HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 20, 17916-17927, May 17, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/20/17916    most recent M110479200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Mukherjee, S. Articles by Hunzicker-Dunn, M. Search for Related Content PubMed PubMed Citation Articles by Mukherjee, S. Articles by Hunzicker-Dunn, M. Social Bookmarking                      What's this?

Aspartic Acid 564 in the Third Cytoplasmic Loop of the Luteinizing Hormone/Choriogonadotropin Receptor Is Crucial for Phosphorylation-independent Interaction with Arrestin2^*

Sutapa Mukherjee, Vsevolod V. Gurevich^§, Anita Preninger^¶, Heidi E. Hamm^¶, Marie-France Bader^**, Asgerally T. Fazleabas, Lutz Birnbaumer^§§, and Mary Hunzicker-Dunn^¶¶

From the Departments of  Cell and Molecular Biology and ^¶ Molecular Pharmacology and Biological Chemistry and the Neuroscience Institute, Northwestern University Medical School, Chicago, Illinois 60611, the ^§ Department of Pharmacology, Vanderbilt University School of Medicine, Nashville, Tennessee 37232, ^** CNRS UPR-2356, Neurotransmission et Sécrétion Neuroendocrine, 5 rue Blaise Pascal, 67084 Strasbourg Cedex, France, and the  Department of Obstetrics and Gynecology, University of Illinois College of Medicine, Chicago, Illinois 60612, and the ^§§ Laboratory of Signal Transduction, NIEHS, National Institutes of Health, Research Triangle Park, North Carolina 27709

Received for publication, October 31, 2001, and in revised form, February 21, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Arrestin2 binding to the active but unphosphorylated luteinizing hormone/choriogonadotropin receptor (LH/CG R) in ovarian follicles is triggered by activation of ADP-ribosylation factor 6 (ARF6) and leads to uncoupling of this receptor from cAMP signaling. We sought to determine how arrestin2 binds to LH/CG R, if binding is of high affinity, and if the receptor also binds arrestin3. Desensitization of intact LH/CG R was equally sensitive to ectopic constructs of arrestin2 that bind other G protein-coupled receptors (GPCRs) either in a phosphorylation-independent or -dependent manner. Intact LH/CG R was not desensitized by ectopic arrestin3 constructs. Surface plasmon resonance studies showed that arrestin2 bound a synthetic third intracellular (3i) LH/CG R loop peptide with picomolar affinity; arrestin3 bound with millimolar affinity. To determine whether Asp-564 in the 3i loop mimicked the phosphorylated residue of other GPCRs, human embryonic kidney (HEK) cells were transfected with wild-type (WT) and D564G LH/CG R. An agonist-stimulated ARF6-dependent arrestin2 undocking pathway to drive desensitization of WT receptor was recapitulated in HEK cell membranes, and ectopic arrestin2 promoted desensitization of WT LH/CG R. However, D564G LH/CG R in HEK cells was not desensitized, and synthetic 3i D564G peptide did not bind arrestin2. Synthetic 3i loop peptides containing D564E, D564V, or D564N also did not bind arrestin2. We conclude that the ARF6-mediated mechanism to release a pool of membrane-delimited arrestin to bind GPCRs may be a widespread mechanism to deliver arrestin to GPCRs for receptor desensitization. Unlike other GPCRs that additionally require receptor phosphorylation, LH/CG R activation is sufficient to expose a conformation in which Asp-564 in the 3i loop confers high affinity binding selectively to arrestin2.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The luteinizing hormone/choriogonadotropin receptor (LH/CG R)¹ is crucial to reproductive function in males and females. Signaling by the receptor regulates spermatogenesis and puberty in males and final follicular maturation, ovulation, and progesterone production by corpora lutea in females. Following activation of the LH/CG R by LH or hCG and resulting activation predominately of stimulatory guanine nucleotide-binding protein (G_s) and adenylyl cyclase (AC), like most G protein-coupled receptors (GPCRs), signaling to G protein-regulated effectors declines or becomes desensitized (1). Desensitization of prototypical ₂-adrenergic receptor and rhodopsin occurs by a two-step mechanism. First the receptor becomes phosphorylated by a G protein-coupled receptor kinase (GRK) (2). That receptor phosphorylation is obligatory for desensitization is evidenced by GRK1 (3) and GRK2 (4) knockout mice that show, respectively, light-dependent retinal degeneration and abnormal heart development and function. The second step in receptor desensitization is the binding of arrestin to the active phosphoreceptor (5). High affinity binding of arrestin is achieved by engagement of two binding sites on arrestin, one that recognizes the active receptor conformation and a second site that interacts directly with the GRK-phosphorylated residue on the receptor (6-8). GPCR phosphorylation is therefore believed to be obligatory for high affinity arrestin binding to active receptor. Arrestin functions in desensitization to sterically hinder the ability of the receptor to interact with its cognate G protein (9). The arrestins are a conserved family of proteins consisting of the visual arrestins (arrestin1 and arrestin4) localized to rod and cone cells, respectively, and the ubiquitous arrestin2 (-arrestin1) and arrestin3 (-arrestin 2) (10).

We have recently shown that homologous desensitization of the LH/CG R requires arrestin2, which is localized to an AC-enriched ovarian follicular membrane fraction (11). This membrane-delimited arrestin2 is released from a putative docking site by a GTP-dependent reaction requiring the activation of the small G protein ADP-ribosylation factor 6 (ARF6) (12). The liberated arrestin2 then binds to the third intracellular (3i) loop of active LH/CG R, resulting in the uncoupling of the receptor from G_s (13). This conclusion is based in part on the ability of neutralizing arrestin antibodies to prevent agonist-dependent desensitization (11) and the ability of a synthetic 3i loop peptide to compete with receptor for endogenous arrestin and block desensitization (13).

