Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

Melatonin, the N-acetyl-5-methoxy metabolite of serotonin, is a hormone known for its ability to adjust the circadian clock and to induce seasonal changes in reproductive physiology (in animals with breeding seasons). Regulation of circadian physiology by melatonin is believed to be mediated primarily by a specific receptor, the melatonin_1a (Mel_1a) receptor, which is enriched in hypothalamic nuclei and in the pars tuberalis of the adenohypohyseal gland. In addition, the effect of melatonin on the circadian clock involves a second receptor that, as opposed to the Mel_1a receptor, has a very low expression level and does not mediate inhibition of neuronal activity in the nucleus suprachiasmaticus, the site of the master clock (Liu et al., 1997). Both the Mel_1a receptor and the second, unidentified receptor are G protein-coupled receptors. The small family of melatonin receptors comprises only three members, two of them, the Mel_1a- and Mel_1b receptors, are present in mammals. The melatonin receptors share little sequence homology with other receptor types, hence, they seem to have emerged early in the course of receptor evolution (for a review, see Dubocovich, 1995).

The pharmacological identification of melatonin receptors has relied largely on the labeling of receptors in brain slices with the agonist radioligand 2-[¹²⁵I]iodomelatonin, which is endowed with high affinity and a low level of nonspecific binding. Because of the highly discrete expression of melatonin receptors, the signaling pathways became amenable to detailed characterization only after cloning and heterologous expression of the recombinant receptor. Signaling by melatonin is sensitive to pertussis toxin (PTX) (White et al., 1987; Carlson et al., 1989) and in accordance with these findings, the recombinant Mel_1a receptor couples to G proteins of the G_i/G_o class, leading to inhibition of adenylyl cyclase (Witt-Enderby and Dubocovich, 1996). Additional signaling pathways activated by the Mel_1a receptor (arachidonic acid release, Ca²⁺-release) were similarly suppressed by PTX, confirming the assumption that the receptor interacts predominantly with G_i (Godson and Reppert, 1997).

The ligand for the Mel_1a receptor, melatonin, is produced in and secreted by the pineal gland. A regular phenomenon in vertebrate physiology, melatonin levels are elevated during the night. The biochemical basis for the circadian changes in melatonin levels was found in the variable activity of serotonin N-acetyltransferase, which catalyzes melatonin production (Gastel et al., 1998). Daylight reduces the stability of the enzyme toward proteasomal degradation. Not only the melatonin production but also the melatonin receptors are regulated in a diurnal rhythm (Tenn and Niles, 1993; Gauer et al., 1994; Neu and Niles, 1997); during the day the level of iodomelatonin binding to membranes from hypothalamic nuclei increases and drops again in the evening, before the hormone level rises. Thus, the melatonin receptor-expressing cell adapts to the day-night rhythm in a fashion that is inversely related to the ambient melatonin levels. Because the melatonin receptor is considered a potential drug target, variations in the expression level may cause therapeutically relevant changes in the response to administered melatonin receptor ligands. We have therefore generated a stable human embryonic kidney (HEK)293 cell line in which the receptor is expressed to a density that presumably occurs in neuronal cells, given the cellular heterogeneity of brain tissue preparations and the distinct distribution of Mel_1a receptors (several hundred femtomoles per milligram of membrane protein, Barrett et al., 1996). We report that if expressed in HEK293 cells, the Mel_1a receptor reveals a marked degree of spontaneous activity. In addition, the Mel_1a receptor exhibits features that are not predicted by the classical model of receptor/G protein (R/G) coupling. For example, whereas signal transduction of the Mel_1a receptor is strongly reduced by PTX, the receptor retains the ability to bind an agonist ligand with high affinity after PTX treatment. We have examined this phenotype and find that the receptor forms a highly stable complex with G_i, such that the G protein  subunit is shielded from PTX. The mode of Mel_1a R/G coupling also results in a high basal activity of the receptor, i.e., the receptor is spontaneously active in the absence of agonists.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. [³⁵S]GTPS (guanosine 5'-(3-O-thio)triphosphate) and [¹²⁵I]iodomelatonin were purchased from NEN (Boston, MA). ()N⁶-3[¹²⁵I](iodo-4-hydroxyphenyl-isopropyl) adenosine was synthesized according to Linden (1984). Guanine nucleotides, adenosine deaminase, protein A, and protein G Sepharose were from Boehringer Mannheim (Mannheim, Germany). 1-O-n-octyl-D-glucopyranoside (octylglucoside, OG), CHAPS (3-[3-cholaminpropyl) dimethyl-ammonio]-1-propane-sulfonic acid) and HEPES were purchased from Biomol (Munich, Germany); suramin was obtained from Research Biochemicals (Natick, MA). The materials required for SDS-polyacrylamide gel electrophoresis were obtained from Bio-Rad (Richmond, CA). Fetal calf serum was purchased from PAA Laboratories (Linz, Austria), Dulbecco's modified Eagle medium (DMEM), nonessential amino acids, -mercaptoethanol, and G418 (geneticin) were obtained from Life Technologies (Grand Island, NY). Melatonin and luzindole were purchased from Tocris (Langford, Bristol, UK). N6-cyclopentyladenosine, PTX, L-glutamine, penicillin G, and streptomycin were purchased from Sigma Chemical Co. (St. Louis, MO). Buffers and salts were purchased from Merck (Darmstadt, Germany).

