Internalization and Down-regulation of Human Muscarinic Acetylcholine Receptor m2 Subtypes ROLE OF THIRD INTRACELLULAR m2 LOOP AND G PROTEIN-COUPLED RECEPTOR KINASE 2*

,

Sequestration/internalization of ␤ 2 -adrenergic receptors seemed to be independent of phosphorylation by GRK2 on the basis of results with ␤ 2 -adrenergic receptor mutants lacking phosphorylation sites or GRK-specific inhibitors (16 -20). On the other hand, the agonist-induced sequestration of hm2 receptors expressed in HEK293 cells is hampered by deletion of the third intracellular loop (I3-loop) which includes the GRK2 phosphorylation sites (21,22). Moreover, 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 kinase activity (23). Recently, Ferguson et al. have reexamined the relationship between the phosphorylation by GRK2 and sequestration of ␤ 2 -adrenergic receptors, demonstrating that phosphorylation by GRK2 (24) or other GRKs (25) facilitates sequestration of ␤ 2 -adrenergic receptors. Phosphorylation facilitates ␤-arrestin binding to ␤ 2adrenergic receptors (26) and thereby appears to enhance sequestration, possibly interacting with clathrin (27), a major protein of coated pits. Pals-Rylaarsdam et al. (8,28) have provided results showing that the phosphorylation by GRK2 of m2 receptors is involved in their internalization as well as in their uncoupling from G proteins in HEK293 cells. These results suggest that the phosphorylation by GRK2 of m2 muscarinic and ␤ 2 -adrenergic receptors may be involved in both internalization and uncoupling through facilitation of their interaction with ␤-arrestin/arrestin 3.
No studies have been carried out on the relation between down-regulation and phosphorylation of G protein-coupled receptors, except that down-regulation of ␤ 2 -adrenergic receptors has been reported to be independent of their phosphorylation by GRK2 (16,18). It is also unclear whether the cellular pathway leading to down-regulation is distinct from that of internalization. If a portion of receptors in clathrin-coated vesicles translocates into lysosomes and is down-regulated, their phosphorylation with GRKs or the deletion of I3-loop should also affect down-regulation. However, if down-regulation occurs by a distinct pathway, receptor phosphorylation may not play a role. Alternatively, both phosphorylated and non-phosphorylated receptors may enter the clathrin-dependent internalization pathway, albeit at different rates. Finally, receptor phosphorylation could affect the rate of translocation between endosomes and lysosomes, or recycling to the cell surface.
Here, we provide evidence that down-regulation as well as internalization of hm2 receptors are facilitated by coexpression of GRK2. Moreover, deletion of I3-loop, which contains the GRK2 phosphorylation sites (22), suppressed rapid internalization and markedly reduced the rate of down-regulation.

Materials-[ 3 H]NMS (specific activity of 71.3 Ci/mmol) and [ 3 H]QNB
(specific activity of 36.4 Ci/mmol) were purchased from NEN Life Science Products; restriction enzymes were from Toyobo Corp. and Takara Shuzo Co., Ltd.; Cy3-conjugated goat anti-mouse IgG antibody was from Jackson Laboratories. cDNA of GRK2 was kindly donated by Dr. R. J. Lefkowitz, mammalian expression vector for hygromycinresistant gene (pSV-hygro) was from Dr. H. Okayama, and mammalian expression vector with neomycin-resistant gene (pEF-neo) and mammalian expression vector pEF-BOS were from Drs. S. Nagata and T. Shimizu. Hybridoma cells expressing 9E10 were obtained from the American Type Culture Collection; Chinese hamster ovary CHO-K1 cells were from the Japanese Cancer Research Resources Bank.
