Introduction

Top Abstract Introduction Methods Results Discussion References

Many kinds of G protein-coupled receptors are known to be phosphorylated by GRKs in an agonist-dependent manner (for a review, see Haga et al., 1994; Premont et al., 1995, Böhm et al., 1997). Muscarinic receptors are also known to be phosphorylated by GRKs. Muscarinic acetylcholine receptor m2 subtypes (m2 receptors) are phosphorylated by GRK2 (beta adrenergic receptor kinase 1) (Kameyama et al., 1993; Richardson et al., 1993), GRK3 (Richardson et al., 1993), GRK5 (Kunapuli et al., 1994), GRK6 (Loudon and Benovic, 1994) and a muscarinic receptor kinase that is the same as or closely related to GRK2 (Haga and Haga, 1990, 1992). The number of phosphorylation sites is reported to be 10 or 11 for muscarinic receptor kinase (Nakata et al., 1994), 4 to 10 for GRK2 and GRK3 (Richardson et al., 1993), 1 to 1.5 for GRK5 (Kunapuli et al., 1994) and 0.6 for GRK6 (Loudon and Benovic, 1994). Muscarinic acetylcholine receptor m3 subtypes (m3 receptors) are also phosphorylated by GRK2 and GRK3 (four phosphorylation sites per receptor) but not by GRK5 and GRK6 (Debburman et al., 1995). Recently, m3 receptors have been reported to be phosphorylated in an agonist-dependent manner by the other kinase that is distinguishable from GRK2 or GRK3 (Tobin et al., 1996). Muscarinic acetylcholine receptor m1 subtypes (m1 receptors) had been reported to be not phosphorylated by GRK2 or endogenous kinase in Sf9 cells under the conditions in which m2 receptors were phosphorylated in an agonist-dependent manner by these kinases (Richardson and Hosey, 1992; Haga et al., 1993), but recently m1 receptors were found to be phosphorylated by GRK2 after removal of unknown factor or factors that copurify with m1 receptors (four or five phosphorylation sites) (Haga et al., 1996). No direct evidence is available for phosphorylation of muscarinic acetylcholine receptor m4 and m5 subtypes (m4 and m5 receptors), as far as we know.

The phosphorylation of G protein-coupled receptors by GRKs is generally thought to be involved in homologous desensitization of receptors. Desensitization of G protein-coupled receptors occurs with three phases: uncoupling from G proteins, sequestration/internalization and down-regulation of receptors. Previously, we have shown that the agonist-dependent phosphorylation and sequestration of m2 receptors expressed in COS-7 cells are facilitated by coexpression of GRK2 and attenuated by coexpression of a dominant-negative mutant of GRK2 (DN-GRK2) that lacks a kinase activity (Tsuga et al., 1994). When we found the relation between sequestration of m2 receptors and their phosphorylation by GRK2, various lines of evidence indicated that phosphorylation by GRK2 and sequestration of beta adrenergic receptors are independent phenomena (Strader et al., 1987; Bouvier et al., 1988; Lohse et al., 1990; Kong et al., 1994). Therefore, we suggested previously that the phosphorylation by GRK2 might have different consequences depending on the species of receptors.

Recently, Ferguson et al. (1995) have shown that overexpression of GRK2 facilitates the sequestration of a beta adrenergic receptor mutant and suggested that phosphorylation by GRKs may play a broader role in agonist-promoted G protein-coupled receptor sequestration than envisaged above. Sequestration of the beta adrenergic receptor mutant was also shown to be facilitated by overexpression of -arrestin (Ferguson et al., 1996), and furthermore, -arrestin/arrestin3 was found to interact with clathrin (Goodman et al., 1996). These results suggest that the sequestration of the beta adrenergic receptor mutant may occur through the following series of events: agonist-dependent phosphorylation of the receptors, binding of -arrestin to the phosphorylated receptors and binding of clathrin to -arrestin. Very recently, Pals-Rylaarsdam and Hosey (1997) also showed that phosphorylation of hm2 receptors facilitated its sequestration as well as desensitization. It is an open question of whether this scheme is also applicable to other G protein-coupled receptors. For G_q-coupled receptors, the relationship between desensitization and phosphorylation by GRKs of alpha-1B adrenergic receptors has been reported (Diviani et al., 1996), but the relation between sequestration and phosphorylation by GRKs has not been reported to our best knowledge. The present article is concerned with the question of the generality of the relation between phosphorylation and sequestration of G protein-coupled receptors, and we specifically attempted to determine whether the sequestration of G_q-coupled receptors is facilitated by phosphorylation with GRKs.

