Increased Receptor Stimulation Elicits Differential Calcium-sensing ReceptorT888 Dephosphorylation*

The calcium-sensing receptor (CaR) elicits oscillatory Ca2+i mobilization associated with dynamic, inhibitory protein kinase C-mediated phosphorylation of CaRT888. While modest CaR stimulation elicits Ca2+i oscillations, greater stimulation either increases oscillation frequency or elicits sustained responses by an unknown mechanism. Here, moderate CaR stimulation (2.5 mm Ca2+o, 10 min) increased CaRT888 phosphorylation (160-kDa mature receptor) 5-fold in CaR stably transfected HEK-293 cells, whereas 3–5 mm Ca2+o treatments were without apparent effect. Treatment with 2 mm Ca2+o caused sustained CaRT888 phosphorylation (≥20 min) and oscillatory Ca2+i mobilization. However, 5 mm Ca2+o increased CaRT888 phosphorylation only briefly while eliciting sustained Ca2+i mobilization, suggesting that greater CaR activation induces rapid CaRT888 dephosphorylation, thus permitting sustained Ca2+i responses. Indeed, 5 mm Ca2+o stimulated protein phosphatase 2A activity and induced CaRT888 dephosphorylation following acute phorbol ester pretreatment, the latter effect being mimicked by CaR-positive allosteric modulators (NPS-R467 and l-Phe). Finally, the phosphatase inhibitor calyculin-A reversed CaR-induced inhibition of parathyroid hormone secretion from bovine parathyroid slices and normal human parathyroid cells, demonstrating the physiological importance of phosphorylation status on parathyroid function. Therefore, high Ca2+o-stimulated protein kinase C acts in concert with high Ca2+o-induced phosphatase activity to generate and maintain CaR-induced Ca2+i oscillations via the dynamic phosphorylation and dephosphorylation of CaRT888.

The calcium-sensing receptor (CaR) 3 regulates parathyroid hormone secretion (PTH) and renal Ca 2ϩ reabsorption and is thus crucial to the control of extracellular Ca 2ϩ (Ca 2ϩ o ) homeostasis (1,2). The parathyroid CaR acts in response to increasing Ca 2ϩ o concentration by mobilizing Ca 2ϩ i release and other intracellular signals to suppress PTH production and release. These effects can be reversed in dispersed bovine parathyroid cells by modulation of protein kinase C (PKC) activity (3)(4)(5)(6)(7)(8)(9)(10).
Of the five predicted PKC consensus sites in human CaR (11), mutation of the intracellular domain residue Thr-888 elicits the greatest effect on the Ca 2ϩ o sensitivity of PI-PLC (12). Phosphomimetic substitution of this site (CaR T888D ) lowers the Ca 2ϩ o sensitivity of the receptor (13) while nonphosphorylatable CaR T888A is more sensitive to Ca 2ϩ o and promotes sustained rather than oscillatory Ca 2ϩ i mobilization (14,15). This is the basis for the hypothesis that dynamic CaR T888 phosphorylation is required for oscillatory Ca 2ϩ i mobilization, an effect that had been proposed previously for other class C G proteincoupled receptors (16 -18).
However, two unresolved issues require attention. One concerns the nature of CaR T888 dephosphorylation and whether receptor activation can modulate phosphatase activity. The second relates to the observation that at maximal or near maximal levels of CaR activity, PKC-mediated inhibition is overcome to elicit sustained elevations in Ca 2ϩ i . Here, we show that heightened CaR activation accelerates CaR T888 dephosphorylation and provide evidence that the rate of dephosphorylation may set the frequency of CaR-induced Ca 2ϩ i oscillations. Thus, we propose that both high Ca 2ϩ o -induced PKC and high Ca 2ϩ oinduced phosphatase activities, acting in concert, are required to produce the cycles of CaR T888 phosphorylation/dephosphorylation necessary to mediate CaR-induced Ca 2ϩ i oscillations.