Thus, for the LH/CG R, like many other GPCRs, desensitization requires the obligatory binding of an arrestin. The LH/CG R, like many GPCRs, is phosphorylated in its cytoplasmic tail upon agonist activation (14, 15). However, in contrast to most GPCRs, phosphorylation of the LH/CG R is not obligatory for desensitization. This conclusion is based (a) in a porcine follicular membrane model on the inability of protein kinase inhibitors to block LH/CG R desensitization (16), the occurrence of desensitization in the presence of millimolar concentrations of the ATP phosphorylation antagonist AMP-PNP (17-19), and the absence of detectable phosphate incorporation into immunoprecipitated LH/CG R under conditions that support LH/CG R desensitization (19) as well as (b) in a heterologous cellular expression model on results showing that truncation of the cytoplasmic tail to remove phosphorylatable serine and threonine residues (14, 15, 20) or mutation of these residues to alanines (21) abolishes detectable receptor LH/CG R phosphorylation (14, 15, 21) but interferes less than 25% with the extent of desensitization (14, 20).

Therefore we sought to compare the extent of desensitization of both the porcine LH/CG R in follicular membranes (a homologous cell background) and the murine LH/CG R stably transfected into human embryonic kidney (HEK) 293 cells (a heterologous cell background) promoted by ectopic constructs of arrestins, which bind to the ₂-adrenergic receptor and rhodopsin either in a constitutively active, phosphorylation-independent or a phosphorylation-dependent manner (6, 22-24). We found that desensitization of both LH/CG Rs in homologous versus heterologous cell backgrounds was equally sensitive to constructs of visual arrestin1 and arrestin2 that bound to other GPCRs in either a phosphorylation-dependent or constitutively active/phosphorylation-independent manner but insensitive to arrestin3. Mutation of Asp-564 to Gly in the 3i loop of the LH/CG R prevented arrestin2 binding and desensitization. Synthetic 3i loop peptides in which Asp-564 was changed to the similarly charged Glu, to Asn, which exhibits similar hydrogen bonding capability but is uncharged, or to nonpolar Val did not compete with the receptor to bind arrestin2. However, a peptide in which the other negatively charged residue in the 3i loop, Glu-557, was changed to Gly retained its ability to bind arrestin2 and blocked desensitization. The specific geometry of Asp-564 is therefore crucial for arrestin2 binding to the active LH/CG R to promote desensitization.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Recombinant arrestins were overexpressed in BL-21 cells and purified (25). Synthetic peptides corresponding to the 3i loop of the LH/CG R (FAVQNPELMATNKDTKIAKK), 3i D564G (FAVQNPELMATNK-GTKIAKK), 3i D564V (FAVQNPELMATNKVTKIAKK), 3i D564N (FAVQNPELMATNKNTKIAKK), 3i D564E (FAVQNPELMATNKETKIAKK), and 3i E557G (FAVQNPGLMATNKDTKIAKK) were synthesized and purified to ~95% purity by high pressure liquid chromatography by the Protein Chemistry Core Facility at Baylor College of Medicine (Houston, TX). Myristoylated (Myr)-ARF6-(2-13) and Myr-ARF1-(2-17) N-terminal peptides were synthesized as described previously (26). AC activity was measured, as detailed in the legend to Fig. 1, either in a sucrose gradient-purified membrane fraction isolated from porcine preovulatory ovarian follicles (18) or in a 10,000 × g membrane pellet fraction prepared from HEK293 cells stably transfected with wild-type (WT) murine LH/CG R or with receptor containing a D564G mutation (27). Reagent sources (11, 13), arrestin immunoprecipitations from ovarian membranes (12), an antibody to an extracellular region of the human LH/CG R (residues 257-271) (28), and an antibody made to the C-terminal 60 amino acids of arrestin3 (residues 350-409) (29) were as described previously. Statistical significance (p  0.05) was determined using Student's t test (30). Synthetic LH/CG R peptides were immobilized to carboxymethyl CM5 BIAcore^TM chips (Amersham Biosciences) according to the manufacturer's instructions. Surface plasmon resonance was performed on a BIAcore 2000 instrument (Amersham Biosciences). For binding to immobilized LH/CG R 3i peptides, recombinant arrestins were diluted in 10 mM Hepes, 0.15 M NaCl, 3.4 mM EDTA, and 0.05% surfactant p20 and injected at 10 µl/min. Kinetic parameters were studied in buffer flow for 6 min. K_d values were calculated with BIAevaluation^TM software. Fitted curves showing statistically significant binding had a ² of <7.5.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Porcine LH/CG R in Follicular Membranes Selectively Binds Arrestin2 with High Affinity to Evoke Desensitization-- Desensitization of the intact LH/CG R is readily detected in a cell-free membrane model using a two-stage reaction. When receptor agonist is present only during the 5-min AC assay (stage 2), maximal LH/CG R-stimulated AC activity is detected (Table I and Fig. 1A, bar 2). LH/CG R desensitization occurs when receptor agonist is added to the stage 1 reaction and is detected as a reduction in maximal hCG-stimulated AC activity (Table I and Fig. 1A, bar 3). Basal AC activity is measured when receptor agonist is omitted from stages 1 and 2. We previously showed that agonist-dependent LH/CG R desensitization results from the agonist-dependent undocking of membrane-delimited arrestin2 (11, 12) and subsequent binding of arrestin2 to the 3i loop of the receptor (13). In the present studies, we initially evaluated the ability of ectopic WT arrestins to promote LH/CG R desensitization of the intact LH/CG R in porcine ovarian follicular membranes when agonist was now omitted from the stage 1 incubation. To this end, follicular membranes were preincubated (30 min at 4 °C, see Fig. 1A, inset) without ectopic arrestins (Fig. 1A, bars 1-3) or with 40 nM ectopic WT arrestins (bars 4-6). For membranes incubated without ectopic arrestins, basal and maximal LH/CG R-stimulated AC activities are represented by bars 1 and 2 in Fig. 1A, and agonist-stimulated LH/CG R desensitization mediated by endogenous arrestin is represented by bar 3 of Fig. 1A. For membranes preincubated with ectopic WT arrestins, membranes were subjected to the stage 1 reaction (30 min at 30 °C) conducted in the absence of hCG followed by a 5-min AC assay (stage 2) conducted in the presence of hCG. Ectopic WT visual arrestin1 and arrestin2 (Fig. 1A, bars 4 and 5) evoked desensitization of hCG-stimulated AC activity, evidenced as a reduction in maximal hCG-stimulated AC activity, while arrestin3 was unable to promote LH/CG R desensitization, consistent with our earlier report (11). Thus, the addition of ectopic WT visual arrestin1 or arrestin2 to the preincubation reaction bypasses the requirement for agonist in stage 1 (which causes undocking of endogenous arrestin; Refs. 12, 13 and 31) and promotes desensitization of the intact LH/CG R. We further evaluated the efficacy of ectopic WT arrestins to promote desensitization of the intact LH/CG R in follicular membranes. Results in Fig. 1B show that a 50% reduction in maximal LH/CG R-stimulated AC activities was achieved with ~10 pM visual arrestin1 and arrestin2, consistent with high affinity binding of these arrestins to the intact LH/CG R in follicular membranes. Arrestin3 did not promote LH/CG R desensitization.

