Retrovirally Mediated Transfer of a G Protein-coupled Receptor Kinase (GRK) Dominant-negative Mutant Enhances Endogenous Calcitonin Receptor Signaling in Chinese Hamster Ovary Cells

G protein-coupled receptor kinases (GRKs) initiate pathways leading to agonist-dependent phosphorylation and desensitization of G protein-coupled receptors. However, the role of GRKs in modulation of signaling properties of native receptors has not been clearly defined. Here we addressed this question by generating Chinese hamster ovary (CHO) cells stably expressing a dominant-negative mutant of GRK2 (DN-GRK2), K220R, using retrovirally mediated gene transfer, and we assessed function of the endogenously expressed calcitonin (CT) receptors. We found that CT-mediated responses were prominently enhanced in CHO cells expressing DN-GRK2 compared with mock-infected control CHO cells with ∼3-fold increases in CT-promoted cAMP production in whole cells and adenylyl cyclase activity in membrane fractions. CT-promoted phosphoinositide hydrolysis was also enhanced in DN-GRK2 cells. The number of CT receptors was increased ∼3-fold in DN-GRK2 cells, as assessed by125I-salmon CT-specific binding, and this was associated with increased CT receptor mRNA levels. These results indicate that DN-GRK2 has multiple consequences for CT receptor signaling, but a primary effect is an increase in CT receptor mRNA and receptor number and, in turn, enhanced CT receptor signaling. As such, our findings provide a mechanistic basis for previous observations regarding agonist-promoted down-regulation of CT receptors and for resistance and escape from response to CT in vitroand in vivo. Moreover, the data suggest that blunting of receptor desensitization by DN-GRK2 blocks a GRK-mediated tonic inhibition of CT receptor expression and response. We speculate that GRKs play a similar role for other G protein-coupled receptors as well.

Stimulation of G protein-coupled receptors (GPCRs) 1 in the plasma membrane triggers two events, activation of G proteinmediated signal transduction pathways and, in parallel, a deactivation or desensitization of signaling. One component of this desensitization, in particular homologous, receptor-specific desensitization, is agonist-dependent phosphorylation of the receptor by specific G protein-coupled receptor kinases (GRKs). This phosphorylation event leads to the recruitment of cytosolic proteins, ␤-arrestins, to the receptor-signaling complex, the uncoupling of receptor from heterotrimeric G proteins, and loss of receptor responsiveness. Recent evidence indicates that GRKs and ␤-arrestins not only promote receptor uncoupling but also may directly participate in GPCR sequestration and the initiation of events leading to clathrin-coated pit-mediated internalization of receptors (for recent reviews see Refs. [1][2][3][4][5]. In addition, GRK activation may be required to initiate certain events unrelated to receptor desensitization (3,4).
At least six different isoforms of the GRK family have been isolated. GRK2, formerly termed ␤ARK 1 , is a widely expressed member of this family and has been shown to phosphorylate various GPCRs (1,2). GRK2 K220R, a dominant-negative GRK2 mutant (DN-GRK2) in which lysine at position 220 has been mutated to arginine to disrupt kinase activity (6), has been used to attenuate desensitization of several GPCR systems such as the ␤ 2 -adrenergic receptor, ␣ 1B -adrenergic receptor, adenosine A2 receptor, thyrotropin receptor, follitropin receptor, and CCR2B receptor (6 -11).
To date, however, relatively little is known about the role of GRKs in desensitization to peptide hormones, in particular, in the regulation and expression of endogenously expressed GPCRs. Thus, most studies of GRKs have involved the use of transfected cells expressing relatively nonphysiological levels of GPCRs. In the current studies, we utilized a model cell, Chinese hamster ovary (CHO) cells, which endogenously expresses calcitonin (CT) receptors, and we evaluated the role of endogenously expressed GRKs on signaling. We stably expressed DN-GRK2 in CHO cells using retrovirally mediated gene transfer and found that CT receptor expression and signaling were markedly enhanced by the DN-GRK2. The results suggest a key role for GRKs in establishing the steady-state level of GPCR expression.

EXPERIMENTAL PROCEDURES
Materials-Salmon calcitonin, human calcitonin, forskolin, and monoclonal M5 anti-FLAG antibody were purchased from Sigma. Antibodies against GRK2 (C-15), GRK-3 (C- 14), GRK5 (C- 20), and GRK6 * This work was supported in part by grants from the National Institutes of Health. 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.
