Parathyroid Hormone Controls the Size of the Intracellular Ca2+ Stores Available to Receptors Linked to Inositol Trisphosphate Formation*

In HEK 293 cells stably expressing type 1 parathyroid (PTH) receptors, PTH stimulated release of intracellular Ca2+ stores in only 27% of cells, whereas 96% of cells responded to carbachol. However, in almost all cells PTH potentiated the response to carbachol by about 3-fold. Responses to carbachol did not desensitize, but only the first challenge in Ca2+-free medium caused an increase in [Ca2+] i , indicating that the carbachol-sensitive Ca2+ stores had been emptied. Subsequent addition of PTH also failed to increase [Ca2+] i , but when it was followed by carbachol there was a substantial increase in [Ca2+] i . A similar potentiation was observed between ATP and PTH but not between carbachol and ATP. Intracellular heparin inhibited responses to carbachol and PTH, and pretreatment with ATP and carbachol abolished responses to PTH, suggesting that the effects of PTH involve inositol trisphosphate (IP3) receptors. PTH neither stimulated detectable IP3 formation nor affected the amount formed in response to ATP or carbachol. PTH stimulated cyclic AMP formation, but this was not the means whereby PTH potentiated Ca2+ signals. We suggest that PTH may regulate Ca2+ mobilization by facilitating translocation of Ca2+ between discrete intracellular stores and that it thereby regulates the size of the Ca2+ pool available to receptors linked to IP3 formation.

Parathyroid hormone (PTH) 1 plays a central role in plasma Ca 2ϩ homeostasis, and many tissues that are not involved in Ca 2ϩ regulation also express PTH receptors (1). Two PTH receptor subtypes (types 1 and 2) have been identified; their cDNA sequences share 70% similarity (2)(3)(4)(5) and identify them as members of a subfamily of G protein-coupled receptors to which the receptors for calcitonin, secretin, ACTH, and glucagon also belong.
It has long been accepted that PTH stimulates an increase in both intracellular cyclic AMP and [Ca 2ϩ ] i in many cell types (6,7). The ability of PTH to activate two signaling pathways was generally assumed to result from its interaction with the two receptor subtypes, an assumption that gained support from the observation that the two pathways could be differentially stimulated by truncated forms of PTH (8). More recently, expression of recombinant receptors has established that type 1 (3,9,10) and type 2 (11) PTH receptors are each alone capable of stimulating increases in both [Ca 2ϩ ] i and intracellular cyclic AMP. Other members of the family of receptors to which the PTH receptor belongs share this ability to independently stimulate cyclic AMP formation and an increase in [Ca 2ϩ ] i (3,(12)(13)(14).
In some cells, the increase in [Ca 2ϩ ] i evoked by PTH is mediated largely by effects of cyclic AMP on Ca 2ϩ entry (7), but in osteoblasts and kidney cells (6) and cells expressing recombinant type 1 (3,9,10) or type 2 (11) PTH receptors, PTH also stimulates release of intracellular Ca 2ϩ stores. The means whereby PTH causes such Ca 2ϩ mobilization is unclear. In some cells, PTH stimulates formation of inositol 1,4,5-trisphosphate (IP 3 ) (3,6), which presumably then causes Ca 2ϩ mobilization via IP 3 receptors. In other situations, PTH and agonists that stimulate IP 3 formation appear to release Ca 2ϩ from different intracellular stores (9,15), and PTH-stimulated Ca 2ϩ mobilization occurs without detectable formation of IP 3 (10,15) and in the presence of antagonists of IP 3 receptors (9, 10). Furthermore, PTH-evoked Ca 2ϩ release appears not to be mediated by ryanodine receptors (9) or by either cyclic ADP-ribose or NAADP ϩ (9), both of which have been shown to stimulate Ca 2ϩ release in other cells (16). In short, the ability of PTH to stimulate Ca 2ϩ mobilization cannot easily be explained by the properties of known intracellular signaling pathways.
In the present study, we examine the mechanisms underlying the effects of PTH on intracellular Ca 2ϩ stores in human embryonic kidney 293 cells stably expressing type 1 PTH receptors (HEK/PTH-R1 cells).