The results of this study provide evidence that coexpression of GRK2 facilitates the sequestration of m2, m3, m4 and m5 receptors and that GRK4 and GRK5 may also facilitate sequestration of m2 receptors.

	Methods

Top Abstract Introduction Methods Results Discussion References

Materials. [³H]NMS (specific activity, 80.4 Ci/mmol) was purchased from DuPont-New England Nuclear (Boston, MA). Mammalian expression vector pSVL was from Pharmacia Biotech (Uppsala, Sweden). Restriction enzymes were from Toyobo (Osaka, Japan) and Takara Shuzo (Kusatsu, Japan). cDNA of GRK2 was kindly donated by Dr. R. J. Lefkowitz (Duke University, Durham, NC). Expression vectors of GRK5 (pCMV5-GRK5) and GRK6 (pCMV5-GRK6) were provided by Drs. J. L. Benovic (Thomas Jefferson University, Philadelphia, PA) and H. Kurose (University of Tokyo, Tokyo, Japan). cDNA of GRK4 was from Dr. R. T. Premont (Duke University, Durham, NC). cDNA of hm2 receptors was provided by Dr. W. Sadée (University of California San Francisco, San Francisco, CA). cDNA of hm1 (Hm1pCD), hm4 (Hm4pCD) and hm5 (Hm5pCDp2) receptors was provided by Dr. T. I. Bonner (National Institutes of Health, Bethesda, MD). cDNA of hm3 receptors was from Dr. J. S. Gutkind (National Institutes of Health, Bethesda, MD). Mammalian expression vector pEF-BOS was from Dr. S. Nagata (Osaka Bioscience Institute, Osaka, Japan).

Cell culture. Syrian hamster kidney BHK-21(C-13) cells were obtained from the Health Science Research Resources Bank (Tokyo, Japan). African green monkey kidney COS-7 cells were from Dr. T. Shimizu (University of Tokyo, Tokyo, Japan). Both cells were cultured in Dulbecco's modified Eagle's medium (Nissui Pharmaceutical, Tokyo, Japan) supplemented with 10% fetal bovine serum (Cansera International, Rexdale, Ontario, Canada), 40 units/ml penicillin G (Meiji Seika Kaisha, Tokyo, Japan) and 40 mg/ml streptomycin sulfate (Meiji Seika Kaisha) at 37°C in 95% air/5% CO₂.

Construction of mammalian expression vectors. Mammalian expression vectors for hm1hm5 receptors were constructed as follows. An NsiI/BamHI fragment of cDNA of hm1 receptors was inserted in the XbaI site of pSVL after the conversion of the end of both fragments to blunt ends (pSVL-hm1). An XbaI fragment of Hm2/pSG5 (Moro et al., 1993) was inserted in the XbaI site of pSVL (pSVL-hm2). A BamHI fragment of pEF-hm3 (Inoue et al., 1995) was inserted in the XbaI site of pSVL after the conversion of the end of both fragments to blunt ends (pSVL-hm3). A BamHI/BglII fragment of cDNA of hm4 receptor was inserted into BamHI site of pBluescript SK⁺ (pBluescript-hm4). The 363-bp SpeI/EcoRI fragment that includes hm4 cDNA between positions 1 and 363 was generated with the polymerase chain reaction using the primers 5-GTGAATTCCATGGCCAACTTCACACCTGT-3 and 5-GTGAGCAACGCCTCCGTCATGAATTCAG-3 and Hm4pCD as the template. An SpeI/PstI fragment of the generated fragment and PstI/EcoRV (1.4 kbp) fragment of pBluescript-hm4 was inserted into SpeI/EcoRV site of pBluescript-Myc-7 (Tsuga et al., 1994). Subsequently an SpeI/EcoRI fragment of this plasmid was inserted in the XbaI site of pSVL after the conversion of the end of both fragments to blunt ends (pSVL-hm4). An EcoRII/PstI fragment (1.7 kbp) of cDNA of hm5 receptor was inserted in the XbaI site of pSVL after the conversion of the end of both fragments to blunt ends (pSVL-hm5). The construction of mammalian expression vectors for hm2 (pEF-hm2) and hm3 receptors (pEF-hm3) was described previously (Tsuga et al., 1994; Inoue et al., 1995).

