Research ArticleGPCR SIGNALING

Agonist-Driven Maturation and Plasma Membrane Insertion of Calcium-Sensing Receptors Dynamically Control Signal Amplitude

See allHide authors and affiliations

Science Signaling  22 Nov 2011:
Vol. 4, Issue 200, pp. ra78
DOI: 10.1126/scisignal.2002208

Abstract

Calcium-sensing receptors (CaSRs) regulate systemic calcium homeostasis in the parathyroid gland, kidney, intestine, and bone and translate fluctuations in serum calcium into peptide hormone secretion, cell signaling, and regulation of gene expression. The CaSR is a G protein (heterotrimeric guanosine triphosphate–binding protein)–coupled receptor that operates in the constant presence of agonist, sensing small changes with high cooperativity and minimal functional desensitization. Here, we used multiwavelength total internal reflection fluorescence microscopy to demonstrate that the signaling properties of the CaSR result from agonist-driven maturation and insertion of CaSRs into the plasma membrane. Plasma membrane CaSRs underwent constitutive endocytosis without substantial recycling, indicating that signaling was determined by the rate of insertion of CaSRs into the plasma membrane. Intracellular CaSRs colocalized with calnexin in the perinuclear endoplasmic reticulum and formed complexes with 14-3-3 proteins. Ongoing CaSR signaling resulted from agonist-driven trafficking of CaSR through the secretory pathway. The intracellular reservoir of CaSRs that were mobilized by agonist was depleted when glycosylation of newly synthesized receptors was blocked, suggesting that receptor biosynthesis was coupled to signaling. The continuous, signaling-dependent insertion of CaSRs into the plasma membrane ensured a rapid response to alterations in the concentrations of extracellular calcium or allosteric agonist despite ongoing desensitization and endocytosis. Regulation of CaSR plasma membrane abundance represents a previously unknown mechanism of regulation that may be relevant to other receptors that operate in the chronic presence of agonist.

Introduction

The calcium-sensing receptor (CaSR) is a G protein (heterotrimeric guanosine triphosphate–binding protein)–coupled receptor (GPCR) that plays a central role in maintaining the serum concentration of ionized Ca2+ over a narrow range, from 1.1 to 1.3 mM, by acutely regulating parathyroid hormone (PTH) secretion and synthesis (13) as well as parathyroid cell proliferation (1). Mutations, polymorphisms, or altered expression of the CASR gene causes Ca2+ handling diseases (4), indicating that the CaSR plays critical roles in the intestine, kidney, and bone to regulate organismal Ca2+ homeostasis (1).

The CaSR is unusual among GPCRs in that it is chronically exposed to agonist during its transit through the cell, from the endoplasmic reticulum (ER) to the plasma membrane. The apparent affinity of the CaSR for Ca2+ is low [effective concentration (EC50) ≈ 3 mM] (1), consistent with its function in sensing serum Ca2+ concentrations. A hallmark of the CaSR, which was recognized before its cloning and functional characterization (2), is the steep dependence on the concentration of extracellular Ca2+ over the physiological range, with estimated Hill coefficients of 3 to 4 for activation of signaling or PTH secretion. Despite chronic exposure to agonist, the CaSR shows relatively weak functional desensitization (1, 5, 6). Western blots or immunohistochemistry of endogenous CaSRs suggests the presence of substantial amounts of immature, intracellular receptors (79). Despite these peculiarities, the CaSR clearly couples to heterotrimeric G protein signaling pathways and interacts with GPCR signaling adaptors including β-arrestins (10, 11).

The CaSR affinity for Ca2+ or allosteric modulators is low (in the micromolar to millimolar range), and current understanding of CaSR structure, function, and regulation derives solely from mutation-induced changes in signaling outputs of lipid, Ca2+, or mitogen-activated protein kinase (MAPK) pathways (12). These signaling pathways exhibit complex regulation, desensitization, and hysteresis independently of the activating GPCR, making it difficult to ascribe alterations in output exclusively to the CaSR. Understanding the factors that regulate the life cycle of the CaSR in a physiological context is critical for the development of strategies to regulate CaSR functional responses. Here, we use total internal reflection fluorescence microscopy (TIRFM) to image live cells and dissect the mechanisms regulating CaSR trafficking to and from the plasma membrane. Our results define a previously unknown mechanism that we term agonist-driven insertional signaling (ADIS), which accounts for the disparate and unique features of CaSR signaling, including the high degree of cooperativity and lack of substantial functional desensitization. ADIS may be a regulatory mechanism that is relevant to both GPCRs and other classes of receptors that signal in the face of chronic agonist exposure.