EXPERIMENTAL PROCEDURES
Materials-The full-length cDNA encoding the human type 1 PTH receptor in the vector, pcDNAI/neo, was a gift from Dr. K. Seuwen (Basel, Switzerland) (5). HEK 293 cells were from the European Collection of Animal Cell Cultures (Porton Down, UK). Human PTH , Rp-8-Br-cAMPS, ionomycin, Xestospongin A, wortmannin, and H89 were from Calbiochem (Nottingham, UK). Fura-2AM and Cascade Blue were from Molecular Probes (Leiden, The Netherlands). Forskolin, heparin, and poly-L-lysine were from Sigma. Tissue culture media, G-418, LipofectAMINE, and OptiMEM were from Life Technologies, Inc. The Biotrak cyclic AMP assay kit and D-myo-[2-3 H] inositol were from Amersham Pharmacia Biotech.
Transfection and Cell Culture-HEK 293 cells were transfected with the expression vector, pRezex-2, containing the complete coding sequence of the human type 1 PTH receptor (5). Cells expressing PTH receptors were selected by screening for those that responded to PTH with an increase in [Ca 2ϩ ] i (see below). Cells stably transfected with the PTH receptor (HEK/PTH-R1) were grown in a Dulbecco's modified Eagle's medium/Ham's F-12 medium supplemented with fetal calf serum (10%), glutamine (2 mM), and G-418 (800 g/ml). Responses to PTH were maintained during at least 40 passages. * This work was supported by Wellcome Trust Grant 039662. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Measurement of [Ca 2ϩ
] i -HEK/PTH-R1 cells were plated (2.5 ϫ 10 6 cells/ml) onto glass coverslips coated with poly-L-lysine and grown for at least 2 days. Confluent cells were loaded with Fura-2 by incubation with Fura-2AM (1 M, 60 min) dissolved in extracellular medium (ECM; 130 mM NaCl, 5.4 mM KCl, 0.8 mM NaH 2 PO 4 , 1.8 mM CaCl 2 , 0.9 mM MgSO 4 , 10 mM glucose, 20 mM Hepes, pH 7.4) at 20°C for 45 min. The cells were then washed, and after 30 min in ECM, they were used for experiments. Previous experiments established that with this loading protocol, Ͼ95% of the Fura-2 fluorescence came from Fura-2 that was both cytosolic and hydrolyzed to its Ca 2ϩ -sensitive form (17). Single cell fluorescence imaging at 21°C using IonVision software, corrections for background fluorescence, and calibration of fluorescence ratios (R 340/380 ) to [Ca 2ϩ ] i were performed as described previously (17). For microinjection with heparin, Fura-2-loaded cells bathed in nominally Ca 2ϩ -free ECM were injected with medium (27 mM K 2 HPO 4 , 8 mM NaH 2 PO 4 , 26 mM KH 2 PO 4 , pH 7.3) containing 10 mg/ml heparin and 1 mg/ml Cascade Blue as a marker for microinjected cells (18). After 30 min, cells were restored to Ca 2ϩ -containing ECM prior to measurement of [Ca 2ϩ ] i .
Measurement of Cyclic AMP Production-Confluent cultures of HEK/ PTH-R1 cells in 6-well plates were preincubated in ECM containing 3-isobutyl-1-methylxanthine (1 mM, 30 min). Agonists were then added, and the incubations were terminated after 10 min by the removal of the medium and the addition of cold 0.1 M HCl in 65% methanol (750 l). After 30 min on ice, the contents of the wells were scraped, washed (750 l acidified methanol) into tubes, and centrifuged (1000 ϫ g, 30 min). The supernatants were dried under vacuum (60°C, 4 h) and stored at Ϫ80°C. A commercial enzyme-linked immunosorbent assay kit using the nonacetylation protocol specified by the manufacturer (Amersham Pharmacia Biotech) was used to determine the cyclic AMP contents of the samples.
