Calcium-sensing Receptor Biosynthesis Includes a Cotranslational Conformational Checkpoint and Endoplasmic Reticulum Retention*

Metabolic labeling with [35S]cysteine was used to characterize early events in CaSR biosynthesis. [35S]CaSR is relatively stable (half-life ∼8 h), but maturation to the final glycosylated form is slow and incomplete. Incorporation of [35S]cysteine is linear over 60 min, and the rate of [35S]CaSR biosynthesis is significantly increased by the membrane-permeant allosteric agonist NPS R-568, which acts as a cotranslational pharmacochaperone. The [35S]CaSR biosynthetic rate also varies as a function of conformational bias induced by loss- or gain-of-function mutations. In contrast, [35S]CaSR maturation to the plasma membrane was not significantly altered by exposure to the pharmacochaperone NPS R-568, the allosteric agonist neomycin, or the orthosteric agonist Ca2+ (0.5 or 5 mm), suggesting that CaSR does not control its own release from the endoplasmic reticulum. A CaSR chimera containing the mGluR1α carboxyl terminus matures completely (half-time of ∼8 h) and without a lag period, as does the truncation mutant CaSRΔ868 (half-time of ∼16 h). CaSRΔ898 exhibits maturation comparable with full-length CaSR, suggesting that the CaSR carboxyl terminus between residues Thr868 and Arg898 limits maturation. Overall, these results suggest that CaSR is subject to cotranslational quality control, which includes a pharmacochaperone-sensitive conformational checkpoint. The CaSR carboxyl terminus is the chief determinant of intracellular retention of a significant fraction of total CaSR. Intracellular CaSR may reflect a rapidly mobilizable “storage form” of CaSR and/or may subserve distinct intracellular signaling roles that are sensitive to signaling-dependent changes in endoplasmic reticulum Ca2+ and/or glutathione.

checkpoint controlling total and plasma membrane expression of WT and mutant CaSRs (21).
Here we examine the very early events in CaSR biosynthesis by monitoring the appearance and maturation of [ 35 S]cysteinelabeled CaSR. Results indicate that [ 35 S]CaSR that accumulates during the pulse label period has undergone cotranslational quality control. CaSR therefore rapidly navigates both generic (glycosylation, disulfide bond shuffling) and specific (helix packing, conformational assessment) quality control checkpoints, and the pharmacochaperone NPS R-568 acts cotranslationally to stabilize [ 35 S]CaSR. CaSR dimers that successfully run the gauntlet enjoy prolonged stability in the ER until release to the Golgi and plasma membrane. Neither membrane-permeant (NPS R-568) nor membrane-impermeant (neomycin) allosteric agonists or Ca 2ϩ are able to facilitate full [ 35 S]CaSR maturation, but truncation of the carboxyl terminus (CT) induces full [ 35 S]CaSR maturation. These results suggest that the CaSR CT is the chief determinant of both the rate of CaSR maturation through the secretory pathway and the subcellular localization of the net cellular complement of CaSR. Such control of the levels of both intracellular and plasma membrane CaSR suggests the exciting possibility of an intracellular signaling role(s) for CaSR.

MATERIALS AND METHODS
cDNA Constructs-All constructs in pEGFP-N1 were generated using PCR primer mutagenesis with Pfu Ultra polymerase (Stratagene) in the background of human CaSR containing an amino-terminal FLAG epitope immediately after the signal sequence (FLAG-CaSR) (23). Point mutations in the WT FLAG-CaSR background (E837I, A843E, and L849P) were generated as described previously (22). The CaSR/mGluR1␣ CT chimera was generated by incorporating a silent PvuI restriction site at the proposed junction in both CaSR and mGluR1␣ constructs, digesting each vector with PvuI, followed by ligation. The final CaSR/mGluR1␣ construct contains WT FLAG-CaSR through residue 869, followed by rat mGluR1␣ residues 849 -1194, followed by EGFP. To rule out a contribution of EGFP to the observed maturation rate, FLAG-CaSR-EGFP (FLAG-CaSR fused after residue 1078 to EGFP) was used as the control. A stop codon was incorporated into FLAG-CaSR by PCR mutagenesis to generate CaSR truncations (CaSR⌬868, CaSR⌬880, CaSR⌬886, CaSR⌬898, CaSR⌬1024, and CaSR⌬1052). All constructs were confirmed by sequencing (Genewiz).