Results

We used live-cell multiwavelength TIRFM and BSEP-CaSR to simultaneously measure net plasma membrane abundance of CaSRs and endocytosis of CaSRs. BSEP-CaSR contains N-terminal additions of an α-bungarotoxin binding site (B) (13) followed by pH-sensitive superecliptic pHluorin (SEP) (14, 15) inserted immediately after the cleaved signal sequence (Fig. 1A). The bungarotoxin binding site on plasma membrane–localized BSEP-CaSR can be labeled with bungarotoxin conjugated to Alexa Fluor 594 (BTx-A594) to monitor endocytosis. The BSEP-CaSR construct also reports net plasma membrane abundance of CaSRs, with minimal contributions from near-membrane receptors in TIRFM because low vesicular pH minimizes intracellular SEP fluorescence (15). In some experiments, we combined TIRFM imaging of plasma membrane–localized BSEP-CaSR with that of an mCherry-tagged phospholipase C–δ (PLC-δ) pleckstrin homology domain (PHD) to monitor PIP2 (phosphatidylinositol 4,5-bisphosphate) hydrolysis or with imaging of intracellular Ca2+ using Fura Red in wide-field mode. This approach enabled resolution of CaSR trafficking at the plasma membrane and signaling. Cells labeled with BTx-A594 showed a rapid increase in net plasma membrane abundance of BSEP-CaSR after exposure to 10 mM Ca2+ without a change in endocytosis rate (Fig. 1B and movie S1). These results are reminiscent of activation of CaSR-mediated signaling, but contrary to the anticipated reduction in net plasma membrane abundance of CaSR due to desensitization and endocytosis. We used an independent method to confirm these results. Cells expressing FLAG-CaSR were labeled with a polyclonal FLAG antibody to block all plasma membrane FLAG epitopes, stimulated with basal (0.5 mM) or maximal (10 mM) Ca2+ (6), and then labeled with a monoclonal FLAG antibody to identify newly exocytosed FLAG-CaSR. The application of 10 mM Ca2+ induced substantially more binding of monoclonal antibody to FLAG-CaSR than that of 0.5 mM Ca2+ (fig. S1A), confirming agonist-induced insertion of new CaSRs at the plasma membrane. In cells exposed to either Ca2+ concentration, polyclonal antibody–labeled CaSRs were endocytosed. We performed a similar experiment with hemagglutinin (HA)–tagged angiotensin receptor (AT1R), using polyclonal and monoclonal HA antibodies and angiotensin stimulation. In contrast to the CaSR, the AT1R showed no labeling with monoclonal antibody but did show substantial endocytosis of polyclonal antibody–labeled receptors, suggesting that exocytosis was not induced by angiotensin (fig. S1B) and therefore may not be a universal feature of GPCR regulation. Ca2+-mediated increase in net plasma membrane abundance of CaSRs was, however, resolved by two independent approaches.

Fig. 1

Agonist-dependent insertional signaling (ADIS) of the CaSR. (A) BSEP-CaSR used for TIRFM imaging. (B) TIRFM images of cells expressing rat BSEP-CaSR labeled with BTx-A594. SEP, green; BTx-A594, red. Scale bar, 50 μm. Plot of normalized intensities of net BSEP-CaSR (black) and BTx-A594 (gray) for individual cells in 10 mM Ca2+ (arrow). Acquisition interval, 30 s. n = 5 cells. (C) TIRFM dose response for Ca2+ of HEK293 cells expressing rat BSEP-CaSR. Ca2+ switched from 0.5 mM Ca2+ to 1.0 (black), 2.5 (purple), 5 (gray), 10 (pink), or 20 (cyan) mM Ca2+ at arrow. Acquisition interval, 10 s; mean ± SD. n = 4 cells. (D) Ca2+–net BSEP-CaSR response curve for rat CaSR [data from (C)] fitted with the Hill equation; EC50, 4.0 ± 0.38 mM; Hill coefficient, 2.7 ± 0.64. (E) Agonist dependence of net BSEP-CaSR as described in (B). Agonists, applied at arrow from 1 mM Ca2+ bath, were 20 mM Ca2+ (pink), 20 mM Mg2+ (blue) or 1 mM Ca2+ plus 2 mM spermine (gray), 300 μM neomycin (purple), and 10 mM phenylalanine (black) or 10 μM NPS R-568 (cyan). Acquisition interval, 30 s; mean ± SD. n = 6 cells. (F) Cells expressing human BSEP-CaSR were labeled with BTx-A594 in 0.5, 5, or 10 mM Ca2+ and imaged by TIRFM in the continued presence of Ca2+. Endocytosis: 0.5 (dark blue), 5 (light blue), and 10 (black) mM Ca2+; net plasma membrane (PM) abundance of BSEP-CaSR: 0.5 (dark gray), 5 (light gray), and 10 (medium gray) mM Ca2+. Mean ± SD. n = 5 cells. (G) Ca2+ responses of net BSEP-CaSR at the PM and intracellular Ca2+ were measured simultaneously in Fura Red–loaded cells. Net PM abundance of BSEP-CaSR is plotted as means ± SD (black), and intracellular Ca2+ responses for individual cells (blue spectrum) are plotted on the same axes to enable resolution of oscillations. n = 5 cells. (H) ADIS model for the CaSR. CaSRs travel to the PM through the secretory pathway (according to the rate constants k1 and k2), are activated by extracellular Ca2+, and then desensitize and are endocytosed and targeted to the lysosome for degradation (according to the rate constant k3). Rates of transition from the ER to the Golgi (k1) and Golgi or TGN to the PM (k2) but not from the PM to the lysosome (k3) are increased by signaling initiated by PM CaSRs. PM CaSR (%) was normalized to the initial point (which was set at 100%), and Net CaSR (%) was calculated by subtracting 100 from all PM CaSR (%) values.