Transfection of mammalian expression vectors and [³H]NMS binding assays. COS-7 and BHK-21 cells were transfected with use of the calcium phosphate method as described previously (Tsuga et al., 1994). COS-7 cells were transfected with 5 µg of expression vectors for receptors (pSVL-hm1 to pSVL-hm5) and 5 µg of expression vectors for GRKs (pEF-GRK2, pEF-GRK2-K220W, pEF-GRK4, pCMV5-GRK5, pCMV5-GRK6) per 10⁷ cells on a 10-cm-diameter dish. BHK-21 cells were transfected with 5 µg of pEF-hm2 or pEF-hm3 and 5 µg of expression vectors for GRKs per 10⁷ cells on 10-cm-diameter dish. In control cells, 5 µg of pEF-BOS was added instead of expression vectors of GRKs. At 3 to 5 hr after the transfection, cells were replated onto 12-well (COS-7) or 6-well (BHK-21) culture dishes. At 40 to 48 hr after transfection, various concentrations of carbamylcholine were added to culture media. The [³H]NMS binding activity of intact cells was measured as described previously (Tsuga et al., 1994). After incubation with carbamylcholine for various times, cells were washed with 1 ml of ice-cold PBS (137 mM NaCl, 2.7 mM KCl, 8.1 mM Na₂HPO₄, 1.5 mM KH₂PO₄, pH 7.5) per well for three times and incubated with 1.2 to 1.6 nM [³H]NMS in HEPES-buffered saline (25 mM HEPES, 113 mM NaCl, 6 mM glucose, 3 mM CaCl₂, 3 mM KCl, 2 mM MgSO₄ and 1 mM Na₂HPO₄, pH 7.4; 0.5 ml for COS-7, 1 ml for BHK-21 per well) at 4°C for 4 hr. After incubation, cells were washed with 1 ml of ice-cold PBS per well three times. After washing, cells were dissolved in 0.3 ml of 1% Triton X-100 (w/v) and mixed with 4.5 ml of Triton-Toluene cocktail containing 0.4% 2,5-diphenyloxazole and 0.01% 1,4-bis-2-(methyl-5-phenyloxazolyl)-benzene, and then radioactivity was measured. Triplicate (BHK-21) or quadruplicate (COS-7) samples were assayed for each point. Data were analyzed with one-way analysis of variance followed by Dunnett's test.

Detection of GRKs by immunostaining. Immunoreactivity of GRKs was detected as follows. At 40 to 48 hr after transfection, cells on two 10-cm-diameter dishes were homogenized with 1 ml of ice-cold HEPES buffer (20 mM HEPES-KOH, pH 8.0, 250 mM NaCl, 5 mM EDTA, 2 mM EGTA, 1 mM phenylmethylsulfonyl fluoride), and supernatant fractions were obtained by centrifugation at 100,000 × g. Supernatant fractions of the cells were subjected to SDS-PAGE followed by electroblotting onto Immobilon transfer membranes (Millipore, Bedford, MA). The membranes were subsequently incubated with a blocking buffer [5% skim milk and 0.1% Tween-20 (w/v) in PBS] for 1 hr at room temperature and then with 1:2000 diluted antisera for 2 hr at room temperature. The anti-GRK2/3 antibodies raised against a fusion protein containing a GRK3 peptide (569-688) linked to the carboxyterminus of GST and anti-GRK5 antibodies raised against a GST-fusion protein containing GRK5 peptide (471-590) were gifts from Dr. H. Kurose. The anti-GRK6 antibodies raised against a GRK6 peptide (525-544) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). After incubation with antisera, the membranes were washed five times with blocking buffer and incubated with the secondary antiserum (goat anti-rabbit IgG coupled to horseradish peroxidase purchased from Jackson Immunoresearch Laboratories, West Grove, PA) in the blocking buffer for 1 hr at room temperature. After removal of the secondary antiserum, the membranes were washed five times with PBS and then incubated with a staining solution [0.8 mg/ml 3,3-diaminobenzidine · 4 HCl (Dojindo Laboratories, Mashiki-Machi, Kumamoto, Japan), 0.4 mg/ml NiCl₂, 0.0025% H₂O₂, 50 mM Tris · HCl, pH 7.5]. The amounts of GRK2 were roughly estimated by visual comparison of the immunostained bands of samples and purified GRK2 obtained from Sf9 cells as described previously (Tsuga et al., 1994).