Measurement of IP 3 Production-HEK/PTH-R1 cells (2.5 ϫ 10 6 cells/ ml) plated onto poly-L-lysine-coated glass coverslips were grown for 48 h in inositol-free Dulbecco's modified Eagle's medium supplemented with 2% fetal calf serum and D-myo-[2-3 H] inositol (2 Ci/ml). Cells were washed twice in ECM and incubated for 15 s with the appropriate agents at 21°C. Reactions were stopped by aspiration and addition of 0.5 ml of 1 mM ice-cold perchloric acid and kept on ice for 30 min. Acid extracts were neutralized and then loaded onto anion exchange columns from which the [ 3 H]IP 3 fraction was eluted (18).
Analysis-For most experiments, responses from 25-30 single cells on a coverslip were averaged. Statistical analyses were applied to the average results derived from independent measurements from several coverslips, where the sample size refers to the number of coverslips. Concentration-effect relationships were fitted to four-parameter logistic equations using least squares curve-fitting routines (Kaleidagraph, Synergy Software).

Ca 2ϩ Mobilization Evoked by Type 1 PTH Receptors-Stimulation of the endogenous muscarinic receptors of HEK 293 cells with carbachol (100 M) evoked an increase in [Ca 2ϩ ] i
irrespective of whether the cells had been transfected with PTH receptors (Fig. 1, A-C). PTH (1 M), however, stimulated an increase in [Ca 2ϩ ] i only in HEK/PTH-R1 cells, and, in keeping with previous reports comparing PTH with agonists that stimulate IP 3 formation (9,15,19), the maximal response to PTH was only 21 Ϯ 5% (n ϭ 7) of that evoked by carbachol (Fig.  1B). In one previous study (19) but not in others (10,11,20,21), it was claimed that HEK 293 cells express type 1 PTH receptors. Our results are consistent with most previous work in demonstrating that PTH evokes Ca 2ϩ mobilization in HEK 293 cells only after transfection with type 1 PTH receptors ( Fig. 1).
Although the sustained phase of the Ca 2ϩ signal evoked by carbachol and the peak amplitudes of the responses evoked by PTH or carbachol were reduced by removal of extracellular Ca 2ϩ (Fig. 1, B and C), both agonists consistently evoked an increase in [Ca 2ϩ ] i in the absence of extracellular Ca 2ϩ (Fig.  1C), and responses to both were abolished after pretreatment with thapsigargin in Ca 2ϩ -free ECM (not shown). We conclude that both PTH and carbachol stimulate release of intracellular Ca 2ϩ stores. Subsequent experiments addressed the mechanisms responsible for this Ca 2ϩ mobilization.
Single cell analyses of HEK/PTH-R1 cells revealed that whereas 96 Ϯ 5% (n ϭ 12 coverslips) of cells responded to carbachol (100 M) with an increase in [Ca 2ϩ ] i , only 27 Ϯ 20% of cells responded to PTH (100 nM). In those cells that responded to PTH, the peak increase in [Ca 2ϩ ] i was 84 Ϯ 16% (n ϭ 11) of that evoked by carbachol in the same cells (Fig. 1D). Single cell analyses of HEK/PTH-R1 cells were used for all subsequent experiments (see "Experimental Procedures").
The magnitude of the responses to maximal concentrations of carbachol and PTH varied between experiments, although the relative amplitudes of the responses to the two stimuli were maintained. For each cell, we therefore expressed all responses as percentages of the response evoked by carbachol (100 M) in normal ECM at the beginning of the experiment. After stimulation with carbachol in normal ECM, cells were allowed to recover for 5 min. Further stimulation then evoked indistin-  2B). From the average responses of each cell in the field, the peak rise in [Ca 2ϩ ] i evoked by PTH after carbachol was more than 10-fold greater than that evoked by PTH alone (160 Ϯ 10%, n ϭ 13; and 12 Ϯ 7%, n ϭ 11, respectively) (Fig. 2C). The interaction between PTH and carbachol was also evident when cells were first stimulated with PTH and then with carbachol; the peak rise in [Ca 2ϩ ] i evoked by carbachol alone was 71 Ϯ 8% (n ϭ 13) of the initial response, and after pretreatment with PTH it increased to 245 Ϯ 10% (n ϭ 8) (Fig. 2, A and C). Similar results were obtained in Ca 2ϩ -containing ECM (Fig. 2C).