mental buffer (20 mM HEPES (pH 7.4), 125 mM NaCl, 4 mM KCl, 0.5 mM CaCl 2 , 0.5 mM MgCl 2 , and 5.5 mM glucose. Experiments were performed at 37°C prior to lysis on ice in radioimmune precipitation assay buffer supplemented with 1 mM N-ethylmaleimide and with protease and phosphatase inhibitors described previously (15). Immunoblotting-CaR T888 phosphorylation was then quantified by semi-quantitative immunoblotting as described previously (15,19), using an affinity-purified phospho-specific anti-CaR pT888 antibody (15). Anti-CaR mouse monoclonal antibody, raised to amino acids 214 -235 (ADD) of the extracellular domain of the human parathyroid CaR was from Affinity Bioreagents (Golden, CO).
Intracellular Ca 2ϩ Assay-Relative changes in intracellular Ca 2ϩ concentration were assayed in CaR-HEK cells grown on glass coverslips as described previously (15,20). Briefly, cells were loaded with 1 M Fura-2/AM in the dark in experimental buffer supplemented with 0.1% bovine serum albumin. Cells were then equilibrated and assayed at room temperature in experimental buffer containing the baseline [Ca 2ϩ ] o appropriate for the ensuing experiment (0.5 mM unless otherwise stated).
Protein Phosphatase 2A (PP2A) Assay-PP2A activity was assayed using a kit (DuoSet IC human active PP2A) from R&D Systems (Abingdon, UK) according to the manufacturer's instructions. Briefly, following experimental treatment and lysis, cellular PP2A was bound to the well of a 96-well plate and assayed for its ability to release phosphate from a PP2A-specific substrate.
Bovine PTH Secretion Assay-Bovine superior parathyroid glands were dissected from freshly killed cattle (Ͻ30 months old), and excess fatty tissue was removed. Three slices (ϳ1 mm) were then made in cross-section and placed in the 0.5 mM CaCl 2 experimental buffer supplemented with 1 mg/ml bovine serum albumin at 37°C for 1 h. A base layer of BioGel P4 (ϳ1 ml) was used to pack a polyprep column attached to a constant perfusion pump and set up in a water bath at 37°C (21). The three preincubated slices were transferred to the column and packed with Sephadex beads (ϳ1 ml), and the mixture perifused with experimental buffer at a rate of 1.5 ml/min. for a 20-min equilibration period. During the experimental period, cumulative samples were collected constantly over 4-min periods on ice. The PTH-containing samples were quickly frozen in cryotubes at Ϫ80°C to minimize hormone breakdown. Analysis of the hormone levels in the samples was carried out using the Immutopics bovine intact PTH enzyme-linked immunosorbent assay kit (Immunodiagnostic Systems Limited, Boldon, Tyne & Wear, UK) according to the manufacturer's instructions.
Preparation of Bovine Parathyroid Particulate Fraction-Following treatment with calyculin A (100 nM) and phorbol 12-myristate 13-acetate (PMA) (1 M) in the experimental buffer described above, parathyroid glands were homogenized in a HEPES buffer (12 mM HEPES (pH 7.6) and 300 mM mannitol) supplemented with 1 mM N-ethylmaleimide and protease and phosphatase inhibitors (15). The homogenate was then centrifuged at 2,500 ϫ g (15 min) with the postnuclear supernatant then centrifuged at 100,000 ϫ g (30 min) to give a particulate protein pellet. Samples were normalized for protein content by assaying according to the method of Bradford (22). The protein equivalency of the subsequent sample loading volumes was demonstrated by temporary staining of blots with Ponceau S and confirmed by immunoblotting with total CaR antibody (ADD).
Human Parathyroid Cell Preparation-Samples from normal parathyroid cell transplants were obtained at neck surgery at the Royal North Shore Hospital and North Shore Private Hospital (St. Leonards, New South Wales, Australia) and the Mater Misericordiae Hospital (North Sydney, New South Wales, Australia). Excess normal parathyroid tissue (ϳ5% of the total), typically taking the form of single 1-1.5 mm diameter "chips," was obtained from parathyroid autotransplants during thyroid surgery as described previously (21). All procedures were performed under guidelines established by the relevant institutional ethics committees, and all patients provided written informed consent for the use of the tissue for experimental purposes.