Stable Cell Lines. Cloning of the human melatonin type 1a (Mel_1a) receptor cDNA using polymerase chain reaction primers selected from the reported Mel_1a receptor sequence and generation of stable cell lines expressing the Mel_1a receptor will be described in detail elsewhere. In brief, human brain mRNA was obtained through polymerase chain reaction amplification, the coding sequence was cloned into the expression vector pcDNA3 (Invitrogen, Carlsbad, CA), and confirmed by DNA sequencing. For stable receptor expression, HEK293 cells were transfected with the human Mel_1a receptor cDNA together with a plasmid carrying the geneticin resistance gene by liposome-mediated transfection. Clones were selected in DMEM supplemented with 10% fetal calf serum and 800 µg/ml of geneticin (G418) and screened for [¹²⁵I]iodomelatonin binding. For the generation of HEK293 cells expressing the human A₁-adenosine receptor see Waldhoer et al. (1998). Transfected HEK293 cells were grown in DMEM containing 10% fetal calf serum, 2 mM L-glutamine, -mercaptoethanol, nonessential amino acids, 100 U/ml penicillin G, and 100 µg/ml streptomycin and 0.2 mg/ml geneticin at 5% CO₂ and 37°C.

Membrane Preparation and Protein Purification. Cells were grown to confluence in 10-cm tissue culture dishes, rinsed once with ice-cold PBS, and scraped off their plastic support. After sedimentation, the cell pellet was resuspended in HME (25 mM HEPES-NaOH, pH 7.5, 2 mM MgCl₂, and 1 mM EDTA) and subsequently deep frozen in liquid nitrogen. Cells were thawed and disrupted by sonication. Membranes were sedimented by centrifugation at 38,000g for 10 min, were resuspended in HME at a protein concentration of 8 to 10 mg/ml and stored in aliquots at 80°C. For experiments in which contaminating nuclear matter had to be discarded, membranes were enriched by differential centrifugation; the supernatant obtained by centrifuging the homogenate at 9,000g, was sedimented at 50,000g.

Radioligand Binding Experiments. Equilibrium binding with [¹²⁵I]iodomelatonin was carried out in a final volume of 25 to 100 µl containing: 50 mM Tris-HCl (pH 8.0), 1 mM EDTA, 5 mM MgCl₂, and 5 to 10 µg membrane protein. The binding reaction was carried out for 60 min at 30°C and terminated by filtration over glass fiber filters using a cell harvester (Skatron, Lier, Norway). [¹²⁵I]Iodomelatonin binding to intact cells was carried out on cells that were detached from their growth support and resuspended in assay tubes. The binding reaction proceeded in DMEM for 90 min at 30°C. Specific [¹²⁵I]iodomelatonin binding was not detectable on nontransfected HEK293 cells and on HEK293 cells transfected with the A₁-adenosine receptor. Nonspecific binding was determined in the presence of 100 µM luzindole and typically amounted to less than 10% of total binding.

Uncoupling of Mel_1a Receptor by G Protein-Selective Antibodies and Suramin. Membranes were prepared from HEK293 cells that had or had not been pretreated with PTX (16 h, 0.1 µg/ml), and from cells that were transfected with a Gq cDNA in a pCIS plasmid (kindly provided by Dr. M. Simon, California Institute of Technology, Pasadena, CA) that drives the overexpression of Gq. Membranes (~10 µg protein/assay or ~25 µg protein/assay if the membranes had been prepared from PTX-treated cells) were preincubated with the indicated amounts of antibody in the presence of 0.2% OG in a volume of 20 µl for 20 min on ice. The binding reaction was started by adding [¹²⁵I]iodomelatonin (300 pM) to give a final OG concentration of <0.05% and was terminated as described above. The nonspecific effect of adding immune globulin was controlled for by performing the binding experiment in parallel with nonimmune IgG.