PTH increased both the maximal amplitude and the sensitivity of the response to carbachol (Fig. 3A). The half-maximal effect (EC 50 ) of carbachol was reduced from 71 Ϯ 6 M (n ϭ 3) in control cells to 25 Ϯ 3 M (n ϭ 3) in the presence of PTH. The EC 50 for the effect of PTH in potentiating responses to carbachol (100 M) was 7.8 Ϯ 0.5 nM (Fig. 3B).
Because most HEK/PTH-R1 cells respond to PTH after carbachol pretreatment, whereas untransfected cells are unresponsive, we conclude that even though few HEK/PTH-R1 cells respond directly to PTH alone, most express functional PTH receptors. Carbachol unmasks a latent ability of PTH to stimulate release of intracellular Ca 2ϩ stores, and PTH increases by 2-3 fold the Ca 2ϩ mobilization evoked by carbachol, which alone stimulates Ca 2ϩ mobilization in all cells. Potentiation of the responses to carbachol by PTH occurred despite [Ca 2ϩ ] i having returned to its basal level ( Fig. 2A) (Fig. 4A). PTH alone also failed to increase [Ca 2ϩ ] i , when added 5 min after carbachol had been removed (Fig. 4A). However, a further challenge with carbachol after addition of PTH evoked a large transient increase in [Ca 2ϩ ] i , equivalent to 135 Ϯ 15% (n ϭ 3) of a standard response to carbachol (Fig. 4A). We conclude that PTH allows carbachol to release intracellular Ca 2ϩ stores that would otherwise be insensitive to carbachol. Similar results were obtained when the order of addition of carbachol and PTH was reversed. PTH caused a substantial mobilization of intracellular Ca 2ϩ stores when added immediately after carbachol had been removed, even though the carbachol itself evoked no increase in [Ca 2ϩ ] i because the carbachol-sensitive Ca 2ϩ stores had already been emptied by prior stimulation (Fig. 4B).
Cross-potentiation of [Ca 2ϩ ] i Signals by ATP and PTH-ATP also stimulates IP 3 formation and Ca 2ϩ mobilization in HEK/ PTH-R1 cells; the Ca 2ϩ mobilization evoked by 10 M ATP was 74 Ϯ 5% (n ϭ 7) of the standard carbachol response (Fig. 5A). Although only 71 Ϯ 7% (n ϭ 6) of cells responded to ATP, in the responsive cells the synergistic interaction with PTH was again evident. PTH added immediately after removal of ATP caused a substantial increase in [Ca 2ϩ ] i ; the response to PTH was 160 Ϯ 30% (n ϭ 4) of the standard carbachol response (Fig. 5A). When PTH was added first, the increase in [Ca 2ϩ ] i evoked by subsequent addition of ATP was increased to 159 Ϯ 28% (n ϭ 3) of the standard carbachol response (Fig. 5B). There was no synergistic interaction between ATP and carbachol (not shown). Carbachol did not increase the fraction of cells that responded to ATP (73 Ϯ 11% before carbachol, and 74 Ϯ 4% after carbachol, n ϭ 4), and in cells that responded to both stimuli, the increase in [Ca 2ϩ ] i evoked by combined application of carbachol and ATP (198 Ϯ 5%) was similar to the sum of the responses evoked by either agonist alone (71 Ϯ 5 and 97 Ϯ 35% for carbachol and ATP, respectively).
Mechanism of Action of PTH: the Role of IP 3 Receptors-Responses to IP 3 are positively cooperative (22). PTH might therefore have potentiated responses to carbachol and ATP by either directly stimulating formation of a subthreshold level of IP 3 or by increasing the amount of IP 3 made in response to ATP or carbachol. Neither explanation seems likely because PTH (10 M) neither directly stimulated detectable IP 3 formation (100 Ϯ 23% of control, n ϭ 6) nor affected the amount made in response to ATP (174 Ϯ 23 and 166 Ϯ 14% in the absence and presence of PTH) or carbachol (246 Ϯ 31 and 266 Ϯ 34%). Ca 2ϩ stimulates IP 3 receptors (22), but this cannot explain the ability of PTH to potentiate responses to carbachol and ATP because the potentiation occurred without an increase in [Ca 2ϩ ] i (Figs. 2, 4, and 5).