Construction of Stable Transfectant Expressing hm2 Receptors and GRK2-The construction of mammalian expression vectors for c-Myc epitope-tagged hm2 receptor (pEF-Myc-hm2) and GRK2 (pEF-GRK2) was described previously (23). CHO-K1 cells (5 ϫ 10 4 cells) were transfected with 18 g of expression vectors of pEF-Myc-hm2 and 2 g of pEF-neo by the calcium phosphate precipitation method (29). Stable transfectants were selected in the presence of 400 g/ml Geneticin (Life Technologies, Inc.) and were subcloned by limiting dilution. Expression of receptors was detected by [ 3 H]QNB binding. The [ 3 H]QNB binding sites in these cells were estimated to be 165 fmol/mg of protein in total homogenate. The transfectants were cultured in F-12 nutrient mixture (Ham's) (Life Technologies Inc.) supplemented with 10% fetal bovine serum (Cansera International Inc.), 40 units/ml penicillin G (Meiji Seika, Kaisha Ltd.), 40 mg/ml streptomycin sulfate (Meiji Seika, Kaisha Ltd.), and 100 g/ml Geneticin at 37°C in 95% air and 5% CO 2 . One of the CHO cell clones expressing hm2 receptors was transfected with 18 g of pEF-GRK2 and 2 g of pSV-hygro, and stable transfectants were selected in the presence of 300 g/ml hygromycin B (Boehringer Mannheim) and subcloned by limiting dilution. Expression of GRK2 was detected with use of Western blotting as described previously (23). The [ 3 H]QNB binding sites of these cells were estimated to be 330 fmol/mg of protein in total homogenate, and expressed amounts of GRK2 were estimated to be 300 -600 fmol/mg of protein in the supernatant by immunostaining with anti-GRK2 antibodies. The transfectants were cultured in F-12 nutrient mixture (Ham's) supplemented with 10% fetal bovine serum, 40 units/ml penicillin G, 40 mg/ml streptomycin sulfate, and 100 g/ml hygromycin B at 37°C in 95% air and 5% CO 2 . A mammalian expression vector for a m2 receptor mutant that lacks a central part of the third intracellular loop (I3-del m2 receptor) was constructed by inserting the NheI-XhoI fragment of the pSG5/ Hm2(d234 -381) (21) into the NheI/XhoI site of pEF-Myc-hm2. I3-del m2 receptors were stably expressed in CHO-K1 cells as described above, and the [ 3 H]QNB binding sites of these cells were estimated to be 260 fmol/mg of protein in total homogenate.
Sucrose Density Gradient Centrifugation Experiments-Sucrose density gradient centrifugation was carried out as described by Harden et al. (30). Semiconfluent CHO cells cultured in a 15-cm diameter dish were treated with 10 Ϫ5 M carbamylcholine for 20 min and then washed three times with 10 ml of ice-cold, phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 , pH 7.5). Washed cells were incubated with 10 ml of serum-free F-12 medium containing 50 g/ml concanavalin A for 20 min on ice, then washed with 10 ml of lysis buffer (1 mM Tris, 2 mM EDTA, pH 7.4), and hypotonically lysed by incubation in 10 ml of lysis buffer for 20 min on ice. After removing the lysis buffer, cells were collected in a small volume of lysis buffer with rubber policeman.  4 , and 1 mM NaH 2 PO 4 , pH 7.4; 0.5 ml/well) at 4°C for 4 h. After incubation, cells were washed three times with 1 ml ice-cold PBS/well. After washing, cells were dissolved in 0.3 ml of 1% Triton X-100 (w/v), mixed with 4.5 ml of Triton-toluene mixture containing 0.4% 2,5-diphenyloxazole and 0.01% 1,4-bis-2-(methyl-5-phenyloxazolyl)benzene, and the radioactivity measured. Quadruplicate samples were assayed for each point. In some experiments, cells were treated with carbamylcholine in the hypertonic medium containing 0.32 M sucrose besides normal constituents. Down-regulation in the hypertonic medium was examined for cells treated with carbamylcholine for 1-4 h, because the incubation for longer than 4 h in the hypertonic medium caused CHO cells to deteriorate.