Cell Culture and [ 35 S]Cysteine Metabolic Labeling of CaSR-HEK293 and opossum kidney (OK) cells were obtained from ATCC and cultured as recommended in minimum Eagle's medium (Mediatech) supplemented with 10% fetal bovine serum and penicillin/streptomycin in 5% CO 2 and used within 25 passages. Cells were plated (10 6 cells/100-mm dish) and allowed to attach overnight before transfection. Each plate was transfected with 6 g of FLAG-CaSR or FLAG-CaSR mutant cDNA plus 18 l of Fugene HD (Roche Applied Science) for 24 h. Cells were starved for 30 min in normal Ca 2ϩ -containing DMEM without cysteine or methionine (Invitrogen). Labeling was initiated by the addition of [ 35 S]cysteine (PerkinElmer Life Sciences; 1075 Ci/mmol) at a final concentration of 100 Ci/ml and methionine to a final concentration of 2.3 mM. Drug treatments were added as indicated for individual experiments, with control samples containing equivalent volumes of the solvent (DMSO). After labeling, plates were rinsed with phosphatebuffered saline, and medium was replaced with minimum Eagle's medium containing 10% fetal bovine serum. Labeled cells were lysed and processed for immunoprecipitation and Western blotting immediately or at variable chase times.
Immunoprecipitation and Western Blotting-Cells were lysed (5 mM EDTA, 0.5% Triton X-100, 10 mM iodoacetamide, plus Complete protease inhibitor tablet (Roche Applied Science) in phosphate-buffered saline) and cleared at 4°C with Sepharose CL-2B (Sigma). Equal amounts of protein (micro-BCA protein assay, Pierce) were immunoprecipitated overnight with M2 anti-FLAG antibody (Sigma) plus anti-GAPDH monoclonal antibody (Abcam) and protein G-agarose (Invitrogen). Samples were eluted in SDS loading buffer containing 100 mM dithiothreitol, incubated at 22°C for 30 min, run on 4 -15% SDS-polyacrylamide gels (Criterion, Bio-Rad) and transferred onto nitrocellulose. [ 35 S]CaSR was detected on an Amersham Biosciences Storm 840 Imager, followed by processing of the blot for total protein. Blots were cut at the 75 kDa marker, and CaSR was detected on the upper portion with rabbit polyclonal anti-LRG antibody (1:2000; custom-generated by Genemed Synthesis, Inc. against LRG epitope residues 374 -391), and the lower part of each blot was probed with rabbit polyclonal anti-GAPDH antibody (1:2000; Abcam). ECL anti-rabbit IgG, horseradish peroxidase-linked F(abЈ) 2 fragment from donkey (GE Healthcare) was the secondary antibody. SuperSignal West Pico Chemiluminescence Substrate (Pierce) was used to visualize proteins to film, followed by scanning to computer and analysis with AlphaEaseFC version 4.0.0 (Alpha Innotech) or by direct chemiluminescence visualization on a FUJIFILM LAS-4000mini luminescent analyzer and processing with Image-Gauge version 3.0.
Data Analysis and Statistics-All experiments were repeated a minimum of 3-5 times (as indicated in individual figure legends). Immunoprecipitation of endogenous GAPDH was used as a loading control. Pixel intensities of [ 35 Tween 20) and fixed (4% paraformaldehyde or MeOH) for 15 min on ice. Cells were blocked in TBS-T, 1% milk, followed by 60 min in TBS-T/monoclonal anti-FLAG-M2-horseradish peroxidase antibody (1:5000; Sigma catalog no. A8592), according to the manufacturer's instructions. The reaction with 3,3Ј,5,5Јtetramethylbenzidine liquid substrate solution (Sigma catalog no. T0440) was stopped with 1 M sulfuric acid, and plates were read at 450 nm. Eight replicates fixed in either paraformaldehyde (plasma membrane receptors) or MeOH (total receptor) were averaged, and background was subtracted (untransfected HEK293 or OK cells fixed with paraformaldehyde or MeOH). Data were normalized to plasma membrane or total expression of FLAG-CaSR, as indicated for specific experiments.  Fig. 1A illustrates the results of a 60-min pulse with [ 35 S]cysteine in HEK293 cells transiently transfected with FLAG-CaSR (36 h) and then treated with DMSO or 10 M MG132 overnight and during the cysteine/methionine starvation period and [ 35 S]cysteine pulse. Lysates were subjected to immunoprecipitation (IP) with monoclonal anti-FLAG plus anti-GAPDH antibodies (for normalization), run on 4 -15% SDS-polyacrylamide gels, and blotted to nitrocellulose membrane, followed by sequential development of the 35 S image and Western blot. The Western blot was cut (indicated by the dotted line) and probed with polyclonal anti-CaSR LRG (top) and anti-GAPDH antibodies (bottom), as described under "Materials and Methods." The blot was reconstructed to illustrate the 35 S image and Western blot (WB) for CaSR and Western blot image for GAPDH (Fig. 1A). Under reducing conditions, the dominant form of monomeric CaSR is ϳ140 kDa (Fig. 1A, CaSRϾ). The maturely glycosylated form is ϳ160 kDa. Incubation with MG132 leads to the appearance of unglycosylated CaSR at ϳ120 