View larger version (79K): [in this window] [in a new window]   Fig. 1.   Porcine follicular LH/CG R desensitization is stimulated by visual arrestin1 and arrestin2 but not by arrestin3. Porcine follicular membranes were preincubated at 4 °C for 30 min with the indicated final concentrations of recombinant arrestins or water. In panels A, B, and C and as depicted in the schematic diagram, membranes were then subjected to a two-stage desensitization reaction at 30 °C (50-µl total volume). The stage 1 reaction was conducted in the presence of 8% ethanol, 25 mM BTP, pH 7.2, 0.4 mM EDTA, 1 mM EGTA, 0.2 mg/ml creatine phosphokinase, 20 mM phosphocreatine, 5 mM MgCl2, 10 µM GTP, 1 mM ATP, and 1 mM [3H]cAMP (~20,000 cpm)) and in the absence or presence of 10 µg/ml BSA or hCG as indicated. The stage 2 reaction was a 5-min AC assay initiated by the addition of 100 µM GTP, [-32P]ATP (~5 µCi, 100-200 cpm/pmol), and 10 µg/ml BSA or hCG. [32P]cAMP generated in stage 2 was purified and quantified (71, 72). Results are means ± S.E. of quadruplicate determinations for a single assay and are representative of three separate assays. The hatched region in panel A reflects the S.E. of maximal hCG-stimulated AC activity. The asterisk indicates that the difference between maximal hCG-stimulated AC activities (solid bar) and desensitized activities are significantly different (p < 0.05). The percent desensitization of LH/CG R was calculated as the percent reduction by receptor agonist or ectopic arrestin of maximal hCG-stimulated AC activity minus basal AC activity. For panels B and C, the open bar reflects basal AC activity when the preincubation contained no exogenous arrestins. The solid bar reflects maximal hCG-stimulated AC activity when the preincubation contained no exogenous arrestins, when BSA was present in stage 1, and when hCG was present in stage 2. The solid symbols reflect AC activity when the preincubation contained the indicated concentrations of ectopic arrestins, when BSA was present in stage 1, and when hCG was present in stage 2. Results are means ± S.E. of quadruplicate determinations for a single assay and are representative of three separate assays. Dotted lines indicate 50% inhibition of maximal hCG-stimulated AC activity above basal AC activity (i.e. ((hCG-stimulated AC activity)  (basal AC activity)/2 + (basal AC activity))). In panel D, 300 µg of follicular membrane protein were incubated in a stage 1 incubation, as described above, in the presence of BSA or hCG, membrane proteins were solubilized in buffer containing 1% Triton X-100 (73), Triton-soluble extract was precleared with protein A/G-agarose, and then immunoprecipitating antibody (15 µl of anti-arrestin2 (Transduction Laboratories, Lexington, KY) or anti-arrestin3 (29)) was added. Washed immune complexes were separated by SDS-PAGE, transferred to nitrocellulose, and then probed with anti-LH/CG R antibody (28). Results are representative of four separate experiments. In panel E, membranes were preincubated with ectopic arrestins (4 °C, 30 min) as in panels A and B, but the stage 1 reaction was omitted; AC reaction was performed for 10 min. Forskolin was present at 10 µM. The rest of the conditions were as described in panel A. % D, percent desensitization; IP, immunoprecipitation; Ab, antibody.

High affinity binding of arrestin to the ₂-adrenergic receptor and rhodopsin requires receptor phosphorylation (32, 33). The receptor-attached phosphate is thought to neutralize a positively charged Arg in arrestin (Arg-175 in visual arrestin1 and Arg-169 in arrestin2), thereby allowing arrestin to bind to active GPCR with high affinity (6, 22-24). Mutations that eliminate the positive charge of this crucial Arg yield arrestins that bind to these GPCRs independent of their phosphorylation state; i.e. in a phosphorylation-independent, constitutively active manner (6, 22, 23). Based on results in Fig. 1B and evidence that LH/CG R desensitization is independent of receptor phosphorylation (16-20, 34), we hypothesized that arrestin2 and visual arrestin1 see a negative charge density on the LH/CG R that mimics phosphorylation and allows high affinity arrestin binding. If this is the case, then phosphorylation-independent/constitutively active arrestin1 and arrestin2 mutants, which bind with high affinity to unphosphorylated rhodopsin or ₂-adrenergic receptor and M2 muscarinic acetylcholine receptor (6, 22-24), should cause LH/CG R desensitization with equivalent potency to "phosphorylation-dependent" arrestin forms. We also hypothesized, based on results in Fig. 1, A and B, either that arrestin3 binding to the LH/CG R requires receptor phosphorylation, in which case phosphorylation-independent/constitutively active arrestin3 mutants should promote LH/CG R desensitization, or that the LH/CG R preferentially recognizes arrestin2 and visual arrestin1 over arrestin3 to promote LH/CG R desensitization.