RNA isolation and amplification of GRK cDNA fragment. Total cellular RNA was isolated from 10⁶ COS-7 or BHK-21 cells using GlassMAX RNA Microisolation Spin Cartridge System (Life Technologies) and then treated with RNase-free DNase I (supplied with the Spin Cartridge System) for 15 min at room temperature. GRK cDNA segments were amplified from 1 µg of COS-7 or BHK-21 total RNA using RT-PCR high (Toyobo) and the specific primers for cDNA of GRK2 (5-ATCTCGAGCGCGGCTTGTTCTTCATCTT-3 and 5-ATTCACGAGGTGGCAGAGACTGTCTTTGA-3, complementary to residues 1558-1580 and 1987-2010 of human GRK2 cDNA, respectively), GRK4 (5-ATTCTAGAGATCCCAGAAGGRSAGA-3 and 5-ATGAATTCAACTTCTCAGAAYACTCCTC-3, complementary to residues 1005-1024 and 1219-1238 of human GRK4 cDNA, respectively), GRK5 (5-GCTCGCTAGCTGCTTCCRGTGGAG-3 and 5-ATGGATCCACTGTGAARGGYGTCA-3, complementary to residues 1450-1468 and 1755-1773 of human GRK5 cDNA, respectively) and GRK6 (5-AGCTGCAGGYRGTGCCCGYGAGGTRAA-3 and 5-AGGAATTCTRCAGTTYCCACAGCAAT-3, complementary to residues 1309-1328 and 1679-1697 of human GRK6 cDNA, respectively). After amplification, 20 µl (one fifth of total) of reaction mixture was subjected to agarose gel electrophoresis. After electrophoresis, the gel was stained with ethidium bromide to detect the amplified cDNA fragments.

	Results

Top Abstract Introduction Methods Results Discussion References

Expression of muscarinic receptors and GRKs in COS-7 and BHK-21 cells. Human m1m5 (hm1hm5) receptors were transiently expressed in COS-7 cells alone or with GRK2, GRK5, GRK6 or DN-GRK2. Hm2 and hm3 receptors were also expressed in BHK-21 cells alone or with GRK2, GRK5 or GRK6. Expression levels of muscarinic receptors on COS-7 cells as assessed as [³H]NMS binding sites were 150 to 300 (hm1), 200 to 500 (hm2), 500 to 800 (hm3), 350 to 550 (hm4) and 200 to 500 (hm5) fmol/mg of homogenate protein. Expression levels of hm2 and hm3 receptors on BHK-21 cells were 200 to 300 and 200 to 500 fmol/mg of homogenate protein, respectively. The endogenous expression of muscarinic receptors was not detected in COS-7 and BHK-21 cells.

View larger version (42K): [in this window] [in a new window]   Fig. 1.   Expression of GRKs was determined by RT-PCR. cDNA fragments of GRKs were amplified from 1 µg of COS-7 or BHK-21 total RNA using specific primers for GRK2 (a), GRK4 (b), GRK5 (c) and GRK6 (d). After amplification, 20 µl (one fifth of total) of the reaction mixtures were subjected to agarose gel electrophoresis. Lanes 1 to 3 of each panel correspond to COS-7 cells, BHK-21 cells and GRK-expressing COS-7 cells, respectively. The expected length of the amplified cDNA fragments is 465, 249, 336 and 405 bp for GRK2, GRK4, GRK5 and GRK6, respectively.

Sequestration of hm2 and hm4 receptors in COS-7 cells. Sequestration of hm2 receptors was estimated by measuring the loss of the [³H]NMS binding activity from the surface of cells that had been treated with carbamylcholine for various times (fig. 2a). The rate of sequestration of hm2 receptors was increased by coexpression of GRK2 and decreased by coexpression of DN-GRK2: t_1/2 values for hm2 receptors treated with 10⁵ M carbamylcholine were estimated to be 37, 14 and 46 min for control and GRK2- and DN-GRK2-expressing cells, respectively.

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Sequestration of [3H]NMS binding sites of COS-7 cells expressing hm2 and hm4 receptors. a and b, COS-7 cells expressing hm2 alone, hm2 plus GRK2 or hm2 plus DN-GRK2. c and d, COS-7 cells expressing hm4 alone, hm4 plus GRK2 or hm4 plus DN-GRK2. Cells were incubated with 105 M (a) and 104 M (c) carbamylcholine for the indicated times or with indicated concentrations of carbamylcholine for 120 min and then subjected to [3H]NMS binding assays at 4°C for 4 hr. Results are shown as mean ± S.D. from three to seven independent experiments. Time course curves were fitted to the following equation: B × exp(0.69t/t1/2) + (100  B). The half-time (t1/2) for hm2 receptors was 37, 14 or 46 min for cells expressing hm2 receptors alone or with GRK2 or DN-GRK2, respectively. The half-time for hm4 receptors was 49, 26 or 103 min for cells expressing hm4 receptors alone or with GRK2 or DN-GRK2, respectively. Significant differences compared with control cells were *P < .05 or **P < .01.