Determination of cAMP Formation and Inositol Phosphates (IP) Accumulation in Intact Cells. HEK293 cells expressing the Mel_1a receptor were grown to confluence in 40-mm culture dishes in DMEM containing 10% fetal calf serum. The cellular ATP pool was labeled by incubating the cells with 2.5 µCi/ml of [³H]adenine for 16 h. After removing free [³H]adenine and rinsing, cells were incubated with DMEM containing rolipram (10 µM) and the assay was started by adding receptor ligands in the absence or presence of forskolin (20 µM). The reaction was stopped after 10 min by aspirating the assay medium and adding 800 µl of 2% perchloric acid with 100 µM cAMP. After neutralizing the cell extract with 80 µl of 4 M KOH, [³H]cAMP was separated from ATP by sequential chromatography on Dowex AG 1X-4W and Alumina columns (Johnson and Solomon, 1991). To monitor the recovery of [³H]cAMP, an internal [³²P]cAMP standard was prepared using recombinant adenylyl cyclase consisting of the two, separately expressed catalytic domains of the type I and type II isoforms (Mitterauer et al., 1998).

Reconstitution of Mel_1a Receptor with Purified G_i To probe the interaction of the Mel_1a receptor with exogenous G protein of the G_i class, membranes were washed with 6 M urea. Two milligrams of membrane protein was taken up in 10 ml 6 M urea/50 mM HEPES (pH 7.5), and after 15 min at 4°C the suspension was centrifuged at 50,000g. The resulting pellet was washed and resuspended in HME. In urea-washed membranes that were stripped off peripheral membrane proteins, ~80% of the iodomelatonin binding was lost. Interaction of the Mel_1a receptor with G_i was tested by measuring iodomelatonin binding and the receptor catalyzed G protein activation. To restore high-affinity iodomelatonin binding, recombinant Gi-3 and purified porcine -dimer at the indicated concentrations were preincubated with urea-washed membranes (~3 µg membrane protein) in the presence of 0.2% OG for 20 min on ice. The binding assay was initiated by adding radioligand and by diluting the preincubation mix about 4-fold. To determine G protein activation by [³⁵S]GTPS binding, preincubation of membranes with G_i-1 in the presence of 0.2% OG on ice was followed by adding receptor ligands in reaction buffer. Ligands were allowed to bind for 15 min at 25°C before the reaction was carried out as described above. To control for the loss of receptor due to urea treatment we also subjected membranes carrying the A₁-adenosine receptor to urea treatment.

[³²P]ADP Ribosylation of G_i and subsequent Immunoprecipitation Using a Gi-Selective Antibody. Membranes (2 mg membrane protein) from native and PTX-pretreated HEK293 cells were solubilized with 10 mM CHAPS in Tris-HCl 50 mM, pH 8.0, and EDTA 1 mM (protein/detergent ratio = 1:4) by stirring on ice for 1 h. The soluble supernatant was collected by centrifugation (75,000g for 15 min) and the volume was reduced over a YM 30 membrane (Amicon) to a protein concentration of ~2.0 mg/ml. PTX was dissolved in preactivation buffer (100 mM DTT, 0.8 M urea, 5 mM CHAPS, and 50 mM potassium phosphate, pH 8.0) and was activated for 25 min at 30°C. One milligram of the soluble membrane extract was subjected to ADP ribosylation by preactivated PTX (5 µg). The reaction was carried out for 1 h at 25°C in a volume of 0.8 ml including 37.5 mM Tris-HCl, 45 mM NaCl, 0.5 mM MgCl₂, 5 mM DTT, 10 µM GDP, 0.4 mg N-(2-chloropropyl)-N,N-dimethyl-ammoniumcholoride, 10 mM thymidine, 0.1 mM NAD, and 450 nCi [³²P]NAD at a specific activity 5.6 pCi/nmol. Proteins were precipitated by adding TCA to a concentration of 10%. The protein pellet was rinsed with acetone, dried under a stream of nitrogen, and resuspended in 250 µl RIPA buffer (50 mM Tris-HCl, pH 8.3, 1% Nonidet P-40, 5 mM EDTA, and 150 mM NaCl). The resuspended protein pellet was incubated with 2.5 µg of an antibody directed against the carboxy terminus of Gi1-3 for 6 h at 4°C. Antibodies were precipitated by the addition of a 1:1 mixture of protein A/protein G-Sepharose (10% of the final volume) for 2 h at 4°C. After centrifugation, the Sepharose was washed four times with RIPA at 0.15% Nonidet P-40. Bound proteins were dissolved by boiling in 60 µl SDS-sample buffer and the entire volume was subjected to gel electrophoresis.

