The association of arrestin-3 with the human lutropin/choriogonadotropin receptor depends mostly on receptor activation rather than on receptor phosphorylation.

Although the involvement of the nonvisual arrestins in the agonist-induced internalization of the human lutropin receptor (hLHR) has been documented previously with the use of dominant-negative mutants, a physical association of the nonvisual arrestins with the hLHR in intact cells has not been established. In the studies presented herein, we used a cross-linking/coimmunoprecipitation/immunoblotting approach as well as confocal microscopy to document the association of the hLHR with the nonvisual arrestins in co-transfected 293 cells. We also used this approach to examine the relative importance of receptor activation and receptor phosphorylation in the formation of this complex. Using hLHR mutants that impair phosphorylation, activation, or both, we show that the formation of the hLHR-nonvisual arrestin complex depends mostly on the agonist-induced activation of the hLHR rather than on the phosphorylation of the hLHR. These results stand in contrast to those obtained with several other G protein-coupled receptors (i.e. the beta2-adrenergic receptor, the m2 muscarinic receptor, rhodopsin, and the type 1A angiotensin receptor) where arrestin binding depends mostly on receptor phosphorylation rather than on receptor activation. We have also examined the association of the nonvisual arrestins with naturally occurring gain-of-function mutations of the hLHR found in boys with Leydig cell hyperplasia or Leydig cell adenomas. Our results show that these mutants associate with the nonvisual arrestins in an agonist-independent fashion.

Internalization of G protein-coupled receptors (GPCRs) 1 is a ubiquitous response that follows agonist activation (reviewed in Refs. [1][2][3][4]. Although GPCRs can be internalized by several distinct pathways (3), the most common and best understood pathway is facilitated by the G protein-coupled receptor kinase (GRK)-catalyzed phosphorylation of GPCRs and the subsequent formation of a complex between the agonist-activated and phosphorylated GPCRs and a family of proteins known as the nonvisual arrestins or ␤arrestins. The nonvisual arrestins (arrestin-2 also known as ␤-arrestin-1 and arrestin-3, also known as ␤-arrestin-2) bind with high affinity to clathrin and to adaptor protein-2 (5,6) and thus target the activated and phosphorylated GPCRs to clathrin coated pits. Once localized to clathrin-coated pits, the GPCRs are internalized by a process that requires the participation of dynamin, a GTPase involved in the fission of clathrin-coated pits (7).
Studies from this and other laboratories have shown that the binding of agonist to the rat, mouse, porcine, or human lutropin receptor (LHR) triggers the internalization of the agonist-receptor complex via clathrin-coated pits by a pathway that can be inhibited with dominant-negative mutants of the nonvisual arrestins and a dominant-negative mutant of dynamin (8 -12). Although the large number of studies on the ␤ 2 -adrenergic receptor emphasize the importance of GPCR phosphorylation on the process of internalization (1)(2)(3)(4), our recent mutagenesis studies with the human (h) LHR suggest that the agonistinduced internalization of this receptor is much more dependent on receptor activation than on receptor phosphorylation (12,13). This conclusion was based on the following findings. First, an activation-competent but phosphorylation-impaired mutant of the hLHR (constructed by mutation of the phosphorylation sites) lengthens the t1 ⁄2 of internalization of the hLHR less than 2-fold (12). Second, two mutations of the second extracellular loop (S512A and V519A) that cause a slight impairment in receptor activation (i.e. a 2-3-fold rightward shift in the EC 50 for cAMP accumulation) lengthen the t1 ⁄2 of internalization 5-7-fold without affecting receptor phosphorylation (13). Third, two other mutants of the hLHR (D405N and Y546F in transmembrane helices 2 and 5, respectively) that cause a more severe impairment in receptor activation (i.e. a 10 -50fold rightward shift in the EC 50 for cAMP accumulation) become resistant to agonist-induced phosphorylation but still lengthen the t1 ⁄2 of internalization 5-7-fold (12). Interestingly, the phosphorylation of the D405N and Y546F mutants and their long t1 ⁄2 of internalization can be rescued by overexpression of GRK2 but only if the phosphorylation sites are intact.
Because dominant-negative mutants of the nonvisual arrestins inhibit the agonist-induced internalization of the hLHR, we hypothesized that the different manipulations described above affected internalization by affecting the association of the hLHR with the nonvisual arrestins. The studies presented here describe a method that can be used to measure the formation of the nonvisual arrestin-hLHR complex and use this method to formally test the hypothesis proposed above.