kDa (f), whereas incomplete reduction of disulfide bonds can lead to resolution of dimers/oligomers on Western blots (3). To confirm the identities of [ 35 S]cysteine-labeled bands, we used HEK293 cells stably expressing FLAG-CaSR to maximize the abundance of maturely glycosylated CaSR. Cells were labeled with [ 35 S]cysteine for 60 min and then chased in unlabeled cysteine plus methionine-containing medium for 24 h to facilitate maturation of [ 35 S]CaSR (Fig. 1B). Lysates were immunoprecipitated with anti-FLAG antibody and eluted with FLAG peptide. Eluates were incubated overnight at 37°C without an addition (Fig. 1B, CONTROL) or with endoglycosidase H (EndoH) or peptide:N-glycosidase F (PNGaseF). Samples were run on 4 -15% gels and blotted as described for Fig. 1A. Both the 35 S image and Western blot (probed with anti-CaSR LRG antibody) illustrate that the 140 kDa band is sensitive, whereas the 160 kDa band is insensitive to endoglycosidase H. In contrast, both the 140 and 160 kDa bands were reduced to the unglycosylated form (120 kDa) by treatment with peptide:N-glycosidase F. The results of Fig. 1 demonstrate that CaSR can be pulse-labeled with [ 35 S]cysteine and appears in the ER as the 140-kDa form, which contains core glycosylation sensitive to endoglycosidase H. CaSR is therefore cotranslationally subjected to quality control surveillance because acute block of the proteasome with MG132 results in an increase in net [ 35 S]cysteine incorporation and appearance of the unglycosylated 120-kDa form. The experiment illustrated in Fig. 1B was performed in cells stably expressing CaSR, with a significant proportion of total receptors in the mature form (Fig. 1B, Western blot (right)). Despite this, less than 50% of [ 35 S]CaSR was processed to the endoglycosidase H-insensitive, peptide:N-glycosidase F-sensitive 160-kDa form after 24 h of chase (Fig. 1B, 35 S image (left)). CaSR maturation in transiently transfected cells is slow and incomplete (Fig. 2). Because stably transfected HEK293 cells exhibit a higher level of maturely glycosylated CaSR (Fig. 1B), we determined whether maturation of [ 35 S]CaSR was more rapid or complete in stably transfected cells. We compared three conditions: HEK293 cells stably expressing FLAG-CaSR (F-CaSR s ), HEK293 cells transiently transfected (24 h) with FLAG-CaSR (F-CaSR t ), and HEK293 cells stably expressing untagged CaSR and transiently transfected (24 h) with FLAG-CaSR (CaSR plus F-CaSR t ). For those conditions requiring it, cells were transiently transfected with equivalent amounts of FLAG-CaSR cDNA (6 g) and cultured for 24 h prior to the experiments. We pulse-labeled with [ 35 S]cysteine for 60 min and chased for various times up to 24 h, followed by cell lysis and IP with anti-FLAG antibody. Results of a representative experiment are illustrated in Fig. 3A. The mature form of CaSR (160 kDa) was observed after 8 h. Fig. 3B illustrates the time     in the presence of NPS R-568, the slope was 0.2 min Ϫ1 , significantly greater than in the presence of DMSO (0.11 min Ϫ1 ) (p Ͻ 0.05)). Because these experiments were performed 24 h posttransfection, the effects are probably the result of NPS R-568 acting on newly synthesized receptors in the ER. The level of CaSR at the plasma membrane at 24 h is variable and is 20 -25% of the level achieved at 72 h (taken as 100%), as assessed by ELISA (Fig. 4E). Such low levels of plasma membrane CaSR may be insufficient to fully activate signaling pathways. To further localize the site of NPS R-568 action, we compared its effects with those of neomycin, a cationic, membrane-impermeant allosteric agonist of CaSR (24). Fig. 4F illustrates results of a 60-min pulse with [ 35 S]cysteine, 24 h after transfection, in the presence of either DMSO (D), 10 M NPS R-568 (NPS), or 300 M neomycin sulfate (neo). As expected, NPS R-568 significantly increased [ 35 S]CaSR generation, whereas neomycin sulfate had no effect. To confirm that the effects of NPS R-568 resulted from specific binding at its allosteric site within the CaSR transmembrane domain, we compared the effects of NPS R-568 treatment on WT CaSR and the mutant CaSR(E837I) (24). Residue Glu 837 is located at the extracellular face of helix 7 of the CaSR transmembrane domain and forms an important salt bridge with NPS R-568 to stabilize its binding (24). CaSR(E837I) is not regulated by NPS R-568 but responds normally to Ca 2ϩ or phenylalanine (24). Of significance here is that synthesis of [ 35 S]CaSR(E837I) was not increased by incubation with NPS R-568 (Fig. 4F). The combined results of Fig. 4 suggest that NPS R-568 binds at its allosteric site within the CaSR transmembrane domain during biosynthesis (i.e. acts as a cotranslational pharmacochaperone) and also modestly increases maturation of CaSR to the plasma membrane at later times (16 -24 h).