Immunofluorescence Confocal Microscopy of hm2 Receptors-CHO cells expressing human c-Myc-tagged m2 receptors were grown overnight on plastic chamber slides (Nunc Inc.). Treatment with various concentrations of carbamylcholine was carried out at 37°C for 10 min. At the end of drug treatment, cells were washed twice with PBS, fixed for 10 min at room temperature with 3.7% paraformaldehyde in PBS, permeabilized in PBS containing 0.25% fish gelatin, 0.04% saponin, and 0.05% NaN 3 . After permeabilization, cells were labeled with anti-Myc monoclonal antibody (9E10) (31) for 1 h, washed four times with PBS, incubated with Cy3 (indocarbocyamine)-conjugated goat anti-mouse secondary antibody, and then washed four times with PBS and once with water. Slides were mounted using Fluoromount G (Fisher Scientific) containing a trace amount of phenylenediamine and stored at 4°C. Samples were visualized using laser scanning confocal microscopy with a krypton-argon laser coupled with a Bio-Rad MRC-600 confocal head attached to an Optiphot II Nikon microscope with a Plan Apo 60 ϫ 1.4 NA objective lens with 1.4 numeric aperture. Cy3 emission was detected with a yellow high sensitivity filter block.

Sequestration of hm2 Receptors as Assessed by Loss of [ 3 H]NMS Binding
Sites from the Cell Surface-CHO cells expressing hm2 receptors with or without GRK2 were treated with carbamylcholine for various times, and then [ 3 H]NMS binding activity of intact cells was measured. greater (21). It should be noted that the portion of sequestered m2 receptors was higher for CHO cells (80%) than for COS-7 cells (40%) or BHK-21 cells (20 -25%).
Many membrane proteins including G protein-coupled receptors have been shown to be internalized through coated vesicles (26,27,32,33), whereas some receptors including m2 muscarinic receptors have also been reported to be internalized via caveolae (34,35). To determine which process is involved in the sequestration of hm2 receptors expressed in CHO cells, we have examined the effect of hypertonic medium on the sequestration, because the hypertonic medium is known to inhibit the internalization through clathrin-coated vesicles but not the internalization through caveolae (33,34,36,37). Sequestration of hm2 receptors in the presence of 10 Ϫ4 M or lower concentrations of carbamylcholine was completely suppressed in the hypertonic medium containing 0.32 M sucrose (Fig. 1E), indicating that hm2 receptors are internalized through clathrincoated vesicles. The inhibition of sequestration by the hypertonic medium was observed whether GRK2 was coexpressed or not, excluding the possibility that the coexpression of GRK2 facilitated the internalization of hm2 receptors through the pathway different from the coated vesicle-mediated pathway.

Assessment of Internalization of hm2 Receptors by Sucrose Density Gradient Centrifugation and Confocal Microscopy-
Sequestration of muscarinic receptors as assessed by the loss of [ 3 H]NMS binding sites from the cell surface is generally thought to represent internalization of receptors in the form of endocytosed vesicles. We confirmed internalization of hm2 receptors expressed in CHO cells with two different methods: sucrose density gradient centrifugation and confocal microscopy. Sucrose density gradient centrifugation was carried out as described by Harden et al. (30). The carbamylcholine-treated cells were incubated with concanavalin A, hypotonically lysed, and then subjected to the centrifugation, which resulted in the separation of two fractions: a heavy membrane fraction containing cell surface membranes and a light fraction containing intracellular vesicles (endosomes). As shown in Fig. 2, the peak of [ 3 H]QNB binding sites shifted from the heavy to light fraction by treatment of cells with 10 Ϫ5 M carbamylcholine for 20 min. This result is consistent with the interpretation that the sequestered [ 3 H]NMS binding sites corresponding to approximately 50% of total hm2 receptors were transferred from cell membranes to light vesicle fractions.