Immunoprecipitation of Mel_1a Receptor. HEK293 membranes expressing the Mel_1a receptor were solubilized using 1% digitonin in 75 mM Tris-HCl (pH 7.4), 12 mM MgCl₂, and 2 mM EDTA and protease inhibitors (5 mg/ml soybean trypsin inhibitor, 5 mg/ml leupeptin, and 10 mg/ml benzamidine). The detergent/protein ratio was 1:1; the mixture was stirred for 3 h at 4°C and centrifuged at 50,000g. For immune precipitation, the digitonin concentration was adjusted to 0.2% and antiserum-536 was added so that the serum made up 1/40 of the reaction volume. Immune complex formation was allowed to proceed for 18 h at 4°C with gentle agitation and during the last 6 h protein A-agarose (10%) was included. The agarose beads were sedimented by centrifugation at 5000g and washed five times with ice-cold buffer including 0.2% digitonin. Bound proteins were resuspended in 0.2% digitonin buffer and incubated with 50 µM guanosine-5'-(,-imido)triphosphate [Gpp(NH)p] for 1 h at 37°C. The supernatant was added to SDS-sample buffer, boiled, and separated by gel electrophoresis.

Immunoblots. Membrane proteins (25 µg/lane) were separated on SDS-polyacrylamide gels (12% acrylamide and 0.16% bisacrylamide) and transferred to nitrocellulose membranes that were probed with antisera recognizing Gi or a purified antibody recognizing Gq (Santa Cruz Biotechnology). The immunostained bands were visualized by enhanced chemiluminescence using an anti-rabbit-IgG antibody conjugated to horseradish peroxidase (Amersham, Arlington Heights, IL).

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

High-Affinity [¹²⁵I]Iodomelatonin Binding to Mel_1a Receptor on Intact Cells. When the human Mel_1a receptor was stably expressed in HEK293 cells and intact cells were subjected to equilibrium binding, the agonist radioligand iodomelatonin bound in a saturable manner and with high affinity (Fig. 1A); in nontransfected cells no specific binding (or cellular uptake) of iodomelatonin was detected. In membranes prepared from Mel_1a receptor-expressing cells (Mel_1a membranes), iodomelatonin bound with similar affinity (Fig. 1B). The total number of Mel_1a receptors detected on the surface of intact cells was about twice the number of receptors retained in isolated membranes. This is presumably due to losses during membrane preparation; furthermore, the data show that the density of receptors in the high-affinity conformation on intact cells was not less than their density in membranes. In addition, Fig. 1A depicts a saturation isotherm obtained on intact cells and in the presence of suramin (0.3 mM); suramin acts as a G protein antagonist (see Freissmuth et al., 1999), but because of its six negative charges, it is virtually membrane impermeable. Iodomelatonin binding to intact cells was not affected by suramin, indicating that the compound did not block the ligand binding site of the Mel_1a receptor. Thus, suramin could be used to inhibit R/G coupling in cell membrane preparations (see below; Fig. 2).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Iodomelatonin binding to isolated membranes from HEK293 cells stably expressing the Mel1a receptor and binding to intact cells. A, iodomelatonin binding was performed to intact cells in suspension in DMEM in the absence () and presence () of 0.3 mM suramin for 1 h at 25°C. The KD value was estimated to be 89 ± 11 pM (n = 3). B, specific iodomelatonin binding to membranes from control () and PTX-treated () cells. Bmax and KD estimates were 732 ± 38 fmol/mg and 56 ± 9 pM for untreated and 526 ± 112 fmol/mg and 93 ± 17 pM for PTX-treated cells (n = 5).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Destabilization of iodomelatonin binding to the Mel1a receptor by guanine nucleotides and by the G protein antagonist suramin. A, binding of iodomelatonin (300 pM) to Mel1a membranes was carried out at 30°C for 60 min in the presence of GTPS at the indicated concentrations with () or without () the addition of detergent (OG, 0.1%). B, inhibition of iodomelatonin binding to the Mel1a receptor by increasing concentrations of suramin was evaluated at various degrees of receptor occupancy (75 and 99%) attained at different iodomelatonin concentrations (, , 100 pM; , , 500 pM). Binding data were obtained in the presence (open symbols) or absence (closed symbols) of 0.1% OG and are given in percentage of the binding in the absence of suramin. IC50 values were estimated by nonlinear curve fitting. C, IC50 estimates were obtained from individual suramin concentration-response curves as in B and were replotted versus the relative receptor occupancy. Curves were generated in native membranes (), in native membranes resuspended in 0.1% OG () and in PTX-treated membranes (). Membranes were labeled with different concentrations of iodomelatonin (range, ~50-700 pM) in the presence of increasing concentrations of suramin, and from the inhibition curve, IC50 values were estimated.