A characteristic of CaSR signaling is the steep dependence on extracellular Ca2+ concentration in the physiological range (1, 6). We propose that agonist-driven increases in the net plasma membrane abundance of CaSRs contribute to functional cooperativity, such that signaling is proportional to the number of functional receptors at the plasma membrane. To test this model, we acquired TIRFM images during step changes in extracellular Ca2+ (1 to 20 mM) and determined steady-state BSEP-CaSR amounts (Fig. 1C). The Ca2+ dose-response relation shows cooperativity [Hill coefficient (n) ≈ 3] with an EC50 ≈ 4 mM, comparable to the EC50 for CaSR activation of signaling (Fig. 1D) (6). Ca2+ is a critical regulator of membrane fusion (1618) and thus may affect insertion of proteins into the plasma membrane independently of CaSR. To confirm that CaSR activation mediated the increase in the plasma membrane abundance of CaSR, we determined whether other CaSR activators had comparable effects. The CaSR is activated by cations or cationic small molecules acting at the orthosteric (such as Mg2+, spermine, or neomycin) or allosteric sites within the extracellular domain (for example, phenylalanine) or in the transmembrane domain (for example, NPS R-568) (5, 19). All orthosteric and allosteric agonists induced a significant increase in the net plasma membrane abundance of BSEP-CaSR, with phenylalanine being the least effective (Fig. 1E). CaSR agonists besides Ca2+ therefore promote insertional signaling, arguing that CaSR activation, rather than increased extracellular Ca2+ concentrations per se, controls plasma membrane amounts of BSEP-CaSR. CaSR endocytosis is apparently insensitive to a step change in the extracellular Ca2+ concentration (Fig. 1B) and not subject to recycling (fig. S2), implying that all changes in the net plasma membrane abundance of CaSR result from alterations in the insertion rate of non–BTx-labeled receptors. To confirm this result, we explicitly determined the Ca2+ dependence of endocytosis by labeling cells in various concentrations of Ca2+ with BTx-A594 and imaging both the net plasma membrane abundance of CaSR and the loss of plasma membrane BTx-A594–labeled receptors at the same concentration of Ca2+ (Fig. 1F). The net plasma membrane amount of BSEP-CaSR was stable over the course of the experiment, but BTx-A594–labeled BSEP-CaSR decreased monoexponentially, with a time constant of ≈6 min regardless of extracellular Ca2+ concentration ([Ca2+]) (0.5 mM [Ca2+], τ = 5.8 ± 0.2 min; 5 mM [Ca2+], τ = 6.2 ± 0.2 min; 10 mM [Ca2+], τ = 6.2 ± 0.2 min; n = 5 cells). Finally, to establish a linkage between net plasma membrane abundance of CaSR and signaling, we loaded cells expressing BSEP-CaSR with Fura Red and monitored both net amounts of CaSR at the plasma membrane and intracellular Ca2+ responses to increases in extracellular Ca2+ concentrations from 0.5 to 10 mM. The average net plasma membrane abundance of BSEP-CaSR increased monotonically between 0.5 and 5 mM Ca2+ (Fig. 1G), whereas intracellular Ca2+ over the same range (plotted separately for each cell) showed the expected induction of Ca2+ oscillations at 2.5 mM Ca2+ (20, 21) followed by steady-state maximal responses at 5 and 10 mM Ca2+, reflecting intracellular Ca2+ store dynamics. Overall, these results argue that net plasma membrane abundance of CaSRs and signaling are responsive to the same range of extracellular Ca2+ concentrations.

We hypothesized that CaSRs are inserted at the plasma membrane as a function of CaSR activation, a mechanism that we termed ADIS (Fig. 1H). We propose that a continual flow of nascent CaSRs to the plasma membrane permits “sampling” of ambient extracellular Ca2+. At the plasma membrane, the CaSR contributes to signaling followed by desensitization, endocytosis, and degradation at the lysosome (aggregate rate k3) (fig. S2). Furthermore, we hypothesized that activation of plasma membrane–localized CaSRs by increased extracellular Ca2+ increases the rate of insertion of CaSRs at the plasma membrane (k2). The ability of a small signal from functional CaSRs at the plasma membrane to trigger insertion of additional receptors would provide a large dynamic range for Ca2+ sensing and substantial signal amplification, with minimal functional desensitization over the narrow, physiological range of extracellular Ca2+. A fundamental requirement of the ADIS model is that signaling by plasma membrane CaSRs precedes and induces insertion of additional CaSRs. To test this prediction, we loaded cells expressing BSEP-CaSR with the Ca2+ indicator Fura Red and acquired images at a higher frequency to resolve early activation events. We observed maximal intracellular Ca2+ responses within 15 s after addition of 10 mM Ca2+, whereas the increase in net plasma membrane abundance of BSEP-CaSR was complete after 1.5 min, arguing that signaling by plasma membrane–localized CaSR initiated the increase in net BSEP-CaSR (Fig. 2A). Likewise, intracellular Ca2+ concentrations dropped rapidly after the cells were returned to 0.5 mM Ca2+, which was followed by a slower decline in the plasma membrane abundance of BSEP-CaSR. We further confirmed that heterotrimeric G protein activation is required for increases in net plasma membrane abundance of BSEP-CaSR by stimulating cells in the continuous presence of gallein, a G protein βγ inhibitor (2224), which attenuated both ADIS and PHD translocation responses (fig. S3).