The method for the preparation of normal human parathyroid cells has been described recently (23). In brief, parathyroid tissue stored overnight in minimum essential medium at 4°C was digested for 20 min in minimum essential medium containing 1 mg/ml collagenase and 0.1 mg/ml DNase I at 37°C with brief oxygenation. The tissue was triturated by repeated passage through the tip of a 5-ml syringe, filtered (200 m pore size nylon filter), sedimented (70 g, 2.5 min), washed (2 ϫ 5 ml of bovine serum albumin-containing minimum essential medium), and finally resuspended in 2 ml bovine serum albumin-containing minimum essential medium. The remaining undigested tissues were returned to medium containing collagenase and DNase I and incubated for a further 20 min at 37°C followed by trituration and centrifugal isolation as above.
Determination of PTH Secretion from Perifused Human Parathyroid Cells-The method for the determination of PTH secretion from normal human parathyroid cells has also been described recently (23) and is similar to the method described above for bovine cells. Briefly, cells were perifused in 4 -5-kDa cut-off gel filtration micro beads with PTH appearing in the void volume. Gel filtration media were pre-equilibrated with physiological saline (125 mM NaCl, 4.0 mM KCl, 1.25 mM CaCl 2 , 1.0 mM MgCl 2 , 0.8 mM Na 2 HPO 4 , 20 mM HEPES (NaOH) (pH 7.4) supplemented with 0.1% D-glucose, 1ϫ basal amino acid mixture (total concentration 2.8 mM) as defined previously (23) and 1 mg/ml bovine serum albumin. 4 -10 ϫ 10 4 cells were loaded onto a 0.4-ml bed volume of Bio-gel P-4 in a small perifusion column, gently covered with 0.4 ml bead volume of Sephadex G-25 and perifused (1.5 ml/min) at 37°C with the physiological saline solution described above in various concentrations of Ca 2ϩ o in the absence or presence of the PP2A inhibitor calyculin A as required. 2-min (3 ml) samples were collected on ice, transferred to dry ice prior to storage at Ϫ80°C, and then analyzed for intact human PTH (1-84) using a third generation, two-site chemiluminescence assay on an Immulite 2000 autoanalyzer.
Statistical Analysis-Data are presented as means Ϯ S.E., and statistical significance was determined by one-way or repeated measures ANOVA or Friedman with appropriate post test, or by unpaired t test, using GraphPad Prism.

RESULTS
Acute treatment of CaR-HEK cells with the phorbol ester PMA (1 M) increased the immunoreactivity of two protein bands of approximate molecular mass 140 and 160 kDa obtained by immunoblotting using the custom-generated phospho-specific polyclonal antibody raised against CaR residue Thr-888 (anti-pCaR T888 ) ( Fig. 1), as shown previously (15). In the current experiments, cells were incubated in PMA (1 M) for various times up to 20 min resulting in a significant, timedependent change in the levels of CaR T888 phosphorylation ( Fig. 1). The anti-CaR pT888 immunoreactivity of the 160 kDa mature protein increased significantly by 2 min (p Ͻ 0.05), reaching a peak by 10 min (p Ͻ 0.001) and declined from 15-20 min (Fig. 1). Similar changes in the phosphorylation of the 140 kDa core glycosylated CaR protein were observed (see supplemental Fig. S1). In response to acute PMA treatment, the level of total CaR was unchanged as demonstrated by immunoblotting with the monoclonal CaR antibody ADD (Fig. 1). Therefore, despite the continued presence of PMA, the induction of CaR T888 phosphorylation was not sustained beyond 10 min.