Xestospongin, which has been shown to block cerebellar IP 3 receptors (23), proved not to be useful in resolving the involvement of IP 3 receptors in the response to PTH. A range of concentrations (10 -50 M) and preincubation periods (10 -40 min) failed to separate direct effects of Xestospongin on the emptying of intracellular Ca 2ϩ stores from inhibition of IP 3 receptors (not shown) (18). Two lines of evidence are, however, consistent with IP 3 receptors mediating the enhanced Ca 2ϩ release evoked by ATP or carbachol with PTH. First, and in contrast with a previous report (10), microinjected heparin (10 mg/ml in the injection pipette) was similarly effective in blocking both the PTH-and carbachol-evoked increase in [Ca 2ϩ ] i (Fig. 6). Second, after combined application of ATP and carbachol to release the IP 3 -sensitive Ca 2ϩ stores, the subsequent response to PTH was almost abolished (down from 160 Ϯ 10 to 25 Ϯ 22%, n ϭ 7).
A plausible explanation for these observations is that PTH generates a signal that causes IP 3 receptors to become more sensitive to IP 3 . The increase in [Ca 2ϩ ] i evoked by PTH alone in some cells might then result from a sufficient basal level of IP 3 to cause Ca 2ϩ mobilization from sensitized IP 3 receptors. The effects of PTH were persistent; even 10 min after removal of PTH, enhanced increases in [Ca 2ϩ ] i were observed after carbachol addition (not shown); and this presumably reflects the slow dissociation of PTH from its high affinity receptor (3). However, the ability of carbachol to enhance the rise in [Ca 2ϩ ] i evoked by PTH was short-lived and reversed within the ϳ10 s taken for the three washes required to completely remove the carbachol-containing medium, in keeping with the relatively low affinity of muscarinic receptors for carbachol and the rapid metabolism of IP 3 .
Cyclic AMP Does Not Mediate the Effects of PTH on Ca 2ϩ Mobilization-PTH stimulates adenylyl cyclase activity in HEK 293 cells with an EC 50 of 20 Ϯ 7 nM (n ϭ 3; not shown); cyclic AMP might therefore have mediated the increases in [Ca 2ϩ ] i evoked by PTH. Cyclic AMP-dependent protein kinase (PKA) phosphorylates IP 3 receptors, and, although the functional consequences differ between cell types, in hepatocytes, which like HEK 293 cells express largely type 2 IP 3 receptors (24), phosphorylation increases the sensitivity of IP 3 receptors to IP 3 (25). Forskolin was previously shown not to mimic the effects of activating type 1 PTH receptors on [Ca 2ϩ ] i (10), but forskolin (5 M) alone only very modestly increased intracellular cyclic AMP (from 0.53 Ϯ 0.05 to 0.91 Ϯ 0.13 pmol/well, n ϭ 3) relative to the increase evoked by PTH (to 3.04 Ϯ 0.05 pmol/well). Similarly modest effects of forskolin on the endogenous adenylyl cyclase activity of HEK 293 cells has been reported previously (26). However, several additional lines of evidence demonstrate that neither cyclic AMP nor PKA mediate the effects of PTH on intracellular Ca 2ϩ stores. First, pretreatment with 8-bromo cyclic AMP (50 M, 5 min) reduced the Ca 2ϩ mobilization evoked by carbachol (Fig. 7A). Second, pretreatment of HEK/PTH-R1 cells with either of two inhibitors of  (28), affected neither the response to PTH alone nor its ability to potentiate the response to carbachol (Fig. 7B). That these inhibitors had effectively blocked activation of PKA is confirmed by their ability to reverse the inhibition of carbacholevoked Ca 2ϩ mobilization caused by 8-bromo-cAMP (Fig. 7A). Third, 3-isobutyl-1-methylxanthine, which by inhibiting cyclic AMP degradation would be expected to potentiate responses to PTH if they were mediated by cyclic AMP, had the opposite effect. Pretreatment with 3-isobutyl-1-methylxanthine (1 mM, 20 min) abolished the potentiation of carbachol responses by either submaximal (6 nM) or maximal (100 nM) concentrations of PTH (not shown). We conclude that the ability of PTH to potentiate carbachol-evoked Ca 2ϩ mobilization is not mediated by cAMP. Indeed PKA, by a mechanism that we have not addressed further, appears to cause desensitization of either muscarinic or PTH receptors.