Fig. 2

CaSR mutants show altered ADIS. (A) HEK293 cells transfected with human BSEP-CaSR (black) or BSEP-CaSRΔ868 (white) and loaded with Fura Red AM were exposed to 10 mM Ca2+. Net PM receptor (circles) and intracellular Ca2+ (squares) were normalized to initial values. Acquisition interval, 15 s; mean ± SD. n = 8 cells. (B) The CaSR mutants L159P (black circles) or R795W (white squares) were labeled with BTx-A594 and exposed to 10 mM Ca2+ as indicated. Maximum changes in net PM abundance of BSEP-CaSR in 10 mM Ca2+ were 16.9 ± 0.5% for the L159P mutant and 9.1 ± 0.6% for the R795W mutant; means ± SD. n = 6 cells. (C) Cells expressing L159P (black) or R795W (white) were loaded with Fura Red AM and stimulated with 30 mM bath Ca2+; both net PM abundance of CaSR (circles) and intracellular Ca2+ (squares) are plotted as means ± SD. n = 8 cells. (D and E) Ca2+ dose–ADIS relations for L159P (D) and R795W (E) at 1 (black), 2.5 (purple), 5 (gray), 10 (blue), 20 (pink), and 30 (cyan) mM Ca2+; means ± SD. n = 4 cells. (F) Ca2+-ADIS response data from (D) and (E). L159P (blue circles) had EC50 of 12.1 ± 0.3 mM; Hill coefficient, 4.7 ± 0.4; maximal response, 46.7 ± 1.0%. R795W (yellow triangles) could not be fitted.

Both truncations and mutations of CaSR alter CaSR trafficking and cause human disease (4). ADIS confers an intimate link between signaling and cellular trafficking of CaSR; thus, we explored the impact of truncation or mutation on ADIS. Because truncation of the CaSR C terminus at Thr868 increases targeting of the CaSR to the plasma membrane (25), we compared net plasma membrane amounts of wild-type and BSEP-CaSRΔ868 by TIRFM, and intracellular Ca2+ with Fura Red (Fig. 2A). In contrast to wild-type BSEP-CaSR, both the net increase in plasma membrane amounts and intracellular Ca2+ responses were slowed and significantly attenuated for BSEP-CaSRΔ868, suggesting that robust induction of ADIS requires the CaSR C terminus.

Many CaSR mutations that cause familial hypocalciuric hypercalcemia (FHH) decrease plasma membrane targeting (26). We generated two well-characterized FHH mutants, Leu159→Pro (L159P) and Arg795→Trp (R795W), and determined their responses to 10 mM Ca2+ (Fig. 2B). Net plasma membrane amounts of both mutants increased in cells exposed to 10 mM Ca2+, although responses were diminished compared to those of wild-type BSEP-CaSR. Endocytosis rates were unaffected by 10 mM Ca2+ (Fig. 2B). Sensitivity to extracellular Ca2+ is reduced for FHH mutants (4, 26), and we therefore tried to elicit ADIS and Ca2+ signaling of L159P and R795W in 30 mM Ca2+ (Fig. 2C). Both mutants mediated intracellular Ca2+ increases, although the activity of R795W was significantly impaired even in 30 mM Ca2+. We next determined the Ca2+ (1 to 30 mM)–ADIS response relations for the L159P (Fig. 2D) and R795W (Fig. 2E) mutants. L159P was less sensitive to extracellular Ca2+ than wild-type CaSR, but was capable of a maximal ADIS response comparable to that of wild-type CaSR (Fig. 2F). The reduced sensitivity of both signaling and ADIS responses to Ca2+ is consistent with the location of Leu159 in the extracellular agonist binding domain (4, 26) and further confirms the intimate coupling between CaSR trafficking and Ca2+ signaling. In contrast, Arg795 is located within the third intracellular loop and likely alters the coupling efficiency to heterotrimeric G proteins rather than affecting Ca2+ sensitivity; thus, the ADIS response was not rescued by increased Ca2+ (Fig. 2, E and F). These results suggest that the ADIS response can be used as a surrogate for signaling.

Prolonged stimulation (>45 min) with 10 mM Ca2+ causes a stable increase in net plasma membrane amounts of BSEP-CaSR (Fig. 3A), suggesting the presence of a substantial intracellular reservoir of CaSRs. The fungal macrocyclic lactone brefeldin A (BFA) stabilizes Golgi-localized Arf1-GDP (guanosine diphosphate) and blocks COPI vesicle assembly, causing collapse of the Golgi complex into the ER and blocking of secretory pathway traffic (27, 28). BFA was used to identify the intracellular compartments that contribute to sustained insertion of CaSRs at the plasma membrane. Pretreatment with BFA attenuated the ADIS and PHD translocation responses to 10 mM Ca2+, whereas BFA treatment after stimulation with 10 mM Ca2+ induced a slow decrease in plasma membrane abundance of BSEP-CaSR (fig. S3). We hypothesized that the largest reservoir of intracellular CaSRs that supports prolonged ADIS is located in pre-Golgi compartments. If so, washout of BFA should lead to increased plasma membrane CaSR amounts as intraorganelle trafficking is restored. Initial rates of recovery during BFA washout and final steady-state amounts of plasma membrane BSEP-CaSR were higher in the presence of 10 mM Ca2+ compared to that in 0.5 mM Ca2+ (Fig. 3B), suggesting that CaSRs are mobilized to the secretory pathway for plasma membrane insertion from the ER.