‡ Recipient of an oversea research fellowship by Japanese Science and Technology Corp. and a postdoctoral fellowship by Uehara Memorial Foundation.
(C-20) were purchased from Santa Cruz Biotechnology, Inc. 125 I-cAMP and myo-[ 3 H]D-inositol were purchased from NEN Life Science Products. GTP␥S was from Roche Molecular Biochemicals. Cell culture media and fetal bovine serum were from Life Technologies, Inc. 125 I-Salmon calcitonin and enhanced chemiluminescence solutions was from Amersham Pharmacia Biotech. Rat brain cDNA was purchased from CLONTECH. pCMVneo/GRK2 K220R was a generous gift from Dr. Jeffrey L. Benovic (Thomas Jefferson University). Rat GRK3 cDNA was kindly provided by Dr. Robert J. Lefkowitz (Duke University) and rat GRK4a and GRK6a cDNAs by Dr. Jean-Marc Elalouf (Gif-sur-Yvette, France). Murine GRK2 cDNA was previously cloned in our laboratory (12). Recombinant rhodopsin kinase was generated in baculovirus and provided by Ryan Adams from Dr. Alexandra Newton's laboratory (University of California, San Diego).
Cells and Cell Culture-CHO 10001 cells derived from a CHO Pro-5 line of cells were kindly provided by Dr. Michael Gottesman, National Institutes of Health (13). Cells were maintained in Ham's F12 medium with 10% fetal bovine serum and antibiotics (50 units/ml penicillin and 50 g/ml streptomycin (Life Technologies, Inc.)) in gelatin-coated 75cm 2 flasks until 80% confluent. COS7 cells were maintained in Dulbecco's modified Eagle's medium with 10% fetal bovine serum and antibiotics.
Construction of Expression Vectors-A pcDNA3/FLAG-tagged DN-GRK2 construct was generated by excising the EcoRI fragment from pCMVneo/GRK2-K220R and subsequent ligation of this blunted fragment by Klenow enzyme (Life Technologies, Inc.) into amino-terminally FLAG-tagged pcDNA3 (Invitrogen) at a blunted NotI site. Retroviral FLAG-tagged DN-GRK2 construct (LFDRNL) was generated by insertion of the blunted BamHI/XhoI fragment of pcDNA3/FLAG-tagged DN-GRK2 into a blunted SalI site of pLRNL vector. All constructs were verified by sequencing. The coding region of murine GRK5 cDNA was isolated from murine S49 lymphoma cell cDNA using PCR primers specific to the murine GRK5 sequence (sense, 5Ј-CAA TGG AGC TGG AAA ACA TCG TGG CC-3Ј; antisense, 5Ј-GAG CCG AAA CTA GCT GCT GCT TCC CGT G-3Ј), and sequencing of the clone was revealed to be identical to the published murine GRK5 cDNA (GenBank TM accession number AF040746). Murine GRK5 cDNA and rat GRK6a cDNA were cloned into amino-terminally FLAG-tagged pcDNA3 at EcoRI/ XhoI sites, and the expression of those constructs were confirmed by immunoblotting with monoclonal M5 anti-FLAG antibody.
Generation of FLAG-DN-GRK2 Retrovirus-FLAG-DN-GRK2 pseudotyped retrovirus was generated by Dr. Atsushi Miyanohara, Vector Development Laboratory, University of California, San Diego. The method for generation of pseudotyped retrovirus with G-glycoprotein of the vesicular stomatitis virus (VSV-G) envelope has been previously described (14). Briefly, with calcium phosphate coprecipitation LFDRNL cDNA and an expression plasmid for amphotropic envelope gene were transfected into a packaging cell line 293GP cells that express Moloney murine leukemia virus gag-pol gene. The amphotropic retroviral vectors generated 48 h after transfection were collected and then filtered through a 0.45-m filter. This amphotropic retrovirus was used to infect another packaging cell line derived from canine thymocyte CF2Th cells in the presence of Polybrene (8 g/ml). The CF2Th packaging cell line stably expresses a tetracycline-inducible vector containing the VSV-G envelope gene, as well as retroviral gag and pol genes. Neomycin-resistant Cf2Th clones were picked after selection in G418 (400 g/ml)-containing medium in the presence of tetracycline, and expression of the inserted FLAG-DN-GRK2 gene was determined by immunoblotting with an anti-FLAG M2 monoclonal antibody (Eastman Kodak Co.). The clone that produced the highest amount of the FLAG-DN-GRK2 gene was then expanded and used for subsequent production of the pseudotyped virus by removal of tetracycline. The culture medium was replaced with fresh medium, and the pseudotyped virus was collected between 24 and 96 h after removal of tetracycline. The collected culture medium was condensed by centrifugation and then filtered through a 0.45-m filter. The VSV-G pseudotyped retrovirus was titered by infecting 3T3 TkϪ cells and HT1080 cells, and infectious virus at a titer 1 ϫ 10 7 colony-forming units/ml was obtained.