7

6

5

5

5

4

3

View this table: [in this window] [in a new window]   TABLE 1 Sequestration of hm2 and hm4 receptors coexpressed with GRKs in COS-7 cells COS-7 cells expressing hm2 or hm4 receptors with or without GRKs were incubated with varying concentrations of carbamylcholine for 120 min. [3H]NMS binding was assayed as described in the text. Proportions of sequestered receptors were determined as decrease of [3H]NMS binding activity on cells treated with 104 M carbamylcholine. Results are given as mean ± standard deviation from three to seven independent experiments. The number of experiments is indicated in parentheses. Dose-response curves for averaged values from three to seven independent experiments were fitted to the following equation: Rmax × EC50/(EC50 + [carbamylcholine]) + (100  Rmax). EC50 values were determined as parameters.

4

7

3

Sequestration of hm3, hm5 and hm1 receptors in COS-7 cells. Sequestration of hm1, hm3 and hm5 receptors expressed in COS-7 cells is summarized in figure 3 and table 2. The rates of sequestration of hm3 and hm5 receptors were not appreciably affected by coexpression of GRK2 or DN-GRK2, whereas the proportion of sequestered receptors increased in GRK2-expressing cells but not in DN-GRK2-expressing cells (fig. 3, a and c). Proportions of sequestered hm3 and hm5 receptors in control cells were 20% to 25% (fig. 3, a and c). Proportions of sequestered receptors in GRK2-expressing cells were much greater for hm5 (51% at maximum) than for hm3 (34% at maximum) (fig. 3, a and c, table 2). Coexpression of DN-GRK2 did not affect the sequestration of hm3 and hm5 receptors, indicating that the sequestration of hm3 and hm5 receptors is not facilitated by the endogenous kinase. The rate of sequestration of hm1 receptors (t_1/2 = 43 min in the presence of 10³ M carbamylcholine) was lower compared with hm3 and hm5 receptors (t_1/2 = 11 and 12 min, respectively), and the proportion of sequestered hm1 receptors (10 ± 5%) was also much lower compared with other muscarinic receptors (fig. 3, e and f). The rate of sequestration and proportion of sequestered hm1 receptors were increased by coexpression of GRK2. Unexpectedly, both the rate and the proportion were also increased by coexpression of DN-GRK2. Differences between cells expressing GRK2 and DN-GRK2 were not statistically significant.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Sequestration of [3H]NMS binding sites on COS-7 cells expressing hm1, hm3 and hm5 receptors. a and b, COS-7 cells expressing hm3 alone, hm3 plus GRK2 or hm3 plus DN-GRK2. c and d, COS-7 cells expressing hm5 alone, hm5 plus GRK2 or hm5 plus DN-GRK2. e and f, COS-7 cells expressing hm1 alone, hm1 plus GRK2 or hm1 plus DN-GRK2. Cells were incubated with 103 M (a, c and e) carbamylcholine for the indicated times or indicated concentrations of carbamylcholine for 120 min (b, d and f) and then subjected to [3H]NMS binding assays at 4°C for 4 hr. Time course curves were fitted to the following equation: B × exp(0.69t/t1/2) + (100  B). The half-time for hm3, hm5 and hm1 receptors was 29, 17 or 43 min for cells expressing receptor alone; 24, 17 or 11 min for cells coexpressing GRK2; and 15, 16 or 12 min for cells expressing DN-GRK2, respectively. Results are shown as mean ± S.D. from three to seven independent experiments. Significant differences compared with control cells were *P < .05 or **P < .01.

View this table: [in this window] [in a new window]   TABLE 2 Sequestration of hm1, hm3 and hm5 receptors coexpressed with GRKs in COS-7 cells COS-7 cells expressing hm1, hm3 or hm5 receptors with or without GRKs were incubated with varying concentrations of carbamylcholine for 120 min. Proportions of sequestered receptors were determined as decrease of [3H]NMS binding activity on cells treated with 104 M carbamylcholine. Results are shown as mean ± standard deviation from three to seven independent experiments. The number of experiments is indicated in parentheses.