RESULTS
CaSR Mutants Encounter a Cotranslational Conformational Checkpoint-The results of Fig. 4 strongly suggest that newly synthesized [ 35 S]CaSR encounters a cotranslational pharmacochaperone-sensitive checkpoint and that the membrane-permeant allosteric agonist NPS R-568 can increase the fraction of newly synthesized CaSR that survive. We next determined whether conformational bias conferred by mutation could affect the rates of [ 35 S]cysteine incorporation into WT CaSR, the gain-of-function mutant A843E, and the loss-of-function mutant L849P. Fig. 5A shows a representative experiment, with both monomer and dimer zones of the blot shown (there was significant residual dimer under reducing conditions in this particular experiment). For WT and both mutants, the net abundance of receptors on the Western blot (24 h accumulation) is consistent with the relative rates of [ 35 S]cysteine incorporation determined over 1 h (i.e. A843E Ͼ WT CaSR Ͼ L849P). The GAPDH portion of the same blot (Fig. 5A) demonstrates that this is not a consequence of differential protein loading. Fig. 5B illustrates the normalized results of 3-5 experiments of the type in Fig. 5A. It is clear that the rate of [ 35 S]cysteine incorporation (slopes of lines) for the two CaSR mutants differs from WT CaSR (*, p Ͻ 0.05 versus WT CaSR), varying over a 10-fold range: A843E (0.12 Ϯ 0.02 min Ϫ1 ) Ͼ WT CaSR (0.07 Ϯ 0.01 min Ϫ1 ) Ͼ L849P (0.012 Ϯ 0.01 min Ϫ1 ). The combined results in Figs. 4 and 5 support the existence of a cotranslational conformational checkpoint in CaSR biosynthesis that rapidly targets for destruction those receptors biased toward the inactive conformation. Fig. 3 suggest that the steady state presence of mature CaSR modestly increases the extent of maturation of newly synthesized receptors, although a significant fraction of receptors remain in the immature form after 24 h. We considered the possibility that signaling by plasma membrane-localized CaSR could affect the extent of maturation but that culture medium Ca 2ϩ concentrations (1.1-1.8 mM) were insufficient to evoke a strong maturation signal. We therefore used the charged allosteric agonist of CaSR, neomycin sulfate (19), to potentiate CaSR activation in normal culture medium. Acute application of neomycin sulfate had no effect on [ 35 S]cysteine incorporation into CaSR during a 1-h pulse at 24 h after transfection (shown in Fig. 4F), when there is 20 -25% of maximal plasma membrane-localized CaSR (Fig. 4E). We therefore initiated [ 35 S]cysteine pulse-chase experiments after 48 h of transfection, when plasma membrane-localized CaSR is nearly maximal (Fig. 4E). Cells were treated without or with 300 M neomycin sulfate during the 60-min [ 35 S]cysteine pulse and the 24-h chase period. Fig. 6A illustrates the [ 35 S]CaSR and West- ]cysteine pulse, however, was significantly increased by neomycin sulfate after 48 h of transfection (143 Ϯ 5.5% of control, p Ͻ 0.05) but not after 24 h of transfection (98.5 Ϯ 14.1, not significant) (Fig. 6C), suggesting that the effect is mediated by plasma membrane-localized CaSR. Both the hydrophobic (NPS R-568) and cationic (neomycin) allosteric agonists therefore increase CaSR cotranslational stability but have limited ability to facilitate CaSR maturation through the secretory pathway. The two allosteric drugs probably act via distinct mechanisms, because NPS R-568 acts cotranslationally on intracellular CaSR, whereas neomycin effects are only observed when CaSR is present at the plasma membrane.