To test these hypotheses, we compared effects on LH/CG R desensitization of ectopic arrestin mutants first at 40 nM (Fig. 1A, bars 7-12) and then at various arrestin concentrations (Fig. 1C). We evaluated the effects of phosphorylation-independent arrestin mutants in which their phosphorylation recognition sites were mutated (R169E in arrestin2 and R175E in visual arrestin1) or in which the regulatory C-tails were deleted (arrestin2-(1-382) and arrestin3-(1-392)) with the effects of the phosphorylation-dependent visual arrestin1-arrestin2 chimera (ABBB, where A is visual arrestin1 and B is arrestin2), and an arrestin2-(1-393) mutant (6, 8, 22, 23). Results (Fig. 1C) showed that phosphorylation-independent R175E visual arrestin1, R169E arrestin2, and arrestin2-(1-382) mutants promoted LH/CG R desensitization with equal efficacy to phosphorylation-dependent visual arrestin1-arrestin2 chimera (ABBB) and arrestin2-(1-393). While the phosphorylation-dependent arrestin constructs have been reported to bind with high affinity only to the phosphorylated forms of other GPCRs, the equivalent high potency of the phosphorylation-dependent and -independent arrestins to cause LH/CG R desensitization supports our hypothesis that these arrestins are able to form a high affinity complex with the LH/CG R in the absence of receptor phosphorylation. Surprisingly the phosphorylation-independent arrestin3-(1-392) mutant (Fig. 1, A and C), like the WT protein (Fig. 1, A and B), did not promote significant LH/CG R desensitization. This result indicates that, rather than requiring LH/CG R phosphorylation to promote arrestin3 binding, arrestin3 simply interacts poorly with LH/CG R to promote desensitization. Taken together these results show that the intact LH/CG R shows strong preference for arrestin2 over arrestin3.

The inability of arrestin3 to promote LH/CG R desensitization suggests that arrestin3 does not bind to the LH/CG R in the follicular membrane model. To test this hypothesis, follicular membranes were incubated under conditions that do not promote LH/CG R desensitization without agonist or in the presence of agonist. Membrane proteins were solubilized, arrestin2 and arrestin3 were immunoprecipitated, and immunoprecipitates were probed for LH/CG R. Results in Fig. 1D show that only arrestin2 immunoprecipitates pulled down the LH/CG R when membranes were incubated with receptor agonist. No detectable LH/CG R was immunoprecipitated with arrestin3 antibody.

Although arrestin2 and visual arrestin1 promoted desensitization of activated receptor when arrestins were present during the stage 1 reaction (which consists of a 30-min incubation of membranes with arrestins, nucleotides, and buffers but without hormone, i.e. in the presence of inactive receptor; see Fig. 1A, inset) and hCG was added only to the 5-min stage 2 AC assay, none of the arrestins at 40 nM promoted desensitization of activated receptor or reduced basal or forskolin-stimulated AC activity (in a 10-min AC assay) when the stage 1 reaction was omitted (Fig. 1E). It is not clear why arrestins need to be present >10 min at 30 °C with inactive receptor to be competent to promote desensitization of active receptor.

We previously showed that arrestin2 binds selectively to the 3i loop peptide of the LH/CG R (13). To determine the actual affinities with which arrestin2 and visual arrestin1 bind to the 3i loop of the LH/CG R, the real-time interaction of arrestins with the 3i peptide immobilized to a carboxymethyl CM5 chip was evaluated by surface plasmon resonance. Sensograms showing fitted curves of the association binding and dissociation of arrestin2 and visual arrestin1 to the 3i peptide are shown in Fig. 2, A and B. Based on rates of association and dissociation, arrestin2 and visual arrestin1 bound to the 3i peptide with picomolar affinity (² < 7.5). Under these same conditions, arrestin3 bound to the 3i peptide with an affinity of 35 mM (² = 127) in a statistically insignificant manner (Fig. 2C).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Arrestin2 and visual arrestin1 exhibit high affinity real-time binding to synthetic LH/CG R 3i peptide. LH/CG R 3i or 3i D564G peptides were immobilized on a BIAcore CM5 chip. Kinetic parameters were determined with the indicated concentrations of arrestin2 (panel A) and visual arrestin1 (panel B) for binding to WT LH/CG R 3i peptide. The association binding of arrestin3 to WT LH/CG R 3i peptide is shown in panel C. The association binding of arrestin2 to LH/CG R 3i D564G peptide is also shown in panel C; equivalent results were obtained for arrestin3 binding to 3i D564G peptide (not shown). For the rest of the details see "Experimental Procedures." Results are representative of three separate experiments.

Taken together these results show that the 3i loop of the LH/CG R selectively binds arrestin2 and visual arrestin1 but does not bind arrestin3. Since the LH/CG R 3i loop was synthesized as an unphosphorylated peptide and since there is no evidence that the LH/CG R is phosphorylated on residues of the 3i loop (14), our results raise the possibility that the 3i loop of the receptor behaves as a constitutively desensitization-competent loop possibly by displaying a negatively charged residue(s) that "replaces" the phosphorylated residue(s) of other receptors allowing for high affinity binding of arrestin2 and visual arrestin1. This hypothesis was tested in a heterologous system in which mutant forms of the receptor were expressed.