Effects of coexpression of GRK4, GRK5 and GRK6 on sequestration of muscarinic receptors in COS-7 cells. Tables 1 and 2 summarize the effect of coexpression of GRK4, GRK5 and GRK6 as well as GRK2 and DN-GRK2 on sequestration of hm1hm5 receptors expressed in COS-7 cells. The sequestration was not affected by coexpression of GRK4, GRK5 and GRK6 except that the sequestration of hm1 and hm3 receptors was significantly decreased by coexpression of GRK5 and GRK4, respectively, and the sequestration of hm2 receptors tended to be facilitated by coexpression of GRK4, GRK5 and GRK6. The reason is not known why sequestration of hm1 and hm3 was decreased by coexpression of GRK5 and GRK4, respectively.

Effects of coexpression of GRKs on sequestration of muscarinic receptors in BHK-21 cells. In previous experiments, we observed that proportions of sequestered m2 receptors were much less for BHK-21 cells compared with for COS-7 cells and that coexpression of DN-GRK2 did not attenuate the sequestration of m2 receptors in BHK-21 cells. These results indicate that BHK-21 cells do not have the endogenous kinase present in COS-7 cells and facilitate the sequestration of m2 receptors. Actually, GRK4 and possibly GRK6 were found to be expressed in COS-7 but not in BHK-21 cells (fig. 1, ad). There is a possibility that the coexpression of GRK4, GRK5 or GRK6 in COS-7 cells does not affect the sequestration of muscarinic receptors because the effect of exogenous kinases overlaps the effect of GRK4, GRK6 or other endogenous kinase in COS-7 cells. Thus, we expressed GRK2, GRK4, GRK5 or GRK6 in BHK-21 cells and examined their effects on sequestration of hm2 and hm3 receptors (fig. 4). Sequestration of hm2 and hm3 receptors expressed in BHK-21 cells was not affected by coexpression of GRK6. On the other hand, coexpression of GRK4 and GRK5 apparently facilitated sequestration of hm2 receptors: the proportions of sequestered hm2 receptors were estimated to be 32% to 34% in cells coexpressing GRK4 or GRK5 in contrast to 25% in control cells, although EC₅₀ values for carbamylcholine did not change (fig. 4, b and c, and table 3). The effect of coexpression of GRK4 and GRK5, however, was smaller than that of coexpression GRK2; 40% of hm2 receptors were sequestered, and EC₅₀ values for carbamylcholine decreased by a factor of 16 with coexpression of GRK2. In contrast to hm2 receptors, the sequestration of hm3 receptors was not affected by coexpression of GRK4 and GRK5 (fig. 4e).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Sequestration of [3H]NMS binding sites on BHK-21 cells expressing hm2 and hm3 receptors. BHK-21 cells expressing hm2 receptors with or without GRK2 (a), GRK4 (b), GRK5 (c) or GRK6 (d) and BHK-21 cells expressing hm3 receptors with or without GRK2, GRK4, GRK5 or GRK6 (e) were incubated with indicated concentrations of carbamylcholine for 120 min and then subjected to [3H]NMS binding assays at 4°C for 4 hr. Results are shown as mean ± S.D. from three independent experiments. Significant differences compared with control cells were *P < .05 or **P < .01.

View this table: [in this window] [in a new window]   TABLE 3 Sequestration of hm2 and hm3 receptors coexpressed with GRKs in BHK-21 cells BHK-21 cells expressing hm2 or hm3 receptors with or without GRKs were incubated with varying concentrations of carbamylcholine for 120 min. Proportions of sequestered receptors were determined as decrease of [3H]NMS binding activity on cells treated with 104 M carbamylcholine. Results are shown as mean ± standard deviation from three independent experiments. Dose-response curves for averaged value from three independent experiments were fitted to the following equation: Rmax × EC50/(EC50 + [carbamylcholine]) + (100  Rmax). EC50 values were determined as parameters.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Previously, we have shown that the sequestration of porcine m2 receptors expressed in COS-7 cells is facilitated by coexpression of GRK2 (Tsuga et al., 1994). In the present study, we extended this observation and showed that sequestration of hm2, hm3, hm4 and hm5 receptors is also facilitated by coexpression of GRK2. Because the facilitation is not observed by coexpression of DN-GRK2 lacking catalytic activity, it is reasonable to assume that the sequestration is facilitated by phosphorylation of these muscarinic receptors with GRK2. In fact, m2 receptors have been shown to be phosphorylated in vivo (Tsuga et al., 1994; Pals-Rylaarsdam et al., 1995) and in vitro (Richardson et al., 1993; Kameyama et al., 1993; Nakata et al., 1994) by GRK2, and m3 receptors have also been shown to be phosphorylated in vitro by GRK2 (Debburman et al., 1995).