Conclusions-We conclude, in keeping with previous work (29), that PTH-mediated Ca 2ϩ mobilization does not result from production of IP 3 ; PTH does not stimulate detectable formation of IP 3 , it has no detectable effect on the IP 3 formation evoked by ATP or carbachol, and the lack of synergy between ATP and carbachol indicates that the IP 3 formed in response to these receptors is incapable of mimicking PTH. Nor are the potentiated Ca 2ϩ signals evoked by PTH mediated by cyclic AMP (Fig. 7), by Ca 2ϩ -induced sensitization of IP 3 receptors (Figs. 2, 4, and 5), or by phosphatidylinositol-3-kinase, because wortmannin (1 M) did not prevent PTH from potentiating responses to carbachol (not shown). Our data are nevertheless consistent with IP 3 receptors being involved in responses to PTH (Fig. 6). We suggest that activation of type 1 PTH receptors stimulates formation of an intracellular messenger that allows more complete emptying of intracellular Ca 2ϩ stores than is possible with maximal concentrations of either ATP or carbachol alone. To account for the larger maximal response to carbachol in the presence of PTH (Fig. 3A), we suggest that PTH may facilitate translocation of Ca 2ϩ between discrete intracellular Ca 2ϩ stores (Fig. 8), such that stores lacking IP 3 receptors are brought into continuity with those expressing them. Such a model would account for the small amount of Ca 2ϩ release evoked by PTH alone (triggered by basal levels of IP 3 ) (Fig. 1C), for the increase in the size of the pool released by maximal concentrations of carbachol or ATP (Figs. 4 and 5B), and for the increased sensitivity of the stores to carbachol (Fig.  3A).
At present we can only speculate on the mechanisms linking PTH receptors to the transfer of Ca 2ϩ between intracellular stores, although it is tempting to suggest that they may be related to the many reports linking small G proteins with Ca 2ϩ translocation between intracellular stores. For example, previous studies had demonstrated that GTP facilitated transfer of Ca 2ϩ between IP 3 -sensitive and -insensitive stores by a process that required GTP hydrolysis and probably involved membrane fusion, the cytoskeleton, and a monomeric G protein (30 -32). Inositol 1,3,4,5-tetrakisphosphate has also been reported to allow IP 3 to more completely empty intracellular Ca 2ϩ stores and to perhaps thereby also promote increased capacitative Ca 2ϩ entry (33,34). Again monomeric G proteins are implicated, because the effects of inositol 1,3,4,5-tetrakisphosphate on IP 3 -evoked Ca 2ϩ mobilization appear to be mediated by a member of the GAP1 family of GTPase-activating proteins (33,35). Indeed monomeric G proteins are involved in every aspect of intracellular membrane trafficking (36), including that between the organelles, endoplasmic reticulum and the Golgi, known to respond to IP 3 (37).
Most extracellular stimuli cause Ca 2ϩ mobilization by stimulating formation of the IP 3 that gates intracellular Ca 2ϩ channels. We suggest that PTH may be the first example of a hormone shown to regulate Ca 2ϩ mobilization by regulating the size of the Ca 2ϩ pool to which IP 3 has access. FIG. 8. A possible mechanism for the effects of PTH on intracellular Ca 2؉ stores. The type 1 PTH receptor, via a mechanism that does not require cyclic AMP or PI-3-kinase, is proposed to allow IP 3 -sensitive Ca 2ϩ stores to become linked to those lacking IP 3 receptors. PTH thereby increases the size of the intracellular Ca 2ϩ stores available to receptors (for ATP and carbachol (CCh)) that stimulate IP 3 formation.