Infection of CHO Cells-CHO cells retrovirally expressing LFDRNL (DN cells) were generated by infection of FLAG-DN-GRK2 retrovirus at a multiplicity of infection (m.o.i.) of 5:1 colony-forming units per cell in the presence of Polybrene (8 g/ml). For mock-infected control cells, CHO cells were infected with LacZ VSV-G pseudotyped retrovirus (LZRNL) that generated from the same pLRNL retroviral backbone (provided by Dr. Miyanohara) (14). As a preliminary experiment to determine the effect of different m.o.i.s on infection efficiency, the cells were infected with LacZ retrovirus at an m.o.i. of 1-10, and then lacZ gene expression was confirmed as assessed by in situ ␤-galactosidase staining. At an m.o.i. of 5 or at higher titer, Ͼ99% of cells were stained 48 h after infection (data not shown).
CHO cells inoculated with retroviruses 48 h after infection were maintained in medium with G418 (600 g/ml) for 2 weeks and 24 neomycin-resistant clones were picked after selection. For DN cells, expression of FLAG-DN-GRK2 was determined by RT-PCR and by immunoblotting with a monoclonal anti-FLAG M2 antibody. One of the DN cells with highest expression of FLAG-DN-GRK2 was chosen for further experiments (DN #1 cells, Fig. 2), and several frozen vials of the DN clone were generated at the third generation of passage after clone selection. For functional studies, the DN clone with highest expression was maintained up to the tenth generation of passage from thawing a frozen vial, and then a new frozen vial of the clone was thawed and used.
Substrate (Tubulin) Phosphorylation Assay for GRK Activity-Two different substrates were used to assay GRK activity as follows: purified "tubulin" and light-dependent phosphorylation of rhodopsin by rhodopsin kinase. Purified tubulin was kindly provided by Drs. Sam Farlow and Lawrence Goldstein (University of California, San Diego), prepared from extracts of freshly isolated bovine brain, and purified to Ͼ99% homogeneity using high pressure liquid chromatography (15). Purified rhodopsin from bovine retina was obtained from Calbiochem.
The phosphorylation reaction was initiated by adding 20 l of kinase reaction buffer (final concentration, 20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 2 mM EDTA, 1 mM dithiothreitol, 50 M ATP, 5,000 cpm/pmol [␥-32 P]ATP) to the immune complexes and 5 l of 1 M taxol-precipitated microtubules or 2 l of a dark adapted suspension of rhodopsin (in the case of tubulin as substrate). After incubation at 30°C for 15 min (in the presence and absence of light in the case of rhodopsin), the reaction was stopped with 5ϫ Laemmli sample buffer and boiled and then resolved by 10% SDS-PAGE. The radioactivities of the phosphorylated proteins were quantitated by an AMBIS radioanalytic imaging system (AMBIS Systems Inc., San Diego, CA).
RT-PCR-Total RNA was prepared by disruption of cells in TRYSOL reagent (Life Technologies, Inc.). Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed as described previously (16). Briefly, total RNA (10 g) was treated with RNase-free DNase I (Life Technologies, Inc.) for 30 min in the presence of RNase guard (Amersham Pharmacia Biotech) to eliminate contamination of genomic DNA. After DNase I pretreatment, RNA was extracted with phenol:chloroform and ethanol-precipitated, vacuum-dried, and then resuspended in RNase-free water. The RNA sample was reverse-transcribed with 200 units of Moloney murine leukemia virus reverse transcriptase (Life Technologies, Inc.), 100 pmol of random hexamer (Amersham Pharmacia Biotech), and 20 units of RNase guard for 1 h at 37°C and was then suspended in distilled water up to 100 l.