Fig. 3

Sustained ADIS results from CaSR signaling–initiated trafficking from the ER. (A) Cells expressing BSEP-CaSR were exposed to 10 mM Ca2+ as indicated. Means ± SD. n = 8 cells. (B) Cells expressing BSEP-CaSR were treated with BFA in 0.5 mM Ca2+ for 30 min before imaging in either 0.5 (purple) or 10 mM Ca2+ (black) bath solution (without BFA). Initial recovery rates were 0.31 ± 0.13% min−1 in 0.5 mM Ca2+ and 9 ± 0.65% min−1 in 10 mM Ca2+. Final steady-state amounts were 117 ± 3.5% in 0.5 mM Ca2+ and 158 ± 7.5% in 10 mM Ca2+. Means ± SD. n = 5 cells. (C) Cells expressing FLAG-CaSR were [35S]cysteine pulse-labeled in the presence of tunicamycin or DMSO. The same blot was probed for [35S]CaSR (left) and total CaSR (right). WB, Western blotting. (D) Cells expressing BSEP-CaSR were treated with tunicamycin in 0.5 mM Ca2+ before imaging in the continued presence of tunicamycin plus 10 mM Ca2+ as indicated. Means ± SD. n = 8 cells. (E) Untransfected hT-HUVECs were loaded with Fluo-4 AM and exposed to tunicamycin (purple) or DMSO (black) for 60 min before and throughout wide-field imaging of intracellular Ca2+. Ca2+ (10 mM) was added as indicated. Means ± SD. n = 6 cells.

The prolonged increase in net plasma membrane amounts of BSEP-CaSR in 10 mM Ca2+ suggests the possibility that signaling is coupled with biosynthesis. CaSR glycosylation is required for ER release and trafficking to the plasma membrane, and tunicamycin blocks plasma membrane localization of CaSR (29). To assess the rapidity of tunicamycin block, we exposed FLAG-CaSR cells to tunicamycin during a brief (30 min) [35S]cysteine-labeling period (Fig. 3C). The immaturely glycosylated form of [35S]CaSR (140 kD) was synthesized in dimethyl sulfoxide (DMSO)–treated cells, but most newly synthesized [35S]CaSR in tunicamycin-treated cells was unglycosylated (120 kD). We treated BSEP-CaSR–expressing cells with tunicamycin and monitored responses to 10 mM Ca2+ by TIRFM. Although tunicamycin did not significantly alter the initial response to 10 mM Ca2+ (Fig. 3D), it did cause a progressive decline in plasma membrane BSEP-CaSR after 10 min in the continued presence of 10 mM Ca2+, suggesting ADIS-mediated depletion of a mobilizable pool of intracellular CaSRs. Finally, we determined whether prolonged Ca2+ signaling in telomerase-stabilized human umbilical vein endothelial cells (hT-HUVECs) with endogenous CaSRs was sensitive to tunicamycin (Fig. 3E). Preincubation with tunicamycin caused a reduced maximal intracellular Ca2+ response to 10 mM Ca2+ and a significant time-dependent decline in the steady-state response compared with DMSO-treated cells. ADIS therefore operates in cells with endogenous CaSRs (fig. S4). These results characterize a sensitive feedback system for the maintenance of a stable population of intracellular CaSRs that can be rapidly mobilized to the plasma membrane in response to CaSR signaling.

Sustained ADIS requires regulated release from a large intracellular pool of transport-ready CaSRs. We recently identified a phosphorylation-regulated arginine-rich motif, 890RRSNVSRKRSSS901, in the proximal carboxyl terminus of CaSR that binds 14-3-3 proteins and contributes to intracellular retention of CaSR (30). C terminus truncation at residue Thr868 (Δ868) and mutation of the arginine-rich motif to alanines (5A; 890AASNVSAAA898) or the phosphomimetic S899D mutant disrupted 14-3-3 protein coimmunoprecipitation with CaSR (Fig. 4A and fig. S5). The S899A mutant showed enhanced association with 14-3-3 proteins (fig. S5). Ca2+-ADIS relations were determined for wild-type and mutant BSEP-CaSRs and fitted with the Hill equation (Fig. 4, B to F). Wild-type CaSR (Fig. 4B) and S899A (Fig. 4C) have comparable EC50s for Ca2+-mediated ADIS, although 14-3-3 immunoprecipitated more S899A than wild-type CaSR (Fig. 4A and fig. S5) and S899A had a lower maximal response, suggesting that CaSR release to the plasma membrane is regulated by phosphorylation at Ser899. In contrast, the maximal responses for the mutants that do not bind 14-3-3 proteins were comparable to that of wild-type (Fig. 4, D and E), whereas the EC50 for Ca2+-induced ADIS was increased for the 5A mutant but was not significantly higher for the S899D mutant. These results implicate phosphorylation-dependent 14-3-3 binding in the Ca2+-dependent release of intracellular CaSR for ADIS. However, none of the mutations attenuated ADIS in a manner comparable to truncation at Thr868 (Fig. 2A), suggesting that additional protein interactions control release. Finally, we compared the ADIS and PHD translocation responses for wild-type BSEP-CaSR and 14-3-3 binding mutants treated with 10 mM Ca2+ (Fig. 4G). For all mutants, signaling as assessed by PHD translocation reflected net BSEP-CaSR at the plasma membrane, defining ongoing CaSR trafficking to the plasma membrane as the essential precursor of sustained signaling.