We have also followed internalization using laser scanning confocal microscopy. CHO cells expressing Myc-tagged hm2 receptors alone or Myc-tagged hm2 receptors together with GRK2 were labeled with anti-Myc monoclonal antibody (9E10) as described previously by Tolbert and Lameh (32). In the absence of agonist, hm2 receptors can be observed only at the cell surface (Fig. 3, A and C). When the cells were treated with 10 Ϫ6 M carbamylcholine for 10 min, vesicles containing hm2 receptors were observed only in cells coexpressing GRK2 (Fig.  3D). In the cells expressing only hm2 receptors, no intracellular vesicles containing hm2 receptors were observed after agonist treatment (Fig. 3B).
These results provide evidence that the sequestration/internalization observed as the loss of [ 3 H]NMS binding sites and the transfer of [ 3 H]QNB binding sites represents the translocation of hm2 receptors from plasma membranes into cytoplasmic vesicles.  (Fig. 4B). In contrast, down-regulation was undetectable in the presence of 10 Ϫ6 M carbamylcholine without GRK2, whereas significant down-regulation occurred with GRK2 coexpression (Fig. 4A). Apparent EC 50 values of carba- mylcholine for the down-regulation of hm2 receptors after 16 h of treatment were estimated to be 0.7 and 6 M for cells with or without coexpression of GRK2 (Fig. 4D). These results provide the first evidence that the down-regulation of G protein-coupled receptors is facilitated by coexpression of GRK2 and suggest that phosphorylation by GRK2 of hm2 receptors is directly or indirectly linked to their down-regulation.

Down-regulation of hm2 Receptors as Assessed by the Decrease in [ 3 H]QNB Binding Sites-The down-regulation of hm2 receptors was assessed as the agonist-induced decrease in
Down-regulation of hm2 receptors, as well as their sequestration, was markedly inhibited in the hypertonic medium (Fig.  4E). When cells were treated with 10 Ϫ4 M carbamylcholine for 4 h, the proportions of down-regulated receptors were 17-18% in the hypertonic medium, in contrast with 39 -49% in the normal medium. The inhibition was observed irrespective of the coexpression of GRK2 or not. The finding that both sequestration and down-regulation were commonly inhibited in the hypertonic medium supports the idea that both sequestration and down-regulation involve the same event, e.g. the internalization through coated vesicles.
Sequestration and Down-regulation of I3-del m2 Receptors-We have stably expressed I3-del m2 receptors in CHO  (22), and I3-del m2 receptors are not phosphorylated by GRK2 (4). I3-del m2 receptors transiently expressed in HEK293 cells have been shown to sequester much less than hm2 receptors (21). Similarly, I3-del m2 receptors in CHO cells failed to sequester significantly upon treatment with carbamylcholine for 1 h (Fig.  5A). The [ 3 H]NMS binding sites were gradually decreased upon prolonged incubation with carbamylcholine, but the rate of loss was much lower for I3-del m2 receptors (t1 ⁄2 ϭ 8.4 h) than for wild type hm2 receptors (t1 ⁄2 ϭ 9.5 min) (Fig. 5C). Fig. 5B shows changes in [ 3 H]QNB binding sites after incubation of cells expressing wild type and I3-del m2 receptors for 16 h with different concentrations of carbamylcholine. Unexpectedly, appreciable loss of [ 3 H]QNB binding sites was observed even for I3-del m2 receptors, although the extent of loss was less for I3-del m2 receptors (44%) than for hm2 receptors (60%). The rate of loss was also much slower for I3-del m2 receptors than for hm2 receptors (t1 ⁄2 ϭ 9.9 versus 2.3 h) (Fig.  5C). These results indicate that the presence of the I3-loop is not required for agonist-induced down-regulation, although it may accelerate the rate of down-regulation. The loss of [ 3 H]QNB binding sites in the presence of 10 Ϫ4 M carbamylcholine occurred in parallel with the loss of [ 3 H]NMS binding sites for I3-del m2 receptors, (t1 ⁄2 ϭ 9.9 and 8.4 h for the loss of [ 3 H]QNB and [ 3 H]NMS binding sites, respectively), in a sharp contrast to the rates for hm2 wild type receptors (t1 ⁄2 ϭ 2.3 h and 9.5 min, respectively) (Fig. 5C). These results indicate that the I3-del m2 receptors are down-regulated as soon as they are lost from the cell surface and that no appreciable amounts of I3-del m2 receptors exist in an internalized form, whereas 40 -60% of hm2 receptors exist in an internalized form (Figs. 5C and 6). DISCUSSION In previous studies (23), we have shown that sequestration of m2 receptors transiently expressed in COS-7 and BHK-21 cells was facilitated by coexpression of GRK2, an effect of which was evident only at low concentrations of carbamylcholine. In the present study, a similar effect of coexpression of GRK2 was observed for the sequestration of hm2 receptors stably expressed in CHO cells. Furthermore, the sequestration assessed as the loss of [ 3 H]NMS binding sites from the cell surface was confirmed to represent the internalization of hm2 receptors from plasma membranes into cytoplasmic vesicles by analyses involving sucrose density gradient centrifugation of membrane fractions and confocal microscopy. The fact that a similar effect was observed in three different cell lines suggests that facilitation by GRK of the internalization of hm2 receptors is a general phenomenon independent of cell species. On the other hand, Pals-Rylaarsdam et al. (8) have argued against the involvement of GRK2 in the internalization of hm2 receptors, based on the finding that the level of sequestration was not affected by coexpression of GRK2 or a DN-GRK2 in a clone of HEK293. They measured the sequestration of hm2 receptors in cells treated with only a high concentration of carbamylcholine (1 mM), and therefore could have missed the effect of GRK2 coexpression. Very recently, these authors have shown that a hm2 receptor mutant with alanine residues in the place of serine/threonine residues in the GRK2 phosphorylation sites was sequestered by a lower extent compared with the wild type receptor, and concluded that sequestration of hm2 receptors was promoted by their phosphorylation (28). As for the effect of coexpression of DN-GRK2, we have also failed to detect any effect on the sequestration of m2 receptors in CHO cells and BHK-21 cells (23) In contrast to wild type hm2 receptors, I3-del m2 receptors (deletion 234 -381), which lack phosphorylation sites by GRK2, failed to internalize rapidly. The simplest interpretation for this finding is that phosphorylation by GRK2 of serine or threonine residues in the I3-loop is a necessary step for rapid internalization. We cannot exclude, however, the possibility that the I3-loop may have other functions. Pals-Rylaasdam et al. reported that a hm2 mutant with a deletion (252-327) in the I3-loop was not phosphorylated by GRK2; yet 50% of the mutant stably expressed in HEK293 cells were sequestered by treatment with 10 Ϫ3 M carbamylcholine for 2 h, although the sequestration of the mutant was less in its extent and slower in its rate as compared with the sequestration of wild type hm2 receptors. Possibly, internalization depends on phosphorylation-independent sites which were deleted from our mutant but not from the 252-327 deletion mutant. Ferguson et al. (24) have shown that overexpression of ␤-arrestin rescues sequestration of ␤ 2 -adrenergic receptor mutant lacking phosphorylation sites, and proposed that the interaction between ␤-arrestin and receptors is essential for internalization and that the internalization is facilitated by but does not require the phosphorylation by GRK2. Both phosphorylation sites and phosphorylation-independent sites in the I3-loop might be involved in the interaction with ␤-arrestin which accelerates internalization.
We have found in the present study that the coexpression of GRK2 facilitates the down-regulation of hm2 receptors by reducing the effective concentrations of carbamylcholine. As the effects of GRK2 coexpression on internalization and downregulation of hm2 receptors were similar to each other, it is tempting to speculate that both internalization and down-regulation involve the same event, e.g. the phosphorylation by GRKs of agonist-bound receptors. To our knowledge, a positive relationship between down-regulation and phosphorylation by GRK2 has not been reported for any G protein-coupled receptors. As for ␤ 2 -adrenergic receptors, receptor mutants lacking phosphorylation sites for GRK2 have been shown to downregulate normally (16,18). It should be noted, however, that these authors have not examined the effect of different concentrations of agonist, and therefore, the ability of GRK2 to reduce the effective concentration might not have been noticed.