Affinity of Mel_1a Receptor for Its G Protein. It is, alternatively, conceivable that the uncoupled, G protein-free form of the receptor can per se bind iodomelatonin with high affinity. Therefore, we probed the R/G interaction. Guanine nucleotide-dependent modulation of agonist binding is a sensitive index for functional R/G coupling. In membranes, guanine nucleotides apparently failed to destabilize iodomelatonin binding. However, the inclusion of detergent in the binding assay unmasked the effect of guanine nucleotides (Fig. 2A). In the presence of detergent (at "lubricant" concentrations below the critical micellar concentration  4 mM CHAPS or 0.1% OG), about 60% of the receptor population was destabilized by guanine nucleotides; GTPS and GTP were as potent and effective as GDP in inhibiting the formation of the high-affinity complex (data not shown). The presence of low detergent concentrations did not significantly affect iodomelatonin binding affinity nor the number of labeled receptors; at higher detergent concentrations the binding was almost completely suppressed by guanine nucleotides, but the number of labeled receptors decreased (not shown). Similarly, in membranes from PTX-treated cells, guanine nucleotides suppressed iodomelatonin binding if detergent was present at low concentrations (data not shown).

Signaling via Mel_1a Receptor Is PTX Sensitive. The ability of guanine nucleotides and of suramin to decrease iodomelatonin binding indicated that high-affinity agonist binding was indeed due to the formation of a ternary complex with the Mel_1a receptor; this ability was retained in membranes pretreated with PTX. We therefore tested whether the Mel_1a receptor would interact with G protein subtypes other than G_i. Figure 3A demonstrates that melatonin accelerated the guanine nucleotide exchange reaction (association of [³⁵S]GTPS) in isolated Mel_1a-membranes, and that addition of detergent enhanced the receptor-dependent association rate (Fig. 3A). PTX abolished agonist-induced G protein activation. To detect a receptor-mediated increase in G protein activation, PTX-treated membranes had to be reconstituted with G_i (not shown).

View larger version (28K): [in this window] [in a new window]   Fig. 3.   G protein activation and effector regulation by the Mel1a receptor stably expressed in HEK293 cells. A, G protein activation in Mel1a membranes before (circles and squares) and after PTX treatment (triangles). The [35S]GTPS association rate was assessed in the presence (open symbols) and absence (full symbols) of 0.1 µM melatonin; y-axis values were obtained by normalizing GTPS binding data to the Bmax values from iodomelatonin saturation isotherms in PTX-treated and untreated membranes. In native membranes [35S]GTPS binding was determined with () and without () the addition of 0.1% OG; the inset shows the agonist promoted increment in [35S]GTPS binding over basal. The experiment shown is representative of three performed. B, [3H]cAMP accumulation was measured in native and PTX-treated cells incubated with rolipram in the presence of the indicated agents (forsk, forskolin 10 µM; mel, melatonin 0.1 µM) for 10 min. [3H]cAMP accumulation is expressed as percentage of basal accumulation, which amounted to 1300 ± 176 cpm. C, accumulation of [3H]IP1 was determined in the presence of LiCl (10 mM) in native and PTX-treated cells; the indicated substances (melatonin 0.1 µM, ATP 30 µM) were added for 30 min. [3H]IP1 accumulation is given in percentage of basal [3H]IP1 accumulation in untreated cells (923 ± 239 cpm). In B and C data are means ± S.E. of at least five individual experiments.