Fig. 4

14-3-3 protein interactions modulate the ADIS response of CaSR. (A) 14-3-3 or FLAG immunoprecipitates from HEK293 cells expressing FLAG-CaSR truncation mutants (Δ898/5A and Δ868), motif mutants (5A), or phosphomutants (S899A/D) were immunoblotted with anti-CaSR antibody. (B to E) Ca2+ bath titration by TIRFM of HEK293 cells expressing wild-type (WT) human BSEP-CaSR (B), S899A (C), 5A (D), and S899D (E) mutants as described in Fig. 1C. Acquisition interval, 10 s. Means ± SD. n = 4 cells. (F) Ca2+ dose–ADIS relations for WT human BSEP-CaSR and 14-3-3 site mutants. Data from (B) to (E) were fitted. WT (black circles; EC50, 2.1 ± 0.5 mM; max, 36.2 ± 0.5%), 5A (white circles; EC50, 6.7 ± 0.4 mM; max, 36.9 ± 1.7%), S899D (white triangles; EC50, 3.2 ± 1.5 mM; max, 38.24 ± 6.7%), S899A (black triangles; EC50, 2.7 ± 0.6 mM; max, 19.2 ± 2.1%). The EC50 of the 5A mutant was significantly different (P < 0.05) from that of WT, and the max of the S899A mutant was significantly different (P < 0.05) from that of WT, as determined by analysis of variance (ANOVA) with a post hoc Dunnett’s test. (G) Cells cotransfected with WT or 14-3-3 binding mutant hBSEP-CaSR and PHD were treated with 10 mM Ca2+ as indicated. Net BSEP-CaSR (black) and PHD (white) TIRFM intensities were plotted as means ± SD. n = 5 cells.

Discussion

ADIS represents a previously unknown adaptation of the canonical GPCR signaling paradigm. CaSR must operate in the chronic presence of extracellular Ca2+ and translate small fluctuations in Ca2+ or allosteric modulators into appropriate cellular responses. The constant, agonist-driven flow of CaSR to the plasma membrane ensures that changes in agonist will be rapidly sensed by newly inserted receptors, and the resulting rapid increase in net plasma membrane abundance of CaSR leads to cooperative activation of signaling pathways. Ongoing CaSR signaling therefore depends on regulated trafficking through the secretory pathway and is coupled to regulation of net cellular CaSR abundance. The ongoing release of CaSR from the ER to the Golgi allows for the accumulation of a pool of mature glycosylated CaSR in Golgi, TGN, or post-TGN vesicles that are poised for rapid insertion into the plasma membrane. Sustained signaling in increased extracellular Ca2+ concentrations requires contributions from ongoing biosynthesis, which mobilizes CaSR from pools in pre-Golgi compartments. CaSR is endocytosed and targeted to the lysosome in an agonist-independent manner without substantial recycling to the plasma membrane. This model is contrary to other regulatory models of GPCR signaling (31, 32) that generally invoke a stable population of receptors at the plasma membrane until exposure to agonist, which triggers activation and desensitization, usually by phosphorylation and β-arrestin–dependent endocytosis, with the possibility for recycling of internalized GPCRs. ADIS, in contrast, requires a large, signaling-regulated pool of intracellular CaSRs, with only a small fraction of total cellular CaSR being localized to the plasma membrane at any given time. Such a model shares features with signaling-induced insertion of aquaporin channels (3335) or glucose transporters (36, 37), and it will be of interest to determine whether the mechanisms regulating plasma membrane insertion of these transmembrane proteins are similar.

The major intracellular reservoir for ADIS resides in the ER, in complexes containing CaSR and 14-3-3 proteins. Additional protein interactions at the CaSR C terminus must contribute to the ADIS response because truncation of the CaSR C terminus at Thr868 reduces the size of the intracellular pool of CaSR (6, 25) and attenuates ADIS. The proximal C terminus of CaSR contains 20 consensus protein kinase sites between residues 862 and 920 (NetPhos2.0) and is evolutionarily conserved (1, 4, 38). This cluster of phosphorylation sites may serve as a “coincidence detector” to permit fine-tuning of the release process, and, as a corollary, no single phosphorylation event or protein interaction should release all intracellular CaSR. Our current data support such a model because the phosphomimic mutation S899D, which abolishes 14-3-3 binding, “tunes” but does not eliminate the ADIS response. A major challenge posed by these results is the identification of additional protein partners that regulate ADIS and dissection of the complex patterns of phosphorylation that lead to altered ADIS kinetics or steady-state changes in net plasma membrane abundance of CaSR.