Membranes of HEK293 Cells Expressing the Mutant Murine D564G LH/CG R Do Not Exhibit Arrestin2-dependent Desensitization-- Asp-564 of the 3i loop of the LH/CG R was tested as a candidate amino acid conferring "phosphorylation"-like properties to the LH/CG R, i.e. for being responsible for high affinity arrestin2 binding. We used a heterologous expression model in which LH/CG R phosphorylation has been demonstrated (15, 21, 35). We first ascertained whether desensitization of murine LH/CG Rs in HEK293 cell membranes required agonist-dependent ARF6 activation, a GTP-dependent event, the consequent undocking of membrane-delimited arrestin, and binding of arrestin to the receptor. Results in Fig. 3A showed that when exogenous GTP was omitted from a stage 1 reaction conducted in the presence of 1 mM AMP-PNP, agonist (hCG) present during the stage 1 reaction did not produce significant (p > 0.05) LH/CG R desensitization (compare solid and dotted bars). However, when GTP was added to stage 1, hCG present during the stage 1 reaction promoted robust LH/CG R desensitization (Fig. 3B, left panel, compare solid and dotted bars). Consistent with this GTP requirement for desensitization of LH/CG Rs in HEK293 cell membranes, preincubation of membranes with synthetic N-terminal Myr-ARF6 peptide, which blocks ARF6 activation (26, 36-38), blocked LH/CG R desensitization, while the corresponding Myr-ARF1 peptide was ineffective (Fig. 3B). The GTP requirement coupled with the ability of inhibitory ARF6 peptide to block desensitization of murine LH/CG Rs in kidney cell membranes indicates that desensitization of the LH/CG R in this heterologous expression model, like that in porcine follicular membranes (12), is dependent on the GTP-binding protein ARF6. LH/CG R desensitization in HEK293 cell membranes was also arrestin-dependent, based on the ability of neutralizing anti-arrestin antibodies (A9C6 plus F4C1), which block arrestin binding to GPCRs (10, 39), but not of nonimmune serum to block agonist-stimulated LH/CG R desensitization (Fig. 3C). These results therefore establish that agonist-dependent desensitization of WT LH/CG R in HEK293 cell membranes is mediated by the binding of a membrane-delimited arrestin to the receptor. As with porcine follicular membranes (see Fig. 1), addition of ectopic WT arrestin2, but not arrestin3, bypassed the requirement for agonist-dependent arrestin undocking and promoted significant desensitization of LH/CG Rs in HEK293 cell membranes (Fig. 3, D and E). These results indicate that the murine LH/CG R exhibits the same preference as the porcine receptor for arrestin2 over arrestin3. Also like the porcine receptor, phosphorylation-independent arrestin2 mutants (R169E arrestin2 and arrestin2-(1-382)) promoted significant desensitization equivalent to that stimulated by ectopic WT arrestin2 (Fig. 3E), while arrestin3 (WT and phosphorylation-independent arrestin3-(1-392)) did not. Taken together these results indicate that desensitization of murine LH/CG R expressed in HEK293 cells closely recapitulates that of endogenous LH/CG R in porcine follicular membranes.

View larger version (43K): [in this window] [in a new window]   Fig. 3.   WT murine LH/CG R stably expressed in HEK293 cells is preferentially desensitized under cell-free assay conditions by arrestin2 over arrestin3. Membranes from confluent HEK293 cells were obtained by homogenizing cells in 10 mM Tris-HCl, pH 7.0, 1.0 mM EDTA, and 27% sucrose with a glass-glass Dounce homogenizer. Homogenate was centrifuged at 1,000 × g for 5 min to remove nuclei and then at 10,000 × g for 30 min. The resulting membrane pellet was resuspended in 10 mM Tris-HCl, pH 7.0, aliquoted, and stored at 70 °C. In panel A, membranes were subjected to a two-stage desensitization incubation as in Fig. 1A except that stage 1 contained 1 mM AMP-PNP and no added GTP. BSA in stages 1 and 2 measures basal AC activity (open bar); BSA in stage 1 and hCG in stage 2 measures maximal hCG-stimulated AC activity (solid bar); hCG in stages 1 and 2 measures desensitization of hCG-stimulated AC activity (dotted bar). Percent reduction of full hCG-stimulated AC activity above basal AC activity when hCG is present in stages 1 and 2 measures the extent of LH/CG R desensitization. The difference between maximal hCG-stimulated AC activity (solid bar) and the dotted bar was not significant (p > 0.05). In panel B, membranes were preincubated for 30 min at 4 °C without or with the indicated final concentrations of synthetic Myr-ARF6-(2-13) or Myr-ARF1-(2-17) peptides (dissolved in 100% Me2SO to ~1 mM) and then subjected to a two-stage desensitization reaction as described in panel A except that stage 1 now contained 1 mM ATP and 10 µM GTP. Results are means ± S.E. of quadruplicate determinations of a single experiment and are representative of three separate experiments. In panel C, membranes were preincubated for 15 min at room temperature, then incubated for 1 h at 4 °C with A9C6 and F4C1 anti-arrestin antibodies or nonimmune serum (NIS) at a final dilution of 1:25 (in 50-µl final reaction volume), and finally subjected to a two-stage desensitization reaction as described in panel B. Results are means ± S.E. of quadruplicate determinations of a single experiment. In panels D and E, membranes were preincubated for 30 min at 4 °C with water or the indicated ectopic arrestins at a final concentration of 40 nM and then subjected to a two-stage desensitization reaction as described in panel B. Results are means ± S.E. of quadruplicate determinations of a single experiment and are representative of three separate experiments. For the rest of the details see the legend to Fig. 1. % D, percent desensitization. *, differences between BSA/hCG versus hCG/hCG incubations are significantly different (p < 0.05).

When HEK293 cells were stably transfected with the murine receptor containing a mutation of the negatively charged Asp residue to Gly in the 3i loop, D564G (Fig. 4A), which for the human receptor results in mildly elevated basal cAMP production (40, 41), hCG present during the stage 1 reaction did not cause arrestin-dependent cell-free LH/CG R desensitization² (Fig. 4B, compare solid and dotted bars). Moreover, none of the ectopic WT or constitutively active/phosphorylation-independent arrestin2 or arrestin3 mutants promoted desensitization (Fig. 4C). These results indicate that mutation of this negatively charged Asp residue in the 3i loop prevents both endogenous (liberated with receptor activation) and exogenous arrestin2 binding to the intact receptor.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   D564G mutant murine LH/CG R stably expressed in HEK293 cells is not desensitized by agonist or exogenous arrestins in the cell-free membrane model. Panel A shows the amino acid sequence of the 3i loop peptide of the LH/CG R connecting transmembrane (TM) domains 5 and 6. The two acidic residues are bold. For panels B and C, membranes from HEK293 cells stably transfected with mutant murine LH/CG R were prepared as described in the legend to Fig. 3. Membranes were preincubated for 30 min at 4 °C with water (panel B) or with the indicated ectopic arrestins (panel C) at a final concentration of 40 nM and then subjected to a two-stage desensitization reaction (in which stage 1 contained 10 µg/ml BSA or hCG as indicated, 10 µM GTP, and 1 mM ATP). Results are means ± S.E. of quadruplicate determinations of a single experiment and are representative of three separate experiments. % D, percent desensitization.