Specificity of G Protein Subtypes Forming a Ternary Complex with Mel_1a Receptor. We then tested whether the Mel_1a receptor engages exclusively G_i when it binds agonists with high affinity and whether it does so even after PTX treatment. We used antibodies selective for the carboxy termini of G protein  subunits to destabilize high-affinity iodomelatonin binding in membranes from untreated and PTX-treated cells (Fig. 4, A and B). The carboxy terminus of the  subunit is a site contacted by the receptor and is, in part, responsible for the selective recognition of G proteins by individual receptors (Conklin et al., 1996). The antibodies employed were directed against the carboxy terminal dekapeptides of _i1-3, _q, and _o. Uncoupling of the Mel_1a receptor required that the antibody be preincubated with membranes in the presence of detergent; if the antibody was added in the absence of detergent a specific inhibitory effect on iodomelatonin binding was not found (data not shown). This observation is consistent with the interpretation that the interaction between receptor and G protein is tight. Figure 4, A and B show the concentration-dependent decrease in iodomelatonin binding by the antibody selective for the carboxy terminus Gi (1-3), which inhibited agonist binding with equivalent potency and efficacy in native and in PTX-treated membranes; the effects of other antibodies were not significant except for the Gq antibody if applied to membranes prepared from cells over-expressing Gq (~5- to 10-fold higher _q level than in vector-transfected controls; not shown); here, the Gq-selective antibody was as effective as the Gi-antibody in reducing iodomelatonin binding. Whereas overexpression of _q led to the interaction of G_q with the Mel_1a receptor, it did not cause an increment in the number of receptors labeled by iodomelatonin. This suggests that there are no spare receptors but that all of the receptors on the cell surface are capable of recruiting G proteins.

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Uncoupling of the Mel1a receptor by G protein  subunit-specific antibodies. Membranes were prepared from HEK293 cells stably expressing the Mel1a receptor before (A) and after (B) PTX treatment, and after transient transfection with the cDNA encoding for the  subunit of Gq (C). Membranes were preincubated with the indicated amounts of antibodies raised against the C terminus of Gi1-3 (), Go (), and Gq/11 (). Binding is expressed in percentage of the binding observed when instead of the specific antibodies, nonimmune IgG was included in the assay at the same amount. After a preincubation of membranes with the antibody in the presence of 0.2% OG for 15 min on ice, the binding reaction was performed at 30°C for 1 h at a final detergent concentration of less than 0.05%. Mean values ± S.E. are given from three to five binding experiments.

Precoupling of Unliganded Receptor. Although the Mel_1a receptor can interact with other G protein subtypes like G_q, this coupling is of minor importance in the formation of a high-affinity ternary complex, because the bulk of ternary complexes is sensitive to disruption by the _i-selective antibody both in control membranes and in membranes prepared from PTX-treated cells. This finding favors the interpretation that the Mel_1a receptor coupled to ADP-ribosylated G_i; however, this is rather unlikely. The alternative interpretation is based on previous observations that indicated that the interaction of an activated receptor with a G protein of the G_i/o class renders the  subunit unavailable to modification by PTX (Tsai et al., 1984). If the unliganded Mel_1a receptor was spontaneously active, this would result in precoupling of the receptor to G_i (in the absence of a receptor agonist) and hence give rise to a sustained Mel_1a R/G complex in which G_i is unavailable to PTX-mediated ADP ribosylation.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Constitutive activity of the Mel1a receptor in membranes of HEK293 cells. A, time course of [35S]GTPS binding to Mel1a membranes in the absence () and presence () of the Mel1a receptor antagonist luzindole (100 µM). The binding assay was performed at 0.175% OG in the presence of 0.3 µM GDP for the indicated intervals at 25°C. B, [35S]GTPS binding to urea-washed Mel1a membranes after reconstitution with increasing concentrations of Gi-1. [35S]GTPS binding was determined in the absence () and presence () of melatonin (0.1 µM) and in membranes that had been heat inactivated (). The binding assay was initiated after preincubation of the membranes with Gi-1 plus 0.2% OG. C, time course of [35S]GTPS binding on reconstitution of Gi-1 with urea-washed Mel1a membranes that had been () or had not been () heat inactivated. Each figure shows a typical experiment of three performed.

Mel_1a Receptor Forms a Stable R/G Complex. Because the Mel_1a receptor displayed basal activity, we conjectured that the resistance of iodomelatonin binding to PTX resulted from the formation of a tightly associated complex of the unoccupied Mel_1a receptor with G_i. Direct evidence to prove this hypothesis was obtained by two experimental approaches.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Recovery of Gi by immunoprecipitation of the Mel1a receptor. Membranes from native and PTX-treated HEK293 cells expressing the Mel1a receptor were incubated with or without melatonin (0.1 µM) for 1 h at 25°C. After solubilization with 1% digitonin, extracts were subjected to immune precipitation using antiserum-536 and protein A-agarose; the precipitates were resuspended and G proteins were dissociated with Gpp(NH)p. The supernatants were subjected to gel electrophoresis and immunoblotted using an antiserum specific for Gi protein  subunits. Membranes prepared from rat brain cortex (20 µg/lane) were applied as a source of G protein standards.

View larger version (40K): [in this window] [in a new window]   Fig. 7.   [32P]ADP ribosylation by PTX of G proteins extracted from HEK293 cell membranes. A, soluble extracts from native and PTX-treated HEK293 cells expressing either the human Mel1a or the A1-adenosine receptor were incubated with preactivated PTX using [32P]NAD as a substrate. From the soluble extracts (~1 mg protein), G proteins were quantitatively immunoprecipitated with a Gi-selective antibody and resolved by gel electrophoresis. Half of the immune precipitates, obtained from PTX-exposed (PTX+) or from nonexposed (PTX) cells, were loaded onto the gel, except in the second lane, where only the 50th part of the sample to the left was applied. Shown is a representative autoradiography of three experiments performed. B, reconstitution of the Mel1a receptor with Gi-3. Iodomelatonin binding to urea-washed Mel1a membranes was performed after reconstitution with increasing amounts of rGi-3 in the presence of a 4-fold molar excess of -dimer purified from porcine brain. This experiment was repeated twice with similar results. C, reconstitution of the Mel1a and A1-adenosine receptor with PTX- and sham-treated Gi-3 (100 nM).

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

The mode of G protein coupling of the Mel_1a receptor is at variance with the predictions of the classical ternary complex model (Hepler and Gilman, 1992), because agonist high-affinity binding is stable in the presence of guanine nucleotides both in intact cells and in cell membranes. High-affinity binding of iodomelatonin to intact cells has previously been observed in various cell types, hence is not due to an artifact arising from clonal selection of a stable cell line (Witt-Enderby and Dubocovich, 1996). Even following pretreatment of HEK293 cells with PTX, which completely abolishes agonist high-affinity binding to typical G_i/o-coupled receptors, a major proportion of the Mel_1a receptor population retains the ability to bind iodomelatonin with unaltered affinity. In the present work, we show that these unusual features arise from a highly stable association of the Mel_1a receptor with G_i.

This conclusion is based on several findings. First, high-affinity agonist binding to the Mel_1a receptor is strictly dependent on G protein coupling; uncoupling (by guanine nucleotides, suramin or G protein-directed antisera) results in a loss of high-affinity agonist binding. In membranes, guanine nucleotides destabilize this R/G complex only on the addition of detergent, which reduces the G protein affinity of the receptor (by about 20-fold at the detergent concentration employed) and enhances receptor-mediated G protein activation. Secondly, in HEK293 cells the Mel_1a receptor interacts preferentially with G_i; this is true even after treatment with PTX, because an antiserum directed against the carboxy terminus of Gi eliminates high-affinity binding in membranes from PTX-treated cells. Thirdly, the receptor is spontaneously active, a property, which, by definition, gives rise to a precoupled state. In the precoupled state, the association of the unliganded receptor with G_i is also tight and this makes the cysteine residue in the _i C terminus unavailable to PTX-mediated ADP ribosylation. The formation of the complex between unliganded Mel_1a receptor and G_i is governed by an affinity of 50 nM. This estimate represents the lower limit of the affinity because it was obtained in reconstitution experiments in the presence of detergent. Nevertheless, the complex between unliganded receptor and G protein is sufficiently stable such that it can be recovered by immunoprecipitation with an antiserum directed against the receptor.

There is an apparent discrepancy in the effects of PTX; whereas pretreatment with the toxin does not abolish high-affinity agonist binding, it completely suppresses agonist-promoted GTPS binding and effector regulation. However, the sensitivity of agonist-promoted GTPS binding is limited by the basal rate of ligand binding to the panoply of other GTP binding proteins that are present in cell membranes. This basal rate of binding is typically reduced by the addition of excess GDP. Under these assay conditions, agonist-promoted binding can, however, only be detected if the receptor acts catalytically (to activate several G proteins) or if the expression level of the receptor is very high. Pretreatment with PTX renders the bulk of G_i unavailable to the receptor leaving only stoichiometric amounts (which escape detection). More importantly, these observations also indicate that (unmodified) G_i has to be present in excess over the receptor to support efficient signaling from the Mel_1a receptor to an effector (e.g., to adenylyl cyclase inhibition).