The current results provide new insight into the factors controlling the Ca2+ dependence of CaSR signaling. The range of apparent affinities for activation by extracellular Ca2+ of endogenous CaSR has been variously attributed to the presence of endogenous allosteric modulators (39) or to differential phosphorylation (40, 41). Here, we define a new mechanism that contributes to the apparent affinity of CaSR for Ca2+, namely, the variable distribution or abundance of interacting proteins that control the stability of the ER-localized, transport-competent CaSR complex. Although 14-3-3 proteins are ubiquitously distributed, the relative abundance of different subtypes can be regulated by cell state or signaling (42), allowing the potential for subtype-specific regulation of the Ca2+-ADIS response relation. CaSR interacts with multiple subtypes of 14-3-3 proteins (fig. S5), and thus, the regulation is likely to be complex. Likewise, differential distribution or abundance of other protein partners may modulate ADIS properties. Dysregulation of CaSR signaling has been implicated in the progression of cancers (43) and in cardiovascular disease (44) and may result from altered abundance of critical interacting proteins or the size of the transport-competent pool of ER-localized CaSR. The ADIS model provides a framework for identifying the critical regulatory factors involved and an assay that permits screening for drugs that normalize ADIS of CaSR.

ADIS may be a relevant regulatory mechanism for other GPCRs. Criteria for candidates using the ADIS mechanism include the presence of a large intracellular pre-Golgi pool of receptors, chronic or extended exposure to agonist during receptor life cycle, relatively low steady-state plasma membrane amounts of receptor (which may manifest as difficulty in measuring plasma membrane amounts by conventional methods including biotinylation or enzyme-linked immunosorbent assays), and rapid or constitutive desensitization and degradation with minimal effects of β-arrestin manipulation. GPCRs that meet at least a subset of these criteria include adrenergic, dopamine, and cannabinoid receptors (4550). The kinetics and magnitude of ADIS-mediated changes in plasma membrane amounts of other GPCRs may vary as a function of distinct interacting proteins or release mechanisms, enabling signal responses and duration to be fine-tuned. The approaches described for dissecting ADIS of CaSR could be applied to other GPCRs.

In summary, we describe a coupled feedback system that maintains signaling of CaSR in the face of chronic exposure to agonist and ongoing desensitization and degradation of plasma membrane–targeted receptors. The challenge now is to identify additional protein partners that stabilize the ER-localized transportable pool of CaSR and the signals that release CaSR to the plasma membrane.

Materials and Methods

Materials

The antibodies used were as follows: mouse monoclonal and polyclonal anti-FLAG antibodies (Sigma); mouse monoclonal (Abcam) and polyclonal (Sigma) anti-HA antibodies; goat polyclonal anti-calnexin antibody (Santa Cruz Biotechnology), rabbit polyclonal anti–14-3-3β/α (Thermo Fisher Scientific) and rabbit anti–14-3-3 isoform panel (AbD Serotec), and mouse monoclonal anti-CaSR antibody (Abcam or Sigma); Alexa Fluor 568 goat anti-mouse antibody and Alexa Fluor 633 goat anti-rabbit antibody (Invitrogen); and fluorescein isothiocyanate (FITC) donkey anti-goat antibody (Jackson ImmunoResearch Laboratories Inc.). Other reagents used were: bungarotoxin–Alexa Fluor 594 (Invitrogen), LysoTracker Red DND-99, Fluo-4 AM, Fura Red AM, probenecid, and pluronic F127 (Invitrogen).

DNA constructs

The N-terminal SEP–CaSR (rat) construct was a gift from J. M. Henley (15); the 13–amino acid α-bungarotoxin binding site (13) was inserted in-frame by reverse polymerase chain reaction (PCR) into an Eco R1 site located upstream of SEP, after residue Asp23 of the rat CaSR sequence. Human BSEP-CaSR (hBSEP-CaSR, in pEGFP-N1) was generated by inserting Kpn I and Age I sites by PCR in human FLAG-CaSR (25) immediately after the FLAG epitope. Oligonucleotides encoding the 13–amino acid α-bungarotoxin binding site flanked with the appropriate restriction sites were annealed, digested, and ligated into FLAG-CaSR (B-CaSR). SEP was generated by PCR using SEP-CaSR (15) as template with primers containing flanking Age I sites. Both SEP and B-CaSR constructs were digested with Age I (R725A, Promega), purified (Qiagen QiaEX II kit 20021), and ligated with T4 DNA Ligase (Promega), generating human BSEP-CaSR. Truncations, 14-3-3 binding site mutants [Δ868, Δ898(5A), 5A, S899A, S899D, S892A, and S892D], and FHH mutants (L159P and R795W) were generated as described previously (26, 30) in the human BSEP-CaSR background. The PLC-δ PHD fragment (amino acids 1 to 174) of GFP-PHD (provided by T. Mirshahi) was amplified by PCR and inserted downstream of the mCherry fluorescent protein (obtained from R. Tsien) in pcDNA3.1 (Clontech) to generate PHD. Coding regions of all constructs were verified by sequencing (GENEWIZ).

Cell culture and transfection

Human embryonic kidney (HEK) 293 cells (American Type Culture Collection) were maintained in minimum essential medium (MEM) supplemented with 10% fetal bovine serum and penicillin/streptomycin in 5% CO2 and transfected with FuGENE HD (Roche Applied Science). HUVECs (Lonza) were maintained in EGM-2 complete medium supplemented with the EGM bullet kit (Lonza Walkersville Inc.) and transfected with Effectene (Qiagen). hT-HUVECs were provided by C. Passananti (University of Maryland, Baltimore, MD) and grown in PromoCell Endothelial Growth Medium 2.