Phosphorylation sites in m2 receptors have been located in the central part of the third intracellular loop (Nakata et al., 1994). Displacement by alanine of serine and threonine residues in the putative phosphorylation sites in hm2 and hm3 receptors has been shown to impair the sequestration of mutated receptors (Moro et al., 1993). These results are consistent with the idea that phosphorylation by GRK2 of serine/threonine residues in the third intracellular loop of m2 and m3 receptors somehow facilitates their sequestration. No direct evidence is available for phosphorylation of m4 and m5 receptors by GRK2. It is likely, however, that m4 and m5 receptors are also phosphorylated by GRK2 at the third intracellular loops, and thereby their sequestration is facilitated because the third intracellular loops of m4 and m5 receptors contain serine and threonine residues, which are flanked by acidic amino acid residues and thought to be phosphorylation sites of GRK2.

Pals-Rylaarsdam et al. (1995) reported that the phosphorylation of hm2 receptors by GRK2 is related to receptor desensitization but not to sequestration, based on findings that the level of sequestration of hm2 receptors expressed in HEK293 cells was not affected by coexpression of GRK2 or DN-GRK2 and that the agonist-dependent sequestration was observed for m2 receptor mutants without GRK-phosphorylation sites. It should be noted, however, that coexpression of GRK2 facilitates the sequestration of hm2 receptors mainly by reducing the effective concentrations of carbamylcholine rather than by increasing the proportions of sequestered receptors, and hence the effect of coexpression of GRK2 is difficult to observe with 1 mM carbamylcholine only, at which concentration the sequestration was observed by Pals-Rylaarsdam et al.

Furthermore, it should be pointed out that the sequestration of m2 and other muscarinic receptors may involve both GRK-dependent and -independent components. The presence of significant sequestration of hm2hm5 receptors in cells coexpressing DN-GRK2 is consistent with this hypothesis. The difference in the sequestration of hm1 and hm2 receptors in cells coexpressing DN-GRK2 may be due to the difference in the GRK-independent sequestration, as in the study by Goldman et al. (1996). Zhang et al. (1996) reported that the sequestration of G protein-coupled receptors may occur by at least two distinct pathways: one mediated by dynamin and -arrestin, and the other not mediated by these factors.

In contrast with hm2hm5 receptors, it is not clear whether there is a relation between phosphorylation by GRK2 and sequestration of hm1 receptors because sequestration of hm1 receptors was apparently facilitated by coexpression of either GRK2 or DN-GRK2. Recently, hm1 receptors were shown to be phosphorylated in vitro by GRK2 (Haga et al., 1996), and phosphorylation sites were located in the central part of the third intracellular loop and include serine and threonine residues, which were shown to be critical for sequestration of m1 receptors (Moro et al., 1993). Agonist-dependent phosphorylation of hm1 receptors by GRK2 could be observed after the removal of unknown factor or factors that copurify with hm1 receptors (Haga et al., 1996). We can speculate that the phosphorylation by endogenous kinase of hm1 receptors and their sequestration are suppressed by a putative inhibitory factor and that the inhibition is relieved by interaction with either GRK2 or DN-GRK2. Another possible explanation is that sequestration of hm1 receptors is facilitated by forming a complex with either GRK2 or DN-GRK2. Both GRK2 and DN-GRK2 are expected to interact with beta gamma subunits. It is interesting to know if whether beta gamma subunits are involved in the internalization of m1 receptors. Regardless of the mechanism, the sequestration of hm1 receptors is different from that of other muscarinic receptors.