The PCR conditions for each GRK isoforms consisted of 35 thermal cycles at 60°C annealing temperature. The condition for CTR C1a consisted of 40 thermal cycles at 55°C annealing temperature and that for GAPDH was 30 thermal cycles at 65°C annealing temperature. To visualize the PCR products, the samples were subjected to electrophore-sis in 3% NuSieve agarose gel or 6% polyacrylamide gel followed by staining with ethidium bromide.
cAMP Assay-The intracellular cAMP content was measured by radioimmunoassay as described previously (17). Briefly, cells were seeded onto gelatin-coated 24-well plates, grown to subconfluency. Growth medium was removed from cells, and cells were equilibrated for 30 min at 37°C in serum-free Dulbecco's modified Eagle's medium containing 20 mM HEPES buffer (DMEH, pH 7.4). Subsequently cells were incubated in fresh DMEH with ligand in the presence of 0.2 mM isobutylmethylxanthine (IBMX), a phosphodiesterase inhibitor. Reactions were terminated by aspiration of medium and addition of 5% trichloroacetic acid. Intracellular cAMP levels were determined by radioimmunoassay (Calbiochem) of trichloroacetic acid extracts following acetylation. Experiments were performed in triplicate.
Adenylyl Cyclase Assay-Adenylyl cyclase activities in membranes prepared from CHO cells were determined as described previously (18). Briefly, 40 g of crude membrane preparation in 100 l of the standard assay mixture (50 mM HEPES, pH 7.4, 1 mM EDTA, 5 mM MgCl 2 , 0.2 mM IBMX, 0.5 mM ATP, 1 mM dithiothreitol, and 10 M GTP) was stimulated by various ligands for 10 min at 30°C, and terminated by being boiled at 90 s. The cAMP levels were determined by radioimmunoassay as described above.
Phosphoinositide Hydrolysis-[ 3 H]Inositol phosphate (IP) formation was assayed as described previously (19). Briefly, cells were seeded onto 6-well plates, grown until subconfluent, and incubated with 3 Ci/ml myo-[ 3 H]D-inositol in culture medium for overnight. Cells were incubated with 10 mM LiCl 2 for 1.5 h and then stimulated with agonists for 10 min. The reactions were terminated with 50% methanol, 50% HCl 2 (0.1 M). Separation of free inositol from inositol phosphates was performed using AG Dowex columns. Aqueous phases of samples were applied to columns and free inositol was removed by two washes of distilled water. Inositol phosphates were eluted out with 2 M ammonium formate plus 100 mM formic acid. 10 ml of scintillation fluid were added and samples were counted in a scintillation counter.

125
I-sCT Binding Assay-Radioligand binding assay was performed by incubating 6 ϫ 10 6 cells/tube for 3 h at 4°C with 125 I-sCT (2,000 Ci/mmol, Amersham Pharmacia Biotech) in the presence or absence of unlabeled sCT (5 M). DMEH containing 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml bacitracin, and 0.1% bovine serum albumin was used to dilute 125 I-sCT. The reaction was stopped by washes three times with ice-cold phosphate-buffered saline containing 0.1% bovine serum albumin and then samples were filtered onto GF/C glass fiber membranes pretreated with 0.1% (w/v) polyethyleneimine for 2 h. Samples were counted in a gamma counter. Binding data were analyzed with the Graphpad Prism statistical fitting software (Graphpad Software Inc., San Diego, CA).

Expression of GRK Isoforms in CHO Cells
-Before studying the role of GRKs in regulating endogenous calcitonin (CT) receptors in CHO cells, we examined the profile of GRK isoforms expressed in the cells. We investigated the expression of GRK isoforms mRNAs by RT-PCR analysis using CHO cell cDNA as a template. PCR primers for GRK2 and GRK3 were designed at the catalytic region of the kinase, whereas those for GRK4, GRK5, and GRK6 were designed at the carboxyl-terminal part of each GRK isoform. Each pair of PCR primers was highly specific for a single corresponding GRK isoform. Murine GRK2 cDNA, rat GRK3 cDNA, rat GRK4a cDNA, murine GRK5 cDNA, and rat GRK6a cDNA were used as controls. Fig.  1A shows the result of RT-PCR for each GRK isoform. PCR products specific for GRK2, GRK5, and GRK6 (572-, 144-, and 117-bp length, respectively) were obtained from CHO cell cDNA, whereas there were no amplified products for GRK3 and GRK4.