MATERIALS AND METHODS
Plasmids and Cells-The preparation and properties of a hLHR-wt expression vector modified with the Myc epitope at the N terminus has been described (12). All hLHR mutants used here were prepared using this vector as template. The preparation, signaling properties, and internalization properties of all these mutants have also been described (12,13). Expression vectors for arrestin-2, arrestin-3, a ␤ 2 -adrenergic receptor tagged with the HA epitope at the N terminus (in pcDNA3.1), and arrestin-3 tagged with GFP at the C terminus (in pEGFP-N1) were kindly provided by Dr. Jeff Benovic (Thomas Jefferson University, Philadelphia, PA). The nonvisual arrestins were modified (using PCR strategies) with the FLAG epitope at the N terminus and subcloned into pcDNA3.1 for expression. Co-transfection experiments revealed that the FLAG-tagged nonvisual arrestin constructs were as effective as their wild-type counterparts in enhancing the internalization of the hLHR-wt. A full-length clone encoding for bovine GRK2 (14) was also kindly provided by Dr. Jeff Benovic (Thomas Jefferson University, Philadelphia, PA), and it was subcloned into pcDNA3.1 for expression. HA-tagged dynamin K44A (15) was kindly provided by Dr. Sandra Schmid (Scripps Research Institute, La Jolla, CA). An expression vector (pcDNA1.1) encoding for a prenylated, kinase-deficient, mutant of GRK2, designated C 20 -GRK2-K220M, (16), was generously provided by Dr. Marc Caron (Duke University, Durham, NC).
293T cells were maintained in Dulbecco's modified Eagle's medium containing 10 mM Hepes, 10% newborn calf serum, and 50 g/ml gentamycin, pH 7.4. Transient transfections were performed using the calcium phosphate method of Chen and Okayama (17). Cells were plated in gelatin-coated plasticware and transfected when 70 -80% confluent with the amounts of plasmid DNA indicated in the figure legends. After an overnight incubation, the cells were washed and used 24 h later.
Measurement of the Association of the hLHR with the Nonvisual Arrestins in Co-transfected Cells-Cells were plated and transfected in 100-mm dishes (in 10 ml of medium) that had been coated with gelatin. At the end of the transfection, the cells were washed, trypsinized, plated in gelatin-coated 100-mm dishes and in gelatin-coated 35-mm wells, and used 24 h later.
The cells plated in 35-mm wells were used to assess 125 I-hCG binding and internalization. The binding of 125 I-hCG was measured by incubating the transfected cells with a saturating concentration (ϳ25 nM) of 125 I-hCG for 1 h at room temperature. These results were used to ensure that the cells were expressing equivalent densities of cell surface receptors and to equalize the amount of receptor to be immunoprecipitated as described below. The internalization of 125 I-hCG was measured during a 50-min incubation of the cells with a concentration of hCG equivalent to the K d (ϳ 2 nM) as described elsewhere (11,18,19). Rates of internalization can be calculated by plotting the internalization index (defined as the ratio of surface to internalized hormone) as a function of time (11,18,19). In all arrestin-3 binding experiments reported herein (using cells co-transfected with dynamin-K44A), the internalization indices for hCG in cells co-transfected with the hLHR-wt or mutants thereof were 0.05-0.1 when measured during a 50-min incubation. This low internalization index measured at this long incubation time is consistent with half-times of internalization in the 150 -200-min range (11,18,19). These results were simply used to confirm that all cells co-transfected with dynamin-K44A were internalizing 125 I-hCG at a slow and comparable rate regardless of the receptor expressed or the amount of arrestin-3 transfected.
The cells plated in the 100-mm dishes were used to measure arrestin-3 binding as follows. The monolayers were washed three times with warm 0.15 M NaCl, 20 mM Hepes, pH 7.4 (Buffer A). After addition of 3 ml of the same buffer, each dish received a 60-l aliquot of vehicle or hCG to give a final concentration of ϳ26 nM (the K d for hCG binding is ϳ2 nM; see Ref. 11) and the cells were incubated at 37°C for the times indicated. At the end of this incubation, each dish received 150 l of a freshly prepared 25 mM solution of dithiobis(succinimidylpropionate) (DSP) in Me 2 SO. The cross-linking reaction was allowed to proceed for 30 min at room temperature while the dishes were rocked (20). After cross-linking the cells were placed on ice, the buffer was aspirated, and the monolayers were washed once with ice-cold Dulbecco's-modified Eagle's medium containing 10 mM Hepes, 10% newborn calf serum, pH 7.4, and then incubated in 5 ml of the same medium for 5 min on ice. Finally, the monolayers were washed two more times with cold 0.15 M NaCl, 20 mM Hepes, pH 7.4, and the cells were lysed during a 30-min incubation on ice with 1 ml of lysis buffer (1% Nonidet P-40, 4 mg/ml dodecyl-␤-D-maltoside, 0.8 mg/ml cholesteryl hemisuccinate in buffer A) supplemented with 40 g/ml Complete ® , EDTA-free protease inhibitor mixture (Roche Molecular Biochemicals). The lysate was clarified by centrifugation, and aliquots (ϳ500 l) containing the same amount of receptor (calculated from the binding experiments done in parallel as described above) were immunoprecipitated with a monoclonal antibody to the Myc epitope (9E10) that had been preabsorbed to agarose-conjugated protein G for 90 -120 min at 4°C as described previously (12). After extensive washing the cross-linked complexes were reduced and eluted by vigorous mixing of the beads in SDS sample buffer with reducing agents for 15 min at room temperature. The eluted material was then resolved on SDS gels and electrophoretically transferred to polyvinylidene difluoride membranes as described elsewhere (21).
The co-immunoprecipitated FLAG-tagged nonvisual arrestins were visualized in the blots during a 1-h incubation with an anti-FLAG M2 monoclonal antibody covalently coupled to horseradish peroxidase used at a final dilution of 1:500. The presence of equal amounts of receptor in the immunoprecipitates was confirmed by developing the blots with an anti-Myc (9E10) monoclonal antibody covalently coupled to horseradish peroxidase used at a final dilution of 1:1000 (this is documented only in Fig. 3). In experiments that utilized GRK2 co-transfections, the transfected GRK2 was visualized with a commercially available primary antibody (C-15 from Santa Cruz Biotechnology) as described previously (22). In some experiments (cf. Fig. 2), the endogenous and/or transfected arrestin-3 were visualized using a 1:2000 dilution of a rabbit polyclonal antibody raised against a glutathione S-transferase fusion protein containing residues 350 -409 of bovine arrestin-3 (23). In all of these cases, the appropriate secondary antibodies covalently coupled to horseradish peroxidase were used at a 1:5000 dilution. All immune complexes were ultimately visualized and quantitated using the Super Signal West Femto Maximum Sensitivity system of detection from Pierce and a Kodak digital imaging system. This image capture system is set up to alert us when image saturation occurs and to prevent us from measuring the intensity of such images.
Measurement of the Association of the ␤ 2 -Adrenergic Receptor (␤ 2 AR) with Arrestin-3 in Co-transfected Cells-293T cells were plated in 100-mm dishes and cotransfected as described above with 10 g of HA-tagged ␤ 2 AR, 2.5 g of dynamin-K44A, 0.25 g of FLAG-arrestin-3, and 5 g of empty vector or 5 g of C 20 -GRK2-K220M. The transiently transfected cells were washed with the buffers described above and incubated with or without 10 M isoproterenol for 5 min at 37°C. The cells were then cross-linked and lysed as described above, except that radioimmune precipitation buffer (1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 150 mM NaCl, 5 mM EDTA, 50 mM Tris-Cl, pH 8.0) with protease inhibitors was used. The lysates were clarified by centrifugation, and aliquots (ϳ500 l) containing the same amount of protein were immunoprecipitated with a monoclonal antibody to the HA epitope (3F10, 1.5 g/sample) that had been preabsorbed to agaroseconjugated protein G for 90 -120 min at 4°C as described above. The rest of this procedure was done exactly as described above for the hLHR. The co-immunoprecipitated FLAG-arrestin-3 and the HAtagged ␤ 2 AR were visualized in the blots during a 1-h incubation with a horseradish peroxidase-conjugated anti-FLAG M2 monoclonal antibody (final dilution ϭ 1:500) or a horseradish peroxidase-conjugated anti-HA (3F10) monoclonal antibody (final concentration ϭ 25 milliunits/ml). Expression of the transfected C 20 -GRK2-K220M was monitored using a monoclonal antibody to GRK2 (3A10 from Jeff Benovic), followed by a secondary antibody coupled to horseradish peroxidase. All immune complexes were ultimately visualized and quantitated using the Super Signal West Femto Maximum Sensitivity system of detection from Pierce and a Kodak digital imaging system. This image capture system is set up to alert us when image saturation occurs and to prevent us from measuring the intensity of such images.
Confocal Microscopy-Cells were plated in eight-chamber coverslip culture vessels coated with polylysine (BioCoat from Becton-Dickinson). They were co-transfected (in a total volume of 400 l) with 400 ng of Myc-hLHR, 100 ng of dynamin-K44A, and 8 ng of arrestin-3-GFP using the methods described above. Two days after transfection, the cells were incubated with or without hCG for 30 min at 37°C as described above for the arrestin-3 binding assays. The medium was removed, and the cells were washed twice with phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 1.4 mM NaH 2 PO 4 , 4.3 mM Na 2 HPO 4 , pH 7.4) and fixed during a 30-min incubation at room temperature with 4% paraformaldehyde (dissolved in phosphate-buffered saline). The fixed cells were washed twice again and then incubated for 1 h at room temperature with phosphate-buffered saline containing 50 mg/ml bovine serum albumin. This solution was removed, and the cells were incubated for another h at room temperature with a 1:100 dilution of the anti-Myc monoclonal antibody (9E10) dissolved in phosphate-buffered saline containing 5 mg/ml bovine serum albumin. After washing three times with phosphate-buffered saline, the cells were incubated for another 1 h at room temperature with a 1:2000 dilution of rhodaminelabeled anti-mouse IgG. Finally they were washed three or four times with phosphate-buffered saline, dried, and mounted in Vectashield mounting medium (Vector Laboratories). The rhodamine-labeled hLHR and the arrestin-3-GFP were visualized with a Bio-Rad confocal microscope at the Central Microscopy Facility of the University of Iowa.
Hormones and Supplies-Human kidney 293T cells are a derivative of 293 cells that express the SV40T antigen (24) and were provided to us by Dr. Marlene Hosey (Northwestern University, Chicago, IL). The 9E10 hybridoma cell line was obtained from the American Type Culture Collection. Purified hCG (CR-127, ϳ13,000 IU/mg) was kindly provided by Dr. A. Parlow and the National Hormone and Pituitary Agency (NIDDK, National Institutes of Health, Bethesda, MD) and purified recombinant hCG 2 was provided by Ares Serono (Randolph, MA). 125 I-hCG was prepared as described elsewhere (25). Partially purified hCG (ϳ3,000 IU/mg) was purchased from Sigma, and it was used only for the determination of nonspecific binding (see above). 125 I-cAMP and cell culture medium were obtained from the Iodination Core and the Media and Cell Production Core, respectively, of the Diabetes and Endocrinology Research Center of the University of Iowa. Concentrated supernatant from the 9E10 cells was prepared by the Hybridoma Facility of the Cancer Center of the University of Iowa. The 9E10 and anti-Flag M2 monoclonal antibodies coupled to horseradish peroxidase were purchased from Roche Molecular Biochemicals and Sigma, respectively. The 3F10 monoclonal antibody to the HA epitope (native or coupled to horseradish peroxidase) were from Roche. Secondary antibodies coupled to horseradish peroxidase or rhodamine were from Bio-Rad and Sigma, respectively. Other cell culture supplies and reagents were obtained from Corning and Invitrogen, respectively. All other chemicals were obtained from commonly used suppliers.

Characterization of an Assay Designed to Measure the Association of the Nonvisual Arrestins with the hLHR at the Cell
Surface-The association of several GPCR binding partners such as GRKs and nonvisual arrestins in intact cells has been difficult to assess using standard immunoprecipitation/immunoblotting approaches but can be readily determined if the complexes are cross-linked prior to immunoprecipitation (20, 26 -28). We encountered the same difficulties in experiments attempting to ascertain the association of the hLHR with the nonvisual arrestins in co-transfected 293T cells, and the only way to reliably detect such complexes was to treat the cells with DSP, a cell-permeable, homobifunctional, cleavable crosslinking agent, prior to immunoprecipitation. Thus, the basic method used below to detect the association of the hLHR with the nonvisual arrestins consisted of co-transfecting 293T cells with a Myc-tagged hLHR (12) and one of the FLAG-tagged nonvisual arrestins. Following incubation with or without agonist, the cells were further incubated with DSP (20) prior to solubilization. The cell lysates were then immunoprecipitated with a monoclonal antibody to the Myc epitope (9E10), and the cross-linked complexes were dissociated by incubation of the immunoprecipitates with SDS sample buffer containing reducing agents. The reduced complexes were then resolved on SDS gels, blotted, and probed with a horseradish peroxidase-conjugated monoclonal antibody to the FLAG epitope (M2) to detect the FLAG-tagged nonvisual arrestins co-immunoprecipitated with the hLHR. Fig. 1 shows that the association of the Myc-hLHR-wt with FLAG-arrestin-3 is minimally detectable in unstimulated cells. When exposed to hCG, however, there is a quick and robust association of these two components. A steady state level of association was attained during a 20 -30-min incubation with a saturating concentration of hCG. When the same experiment was done in cells co-transfected with dynamin-K44A (a condition that blocks the agonist-induced internalization of the hLHR; see Refs. [11][12][13], the time course of formation of the complex was somewhat slower, but, at steady state, the magnitude of the agonist-induced increase in the amount of complex formed was higher. In several experiments similar to that shown in Fig. 1, the amount of FLAG-arrestin-3⅐Myc-hLHR-wt complex formed (quantitated as described under "Materials and Methods") during 30 min with hCG increased 4 -10-fold and 20 -50-fold in cells co-transfected without or with dynamin-K44A, respectively. Because the aim of the experiments presented here was to compare the association of the nonvisual arrestins with mutants of the hLHR that are internalized at different rates, we wanted to ensure that changes in the formation of the receptor-arrestin-3 complexes were not affected by the different rates of internalization of the mutants in question. Thus, all subsequent experiments were done in cells that were co-transfected with dynamin-K44A, because the ligandinduced internalization of any of the receptor mutants tested herein was virtually abolished during the 30-min time period used to measure arrestin association (see "Materials and Methods"). Fig. 2 shows a representative experiment in which the association of the Myc-hLHR-wt with FLAG-arrestin-3 was measured in cells co-transfected with varying amounts of FLAG-␤arrestin-3. This experiment shows that the formation of this complex is agonist-dependent and that its levels increase as a function of the levels of FLAG-arrestin-3 expressed (compare panels A and B). The minimal amount of FLAG-arrestin-3 transfected that resulted in detectable formation of the complex was 0.03 g of plasmid/100-mm dish (Fig. 2B). To gain a better understanding of the extent of arrestin-3 overexpression in the transfected cells, we performed experiments in which an antibody to arrestin-3 (23) was used to compare the levels of total arrestin-3 (i.e. endogenous ϩ transfected) present in lysates of untransfected cells and of cells transfected with the lowest amount of FLAG-arrestin-3 shown in Fig. 2A. The results of these experiments (Fig. 2C) showed that the total amount of arrestin-3 expressed in these cells was 4.6 Ϯ 0.3-fold higher (mean Ϯ S.E. of six determinations) than the amount of endogenous arrestin-3 present in untransfected cells.
Because the amount of Myc-hLHR⅐FLAG-arrestin-3 complex did not appear to saturate at the highest levels of plasmid used in Fig. 2, we performed additional experiments in which the formation of this complex was measured in hCG-stimulated cells that had been co-transfected with a constant amount of Myc-hLHR-wt, dynamin-K44A, and a wider range of FLAGarrestin-3. The results of a representative experiment are shown in Fig. 3, and the quantitative analysis of several experiments is shown in Fig. 4. These results clearly show that the agonist-dependent binding of arrestin-3 to the hLHR is saturable. Half-maximal saturation is attained when the cells are transfected with 1 Ϯ 0.4 g of FLAG-arrestin-3/100-mm dish (Fig. 4, n ϭ 3), and the binding saturates when the cells are transfected with 5 g of FLAG-arrestin-3/100-mm dish. In parallel experiments (done using cells that were co-transfected without dynamin-K44A), we showed that a similar range of FLAG-arrestin-3 enhanced the internalization of hCG in a saturable fashion (Fig. 4). The enhanced internalization of hCG saturated in cells transfected with 1 g of FLAG-arrestin-3/ 100-mm dish and the half-maximal effect was attained in cells transfected with 0.07 Ϯ 0.02 g of FLAG-arrestin-3/100-mm dish (Fig. 4, n ϭ 3). We did not seek an explanation for the lower amount of arrestin-3 needed for half-maximal internalization (when compared with that needed for receptor binding), but this finding is likely to be a reflection of the need for arrestin-3 to bind to at least three different proteins (i.e. the LHR, clathrin, and adaptor protein-2) during the process of internalization. Thus, whereas the internalization assay measures a composite of three different K d values, (i.e. the binding of arrestin-3 to the hLHR as well as the binding of the arrestin-3-hLHR complex to clathrin and to adaptor protein-2), the binding assay measures a single K d (i.e. that of the arrestin-3-hLHR interaction).
Association of Arrestin-3 with Mutants of the hLHR-We have previously characterized the effects of a number of hLHR mutations (naturally occurring and laboratory-designed) on the agonist-induced phosphorylation and internalization of the hLHR (12,13). In the experiments described below, we measured the ability of these mutants to associate with FLAGarrestin-3. A representative experiment is shown in Fig. 5, and quantitative data obtained from several experiments are shown in Fig. 6.
The extent of basal and agonist-induced association of FLAG-␤-arrestin-3 with Myc-hLHR-5S/A (an activation-competent but phosphorylation-deficient mutant of the hLHR produced by the simultaneous mutation of 5 serine residues present in the C-terminal tail to Ala; see Ref. 12) is comparable with the extent of association of FLAG-arrestin-3 with Myc-hLHRwt. The t1 ⁄2 of internalization of agonist mediated by the hLHR-5S/A mutant (ϳ30 min) is also comparable with that of the hLHR-wt (ϳ20 min) (12). Three naturally occurring constitutively active mutants of the hLHR (L457R, D578Y, and D578H) display a constitutive (i.e. agonist-independent) level of FLAGarrestin-3 association that is comparable with, or higher than, that detected in agonist-stimulated cells expressing the hLHRwt. The association of FLAG-arrestin-3 with these mutants was not further increased (D578Y or D578H), and in one case (L457R) it was actually decreased by agonist stimulation (L457R). These three constitutively active mutants internalize 125 I-hCG with a shorter t1 ⁄2 than hLHR-wt (i.e. 3-7 min; see Ref. 12). Two laboratory-designed mutants that are activation-and phosphorylation-impaired (D405N and Y546F, see Ref. 12) display a ϳ90% reduction in the agonist-dependent association with FLAG-arrestin-3. These two mutants also internalize hCG with very long t1 ⁄2 (120 -150 min; see Ref. 12). Two additional laboratory designed mutants that display a less pronounced impairment in agonist-induced activation and are phosphorylated normally in response to agonist stimulation (S512A and V519A; see Ref. 13) also display a reduction in the agonist-dependent association with FLAG-␤-arrestin-2 that correlates with a very long half-time (120 -160 min) of internalization (13). The amount of FLAG-arrestin-3 associated with the S512 and V519A mutants after agonist activation is reduced to ϳ5 and ϳ40% of control, respectively (Fig. 6). Finally, we also tested the binding of FLAG-arrestin-3 to two additional laboratory-designed mutants (F515A and T521A) that display enhanced sensitivity to agonist stimulation (i.e. the EC 50 for the hCG-induced increase in cAMP accumulation is slightly lower than that for hLHR-wt). The agonist-induced phosphorylation of these mutants is normal, but they internalize hCG with slightly shorter half-times than hLHR-wt (10 -12 min; see Ref. 13). As summarized in Fig. 6, the agonist-provoked binding of FLAG-arrestin-3 to these two mutants was comparable with that detected with hLHR-wt (F515A) or reduced by ϳ20% (T521A). Additionally, note that changes in arrestin-3 binding described above cannot be explained by differences in the expression of the different hLHR mutants at the cell surface because the density of cell surface receptors in 293T cells transiently transfected with any of these mutants is comparable with that of cells expressing hLHR-wt (12,13). This is also documented below by confocal microscopy with several of the mutants used (cf. Fig. 7).
The association of arrestin-3 with the hLHR-wt, the activation-competent but phosphorylation-negative mutant (5S/A), a constitutively active mutant (D578H), and a signaling-and phosphorylation-impaired mutant (D405N) was also monitored by confocal microscopy in cells co-transfected with an arrestin-3-GFP construct and the hLHR-wt or the appropriate Myctagged LHR mutants (Fig. 7). In cells co-transfected with arrestin-3-GFP and the hLHR-wt or hLHR-5S/A, the receptor is mostly localized at the plasma membrane, whereas the arrestin-3-GFP is diffusely distributed throughout the cytoplasm. Addition of hCG to these cells results in the recruitment of arrestin-3-GFP to the cell surface where it co-localizes with the receptor. In cells expressing hLHR-D405N, the receptor and arrestin-3 are also localized to the plasma membrane and cytosol, respectively, but there is no redistribution of arrestin to the plasma membrane when these cells are incubated with hCG. Finally, arrestin-3 and the D578H mutant are co-localized at the plasma membrane in cells incubated with or without hCG. These results are in complete agreement with the more quantitative analysis summarized in Figs. 5 and 6.
We have shown previously that co-transfection with GRK2 rescues the impairment in hCG-induced phosphorylation and internalization displayed by the D405N and Y546F mutants but only if their phosphorylation sites are intact (12). To deter- The internalization of 125 I-hCG (white squares) was measured in parallel experiments in cells co-transected using the same conditions described in the legend to Fig. 3 but without dynamin-K44A. The amount of surface-bound and internalized 125 I-hCG were determined during a 5-min incubation of the transfected cells with 2 nM 125 I-hCG as described elsewhere (11,18,19) and under "Materials and Methods" and plotted as a function of the amount of FLAG-arrestin-3 transfected. Each point shows the average Ϯ S.E. of three independent transfections. The line shown is the best fit of the data points calculated by nonlinear regression using the Prism software package (GraphPad Software). mine whether these changes in internalization were associated with changes in the ability of these mutants to bind the nonvisual arrestins, we performed FLAG-arrestin-3 association assays under the same conditions. Representative blots of these experiments are shown in Fig. 8, and a summary of the quantitation of several experiments is shown in Fig. 9. The results presented show that GRK2 co-transfections do not enhance the agonist-provoked association of FLAG-arrestin-3 with the hLHR-wt. More importantly, these experiments show that GRK2 co-transfection rescues the impairment in hCGstimulated FLAG-arrestin-3 association displayed by the D405N and Y546F mutants, but only if the phosphorylation sites are intact.
Association of Arrestin-3 with the ␤ 2 AR-The results presented above show that the binding of the nonvisual arrestins to the hLHR is dependent mostly on receptor activation rather than on receptor phosphorylation. This is an unexpected finding, in view of previous reports showing that the in vitro association of nonvisual arrestins to other GPCRs (such as the ␤ 2 AR and the m2 muscarinic receptor) is dictated mostly by receptor phosphorylation rather than by receptor activation (29 -33). Because the binding of the nonvisual arrestins to these two receptors has been measured in vitro, using a reconstituted system, it is important to show that the methodology used here can reproduce the results previously obtained with one of these two receptors in vitro. This issue was addressed by using the co-transfection/cross-linking/coimmunoprecipitation approach described above to measure the agonist-dependent association of FLAG-arrestin-3 with the HA-tagged ␤ 2 AR in co-transfected cells. To test the importance of agonist-induced ␤ 2 AR phosphorylation in this process, this experiment was also done in cells co-transfected with C 20 -GRK2-K220M, a dominant-negative inhibitor of GRK2, that was previously shown to inhibit the agonist-promoted phosphorylation and internalization of the ␤ 2 AR in 293 cells (16). The results of these experiments (Fig. 10) clearly show that we can detect an agonist-dependent association of arrestin-3 with the ␤ 2 AR and that, as expected from the results of others (see above), this association can be blocked by interfering with the agonist-induced phosphorylation of the ␤ 2 AR. DISCUSSION Although the involvement of the nonvisual arrestins on the agonist-induced internalization of the hLHR has been documented by the use of dominant negative mutants, (11,12,18), the association of the nonvisual arrestins with the hLHR in intact cells has, until now, not been studied in detail. The only documentation of a physical association of the nonvisual arrestins with the LHR was provided by the finding that a synthetic peptide corresponding to the third intracellular loop of the porcine LHR can bind recombinant arrestin-2 in vitro (35). The studies presented herein describe a method that can be used to study the association of the hLHR with the nonvisual arrestins in intact cells and utilize this method to ascertain the relative importance of receptor activation and phosphorylation in the formation of the hLHR-nonvisual arrestin complex.