Modulation of CaSR Biosynthesis by Plasma Membrane CaSR-Data in
Extracellular Ca 2ϩ is an important regulator of cell function, and normal cell culture medium contains 1.1-1.3 mM Ca 2ϩ plus contributions from added serum. CaSR activation by extracellular Ca 2ϩ is highly cooperative (25), with EC 50 of 3 mM for activation of PLC␤ (19,26), and thus CaSR in normal culture medium might be expected to be partially activated and/or desensitized. We considered the possibility that the intracellular pool of immature receptors may serve as a reservoir of readily releasable CaSR. We compared the effects of varying medium Ca 2ϩ on maturation of CaSR by pulse-chase analysis of cells transiently or stably expressing FLAG-CaSR. Cells were pulsed with [ 35 S]cysteine for 60 min in normal medium, followed by a chase period of up to 24 h in varying medium Ca 2ϩ , either normal DMEM or DMEM containing 0.5 or 5 mM Ca 2ϩ . Results of individual experiments are plotted in Fig. 6D. Neither low nor high Ca 2ϩ medium altered the basic features of [ 35 S]CaSR maturation, suggesting that the orthosteric agonist Ca 2ϩ is not the prime regulator of CaSR maturation to the plasma membrane.
The CaSR CT Dictates CaSR Maturation Rate-[ 35 S]Cysteine pulse-chase analysis of CaSR biosynthesis suggests slow and incomplete maturation. Both allosteric modulators, such as neomycin or NPS R-568 and the orthosteric agonist Ca 2ϩ have a limited ability to enhance the [ 35 S]CaSR maturation rate and/or extent. We therefore considered the possibility that slow maturation (i.e. ER retention) is a physiological requirement for normal CaSR function. Many membrane proteins contain retention, targeting, and trafficking sequences within their CTs. The CaSR CT is large (215 residues) and unique, and we tested whether it contributed to the significant intracellular retention of CaSR. We first compared WT CaSR-EGFP (chimera of full-length CaSR linked at the extreme carboxyl terminus to EGFP) with the CT chimera CaSR/mGluR1␣-EGFP, containing WT CaSR sequence through CT residue 869 and rat mGuR1␣ residues 849 -1194 followed by EGFP. Fig. 7A illustrates results of a 60-min [ 35 S]cysteine pulse followed by a 24-h chase period, and Fig. 7B illustrates the averaged results of three independent experiments. The carboxyl-terminal chimera FLAG-CaSR/mGluR1␣-EGFP undergoes full maturation over the 24-h period, with the crossover point (50% each 140-and 160-kDa forms) at 8 h of chase, whereas the WT FLAG-CaSR-EGFP shows the limited maturation of WT CaSR documented in earlier experiments using FLAG-CaSR. These results suggest that the resistance to maturation of [ 35 S]CaSR is defined by the CaSR CT. It is possible, however, that the mGluR1␣ CT contains maturation-promoting elements. We therefore compared maturation of WT CaSR with the CaSR truncation CaSR⌬868 using a 60-min [ 35 S]cysteine pulse and 24-h chase period (Fig. 7,  C and D). For comparison in Fig. 7C, we also treated full-length FLAG-CaSR with NPS R-568, which, as expected (see Fig. 4B), did not elicit full maturation of WT CaSR. Both the individual experiment (Fig. 7C) and the averaged results of three independent experiments (Fig. 7D) show progressive maturation of CaSR⌬868 without a significant lag period after the end of the [ 35 S]cysteine pulse, achieving 50% maturation after 16 h. The combined results of Fig. 7, A and D, argue that determinants within the CaSR CT distal to residue 868 facilitate ER retention of a significant fraction of newly synthesized [ 35 S]CaSR. To test whether the CaSR CT retention determinants also operate in cells that express endogenous CaSR, we transiently transfected proximal tubule OK cells with either WT CaSR or CaSR⌬868 and examined plasma membrane targeting using ELISAs targeted against an extracellular epitope (FLAG) as a surrogate for maturation of glycosylation. CaSR⌬868 showed significantly higher plasma membrane localization at comparable levels of total protein expression than WT CaSR in both HEK293 and OK cells (Fig. 7E, black bars), suggesting that the determinants of retention within the CaSR CT are functional in both cell types and significantly impact plasma membrane localization of CaSR.