When hm2 receptor-expressing cells were treated with 10 Ϫ4 M of carbamylcholine, hm2 receptors were rapidly internalized with t1 ⁄2 of 9.5 min and slowly down-regulated with t1 ⁄2 of 2.3 h. Thus, approximately 60% of receptors were down-regulated, 30% were in an internalized form, and 10% remained in the cell surface after a 16-h incubation (see Fig. 5C). In contrast, I3-del m2 receptors were lost from the cell surface and down-regulated with slower rates of t1 ⁄2 ϭ 8.4 and 9.9 h, respectively, so that approximately 60% of receptors were down-regulated, no appreciable receptors were detectable in an internalized form, and 40% remained in the cell surface after a prolonged incubation (see Fig. 5C). These results indicate that down-regulation may occur without the I3-loop. However, the I3-loop is necessary for rapid internalization and accumulation of internalized receptors.
In Fig. 6, we have presented a tentative schema for the relationship between internalization and down-regulation of hm2 receptors. We assume in this schema that agonist-bound receptors are rapidly internalized and that internalized receptors are slowly down-regulated. This schema explains the present findings that both rapid internalization and down-regulation in the presence of low concentrations of carbamylcholine are accelerated in parallel by coexpression of GRK2; this explanation is based on the assumption that the amounts of phosphorylated hm2 receptors are increased by coexpression of GRK2, the rate of internalization is limited by the concentration of phosphorylated receptors, and the rate of down-regulation is limited by concentrations of internalized receptors. The finding that both sequestration and down-regulation are inhibited in the hypertonic medium supports the scheme and suggests that the rapid internalization occurs via coated vesicles. We, however, cannot exclude the possibility that the downregulation occurs through multiple pathways. In contrast to hm2 receptors, the I3-del m2 receptors are lost from the cell surface and down-regulated with similar slow rates (see Fig.  5C), indicating that no appreciable amounts of receptors exist in an internalized form. It is possible that hm2 receptors may down-regulate via two independent pathways, the I3-looprequiring and I3-loop-independent pathways, which do and do not involve the rapid internalization, respectively. The I3-looprequiring and I3-loop-independent pathways may represent the coated vesicle-mediated and coated vesicle-independent pathways, respectively. This interpretation is consistent with the results that the internalization of hm2 receptors caused by 10 Ϫ4 M carbamylcholine was completely suppressed in the hypertonic medium but the down-regulation was only partly suppressed, and that the proportions of down-regulated hm2 receptors in the hypertonic medium were similar to those of down-regulated I3-del hm2 receptors in normal medium (compare Figs. 4E and 5C). Another interpretation is that downregulation of hm2 receptors occurs through a single step involving internalized vesicles and that the internalization step proceeds rapidly for hm2 receptors with intact I3-loop but greatly slows down and becomes the rate-limiting step for down-regulation for I3-del hm2 receptors. At present, the question remains open whether down-regulation of hm2 receptors occurs through a single route via internalized receptors or through multiple independent pathways.
In the present study, we have shown that both internalization and down-regulation of hm2 receptors are facilitated by coexpression of GRK2, and that the I3-loop is necessary for FIG. 6. Schema for a possible relationship between internalization and down-regulation of hm2 receptors. The rates of internalization (9.5 min and 8.4 h) and down-regulation (2.3 and 9.9 h), and proportions of cell surface receptors, internalized receptors, and downregulated receptors were estimated for cells treated with 10 Ϫ4 M carbamylcholine for 16 h or more (Fig. 5C). R, receptor. rapid internalization but not necessary for down-regulation, although the rate of down-regulation is reduced in its absence.