The finding that sequestration of hm2 and hm4 receptors is attenuated by coexpression of DN-GRK2 suggests that there is at least an endogenous kinase that phosphorylates hm2 and hm4 receptors and facilitates their sequestration. In fact, hm2 receptors were phosphorylated by the endogenous kinase in an agonist-dependent manner (Tsuga et al., 1994). The relevant kinase is not likely to be GRK2 because (1) GRK2 could not be detected with immunostaining or RT-PCR and (2) the sequestration of hm3 and hm5 receptors was facilitated by expression of GRK2 but not attenuated by DN-GRK2. GRK5 is also not likely to be the relevant kinase because GRK5 also could not be detected by immunostaining or RT-PCR. On the other hand, we could detect expression of GRK4 in COS-7 cells, but not in BHK-21 cells, with RT-PCR. Coexpression of GRK4 facilitated the sequestration of hm2 receptors expressed in BHK-21 cells but only slightly facilitated the sequestration of hm2 receptors in COS-7 cells. In contrast, the sequestration of hm3 receptors, which were not affected by DN-GRK2, was not facilitated by coexpression of GRK4 in either COS-7 or BHK-21 cells. GRK6 was also detected in COS-7 cells with RT-PCR. However, coexpression of GRK6 in BHK-21 cells did not facilitate sequestration of hm2 receptors, suggesting that as previously reported, hm2 receptor may not be a good substrate for GRK6 (Loudon and Benovic, 1994). These results are consistent with the idea that GRK4 is an endogenous kinase in COS-7 cells that phosphorylates hm2 and hm4 receptors and facilitates their sequestration, although we cannot exclude the possibility that there are additional endogenous kinases, including GRK2.

Sequestration of hm2, but not hm3, receptors expressed in BHK-21 cells was found to be facilitated by coexpression of GRK5 but not by GRK6. These results have some correlation with the extent of in vitro phosphorylation; the number of phosphorylation sites in hm2 receptors is estimated to be 4 to 10 for GRK2 (Richardson et al., 1993; Nakata et al., 1994), 1.5 for GRK5 (Kunapuli et al., 1994) and 0.5 for GRK6 (Loudon and Benovic, 1994), and those for hm3 receptor are estimated to be 2 to 4 for GRK2 and none for GRK5 or GRK6 (Debburman et al., 1995). GRK5-phosphorylation sites in hm2 receptors have not been determined, but they are likely to reside in the third intracellular loop, including many serine and threonine residues, because substrate specificity of GRK2 and GRK5 appears to be very similar (Fredericks et al., 1996).

Recently, an agonist-dependent sequestration of a beta adrenergic receptor mutant expressed in HEK293 cells was shown to be facilitated by overexpression of GRK2 (Ferguson et al., 1995), GRK3 or GRK5 (Ménard et al., 1996). Although the sequestration of wild-type beta adrenergic receptors has been reported to not be linked with their phosphorylation by GRKs, the recent report indicated that the mutation of GRK-phosphorylation sites partially attenuated their agonist-dependent sequestration. The sequestration of beta adrenergic receptors is likely to be composed of GRK-dependent and -independent components. Furthermore, dopamine D₂ receptors expressed in COS-7 cells were found to become sequestered in an agonist-dependent manner when GRK2 or GRK5 was coexpressed, whereas virtually no sequestration was observed in the absence of coexpression of GRKs (Ito and Haga, 1996). The sequestration of dopamine D₂ receptors in this system appears to be composed of only GRK-dependent process. These results, together with the present results, indicate that (1) the agonist-dependent sequestration of G protein-coupled receptors may be composed of GRK-dependent and -independent components, (2) the relevant GRK species may be different depending on receptor species and (3) GRK2 is involved in the sequestration of receptors linked to either G_s (beta adrenergic), G_i/G_o (muscarinic m2, m4 and dopamine D₂) or G_q (muscarinic m3 and m5).

Recently, it was proposed that the phosphorylation by GRK2 facilitates the sequestration of phosphorylated receptors by increasing their affinity for -arrestin, and the expression level of GRKs and -arrestin determines kinetics of sequestration of beta adrenergic receptors (Ferguson et al., 1996; Ménard et al., 1997). Furthermore, Goodman et al. (1996) demonstrated that -arrestin/arrestin3 interacts with clathrin, a major protein of coated vesicles. These results suggest that the phosphorylation by GRK2 of m2 muscarinic and beta adrenergic receptors may be involved in both internalization and uncoupling through facilitation of their interaction with -arrestin/arrestin3. In fact, -arrestin has been shown to interact with the phosphorylated form of m2 receptors (Gurevich et al., 1993). Very recently, Schlador and Nathanson (1997) showed that desensitization and sequestration of m2 receptors were synergistically enhanced by GRK2 and -arrestin-1. This result suggests that the sequestration of muscarinic as well as beta adrenergic receptors is also dependent on the interaction of -arrestin with phosphorylated receptors.

We propose that the relation between the agonist-dependent phosphorylation and facilitation of receptor sequestration is not limited to m2 receptors but is generalized to include at least four muscarinic receptors, beta adrenergic receptors and dopamine D₂ receptors, and that GRK4 and GRK5 in addition to GRK2 may be involved in this relation.