Expression of GRK proteins was studied by Western blot analysis (Fig. 1B). Relatively abundant expression of GRK2 and GRK6 proteins was demonstrated in CHO cells; limited amount of GRK5 protein expression was detectable in the cells. GRK3 protein expression was not detected in CHO cells, consistent with the result from RT-PCR. A specific antibody against rodent GRK4 protein has not been available; thus expression of GRK4 protein was not investigated.
Stable Expression of DN-GRK2 in CHO Cells-We generated a CHO cell line stably expressing DN-GRK2 by infecting the cells with a retrovirus-expressing FLAG-tagged DN-GRK2 and generated a mock-infected control CHO cell line with a LacZ retrovirus in the same backbone vector (LZNRL). Twenty four neomycin-resistant clones were obtained after G418 (500 g/ ml) selection of either DN-GRK2 retrovirus-infected cells or mock-infected cells. Fig. 2A shows immunoreactivities of FLAG-tagged DN-GRK2 in CHO cells. Whole cell lysates obtained from CHO cells were immunoprecipitated with anti-GRK2 antibody and were subsequently immunoblotted with a monoclonal anti-FLAG M5 antibody. Eighty-kDa proteins corresponding to the molecular weight of FLAG-tagged DN-GRK2 were detected in three different DN-GRK2 expressing cells (Fig. 2, DN #1-#3), whereas there was no 80-kDa protein in mock-infected cells (Fig. 2, Mock). DN 1 clone, the DN-GRK2 expressing cell lines with the highest expression, was used for all further functional studies; DN 2 and DN 3 were used in certain studies. As shown in Fig. 2B, this clone showed enhanced expression of immune reactive "GRK2," which represents both endogenous protein and the cross-reacting DN construct as well. The expression of the DN protein is approximately 3-fold greater than in the mock or parental cells. Other immunoblotting studies revealed no change in expression of GRK6 (data not shown).

FIG. 1. Profile of GRK isoforms in CHO cells.
Whole cell lysates from either CHO cells (50 g of protein) or COS7 cells transiently expressing each GRK isoform (10 g of protein) were immunoblotted with specific polyclonal antibodies against GRK2, GRK3, GRK5, and GRK6. RT-PCR was performed using 500 ng of total RNA isolated from CHO cells at 35 thermal cycles. PCR primers were designed based on rat GRK2, rat GRK3, rat GRK4a, murine GRK5, and rat GRK6a cDNA sequences, and those cDNAs were used for positive controls of PCR.
Substrate Phosphorylation by GRK2 Activity Obtained from CHO Cells-To investigate GRK-dependent phosphorylation activities in CHO cell lines, an in vitro phosphorylation study was performed using either purified tubulin or rhodopsin as a substrate. The highly purified tubulin preparation that we used was free of endogenous tubulin kinase activity (data not shown). Post-nuclear fractions from mock-infected cells (Mock) or DN-GRK2 expressing (DN) cells were immunoprecipitated with anti-GRK2 antibody, and the immune complexes were used as sources for kinase activity. As shown in Fig. 3A, we detected time-dependent phosphorylation of tubulin in both Mock cells and DN cells. When the cell lysates were immunoprecipitated with nonimmune rabbit IgG, very little phosphorylation was detected in the two cell lines (data not shown). Fig.  3B shows the radioactivities of phosphorylated tubulin from four independent assays. The result shows that GRK2-dependent phosphorylation was inhibited by approximately 40% in DN cells compared with Mock cells. Further evidence for inhibition of GRK-dependent phosphorylation by DN expression was obtained in studies assessing light-dependent phosphorylation of rhodopsin, in which the intensity of phosphorylation was inhibited about 70% (Fig. 3C). Thus the DN-GRK2 acts as a functional inhibitor of GRK activity.