Concentration Dependence of Ca 2ϩ o Effect on CaR T888 Phosphorylation-CaR-HEK cells were next exposed to various concentrations of Ca 2ϩ o (0.5-5 mM) for 10 min, and the resulting level of CaR T888 phosphorylation was quantified using the anti-CaR pT888 antibody. As the Ca 2ϩ o concentration increased from 0.5 to 2.5 mM, CaR T888 phosphorylation also increased, reaching a maximum at 2.5 mM (p Ͻ 0.001; Fig. 2). As the Ca 2ϩ o concentration increased between 2.5-5 mM however, CaR T888 phosphorylation declined to reach baseline levels following 10 min of exposure to 5 mM Ca 2ϩ o ( Fig. 2; see supplemental Fig. S2 for replicates). In consequence, there was a biphasic concentration response relationship between Ca 2ϩ o and 160-kDa protein CaR T888 phosphorylation. In contrast, the amount of 140 kDa, core-glycosylated CaR phosphorylated on Thr-888 increased from 0.5 to ϳ2 mM Ca 2ϩ o and remained at that level from 2-5 mM Ca 2ϩ o (supplemental Fig. S3). Total CaR abundance was not affected by the acute Ca 2ϩ o treatments as confirmed by Western blotting using total CaR antibody (Fig. 2).

Ca 2ϩ o -induced CaR T888 Phosphorylation as a Function of Time-Although it is possible that Ca 2ϩ
o -induced P160 CaR T888 phosphorylation is paradoxically inhibited at higher Ca 2ϩ o concentrations, an alternative explanation for the biphasic concentration-effect relationship ( Fig. 2) is that disproportionate dephosphorylation occurs at Ca 2ϩ o concentrations Ͼ2.5 mM. That is, above 2.5 mM Ca 2ϩ o , phosphatase activity is increased with respect to PKC activity such that CaR T888 phosphorylation levels return to baseline sooner, i.e. by ϳ10 min. To test this possibility, cells were incubated at various times up to 20 min in the presence of either 2 mM Ca 2ϩ o , i.e. close to the concentration that gave the greatest apparent phosphorylation of the 160-kDa band, or, 5 mM Ca 2ϩ o , the concentration that appeared to have no effect at the time point tested in Fig. 2.
Treatment of CaR-HEK cells with 2 mM Ca 2ϩ o caused a rapid increase in 160 kDa CaR T888 phosphorylation, reaching significance after only 30 s of exposure (p Ͻ 0.01; Fig. 3A). This level of phosphorylation of the 160-kDa band was largely maintained throughout the time course with the band still being significantly more phosphorylated than control after 20 min (p Ͻ 0.05). Phosphorylation of the 140 kDa band exhibited a similar time-course (supplemental Fig. S4A). Total CaR abundance remained constant at each time point tested (Fig. 3A). Therefore, 2 mM Ca 2ϩ o elicited rapid and sustained CaR T888 phosphorylation in CaR-HEK cells. Using the ratiometric fluorescent dye Fura-2, an increase in Ca 2ϩ o from 0.5 mM baseline to 2.5 mM was shown to elicit a rapid increase in Ca 2ϩ i levels in the form of oscillations (Fig. 3B, see also supplemental Fig. 5).
In contrast, exposure of the cells to 5 mM Ca 2ϩ o for various time points up to 20 min revealed a different time-course of response for CaR T888 phosphorylation (Fig. 4A). Initially, exposure of the cells to 5 mM Ca 2ϩ o resulted in a rapid increase in 160 kDa CaR T888 phosphorylation, with significance achieved  after only 30 s (p Ͻ 0.01). The response peaked by 2 min (p Ͻ 0.001) but then declined until by 10 min, it was significantly lower than the maximal level of phosphorylation achieved (p Ͻ 0.05) and was indistinguishable from zero time control levels. The 140 kDa intracellular CaR exhibited a similar time-dependent increase in phosphorylation, being significantly different from baseline after 30 s (p Ͻ 0.01) and peaking by 2 min (p Ͻ 0.001), but unlike the 160-kDa band, it was still phosphorylated significantly even after 20 min (supplemental Fig. S4B). Again, the level of CaR abundance (total CaR) remained constant at each time point tested (Fig. 4A). It appears therefore that for the mature form of CaR, stimulation with Ca 2ϩ o induces rapid feedback phosphorylation of the receptor at all concentrations of Ca 2ϩ o up to 5 mM. However, stimulation with Ca 2ϩ o may also induce dephosphorylation of the CaR where the relative rates of phosphorylation/dephosphorylation are determined by the extent of CaR activation. In Fura2-loaded cells, 5 mM Ca 2ϩ o induced a rapid increase in Ca 2ϩ i levels, which was much larger than for the response to 2.5 mM Ca 2ϩ o (ϩ469%, p Ͻ 0.05), and the individual cellular responses were largely sustained/nonoscillatory (Fig. 4B).
Effect of Positive Allosteric Modulation on CaR T888 Phosphorylation-In light of the possibility that orthosteric stimulation of the receptor, i.e. with Ca 2ϩ o , can induce both CaR T888 phosphorylation and dephosphorylation at different rates, the effects of the positive allosteric modulator NPS-R467 (24) and the less potent CaR positive allosteric modulator L-phenylalanine (25) were next investigated. Incubation in 1.2 mM Ca 2ϩ o for 2, 5, or 10 min elicited only marginal CaR T888 phosphorylation of the 160-kDa band (Fig. 5A) as shown earlier (Fig. 2). However, when cotreated with NPS-R467, a significantly greater level of CaR T888 phosphorylation was seen after 2 min (p Ͻ 0.05) most likely as the calcimimetic sensitizes the CaR to the ambient Ca 2ϩ o (24). Following exposure to 5 min NPS-R467, the level of phosphorylation of the 160-kDa band declined, returning to baseline by 10 min (Fig. 5A). Together with the data from the concentration-effect experiments, this would suggest that addition of the calcimimetic NPS-R467 to the 1.2 mM Ca 2ϩ o -containing buffer elicits a response that is rapid and transient similar to the response elicited by 5 mM Ca 2ϩ o alone. In contrast, while cotreatment with 10 mM L-Phe (in the presence of 1.2 mM Ca 2ϩ o ) failed to induce significant CaR T888 phosphorylation (160-kDa mature protein) after 2 min, phosphorylation levels were raised significantly after 10 min, consistent with the more sustained response evoked by 2 mM Ca 2ϩ o alone (see Fig. 3). In Fura2-loaded cells, NPS-R467 induced a rapid and sustained increase in Ca 2ϩ i levels (in the presence of 1.5 mM Ca 2ϩ o ; Fig. 5B) similar to the response to 5 mM Ca 2ϩ o . As with calcimimetic cotreatment, Ca 2ϩ i concentration increased rapidly following the addition of 10 mM L-Phe (in the presence of 1.5 mM Ca 2ϩ o ), compared with the response seen in the presence of 1.5 mM Ca 2ϩ o alone (Fig. 5C). However the "global" change was lower than for the response to 1 M NPS-R467 (R467 response, 829% greater than L-Phe response, p Ͻ 0.001) and, significantly, the individual cellular responses were mostly slow Ca 2ϩ i oscillations, consistent with previous studies (26). Therefore, whereas at 1.5 mM Ca 2ϩ o , NPS-R467 (1 M) mimicked the   i mobilization, 10 mM L-Phe appeared to mimic more closely the CaR T888 phosphorylation and Ca 2ϩ i oscillatory responses to 2-2.5 mM (moderate) Ca 2ϩ o concentration. Consistent with this idea, the mutant receptor CaR T888A exhibits not oscillatory but sustained responses to L-Phe (supplemental Fig. 6) or moderate Ca 2ϩ o concentrations (14,15). CaR-induced Dephosphorylation of CaR T888 -Next, it was investigated whether CaR activation can stimulate CaR T888 dephosphorylation directly. To test this hypothesis, CaR-HEK cells were initially exposed to PMA (1 M in 0.5 mM Ca 2ϩ ocontaining buffer) for 10 min to induce maximal phosphorylation of CaR T888 (with 250 nM GF109203X added for the final 30 s) and then incubated for a further 30 s in the absence of PMA but again in the presence of the PKC inhibitor GF109203X (27) in buffer at a Ca 2ϩ o concentration of either 0.5, 1.2, 2.2, or 5 mM (Fig. 6). GF109203X was used to prevent further phosphorylation thereby allowing endogenous phosphatase activity to dephosphorylate the CaR T888 as described previously (15). Incubation in any of the PMA-free Ca 2ϩ o -containing buffers for the final 30 s resulted in a reduction in mature 160 kDa CaR T888 phosphorylation (Fig. 6) and in the 140-kDa core-glycosylated form (supplemental Fig. S7A), presumably due to endogenous phosphatase activity. Exposure to 5 mM Ca 2ϩ o significantly enhanced the dephosphorylation of P160 CaR T888 (p Ͻ 0.05; Fig. 6A) and also the P140 CaR T888 (supplemental Fig. S7B). By eliminating any further CaR-induced phosphorylation of CaR T888 in the final 30 s using the PKC inhibitor, this experiment reveals that exposure to high Ca 2ϩ o concentration elicits CaR T888 dephosphorylation. Thus, CaR T888 dephosphorylation is not "passive" but also dependent on receptor activation (15).
To determine whether the concentration-dependent, Ca 2ϩ oinduced decrease in CaR T888 phosphorylation was indeed a result of Ca 2ϩ o acting upon the CaR, and not due to a CaRindependent pathway, the effects of CaR-positive allosteric modulators on CaR T888 dephosphorylation were examined. The amino acid L-Phe (10 mM) was without effect in the presence of 1.2 mM Ca 2ϩ o (Fig. 6A) but elicited significant CaR T888 dephosphorylation in the presence of 2.2 mM Ca 2ϩ o (Fig. 6B). Similarly, the calcimimetic NPS-R467 (1 M) also decreased CaR T888 phosphorylation in 2.2 mM Ca 2ϩ o -containing buffer ( Fig. 6C; supplemental Fig. 7C). These acute treatments had no effect on the level of total-CaR present within the cells (Fig. 6). These data indicate that the increased rate of CaR T888 dephosphorylation induced by elevated Ca 2ϩ o concentration is dependent on CaR activation. Next, to confirm whether high Ca 2ϩ o concentration directly activates PP2A, CaR-HEK cells were incubated for up to 5 min in buffer containing 0.5 or 5 mM Ca 2ϩ o and then lysed and processed for the human active PP2A assay. Indeed, high Ca 2ϩ o concentration increased PP2A activity significantly within 2 min, an effect that was sustained until 5 min (Fig. 6D).
Effect of Phosphatase Inhibition on PTH Secretion from Bovine Parathyroid Slices-We next assessed whether CaR T888 phosphorylation status has a functional impact upon PTH secretion and thus may contribute to the modulation of calcium homeostasis in vivo. Therefore, we investigated the effect of administering calyculin A on the secretion of PTH from freshly isolated bovine parathyroid glands. The PP1/PP2A protein phosphatase inhibitor calyculin-A was chosen as it had been previously found to promote net CaR T888 phosphorylation in PMA-treated CaR-HEK cells (15). Slices of bovine parathyroid gland were perifused (21) with 1.2 mM Ca 2ϩ o , a concentration that partially activates the CaR and submaximally suppresses PTH secretion (1-3; see also supplemental Fig. 8), in the presence of the PP1/PP2A phosphatase inhibitor calyculin-A (100 nM). Following removal of calyculin A, suppression of PTH secretion was observed within 20 min (p Ͻ 0.01; Fig. 7), whereas in the continued presence of calyculin A, PTH secretion was maintained (p Ͻ 0.05 versus slices in which calyculin was removed; Fig. 7A). In separate parathyroid slices incubated (as described under "Experimental Procedures") in the absence or presence of 100 nM calyculin (plus 1 M PMA), significant CaR T888 phosphorylation was observed after both 10-and 30-min time points (Fig.  7B). Thus, the maintenance of PTH secretion in the presence of calyculin A is consistent with the hypothesis that blocking CaR T888 dephosphorylation deactivates the inhibitory impact of the receptor on PTH secretion. This was further demonstrated in Fura2-loaded parathyroid cells in which calyculin A suppressed high Ca 2ϩ o -induced Ca 2ϩ i mobilization (Fig. 7C). Human Parathyroid Cell Experiments-Finally, we also tested the effect of calyculin A on high Ca 2ϩ -induced suppression of PTH secretion from perifused normal human parathy-roid cells. In the absence of calyculin A, PTH secretion from these cells was reversibly stimulated by low Ca 2ϩ o concentrations (0.8 or 1.0 mM) and suppressed at higher Ca 2ϩ o concentrations (1.2 or 1.4 mM) as expected (1,2). In three similar experiments, in the presence of elevated Ca 2ϩ o , 100 nM calyculin A induced a progressive increase in PTH secretion over 10 -15 min (0.95 Ϯ 0.19 fg min Ϫ1 cell Ϫ1 PTH secretion in 1.4 mM Ca 2ϩ o before calyculin A treatment, 5.26 Ϯ 0.77 at the peak of the calyculin response; p Ͻ 0.01 by unpaired t test), which slowly recovered to baseline levels upon return to the control solution.

DISCUSSION
PKC-mediated modulation of class C G protein-coupled receptor signaling is emerging as a significant determinant of receptor function for these receptors. It was first shown that mGluR5-induced intracellular Ca 2ϩ oscillations could be abolished by PKC activators (16)(17)(18). Initially, mGluR5 T840 , located in the juxtamembrane region of the intracellular domain of the receptor, was proposed to be the PKC site that mediates this response (16). However, mutation of mGluR5 T840 did not prevent oscillations (28), whereas the adjacent residue mGluR5 S839 has since been identified as the PKC site most likely responsible (29). Given the sequence and structural homology between the CaR and the mGluRs (1,11), it is perhaps not surprising that a number of functional features of mGluR signaling also apply to the CaR. Having previously shown that CaR T888 is subject to agonistinduced phosphorylation (15), we demonstrate here, for the first time, agonist-induced dephosphorylation of a class C G protein-coupled receptor. Such receptor-elicited dephosphorylation of this protein kinase C site could help explain the control of signal oscillation frequency observed in CaR (30), given that CaR T888 dephosphorylation appears to increase oscillation frequency and/or initiate sustained Ca 2ϩ i mobilization in response to heightened CaR activation. In contrast, Ca 2ϩ i oscillation frequency in mGluR5-transfected cells is largely independent of agonist concentration (18). Therefore, it is possible that mGluR5 dephosphorylation is not agonist concentration-dependent to the same extent as CaR. Neverthe-

less, calyculin-A decreases Ca 2ϩ
i oscillation frequency in both mGluR5-and CaR-transfected cells (15,18) and thus phosphatase activity is a signaling determinant for both receptors. While the identity of the protein phosphatase responsible for CaR T888 dephosphorylation was not determined here, we have shown previously that CaR T888 dephosphorylation can be inhibited by the PP1/PP2A inhibitors calyculin-A and endothall, with the PP1 inhibitor tautomycin being relatively ineffective and that PP2A colocalizes with phosphorylated CaR in CaR-HEK cells (15). Furthermore in the current study, elevated Ca 2ϩ o concentration stimulated PP2A activity (Fig. 6D). Thus together, these data suggest strongly that PP2A is at least partly responsible for CaR T888 dephosphorylation.