The method described here (Figs. 1-4) relies on the use of co-transfected cells to measure the association of FLAG-arrestin-3 and Myc-hLHR-wt (or mutants thereof), but this method should be readily applicable to measuring the interaction of the nonvisual arrestins with other GPCRs as documented here with the ␤ 2 AR (Fig. 10) and by others with the angiotensin type 1A receptor (34). The reliance of this method on co-transfection strategies to measure the formation of this complex is similar to that of other in vivo approaches that also utilize co-transfected cells and a combination of immunological and microscopic procedures to co-localize epitope-tagged GPCRs and fluorescently tagged nonvisual arrestins (Fig. 7 and Refs. 36 -41). This microscopy approach can be readily used to accurately track the location of the GPCR-nonvisual arrestin complex among different subcellular compartments, but the formation of the complex is difficult to quantitate (38,39). Our assay is done under conditions where the trafficking of the GPCR-nonvisual arrestin complex is prevented (by co-transfection of the cells with dynamin-K44A), and it has the advantage of the ease of quantitation of the amount of complex formed using conventional image analysis. The use of co-transfected cells also allows for measurements of the association of FLAG-tagged nonvisual arrestins with GPCRs over a wide range of concentrations of FLAG-tagged nonvisual arrestins that result in an enhancement of GPCR internalization (Figs. 3 and 4). Because an anti-FLAG monoclonal antibody that is already covalently linked to horseradish peroxidase is commercially available, the use of the FLAG-tagged arrestins also allows for a single step detection of the co-immunoprecipitated arrestins that results in single prominent bands (Fig. 1) that can be readily quantitated using conventional image analysis software. Finally, the results of the cotransfection/cross-linking/coimmunoprecipitation approach described here agree well with those obtained when the association of the hLHR and arrestin-3 is ascertained by confocal microscopy (Figs. 5-7).
The method validated here was initially set up because of our interest in ascertaining the relative importance of hLHR activation and phosphorylation in the formation of this hLHRnonvisual arrestins arrestin complex. This question was raised because the hCG-induced internalization of the hLHR is dependent mostly on receptor activation rather than phosphorylation (12,13). Because this process can be inhibited with dominant-negative mutants of the nonvisual arrestins (11,12,18), we speculated that the formation of the hLHR-nonvisual arrestin complex would be more dependent on hLHR activation than on phosphorylation. A relative lack of importance of receptor phosphorylation on the binding of arrestin-3 to the hLHR is readily illustrated here by the behavior of three different types of mutants (Figs. 5-7). First, a variant of the hLHR with mutated phosphorylation sites (designated hLHR-5S/A) that responds properly to agonist-activation but displays a ϳ90% reduction in agonist-induced phosphorylation (12) displays levels of basal and agonist-dependent arrestin-3 binding that are 293T cells (plated in eight-chamber coverslip culture vessels) were co-transfected (in a total volume of 400 l) with 400 ng of Myc-hLHR or mutants thereof, 100 ng of dynamin-K44A, and 8 ng of arrestin-3-GFP. The transfected cells were washed and incubated with (26 nM) or without hCG at 37°C for 30 min. The cells were fixed, and the receptors (in red) were visualized using an anti-Myc monoclonal antibody (9E10) and a rhodamine-conjugated anti-mouse antibody. Arrestin-3-GFP is shown in green, and co-localized components are shown in yellow. The cells were observed and analyzed using a Bio-Rad confocal microscope at the Central Microscopy Facility of the University of Iowa.
FIG. 8. Effect of GRK2 co-transfections on the hCG-induced binding of FLAG-arrestin-3 to the Myc-hLHR-wt and mutants thereof. 293T cells (plated in 100-mm dishes) were transiently cotransfected with a constant amount of the hLHR-wt or mutants thereof (10 g/dish), dynamin-K44A (2.5 g/dish), and FLAG-arrestin-3 (0.25 g/dish). Lanes labeled Ϫ and ϩ GRK2 are from cells that were cotransfected with the indicated constructs plus 5 g of pcDNA3.1 or GRK2, respectively. The transfected cells were washed and incubated with (26 nM) or without hCG at 37°C for 30 min prior to cross-linking and immunoprecipitation as described under "Materials and Methods." The results presented show the relevant areas of the blots of a representative experiment in which aliquots of the different lysates (ϳ18 l of lysate for panel A and ϳ500 l of lysate for panel B) containing equivalent amounts of receptor (measured by 125 I-hCG binding as described under "Materials and Methods") were used to measure the total amount of FLAG-arrestin-3 expressed (panel A), total GRK2 expressed (panel B), or the amount of FLAG-arrestin-3 immunoprecipitated with the 9E10 monoclonal antibody. similar to those of the hLHR-wt. Second, two individual single point mutations of the second extracellular loop (S512A and V519A) that induce a slight impairment in agonist-induced activation without affecting agonist-induced phosphorylation (13) induce a ϳ60% (V519A) to ϳ90% (S512A) impairment in hCG-dependent arrestin-3 binding. Third, two individual sin-gle point mutations (D405N in transmembrane helix 2 and Y546F in transmembrane helix-5) that induce a pronounced impairment in agonist-induced activation agonist-induced phosphorylation (12) induce a decrease in the hCG-induced arrestin-3 binding that is comparable in magnitude (ϳ90%) to that detected in one of the activation-impaired but phosphorylation-competent mutants (i.e. S512A). The effects of these mutations on the internalization of hCG closely parallel their effects of arrestin-3 binding. Thus, the 5S/A mutant internalizes hCG with a t1 ⁄2 (ϳ30 min) that is comparable with that of the hLHR-wt (ϳ 20 min), whereas the S512A, V519A, D405N, and Y546F mutants all internalize hCG with very long t1 ⁄2 values (120 -150 min) (see Refs. 12 and 13).
An involvement for the agonist-induced hLHR phosphorylation on the binding of arrestin-3 can only be demonstrated when receptor activation is greatly impaired. Thus, the profound reduction in the hCG-induced binding of arrestin-3 to the hLHR induced by mutations that impair activation and phosphorylation (i.e. hLHR-D405N and hLHR-Y546F) can be rescued by overexpression of GRK2 (Figs. 8 and 9), a manipulation that rescues the hCG-provoked phosphorylation of the hLHR (12). Again, these changes correlate perfectly with the effects of GRK2 on the internalization of hCG mediated by these mutants (12).
The L457R, D578Y, and D578H are naturally occurring mutants associated with Leydig cell hyperplasia and precocious puberty (L457R and D578Y, see Refs. [42][43][44][45] or Leydig cell adenomas and precocious puberty (D578H, see Ref. 46). In agreement with their constitutive activity and relative refractoriness to further stimulation when using cAMP accumulation as an index of activation (12,47), these three naturally occurring mutants also display a remarkably elevated binding of arrestin-3 in the absence of hCG and are relatively insensitive to further stimulation (Figs. 5-7). The enhanced agonist-independent association of these mutants with nonvisual arrestins may help explain their enhanced rates of agonist-independent and -dependent internalization (12). Finally, because the formation of GPCR-nonvisual arrestin complexes are thought to be mediators of the GPCR-induced activation of mitogenic pathways involving Src or mitogen-activated protein kinases (27,28,48), it is possible that the constitutive activity of the L457R, D578Y, and D578H mutants toward arrestin-3 binding contributes to the Leydig cell hyperplasia or neoplasia found in the individuals harboring these mutations.
Although the relative importance of phosphorylation and activation in the process of internalization has been examine with many GPCRs, up until now the relative importance of phosphorylation and activation in the formation of GPCR-arrestin complexes has been examined only with four GPCRs: rhodopsin, the ␤ 2 AR, the m2-muscarinic receptor, and the type 1A angiotensin receptor. The binding of arrestin to these four GPCRs depends mostly on receptor phosphorylation rather than on receptor activation (29 -34). Rhodopsin does not undergo internalization, but, like nonvisual arrestin binding, the agonist-induced phosphorylation of the ␤ 2 AR (16,49), the m2muscarinic receptors (50), and the type 1A angiotensin receptor (34) have also been shown to be required for internalization. In contrast, the data presented here show that the formation of the hLHR-nonvisual arrestin complex is independent of the agonist-induced phosphorylation of the hLHR and it depends mostly on the agonist-induced activation of this receptor. These results agree with our previous data showing that a functional effect of nonvisual arrestin binding to the hLHR (i.e. internalization) is also more dependent on the agonist-induced activation rather than the phosphorylation of the hLHR (12, 13). Thus, the role (if any) played by the phosphorylation of the hLHR remains elusive. 293T cells (plated in 100-mm dishes) were cotransfected with 10 g of HA-tagged ␤ 2 AR, 2.5 g of dynamin-K44A, 0.25 g of FLAG-arrestin-3, and 5 g of empty vector or 5 g of C 20 -GRK2-K220M. The transiently transfected cells were washed and incubated with or without 10 M isoproterenol for 5 at 37°C. The cells were then cross-linked, lysed, and immunoprecipitated with an anti-HA monoclonal antibody (3F10) as described under "Materials and Methods." The results presented show the relevant areas of the blots of a representative experiment in which aliquots of the different lysates (ϳ18 l of lysate for panel A and ϳ500 l of lysate for panels B and C) containing equivalent amounts protein were used to measure the total amount of GRK2 expressed (panel A) or the amount of HA-␤ 2 AR (panel B) or FLAG-arrestin-3 (panel C) immunoprecipitated with the 3F10 monoclonal antibody as described under "Materials and Methods."