Stimulation of cAMP Production in CHO Cells by Calcitonin (CT)
Agonists-Calcitonin is known to activate the G s /adenylyl cyclase and G q /PI hydrolysis-Ca 2ϩ signaling systems (20). Previous studies have documented that various CHO isolates express CT receptor C1a mRNA and CT receptor C1a-like receptors (21,22). We investigated whether functional CT receptor was expressed in CHO 10001 cells and, if so, whether DN-GRK2 expression affected CT-stimulated cAMP responses. Both Mock cells and DN cells were incubated with salmon CT (sCT) or human CT (hCT) in the presence of IBMX (0.2 mM) for 10 min, and whole cell cAMP production was assayed. As shown in Fig. 4A (Fig. 4, A and B). Thus the DN-GRK2 increased maximal response without substantially altering EC 50 .
To determine if stable expression of DN-GRK2 altered either the rate of cAMP production or the time to peak response after CT stimulation, Mock cells and DN cells were assayed either in the presence or in the absence of IBMX (0.2 mM) at several times after sCT (1 nM) stimulation. In Mock cells, cAMP levels in the presence of IBMX increased linearly up to 10 min after sCT stimulation, reached a maximal level by 20 min, and then gradually decreased to a low level by 60 min (Fig. 5, A and B). Without IBMX, the level of cAMP in Mock cells was maximal by 10 min but was only one-third of that in the presence of IBMX. In DN cells, the cAMP level in the presence of IBMX increased linearly up to 10 min after sCT stimulation and reached a maximal level at 30 min. The cAMP level in DN cells without IBMX reached a maximal level at 20 min and gradually decreased by 30 min. The peak cAMP level in DN cells without IBMX was 80% of that in the cells with IBMX. As shown in Fig.  5B, expression of the DN-GRK2 appeared to increase the rate of cAMP formation relative to that of Mock cells. In addition, retroviral DN-GRK2 expression did not alter basal cAMP levels or forskolin-stimulated cAMP levels. ( Fig. 6). In Mock cells, sCT-induced cAMP response following sCT treatment was approximately one-half of the response following incubation with vehicle; thus, sCT treatment homologously desensitized sCT-induced cAMP response in the control cells as well as in three separate clonal isolates of cells expressing construct lacking DN-GRK. In DN cells, however, sCT treatment did not alter the sCT-induced cAMP response; thus, desensitization of sCT-induced cAMP production was impaired in DN cells. In contrast to sCT-induced cAMP response, forskolin ( Fig. 7, both sCT and hCT stimulated adenylyl cyclase activity in Mock cells (4.5 Ϯ 1.0and 2.2 Ϯ 0.5-fold over basal for 1 M sCT and 1 M hCT, respectively). In DN cells, the stimulation of adenylyl cyclase by sCT and hCT was significantly increased compared with Mock cells (11.5 Ϯ 1.3-and 6.7 Ϯ 0.7-fold over basal for 1 M sCT and 1 M hCT, respectively). The result strongly suggest that potentiation of CT-induced adenylyl cyclase activity in DN cells is not secondary to changes in either G proteins or adenylyl cyclase as the response to GTP␥S or forskolin was unchanged by DN-GRK2 expression.

Impairment of Desensitization of sCT-induced cAMP Production by DN-GRK2 Expression-To
Enhancement of CT-increased IP Formation by DN-GRK2 Expression-Previous reports have shown that CT couples to G q -dependent signaling pathways as well as to the G s -dependent signaling pathway (23). To investigate whether DN-GRK2 expression also potentiated G q -dependent signaling, we assayed CT-increased IP formation. Although in Mock cells, we could not reproductively detect an sCT (1 M)-mediated increase in IP formation, in DN cells, sCT (1 M) stimulated approximately a 2-fold increase over basal (Fig. 8). ATP (10 M) also stimulated a 2-fold increase over basal in DN cells. Therefore, the data suggest that DN-GRK2 expression enhances G q -dependent CT signaling and that of another G q -dependent receptor (P 2 Y receptor).