To confirm that changing the Asp residue in the 3i loop of the LH/CG R to a Gly prevents arrestin2 binding to the receptor, we used a second approach. We previously reported that preincubation of follicular membranes with a synthetic peptide corresponding to the 3i loop of the porcine LH/CG R competed with intact receptor for endogenous arrestin2 liberated by agonist-dependent receptor activation and blocked desensitization (13). We established the validity of this assay to detect arrestin binding to the 3i loop of the intact LH/CG R based on the ability of ectopic arrestin2 to bind to the synthetic 3i peptide and restore agonist-dependent LH/CG R desensitization induced by endogenous arrestin2 (13). In the following experiments we used this assay to determine whether or not a synthetic 3i peptide in which Asp-564 was changed to Gly bound endogenous arrestin2 and blocked desensitization of the intact receptor. Results showed that unlike the synthetic WT 3i peptide, which competed with receptor for endogenous arrestin2 and blocked desensitization of the LH/CG R in porcine follicular membranes (Fig. 5A, compare solid and dotted bars of the middle panel), the synthetic D564G 3i peptide did not block LH/CG R desensitization and thus did not bind the endogenous arrestin2 (Fig. 5A, right panels). Moreover, while arrestin2 avidly bound to WT 3i peptide coupled either to the BIAcore CM5 chip (see Fig. 2) or to CNBr-activated Sepharose 4B (Fig. 5B), arrestin2 did not bind to the D564G 3i peptide coupled either to BIAcore CM5 chip (Fig. 2C) or to CNBr-activated Sepharose 4B (Fig. 5B). These two different experimental approaches using the synthetic 3i D564G peptide combined with the results using the mutant D564G LH/CG R in HEK293 cells show that the charged Asp residue in the 3i loop of the LH/CG R is necessary for arrestin2 binding. However, the inability of constitutively active/phosphorylation-independent arrestin2 mutants to promote desensitization of the mutant D564G LH/CG R in HEK cell membranes (see Fig. 4C) suggests that Asp-564 does not simply substitute for the phosphate group of other GPCRs.

View larger version (46K): [in this window] [in a new window]   Fig. 5.   Synthetic porcine 3i D564G peptide neither blocks LH/CG R desensitization in porcine follicular membranes nor binds arrestin2. In panels A, C, and D porcine follicular membranes were preincubated for 30 min at 4 °C with water or the indicated final concentrations of synthetic 3i LH/CG R peptides and then subjected to a two-stage desensitization reaction (with BSA or hCG in stage 1 as indicated). Results are means ± S.E. of quadruplicate determinations of a single experiment and are representative of three separate experiments. The hatched region in panels C and D reflects the S.E. of maximal hCG-stimulated AC activity when the preincubation contained water. In panel B, WT 3i or 3i D564G LH/CG R peptides were coupled to CNBr-activated Sepharose 4B (13), resin was incubated with arrestin2 and extensively washed, and then bound arrestin2 was eluted and detected by SDS-PAGE and Western blotting using anti-arrestin2 antibody (Transduction Laboratories). Results are representative of three separate experiments. For the rest of the details see the legend to Fig. 1. % D, percent desensitization. *, differences between BSA/hCG versus hCG/hCG incubations are significantly different (p < 0.05).

We next sought to determine whether the requirement for Asp-564 to bind arrestin2 was based on its negative charge, its hydrogen bonding capability, or its basic geometry. We also determined whether the other negatively charged residue in the 3i loop of the LH/CG R, Glu-557, was required for arrestin2 to bind to the receptor to promote desensitization. To this end, synthetic 3i LH/CG R peptides were evaluated for their ability to compete with the intact LH/CG R in follicular membranes for binding of endogenous arrestin2 as described above. Results in Fig. 5C show that when Asp-564 was conservatively changed to the negatively charged Glu, this 3i D564E peptide did not bind arrestin2 to block desensitization. Also, when Asp-564 was changed to Asn, which is uncharged but exhibits hydrogen bonding capabilities similar to Asp, or to the nonpolar Val, neither the 3i D564N peptide (Fig. 5D) nor 3i D564V peptide (Fig. 5C) bound arrestin2 and blocked desensitization. Finally replacement of the other negatively charged residue in the 3i loop, Glu-557, with the hydrogen bond-accepting but nonpolar Gly yielded a peptide 3i E557G that retained its ability to bind arrestin2 (Fig. 5D). This result indicates that the binding of arrestin2 to the 3i loop peptide is not just dependent on the overall charge of this peptide since neutralization of the negatively charged Glu does not disrupt its binding to arrestin2. Taken together these results show that the specific geometry of Asp-564, rather than its negative charge or its hydrogen bonding capabilities, is crucial for arrestin2 binding to the active LH/CG R to promote desensitization.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

It is well established that phosphorylation of most GPCRs facilitates the binding of arrestins and consequent receptor desensitization. The requirement of receptor phosphorylation for desensitization was originally demonstrated for rhodopsin (42, 43), subsequently for muscarinic receptors (44), and more recently for many GPCRs (45). Phosphorylated GPCRs bind arrestins with high (picomolar) affinity (25). High affinity binding of visual arrestin1 and arrestin2 to GPCRs such as rhodopsin and ₂-adrenergic receptor appears to require engagement of two primary binding sites in the N-domain of arrestin, an activation-recognition site that recognizes the active conformation of the receptor and a phosphorylation-recognition site that directly interacts with GRK-phosphorylated residues (6-8). Engagement of both sites results in mobilization of a secondary site located in the C-domain and leads to high affinity binding of arrestin to the GPCR (6, 24). The basal (inactive) state of arrestin is stabilized by several intramolecular interactions (6, 24). One of these is the interaction of a positively charged Arg (Arg-175 in visual arrestin1 and Arg-169 in arrestin2) with neighboring negatively charged residues in the "polar core" (6, 22-24) as depicted in Fig. 6A for WT arrestin. Association of the negatively charged receptor-attached phosphates with this critical Arg is believed to neutralize the positive charge of Arg, disrupting the polar core and allowing arrestin to bind to the GPCR with high affinity (22), as depicted in Fig. 6B. Mutations eliminating the positive charge of Arg-175 in visual arrestin1 or Arg-169 in arrestin2 yield phosphorylation-independent, constitutively active forms of arrestin that no longer require GRK-catalyzed phosphorylation of rhodopsin, ₂-adrenergic receptor, or M2 muscarinic acetylcholine receptor to achieve high affinity binding (6, 22, 23) as depicted in Fig. 6C. The polar core also includes Arg-393 of the C-terminal tail of arrestin2 (22, 24, 46); therefore, truncation of arrestin2 at residue 382 yields a constitutively active arrestin that binds active ₂-adrenergic receptor independent of its phosphorylation state (6). However, truncation of arrestin2 at residue 393 does not remove any residues that contribute to the polar core; therefore, arrestin2-(1-393) behaves like WT toward ₂-adrenergic receptor (6). The mechanism of arrestin3 activation appears to be similar (47), and its C-terminal truncation mutant arrestin3-(1-392) avidly binds to active ₂-adrenergic receptor and M2 muscarinic acetylcholine receptor independent of their phosphorylation states (25).