It has long been known that an activated receptor prevents ADP ribosylation of G protein  subunits by PTX (Tsai et al., 1984). To the best of our knowledge, it has, however, not yet been documented that an unliganded receptor can effectively prevent access of PTX to Gi. The stoichiometric levels of unmodified Gi that persist after PTX treatment suffice to support high-affinity agonist binding to the Mel_1a receptor. In contrast, an abundant pool of G proteins is indispensable for productive signaling of the receptor because regulation of adenylyl cyclase or phospholipase C is abolished by PTX treatment. We stress that the marked avidity with which the Mel_1a receptor binds to G_i is not a consequence of a very high receptor expression level in HEK293 cells. First, the levels of Mel_1a receptor in our HEK293 cell clone (~0.7 pmol/mg) is not substantially higher than that observed in membranes prepared from brain nuclei (~0.3 pmol/mg; Barrett et al., 1996) if the cellular heterogeneity in these preparations is taken into account. Secondly, a similar resistance of the Mel_1a receptor to guanine nucleotide modulation has been observed in membranes prepared from various brain regions (Rivkees et al., 1989; Morgan et al., 1996), as well as in membranes from NIH3T3 cells expressing the cloned receptor (Nonno et al., 1998). Thirdly, we have previously examined the A₁-adenosine and the D₂-dopamine (G_i/o-coupled) receptors in HEK293 cells that were expressed to levels that were either lower or higher (between 0.2 and 4.0 pmol/mg) than that obtained with the Mel_1a receptor cell clone. High-affinity agonist binding to these receptors was readily suppressed at low guanine nucleotide concentrations and was completely abolished by PTX (Waldhoer et al., 1998). Finally, even a fusion protein composed of the human A₁-adenosine receptor and Gi in which the receptor is forced into close contact with the  subunit does not reproduce the features of Mel_1a receptor/G protein coupling (Waldhoer et al., 1999). We therefore conclude that the tight association of the Mel_1a receptor and Gi is caused by intrinsic properties of the receptor and that it is independent of the expression level.

Several receptors endowed with constitutive activity (e.g., the ₂-adrenergic receptor, the 5-hydroxytryptamine_2C receptor, and a mutated form of the _1B-adrenergic receptor) form ternary complexes that display limited sensitivity to guanine nucleotides (Neubig et al., 1988; Jagadeesh et al.,1990; Westphal and Sanders-Bush, 1994, Kjelsberg et al., 1992) However, a guanine nucleotide-resistant ternary complex does not necessarily result in constitutive activity of the receptor. In the avian -adrenergic receptor, for example, a stretch of the extended carboxy terminus confers guanine nucleotide resistance and restrains the spontaneous activity of the receptor; a truncated form of this receptor is constitutively active and becomes sensitive to guanine nucleotide modulation (Hertel et al., 1990; Parker and Ross, 1991). Similarly, guanine nucleotide modulation of the brain A_2A-adenosine receptor is restored by partial proteolysis of the receptor protein (Nanoff et al., 1991). Like these receptors, the Mel_1a receptor possesses an extended carboxy terminus; based on this analogy, it is attractive to speculate that the carboxy terminus is the candidate domain mediating the high stability of the R/G complex.

The level of Mel_1a receptor stably expressed in HEK293 cells is reasonably similar to the density that is expected to occur in individual hypothalamic nerve cells; hence, it is highly probable that the constitutive activity of the receptor is physiologically relevant because the neuronal activity will vary with the fluctuations in the receptor level, independently of the ambient melatonin concentration. Because the receptor expression follows a circadian rhythm (Tenn and Niles, 1993; Neu and Niles, 1997), the effect of melatonin receptor ligands in adjusting the internal clock will be different with the time of the day; in rodents, for example, melatonin or melatonin agonists advance the phase of the circadian clock only when administered during a time window preceding the onset of locomotor activity (Van Reeth et al., 1997). Based on our observations, we postulate that melatonin antagonists rather than agonists will alter the circadian rhythm at other time points, e.g., at noon, when the Mel_1a receptor density in target cells increases, and during the night, when the hormone levels rise. We also propose that inverse agonists will be effective in adjusting the circadian clock because they not only block the actions of melatonin but also eliminate signaling resulting from the spontaneous receptor activity.