Up-regulation of sCT-specific Binding Sites by DN-GRK2 Expression-To determine whether the enhanced CT responses in DN cells were dependent on the number of CT receptors, we performed a saturation binding study of the radioligand 125 I-sCT using intact Mock cells and DN cells (Fig. 9). Whole cells were incubated with CT ligands for 3 h in chilled medium containing 0.1% bovine serum albumin and protease inhibitors to prevent agonist-induced internalization and peptide degradation. Binding affinity to 125 I-sCT was not significantly different between both cell lines (K d values: for Mock cells, 24.1 pM, 95% CI 8.0 -40.2 pM, for DN cells, 22.3 pM, 95% CI 7.3-37.4 pM). The maximal number of CT specific binding sites in DN cells was, however, 2.5-fold larger than that found in Mock cells (610 Ϯ 55 sites/cell versus 240 Ϯ 40 sites/cell, p Ͻ 0.005). The level of specific binding sites for sCT in DN cells was comparable to the enhancement of sCT-promoted cAMP production in the cell line relative to that in Mock cells. The results of the whole cell binding studies demonstrate that stable expression of DN-GRK2 up-regulates CT receptor expression in CHO cells.
CT Receptor C1a mRNA Expression-To investigate further whether up-regulation of CT receptors was dependent on CT receptor mRNA expression, the relative abundance of CT receptor mRNA expression was examined by RT-PCR. Since other CHO cell strains have been reported to express CT receptor C1a (21,22), expression of CT receptor C1a mRNA in our CHO cell lines was examined using PCR primers that were designed based on rodent CT receptor C1a cDNAs. As shown in Fig. 10, a 253-bp PCR product obtained with primers specific to C1a was detectable in parental and Mock CHO cells using 40 thermal cycles (not detectable using 35 thermal cycles), whereas higher expression of the amplified product was shown in DN cells, and a signal could be detected using 35 thermal cycles (data not shown). Sequencing of the 253-bp PCR product obtained from rat brain cDNA was identical to the published rat CT receptor C1a sequence (GenBank TM accession number L13041); sequencing of the cloned PCR products obtained from CHO cell lines revealed 90.5% identity with the rat CT receptor C1a mRNA. A housekeeping gene GAPDH mRNA was used as an internal standard for RT-PCR amplification, and the expression levels of GAPDH in Mock cells and DN cells were similar. Although the RT-PCR assay was only semi-quantitative, the data strongly suggest that DN cells express more CT receptor C1a mRNA than do Mock cells. DISCUSSION Calcitonin is an important hormone in the regulation of serum calcium levels and bone mineral density through its effects on bone resorption and renal calcium excretion. Previous studies have not clearly defined the mechanisms involved in the regulation of CT receptor expression, in particular, agonist-mediated desensitization and down-regulation of these receptors (24 -26). In the current studies, we used CHO 10001 cells as a model to examine CT receptor expression by GRKs. We found that retrovirally mediated gene transfer is a useful means to establish stable expression of DN-GRK2. Stable expression of DN-GRK2 inhibited GRK2-mediated substrate phosphorylation by approximately 40% and impaired CT-mediated desensitization of cAMP formation. Potentiation of cAMP generation in DN-GRK2-expressing cells, however, appeared to be much greater than what would have been expected from the loss in GRK2 phosphorylating activity and appeared to relate to the increase in CT receptor number. Enhancement of CT receptor signaling by DN-GRK2 expression was also observed in the phosphoinositide pathway. The up-regulation of CT receptor number in DN-GRK2-expressing cells was associated with an increase in mRNA for CT receptor C1a. These data suggest that the CT receptor, one of the Class II family of GPCRs, can be regulated by GRKs and that CT receptor mRNA and protein expression are negatively influenced by GRK activity.
The GRK isoforms that regulate CT receptor signaling were not precisely defined in our studies, but the data suggest that GRK2 may play an important role for desensitization of the receptor. Recent studies reveal that various GPCRs can be regulated by multiple GRKs in vitro (e.g. Refs. 9, 10, and 27). The secretin receptor, another of the class II GPCRs, can be desensitized by expression of GRK5 as well as that of GRK2 or GRK3 (27). In the CHO cells that we used, in addition to GRK2, GRK6 and GRK5 are also expressed. Although we found that the K220R DN-GRK2 construct inhibited GRK2-mediated activity (Fig. 3), we cannot rule out an effect of the DN construct on activity of GRK5 or GRK6 as well. Indeed, based on structural similarities, to be described below, we believe this is quite likely.
Previous workers have employed several different techniques to modulate GRK expression (e.g. Refs. [1][2][3][4][5]. Many previous studies have involved overexpression of GRK isoforms. The more limited efforts designed to inhibit GRK expression have included use of a carboxyl-terminal fragment from GRK2 (termed ␤ARKct), antisense oligonucleotides, and studies in cells or tissues from mice with a knockout of a GRK isoform. We believe that use of a retrovirally expressed DN construct offers advantages relative to those other methods. Thus, for example, ␤ARKct is an inhibitor of G␤␥ and some of its actions in cells may be attributable to G␤␥-dependent, but GRK-independent, events (3,4). Moreover, because G␤␥ is not involved in the action of all GRK isoforms, studies with ␤RKct will not assess the role of G␤␥-independent isoforms. Antisense oligonucleotides, which are generally isoform-selective, are primarily useful in acute experiments and not for generation of stably inhibited cells. Material from knockout animals is limited thus far to murine tissues and generally only from heterozygotic animals that have loss of expression of a single type of GRK. We believe that the retrovirally engineered DN-GRK2 construct provides a useful complementary approach to those other methods because of its theoretical ability to block function of multiple GRK isoforms (the K220R in GRK2 represents a conserved region in the catalytic domain ATP-binding site of all GRK isoforms), its potential utility to generate stably ex-pressing cells, such as those we have used here, and the rather widespread tropism of the retroviral vector. Since CHO cells are often used for the heterologous expression of GPCRs, the stable cell line that we have developed may prove useful to examine GRK-mediated regulation of transfected GPCRs.
Our findings in the DN-GRK CHO cells strongly suggest that GRKs not only regulate receptor desensitization but also play a role in the steady-state level of receptor expression. Activity of GRK to regulate CT receptor desensitization presumably reflects phosphorylation of one or more of the 8 serine/threonine residues in the carboxyl-terminal portion of C1a receptors. Mutagenesis and related approaches will be necessary to define the precise sites that are regulated by the GRKs. Although such sites likely are involved in GRK-mediated desensitization of the receptors, it is not clear how receptor phosphorylation by GRK would regulate receptor expression. Presumably, GRKs impact on the CT receptor life cycle through events subsequent to receptor phosphorylation and internalization. The evidence that DN-GRK2 cells have an increase in CT receptor mRNA suggests that GRKs regulate the transcription and/or turnover of CT receptor mRNA. In this regard, it is of further interest that although the focus of the current studies is on CT receptor, ATP-mediated phosphoinositide hydrolysis, presumably via one or more P2Y receptors, was also enhanced in the DN cells (Fig. 8). Perhaps GRKs are able to influence expression of multiple types of GPCRs.
In summary, the current studies with CHO cells show that these cells express CT receptor C1a and GRK2, GRK5 and GRK6, but not GRK3 and GRK4. We found that stable expression of DN-GRK2 by retrovirally mediated gene transfer inhibited GRK2-promoted substrate phosphorylation, potentiated CT receptor signaling (both cAMP generation and phosphoinositide hydrolysis) in CHO cells, and blunted desensitization of CT receptors. Moreover, DN-GRK2-expressing CHO cells had an up-regulation in expression of CT receptors and receptor mRNA. The findings thus provide a mechanistic explanation for previous observations regarding agonist-mediated down-regulation of CT receptors, a phenomenon that has been implicated in resistance and escape from response to CT (25,26). In addition, the data indicate that GRKs are involved not only in desensitization of the CT receptor but also in the regulation of CT receptor expression. We speculate that through their effects on receptor phosphorylation, GRKs are able to inhibit expression of CT receptor mRNA and protein in CHO cells and perhaps more generally of GPCRs in other cell types as well. FIG. 10. Expression of CT receptor mRNA in CHO cells. RT-PCR was performed using 0.2 g of total RNA isolated from CHO cells (three clonal isolates each of Mock and DN cells) at 40 thermal cycles for CT receptor C1a (CTR C1a) and 0.1 g of total RNA at 30 thermal cycles for GAPDH. PCR primers were designed based on murine CT C1a receptor and rat GAPDH cDNAs. Upper panel, PCR products (equivalent to 500 ng of total RNA) were resolved on a 6% polyacrylamide gel, Lower panel, PCR products (equivalent to 2 g of total RNA) were resolved on a 3% NuSieve agarose gel. Rat brain RNA yielded bands of identical size (data not shown).