To isolate the region of the CaSR CT mediating ER retention, we generated and transiently expressed a range of CaSR CT truncations (in the same FLAG-CaSR background), followed by anti-FLAG IP and Western blotting. Fig. 8A illustrates a representative Western blot. Progressive truncation of the CaSR CT not only affects maturation but also affects CaSR stability, and Fig. 8A therefore illustrates two different exposures of the same blot to permit accurate quantitation of mature and immature bands of monomeric CaSR. Fig. 8B illustrates the averaged results of four independent transfections using a wider range of truncations. Truncations shorter than CaSR⌬898 show a significant increase in maturely glycosylated CaSR, indicating that residues critical to ER retention lie between CaSR⌬868 and CaSR⌬898. This region contains an extended arginine-rich motif as well as numerous phosphorylation sites that may serve as protein interaction sites for regulated ER retention. D, the normalized abundances of immature (black symbols) and mature (white symbols) FLAG-CaSR (circles) or CaSR⌬868 (triangles) were quantified as described and plotted as mean Ϯ S.D. (n ϭ 3). Statistical significance (p Ͻ 0.05 (*) or p Ͻ 0.005 (**)) was determined at each time point relative to FLAG-CaSR. E, ELISA to quantify relative plasma membrane (black bars) or total (white bars) expression of full-length FLAG-CaSR or the truncation FLAG-CaSR⌬868 in HEK293 and OK cells. Cells were transfected for 48 h prior to ELISA assay as described under "Materials and Methods." Data for each cell type were normalized to WT FLAG-CaSR plasma membrane or total abundance. Statistical significance (**, p Ͻ 0.005) was determined relative to FLAG-CaSR plasma membrane abundance in the same cell type.

DISCUSSION
In this report, we describe the general features of CaSR biosynthesis using [ 35 S]cysteine pulse-chase methods modified to facilitate sufficient label incorporation into nascent CaSR. Several features of [ 35 S]CaSR synthesis and maturation are notably distinct from results obtained for other GPCRs, including opioid (4,27), vasopressin (10,28), bradykinin (29), and luteinizing hormone (30,31) receptors. [ 35 S]CaSR is stable in the ER form (ϳ140 kDa), and maturation to the endoglycosidase H-resistant form (ϳ160 kDa) is generally observed 16 h after the [ 35 S]cysteine pulse. Once initiated, maturation does not go to completion but rather reaches a ratio of immature/mature forms of Յ50% for times up to 48 h. These properties are in sharp contrast to the generally immediate and rapid maturation of many newly synthesized GPCRs (27)(28)(29)(30)(31). It is unlikely that this large store of immature [ 35 S]CaSR represents misfolded protein awaiting degradation because ER-associated degradation is rapid (e.g. ␦-opioid (10) and vasopressin V1b/V3 (28) receptors undergo either maturation or degradation within 4 h of synthesis). A trivial explanation for the current results is that heterologous expression of CaSR limits maturation by saturat-ing the secretory pathway and/or as a result of the absence of cell type-specific chaperones. Two factors argue against this explanation. First, the CT chimera CaSR/mGluR1␣ and the CaSR⌬868 truncation undergo maturation without a lag, suggesting a specific retention mechanism. Second and more compelling is the documented presence of intracellular CaSR in a variety of cell types having endogenous expression, including keratinocytes (32,33), where an intracellular role for CaSR has been suggested (34), kidney cells (35,36), neurons (37), and glia (38). Enhanced plasma membrane targeting is observed for the truncation mutant CaSR⌬868 in HEK 293 cells and OK cells (opossum kidney proximal tubule cells, which express endogenous CaSR (39)), suggesting common mechanisms mediating intracellular retention of full-length CaSR.
Given that intracellular CaSR is a physiologically relevant form of the receptor, there are two potentially non-exclusive roles for the large, stable intracellular population of immature [ 35 S]CaSR. Mobilization of nascent intracellular CaSR may allow more rapid alterations in plasma membrane CaSR levels in response to cellular signaling than regulation at the transcriptional or even translational levels. Cells are chronically exposed to extracellular Ca 2ϩ , and a stable intracellular pool of CaSR may represent an adaptive mechanism for sensing dynamic changes in extracellular Ca 2ϩ in the face of chronic desensitization. The inability of either low (0.5 mM) or elevated (5 mM) Ca 2ϩ to facilitate maturation of [ 35 S]CaSR argues against this possibility, although mobilization of CaSR may result from non-CaSR signaling. A second possibility is that intracellular CaSR may subserve distinct cellular signaling functions from that of plasma membrane CaSR, as has been demonstrated for the related Family C GPCRs, mGluR1 and mGluR5 (40 -42), and Family A GPCRs, including ␣1 adrenergic (43), estrogen-sensitive GPR30 (44), and apelin, angiotensin AT 1 , and bradykinin B 2 (45) receptors. CaSR fulfills the criteria for a GPCR with the potential for intracellular signaling (i.e. CaSR is stably expressed within these intracellular compartments, and it has access to its agonist, Ca 2ϩ , which is concentrated within the lumen of the ER, Golgi, and nuclear envelope compartments). Further, the ER lumen contains not only Ca 2ϩ but glutathione, which is an allosteric activator of CaSR (46). The present data also argue that active mechanisms are invoked to retain CaSR at significant levels within the ER. The maturation profiles of both the CaSR/mGluR1␣ chimera and CaSR⌬868 suggest that the CaSR CT distal to residue 868 actively participates in ER retention. Fine mapping of the maturation of CaSR truncations suggests that the interactions mediating ER retention are localized between Thr 868 and Arg 898 of the proximal CT. Numerous studies have characterized CaSR CT truncations (e.g. see Refs. [47][48][49], identifying roles for the proximal CT in signaling and plasma membrane targeting. The current work identifies a novel contribution of the proximal CT to ER retention. Despite the large and unique features of the CaSR CT, few interacting proteins have been identified, probably because traditional yeast two-hybrid screening approaches are not optimized for identification of protein interactions regulated by phosphorylation. The proximal CaSR CT distal to residue 868 is rich in potential phosphorylation sites for protein kinases C and A as well as casein kinases FIGURE 8. Residues between Thr 868 and Arg 898 control ER retention of CaSR. A, HEK293 cells were transiently transfected with 1 g of cDNA (6-well plates) for 48 h with FLAG-CaSR or various CT truncations (CaSR⌬868, CaSR⌬880,, CaSR⌬886, CaSR⌬898, CaSR⌬908, or CaSR⌬1024), followed by cell lysis, IP with anti-FLAG antibody, and Western blotting. Blots were probed with anti-CaSR antibody, and the interval image capture mode of the FUJIFILM LAS-4000mini luminescent analyzer was used to capture 15 images at 1-min intervals. Images that were optimal for individual truncations were used for quantitation of the immature and mature monomeric forms of CaSR (which varied in mass, depending upon the degree of truncation). B, plot of the fraction of total CaSR in mature (filled circles) or immature (open circles) forms for WT FLAG-CaSR and truncations (CaSR⌬868, CaSR⌬880, CaSR⌬886, CaSR⌬898, CaSR⌬908, CaSR⌬1024, and CaSR⌬1052). Data were calculated as the percentage of total CaSR protein for each truncation (WT CaSR has 1078 residues). Error bars, S.D.
I and II, Akt, and GSK3␤ (NetPhos 2.01), suggesting that protein interactions with the CaSR CT may be differentially regulated by cellular signaling. ER luminal Ca 2ϩ and/or glutathione levels vary as a function of cellular signaling (50,51) and redox status (52,53), and ER-retained CaSR may therefore play a physiological role in integrating these diverse and dynamic signals. Careful dissection of the properties of plasma membrane versus intracellular CaSR-mediated signaling will be required to validate this hypothesis.
CaSR is regulated by a variety of endogenous and pharmacological allosteric modulators, including amino acids and glutathione, polyamines, and polycationic antibiotics, targeted to site(s) on the ECD, and several classes of allosteric agonists and antagonists targeted to site(s) within the heptahelical domain (reviewed in Ref. 19). We have previously shown that NPS R-568 can rescue both WT CaSR and some loss-of-function mutants identified in familial hypocalciuric hypercalcemia/ neonatal severe primary hyperparathyroidism patients, increasing both total and plasma membrane-targeted levels of receptor protein and function (21,22). Allosteric agonists may have similar effects in vivo because uremic rats treated with cinacalcet (54) as well as first/second generation allosteric agonists (calcimimetics) NPS R-568 (55), Amgen R-568 (56,57), or AMG-641 (58) show reduced parathyroid gland hyperplasia, vascular calcification, and remodeling as a result of both enhanced activation and expression of CaSR in relevant tissues (54 -58). The current results explicitly define the mechanism of NPS R-568 action as cotranslational stabilization of newly synthesized CaSR. Results suggest that CaSR navigates all generic (glycosylation, disulfide bond shuffling) and specific (helix packing, conformational assessment) quality control checkpoints cotranslationally, and the rate of appearance of [ 35 S]CaSR can be taken as a measure of the relative stability conferred by ambient conditions in the ER. Results with NPS R-568 on WT CaSR were recapitulated with CaSR mutations, with the rate of [ 35 S]cysteine incorporation of the representative gain-of-function mutant A843E Ͼ WT CaSR Ͼ the loss-of-function mutant L849P. Targeting of misfolded receptors to the ERAD pathway occurs cotranslationally because the presence of MG132 during the [ 35 S]cysteine labeling period induces the appearance of the unglycosylated form of the receptor. Interestingly, the continued presence of NPS R-568 during the chase period, although increasing net CaSR protein, did not eliminate the lag to initiation of maturation of [ 35 S]CaSR. Overall, these results suggest that NPS R-568 acts cotranslationally to stabilize nascent [ 35 S]CaSR but is not the dominant regulator of CaSR maturation. This is in contrast to the pharmacochaperone effects on rescue of vasopressin and gonadotrophin-releasing hormone receptors, where exposure of cells to pharmacochaperones after biosynthesis is able to rescue misfolded receptors to the plasma membrane (9,59,60).
The conformational checkpoint may be a unique and necessary step in the biosynthesis of CaSR, which is exposed to an agonist-rich compartment during biosynthesis. Misfolded CaSRs which are not able to achieve the active conformation may be unable to participate in the protein interactions required for maturation through the secretory pathway. ERbased conformational sampling has been reported for AMPA (␣-amino-3-hydroxy-5-methyl-4-isoxazolepriopionate) and kainate receptor channels, which progress through the range of normal channel gating motions in the ER prior to release to the plasma membrane (60,61). Mutant receptors biased toward either the open or closed conformations of the agonist binding domain have significantly different ER exit rates, and optimal ER exit occurs when the channels reversibly achieve the closed cleft state normally stabilized by agonist (61). These studies are of considerable interest because the extracellular agonist binding domains of AMPA/kainate channels are homologous to the CaSR ECD (62), and similar constraints on stability and/or ER exit of CaSR may apply.
The combined results of Figs. 3 and 6 suggest a second level of allosteric modulation of CaSR cotranslational stability, resulting from activation of plasma membrane-localized, maturely glycosylated CaSR. Neomycin and related aminoglycoside antibiotics have been shown to activate CaSR signaling with potencies positively correlated with the number of cationic charges (63). In the present studies, neomycin had no effect on the rate of [ 35 S]CaSR synthesis after 24 h of transient transfection (i.e. at low levels of plasma membrane CaSR), in sharp contrast to the effects of the hydrophobic allosteric agonist NPS R-568. However, neomycin significantly (ϳ50%) increased [ 35 S]CaSR when applied after 48 h of transfection, when Western blots and ELISAs indicate significant levels of maturely glycosylated plasma membrane-localized CaSR. Neomycin is therefore not a pharmacochaperone but rather activates CaSR signaling pathways. A possible candidate is the MAPK pathway, which is robustly activated by CaSR (26). Phosphorylation of ERK1/2 has been shown to lead to Mnk1 phosphorylation, leading to enhanced translation initiation (64). Further studies are required to determine whether neomycin enhances translation initiation or mediates cotranslational stabilization of nascent [ 35 S]CaSR by as yet undefined pathways, but the current results suggest that cellular CaSR abundance may be tuned by CaSR signaling.
In summary, we have characterized the early steps in CaSR biosynthesis using [ 35 S]cysteine pulse-labeling approaches. CaSR stability is regulated cotranslationally by a conformational checkpoint that targets to ERAD receptors biased toward the inactive conformation. Receptors can be stabilized by NPS R-568, a cotranslational pharmacochaperone. CaSR maturation is actively regulated by determinants in the CaSR CT, and neither the orthosteric agonist Ca 2ϩ nor the allosteric agonists NPS R-568 or neomycin are able to significantly reduce the lag prior to the initiation of maturation or drive complete maturation. These results strongly suggest that a significant fraction of CaSR is actively retained in the ER and lead to the intriguing possibility that CaSR serves a unique role(s) in an as yet undefined subcompartment of the ER. The challenges now are to characterize the potentially unique contributions of intracellular CaSR to cellular physiology and to identify the protein partners mediating CaSR retention.