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Schematic models depicting high affinity arrestin binding to the 2-adrenergic receptor versus the LH/CG R. In panel A, WT arrestin with its stable polar core (depicted by  and +) does not bind unphosphorylated 2-adrenergic receptor. The association of negatively charged phosphate (P) on phosphorylated 2-adrenergic receptor with positively charged Arg(R) on arrestin allows for high affinity arrestin binding (panel B). Mutations eliminating the positive charge of this Arg create constitutively active (CA) arrestin that binds with high affinity to unphosphorylated 2-adrenergic receptor (panel C). For the LH/CG R, both WT (panel D) and constitutively active (panel E) arrestin bind to the 3i loop of unphosphorylated receptor, which contains a critical Asp (D). Mutation of this Asp to Gly (G) yields a receptor that no longer binds either WT or constitutively active arrestin (panel F). Replacement of this Asp with Glu (E), Val (V), or Asn (N) yields a receptor that no longer binds WT arrestin (panels G-I).

Unlike most GPCRs, the LH/CG R does not require phosphorylation to become desensitized as reviewed in the Introduction. Consistent with evidence that phosphorylation is not obligatory for LH/CG R desensitization, we have shown that desensitization of the active porcine LH/CG R is equally sensitive to phosphorylation-dependent arrestin2 constructs as depicted in Fig. 6D and to constitutively active/phosphorylation-independent arrestin2 mutants as depicted in Fig. 6E. In agreement with these results, arrestin2 binds with picomolar affinity to a synthetic peptide corresponding to the 3i loop of this receptor. In the presence of receptor agonist, this 3i peptide competes with the intact LH/CG R in follicular membranes for endogenous arrestin2 and blocks agonist-dependent desensitization (13). Thus, arrestin2 binds active LH/CG R in follicular membranes with high affinity without the apparent need for receptor phosphorylation. The 3i loop of the LH/CG R can also bind visual arrestin1 with a similar high affinity. Like the porcine receptor in follicular membranes, the murine LH/CG R expressed in HEK293 cells is also desensitized by both ectopically added phosphorylation-dependent WT arrestin2 as well as phosphorylation-independent arrestin2 constructs. These results importantly show that the characteristics we have demonstrated for the porcine receptor, which most closely resembles the human receptor, also apply to the more distant murine receptor (48).

Asp-564 in the 3i loop is crucial for arrestin2 binding to the LH/CG R as depicted in Fig. 6, F-I; however, this negatively charged residue does not simply substitute for the phosphate group of other GPCRs since mutation of the negatively charged Asp-564 to the uncharged Gly in the D564G mutant LH/CG R does not yield a receptor that can be desensitized by phosphorylation-independent/constitutively active arrestin2 mutants. Consistent with this conclusion, substitution of Asp-564 with another negatively charged residue (Glu) or replacement with Asn, a residue that exhibits hydrogen bonding capability similar to Asp, does not mimic the Asp residue. Replacement of the other negatively charged residue in the 3i loop, Glu-557, with Gly does not affect the ability of this peptide to bind arrestin2 since Asp-564 is retained. These results suggest that Asp-564 must serve a second function in addition to neutralizing the polar core of arrestin to allow high affinity binding of arrestin2 to the 3i loop of the active LH/CG R. Conceivably Asp-564 not only may function to neutralize the polar core of arrestin2 but also may be required for arrestin2 to recognize the active conformation of the receptor. Asp-564 does not reside within the DRY motif reportedly necessary for arrestin and G_s binding by some GPCRs (49, 50). It is not known whether the inability of D564G LH/CG R to bind arrestin2 contributes to the mild constitutive activity exhibited by this receptor or whether the constitutive activity of this receptor is solely due to the formation of a partially activated conformation as a result of this mutation (40, 41, 51, 52). However, the inability of the receptor to bind arrestin2 does not automatically yield a constitutively active receptor since receptor containing D564E neither binds arrestin2 nor exhibits constitutive activity (40, 52). The homologous Asp residue is conserved in the glycoprotein hormone and cannabinoid receptors but is generally replaced with a charged Glu residue in most other GPCRs (52). Mutation of the homologous Asp to Gly in thyroid-stimulating hormone but not follicle-stimulating hormone receptors also leads to mildly constitutively active receptors (51, 53, 54). Interestingly some earlier reports using heterologous expression models purported a role for receptor phosphorylation in LH/CG R desensitization (14, 34), while others did not (20). Based largely on the ability of GRKs to promote desensitization of follicle-stimulating hormone receptor and thyroid-stimulating hormone receptor in transfected cell models, these receptors are believed to depend on phosphorylation for desensitization (55, 56). However, it is possible that these GPCRs that have this conserved Asp, like the LH/CG R, also bind arrestin2 in a phosphorylation-independent manner.

