially blocked by specific inhibitors of either PP1 or PP2A. TP phosphatase activity did not have an absolute requirement for divalent cations and was found primarily in cytosolic fractions of the cell. These pharmacological data are consistent with the notion that TP are dephosphorylated by PP1- and PP2A-like protein phosphatases. Regulating the phosphorylation state of TP may modulate tissue responsiveness to TxA₂.