Total internal reflection microscopy

Cells grown on 22-mm2 glass coverslips coated with human fibronectin (Millipore) were transfected with the indicated constructs. After 72 hours, the medium was replaced with 0.5 mM Ca2+ bath solution (20) plus 0.1% bovine serum albumin and incubated for 30 min. Cells were labeled with BTx-A594 (5 μg/ml; 5 min at room temperature) in a 0.5 mM Ca2+ bath solution before imaging. All bath solutions were osmolality-matched, and drug responses were compared with vehicle. Gallein (15 μM), BFA (100 ng/ml), and tunicamycin (5 μg/ml) were solubilized in DMSO and applied to cells as indicated in the figure legends. TIRFM images were collected on a Nikon TE2000E microscope equipped with a 60×, 1.45 numerical aperture (NA) objective (Perfect Focus) and a TIRF-2 illuminator, using laser lines at 488 and 594 nm. TIRFM angles were optimized for each wavelength with customized software, and data collection was automated with μManager (National Institutes of Health). Images were captured with a Photometrics CoolSNAP HQ2 charge-coupled device (CCD) camera with appropriate emission filters mounted on a motorized shutter wheel (Ludl). For intracellular Ca2+ imaging, cells were incubated with 2.5 μM Fluo-4 AM or 5 μM Fura Red AM and an equal volume of pluronic F127 plus 2.5 mM probenecid in 0.5 mM Ca2+ for 30 min (at 37°C and 5% CO2) and then washed and incubated for 30 min at room temperature before imaging by means of a 488-nm laser line with angle set for wide-field acquisition.

Image and data analysis

TIRFM images were analyzed with ImageJ software (National Institutes of Health). Averaged mean surface intensities for regions of interest for individual cells were normalized to the first value in the captured sequence. Background intensities were subtracted with the Subtract Background function in ImageJ. Normalized data from individual cells were averaged over several independent experiments and analyzed and plotted with SigmaPlot v.11. Ca2+ dose–ADIS response curves were generated by fitting the raw, normalized imaging data with a four-parameter sigmoid curve: f = y0 + [(max × tn)/(cn + tn)]. The fitted maximum at each [Ca2+] was then fitted with a four-parameter Hill equation to extract EC50 and the Hill coefficient (n): F = Baseline% + [(Max% × [Ca2+]n)/(EC50n + [Ca2+]n)].

Immunoprecipitation and Western blotting

HEK293 cells transfected with 1 μg of total DNA per 35-mm well were cultured for 3 days and processed for immunoprecipitation with M2 anti-FLAG antibody (FLAG-CaSR) or 14-3-3 antibody (Abcam) plus protein G–agarose (Invitrogen) as previously described (30). Reduced samples (loading buffer plus dithiothreitol, 30 min at room temperature) were electrophoresed on 4 to 15% gradient tris-HCl gels (Bio-Rad), and blots were probed with polyclonal anti-CaSR antibody [LRG epitope, custom-generated by Genemed Synthesis Inc.; horseradish peroxidase (HRP)–conjugated secondary antibody for ECL, GE Healthcare] or polyclonal anti–14-3-3 antibodies (AbD Serotec). Proteins were visualized with SuperSignal West Pico Chemiluminescence Substrate (Pierce) on a Fujifilm LAS-4000mini Luminescent Image Analyzer.

[35S]Cysteine labeling

HEK293 cells were transfected with FLAG-CaSR for 48 hours and starved of cysteine and methionine for 1 hour and then [35S]cysteine was added with 2.3 mM methionine for 30 min. DMSO or tunicamycin (5 μg/ml) was present continuously from the initiation of cysteine starvation. Cells were lysed and immunoprecipitated with M2 anti-FLAG antibody (Sigma) and processed for Western blotting as previously described (25). [35S]CaSR was detected with a Molecular Dynamics Storm Imager followed by immunodetection with anti-CaSR antibody.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/4/200/ra78/DC1

Fig. S1. Agonist-driven insertion of CaSRs but not AT1Rs.

Fig. S2. CaSRs do not recycle to the PM after stimulation but are targeted to the lysosome for degradation.

Fig. S3. ADIS requires signaling by PM CaSR and release from secretory compartments.

Fig. S4. Endogenous CaSRs in HUVECs undergo ADIS.

Fig. S5. CaSR interactions with 14-3-3 proteins are modulated by C-terminal mutations and encompass different 14-3-3 subtypes.

Movie S1. TIRFM of BTx-A594–labeled BSEP-CaSR stimulated with extracellular Ca2+.

References and Notes

Acknowledgments: We thank R. L. Henley for rat SEP-CaSR, T. Mirshahi for GFP-PHD, R. Tsien for mCherry, A. Passaniti for HUVECs stably expressing hTERT, and D. J. Carey, T. Mirshahi, J. Robishaw, and W. Schwindinger for helpful discussions during the course of this work. Funding: Supported by NIH R01 GM77563 and funds from the Geisinger Clinic. Author contributions: M.P.G., A.S., and A.C. performed the experiments. M.P.G., A.S., A.C., and G.E.B. analyzed the data. G.E.B. prepared the manuscript. Competing interests: The authors declare that they have no competing interests.
View Abstract

Stay Connected to Science Signaling

Navigate This Article