Despite their widespread usage in studies of CaR activity, HEK-293 cells are merely a model for endogenous CaR expression e.g. by parathyroid cells. Nevertheless, the CaR responses in both cell systems show a significant number of similarities including inositol 1,4,5-trisphosphate-generated Ca 2ϩ i mobili-zation (including Ca 2ϩ i oscillations), suppression of cAMP formation, mitogen-activated protein kinase (MAPK) activation, and a lack of electrical excitability (2). Furthermore, when HEK-293 cells are transfected with the PTH and CaR genes, CaR activation increases the instability of the PTH mRNA transcript, just as for parathyroid cells (31). Finally, this and our previous study (15) demonstrate that in both cell types, CaRinduced Ca 2ϩ i mobilization can be attenuated reversibly by PKC activation or PP1/PP2A inhibition and that CaR T888 is phosphorylated upon PKC activation.
After finding evidence of a role for a calyculin A-sensitive phosphatase in CaR-HEK cells, we explored the functional impact of calyculin A on PTH secretion from perifused bovine parathyroid slices and normal human parathyroid cells. CaR stimulation suppresses PTH secretion (2), but it was demonstrated here that cotreatment with calyculin A prevented and even reversed high Ca 2ϩ o -induced suppression of PTH secretion (Figs. 7 and 8). Subsequent removal of calyculin A restored the normal Ca 2ϩ o -induced suppression of PTH secretion. The mechanism responsible for 140 kDa CaR T888 phosphorylation is unclear. Deglycosylation of the 140-kDa CaR band using endoglycosidase H (32, 33) reveals it to be the high mannose/core-glycosylated form of the CaR, whereas the 160-kDa CaR band is only susceptible to deglycosylation using PNGaseF, indicative of mature glycosylation (32,33). Indeed, only the 160-kDa protein can be detected on the cell membrane (34), albeit as a small fraction of the total 160-kDa CaR pool. The ability of Ca 2ϩ o to induce phosphatidylinositol hydrolysis in cells expressing various glycosylation-deficient CaR mutants declines in proportion to the reduction in cell surface expression of the mutant CaRs (33). Similarly, tunicamycin, which prevents mature glycosylation of the CaR (32,35), decreased CaR-induced phosphatidylinositol hydrolysis (35). Together, these findings suggest that plasma membrane expression is required for CaR activity and that the phosphorylation of the 140-kDa CaR protein is likely to occur secondary to 160 kDa CaR activation on the membrane. However, we cannot exclude  the possibility that the immature 140-kDa form of the CaR is activated intracellularly, for example, in response to elevated intravesicular Ca 2ϩ levels, thereby inducing its own PKC-mediated phosphorylation. It should be noted in this regard that whereas the Ca 2ϩ o concentration-dependence of 160 kDa CaR T888 phosphorylation was clearly biphasic, CaR T888 phosphorylation of the 140-kDa band was sustained even after 20 min in the presence of 5 mM Ca 2ϩ o (supplemental Fig. 5). Therefore, the phosphatase responsible for dephosphorylation of the 160-kDa protein was less effective at, or had less access to, the 140-kDa protein. Thus, the CaR-stimulated phosphatase activity we report here is capable of dephosphorylating 160-kDa CaR T888 in response to CaR activation, but not 140-kDa CaR. It should be noted that if dephosphorylation of 160-kDa CaR occurs mostly, if not exclusively, at the membrane, then the current data may have underestimated the magnitude of the effect at the membrane as the immunoblots here include all p160 CaR, whether derived from the plasma membrane or intracellular organelles.
Together, the data presented in the current study demonstrate a biphasic Ca 2ϩ o concentration-dependence for CaR T888 phosphorylation in CaR-HEK cells that is most likely due to a disproportionate rise in CaR T888 dephosphorylation at higher levels of receptor stimulation, thus permitting more Ca 2ϩ i mobilization. Although the calyculin A-sensitive phosphatase responsible for mediating this effect has not yet been identified, this work highlights the functional importance of phosphataseregulatory mechanisms in the control of CaR function and the findings from the experiments on bovine parathyroid slices suggest that these effects may also be physiologically important for the regulation of calcium homeostasis in vivo.