Both the porcine LH/CG R of ovarian follicles and the murine receptor expressed in HEK293 cells interact only with arrestin2 and not with arrestin3 to promote desensitization. These results were obtained using a membrane model that allows us to study receptor desensitization distinct from receptor internalization. We have also shown that arrestin3 antibody, in contrast to arrestin2 antibody, does not immunoprecipitate the active LH/CG R from porcine follicular membranes. This result most likely reflects the absence of arrestin3 from porcine follicular membranes as we have previously shown that follicular membranes contain readily detectable levels of arrestin2 but not of arrestin3 using F4C1 anti-arrestin antibody that recognizes all mammalian arrestins (11). However, overexpression of either arrestin2 or arrestin3 promotes LH/CG R internalization (57-60), and chemical cross-linking studies have shown that FLAG-tagged arrestin3 can associate with the human LH/CG R (61). These results suggest that the LH/CG R exhibits distinct arrestin binding sites for desensitization versus internalization and that, in contrast to the arrestin2-selective binding site in the 3i loop of the active LH/CG R that mediates desensitization, the apparent binding site(s) necessary for receptor internalization is less selective for the nonvisual arrestins. Recent studies show that some GPCRs exhibit selective binding sites and/or preferences for the arrestins leading to receptor desensitization or internalization. For example, arrestin3 preferentially binds to the 3i loop of the chemokine receptor CXCR4 to promote receptor internalization and extracellular-regulated kinase activation, while receptor desensitization (to G_i) requires binding of either arrestin2 or arrestin3 to the C-terminal tail (62). And ₂-adrenergic receptor sequestration appears to be mediated by arrestin3 and not arrestin2 based on results in fibroblasts derived from arrestin3 and arrestin 2 knockout mice (63) and in cells where the nonvisual arrestins are not overexpressed (64). The interaction of the LH/CG R with arrestin2 and not with arrestin3 for desensitization adds to a growing number of functional differences between the two nonvisual arrestins. These include the regulation of arrestin2 but not arrestin3 interaction with clathrin by phosphorylation (65) and the function of arrestin3 and not arrestin2 as a scaffold for c-Jun amino-terminal kinase 3 activation (66).

For the native LH/CG R present in ovarian follicles, ARF6 activation downstream of the activated receptor serves as the trigger to release a pool of membrane-delimited arrestin2, which binds to active receptor and promotes LH/CG R desensitization. Cell-free desensitization of the murine LH/CG R expressed in kidney cells, like that of the porcine receptor in ovarian follicular membranes (12), also requires the GTP-ARF6-dependent pathway to liberate membrane-docked arrestin2. The presence of the ARF6-regulated arrestin2 undocking mechanism in HEK293 cells (that do not normally express LH/CG R) as well as the ability of LH/CG R agonist to promote ARF6 activation in HEK293 cells indicates that ARF6 activation to liberate membrane-docked arrestin2 and thereby mediate desensitization may be a widespread, earlier unappreciated mechanism to deliver arrestin to GPCRs. Consistent with this notion, ARF activation also occurs upon agonist-dependent activation of the follicle-stimulating hormone (28), M3 muscarinic acetylcholine (67), fMet-Leu-Phe (68), and H1 histamine and B2 bradykinin (69) receptors, and the involvement of ARF6 in ₂-adrenergic receptor endocytosis was recently reported (70). While most other GPCRs additionally require receptor phosphorylation to achieve high affinity arrestin binding leading to desensitization, the LH/CG R does not. The absence of the phosphorylation constraint for arrestin2 binding to the LH/CG R, present in so many other GPCRs, suggests either that continued cAMP signaling is detrimental to and/or alternate pathways initiated by arrestin2-LH/CG R interaction are critical for continuation of the species. Understanding the biological significance of the LH/CG R-arrestin2 interaction to reproductive success is a goal of future studies.

	ACKNOWLEDGEMENTS

We gratefully acknowledge the gift of pan-arrestin antibodies from Dr. Larry Donoso (Wills Eye Hospital Research Division, Philadelphia, PA) and Krzysztof Palczewski (University of Washington School of Medicine, Seattle, WA), the gift of arrestin3-selective antibody from Dr. Jeffrey Benovic (Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA), and the generation of the stable cell lines expressing WT and mutant LH/CG Rs by Drs. Xi Xhu (Department of Pharmacology and Neurobiotechnology Center, Ohio State University, Columbus, OH) and Mariel Birnbaumer (Laboratory of Signal Transduction, NIEHS, National Institutes of Health, Research Triangle Park, NC). We thank Dr. Evelyn T. Maizels for critically reading the manuscript.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants R01 HD 38060 (to M. H. D.), R01 EY 11500 and GM 63097 (to V. V. G.), and R01 EY 10291 and EY 06062 (to H. E. H.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Current address: Dept. of Pharmacology, Vanderbilt University School of Medicine, 422 Robinson Research Bldg., Nashville, TN 37232.

^¶¶ To whom correspondence should be addressed: Dept. of Cell and Molecular Biology, Northwestern University Medical School, 303 East Chicago Ave., Chicago, IL 60611. Tel.: 312-503-8940; Fax: 312-503-0566; E-mail: mhd@northwestern.edu.

Published, JBC Papers in Press, February 26, 2002, DOI 10.1074/jbc.M110479200

² Membranes prepared from cells containing D564G mutant LH/CG R variably exhibit up to ~30% agonist-dependent cell-free LH/CG R desensitization (27). However, this desensitization is arrestin-independent and is not inhibited by neutralizing anti-arrestin antibodies A9C6 plus F4C1 (in the same experiments in which desensitization in WT cells is blocked by these anti-arrestin antibodies; not shown), consistent with our evidence that the 3i peptide containing D564G does not bind arrestin2 and does not block LH/CG R desensitization.

	ABBREVIATIONS

The abbreviations used are: LH/CG R, luteinizing hormone/choriogonadotropin receptor; WT, wild type; hCG, human chorionic gonadotropin; GPCR, G protein-coupled receptor; HEK, human embryonic kidney; ARF, ADP-ribosylation factor; 3i, third intracellular; AC, adenylyl cyclase; GRK, G protein-coupled receptor kinase; AMP-PNP, adenylylimidodiphosphate; Myr, myristoylated; BTP, 1,3-bis[tris(hydroxymethyl)methylamino]propane; BSA, bovine serum albumin.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES