Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

5-HT₇ Receptors are G protein-coupled receptors with seven transmembrane segments, (Bard et al., 1993; Ruat et al., 1993; Shen et al., 1993), and exist in three alternatively spliced isoforms (Heidmann et al., 1997; Jasper et al., 1997; Stam et al., 1997; Heidmann et al., 1998). The predominant isoform, 5-HT_7A, is localized in discrete limbic regions (thalamus and hypothalamus) of the brain as well as peripheral tissues, including coronary artery and certain gastrointestinal tissues (Meyerhof et al., 1993; To et al., 1995; Leung et al., 1996). Multiple physiological roles have been proposed for the receptor, such as regulation of circadian rhythms [because of its presence in the suprachiasmatic nucleus of the hypothalamus (Lovenberg et al., 1993; Kawahara et al., 1994)], mood and emotions [because of its limbic location and its high-affinity interactions with antipsychotics and antidepressants (Adham et al., 1998; Meyerhof et al., 1993; To et al., 1995)], and also migraine genesis [because of its smooth muscle relaxing activity (Leung et al., 1996; Cushing et al., 1996)]. At cellular level, the 5-HT_7A receptor, when expressed heterologously in mammalian cells, positively modulates adenylyl cyclases (Bard et al., 1993; Ruat et al., 1993; Shen et al., 1993; Adham et al., 1998) via activation of G_s subtypes of G proteins. We confirmed here that the human 5-HT_7A receptor as expressed heterologously in human embryonic kidney (HEK) 293 cells mediates the cholera toxin-sensitive cAMP production. At the same time, we observed that most receptors (> 60%) bound 5-HT with high affinity. Classically, high-affinity agonist binding represents the phenotype for G protein-bound receptors, but was unusual for 5-HT_7A in its marked abundance and insensitivity to 5'-guanylylimidodiphosphate (GppNHp), as observed here. In this study, we explored whether phospholipid metabolites and similar amphipathic compounds could modulate the 5-HT_7A receptor, a membrane-embedded receptor. Such possibilities were hinted at by earlier reports that oleamide and oleic acid influenced 5-HT_7A receptors expressed in HeLa cells (Hedlund et al., 1999). In addition, our preliminary experiments showed that fatty acid-free bovine serum albumin (FF-BSA), a scavenger of lipid metabolites, markedly affected binding of [³H]5-HT, but not [³H]mesulergine (antagonist), to 5-HT_7A.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

The cDNA for the human 5-HT_7A receptor has been cloned into a PCI-Neo mammalian expression vector. The vector was used for transfection of HEK293 cells, using Ca²⁺ phosphate precipitation techniques. Cells were selected for a month in the presence of G-418 (400 µg/ml). Transcripts for 5-HT_7A were robustly detected with reverse transcription-polymerase chain reaction (using 3' rapid amplification of cDNA ends reaction) from transfected but not naive cells. Cell membranes were prepared by standard procedures including homogenization and differential centrifugation as described elsewhere (Pregenzer et al., 1993). Binding of radioactive ligands was measured in membranes expressing recombinant receptors, using filtration techniques as described elsewhere (Pregenzer et al., 1993). Briefly, [³H]mesulergine and [³H]5-HT binding were measured in the medium containing 100 mM NaCl, 2 mM MgCl₂, 1 mM EDTA, 20 mM HEPES/Tris, pH 7.4, the radioactive ligand at varying concentrations (0.1 to 20 nM for typical binding profiles), and 20 µg of membrane protein, in a total volume of 500 µl at 23°C for 60 min. Reaction mixtures were filtered over Whatman GF/B filters under vacuum. Filters were washed three times with 4 ml of ice-cold 50 mM Tris/HCl buffer, pH 7.4. Nonspecific binding was estimated in the presence of excess unlabeled mesulergine (10 µM). Competition experiments with [³H]mesulergine (2 nM) were carried out in the presence of test compounds at various concentrations under the same conditions.

GTP³⁵S binding was measured with the use of the procedures reported earlier (Chabert et al., 1994; Pregenzer et al., 1997), in the medium containing 25 mM HEPES, pH 8.0, 100 mM NaCl, 1 mM EDTA, 3 mM MgCl₂, 0.5 mM dithiothreitol, 0.003% digitonin, 2 nM GTP³⁵S (5 to 3 × 10⁵ cpm/assay), and 10 µg of membrane protein in a volume of 200 µl. Digitonin was used at trace amounts here, solely to make the membranes permeable to GTP³⁵S. Test ligands were included at 10 µM, unless indicated otherwise. Membranes were preincubated with 100 µM 5'-adenylylimidodiphosphate for 30 min, 10 µM GDP for 10 min on ice, and then were added to the rest of reaction components. In some experiments, membranes were treated with N-ethylmaleimide (NEM) at 100 µM, and excess NEM was neutralized with -mercaptoethanol at the end of 30-min incubation. Reaction mixtures were incubated at 30° for 30 min and were filtered over Whatman GF/B filters under vacuum. Filters were washed three times with 4 ml of ice-cold buffer containing 100 mM NaCl, 20 mM Tris/HCl, pH 8.0, and 25 mM MgCl₂. Agonist-induced GTP³⁵S binding was obtained by subtracting that observed without agonists. The binding data were analyzed using nonlinear regression method (Sigma Plot), and presented with mean values ± S.E.

Cellular changes in cAMP were measured using a FlashPlate assay kit from PerkinElmer Life Science Products (Boston, MA). Briefly, cells were grown in a 96-well plate to about 80%confluence, washed three times with phosphate-buffered saline, and treated with test ligands for 30 min. cAMP in cell lysates were measured using the competition between ¹²⁵I-cAMP and nonradioactive antigen for a fixed number of antibody binding sites in microplates coated with solid scintillant.

GTP³⁵S-bound G subunits were identified following the method described elsewhere (Okamoto et al., 1992) with a modification (Alberts, et al., 1999): the activation of receptors with serotonin in the presence of GTP³⁵S before membrane solubilization with detergents. Briefly, membranes were incubated in the presence of GTP³⁵S (4 nM) and serotonin (10 µM) under conditions identical to those for GTP³⁵S binding as described above. Treated membranes were solubilized with an equal volume of a buffer containing 100 mM Tris/HCl, pH 8.0, 10 mM MgCl₂, 100 mM NaCl and 0.6% CHAPS for 30 min on ice, and were diluted to a final CHAPS concentration of 0.125%. An aliquot of the mixtures (typically 300 µl) was transferred to a well in a 96-well plate that had been coated successively with goat anti-rabbit antibodies (1:100 dilution), bovine serum albumin (5 mg/ml), and one of the affinity-purified rabbit antibodies for various G subunits (1:200 dilution). After incubation for 1 h at room temperature and washing, each well was counted for ³⁵S using a standard scintillation cocktail and a -counter. The antibodies we used here include those specific for G_i (the C-terminal sequence, 345-354), G_s (the C-terminal sequence, 385-394), G _q/11 (the common C-terminal sequence, QLNLKEYNLV), or G₁₃ (the sequence 367-377) from Calbiochem (San Diego, CA). The mouse monoclonal antibody raised against bovine G_o protein was obtained from Chemicon (Temecula, CA). Agonist-induced GTP³⁵S binding was computed by subtracting the level observed without test agonists. Stock solutions were prepared in 0.1% ascorbic acid for monoamines, in dimethyl sulfoxide for lysophospholipids, and in alcohol for fatty acids and oleamide. Dilutions were made in borosilicate glass tubes, and final concentrations of dimethyl sulfoxide or alcohol were less than 0.5%. CHAPS and FF-BSA were prepared in the assay buffer. Reactions were carried out in polypropylene microtubes.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Binding of [³H]5-HT and [³H]mesulergine (antagonist) to the human 5-HT_7A receptor, when expressed heterologously in HEK293 cells, fitted well to a one-site binding model (linearity) with dissociation constants (K_D) of 1.5 ± 0.1 and 4.3 ± 0.1 nM, respectively, and maximal binding values of 1.7 ± 0.2 (Fig. 1) and 2.6 ± 0.1 pmol/mg of protein, respectively (Fig. 1). With naive HEK293 cells, we observed no detectable levels of [³H]mesulergine and [³H]5-HT binding under the same conditions. We also carried out competition experiments with [³H]mesulergine at 4 nM and 5-HT at various concentrations (Fig. 1). The displacement data fitted to a model of two binding sites, high-affinity sites with a K_I value of 1.8 ± 0.3 nM (similar to the K_D for 5-HT), accounting for 64% of [³H]mesulergine binding sites, and low-affinity sites with a K_I value of 93 ± 4 nM. 5-Carboxamidotryptamine maleate (5-CT), another agonist, displayed a similar biphasic profile, with high-affinity sites with a K_I value of 1.6 ± 0.2 nM accounting for 65% of [³H]mesulergine binding sites and low-affinity sites with a K_I value of 76 ± 3 nM (Table 1). This indicates that more than 60% of the total receptors, when estimated from maximal [³H]mesulergine (an antagonist) binding, displayed the high-affinity 5-HT binding. Classically, this phenotype represents G protein-associated receptors but was unusual for 5-HT_7A in its marked abundance. Several antagonists, on the other hand, displaced [³H]mesulergine binding in monophasic patterns (e.g., methiothepin, lisuride, and metergoline, with K_I values of 0.25 ± 0.04, 0.33 ± 0.05, and 0.97 ± 0.05 nM, respectively) (Table 1). GppNHp (10 µM), a nonhydrolyzable GTP analog, often uncouples receptor-G protein interactions, but it produced no appreciable changes in the relative populations of low- and high-affinity sites for 5-HT, as shown in the biphasic displacement pattern of [³H]mesulergine by 5-HT (Fig. 1). On the other hand, GppNHp abolished high-affinity dopamine binding to the D1 dopamine receptor (a prototypic G_s coupled receptor) when expressed in the same cell line (HEK293 cells) at a similar receptor density, 2.5 ± 0.2 pmol/mg protein. Dopamine reduced [³H]SCH23390 (antagonist with a K_D value of 0.66 ± 0.05 nM) binding to D1 receptors by 12 ± 2% with a K_I value of 53 ± 16 nM, but this high-affinity site disappeared in the presence of GppNHp, leaving only low-affinity sites with a K_I value of 7576 ± 574 nM (Fig. 1).

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Comparison of high- and low-affinity binding of 5-HT to 5-HT7A and of dopamine to D1 receptors expressed heterologously in HEK293 cells. A, we analyzed [3H]5-HT or [3H]mesulergine binding to the human 5-HT7A receptor at various concentrations, using Scatchard plot. B and C, competition experiments were carried out with [3H]mesulergine at 4 nM and 5-HT at 0.1 to 1000 nM in cell membranes expressing 5-HT7A receptors and with [3H]SCH23390 at 2 nM and dopamine at 10 to 50,000 nM in cell membranes expressing D1 receptors. [3H]Mesulergine binding was blocked by 5-HT () in a biphasic manner (solid line) with a KI value of 1.8 ± 0.3 (high affinity) accounting for 64% of the total [3H]mesulergine binding and with a KI value of 93 ± 4 nM (low affinity). Treatment with 10 µM GppNHp did not alter the biphasic profile (). [3H]SCH23390 binding was inhibited by dopamine in a biphasic manner (solid line) with a KI value of 53 ± 16 (high affinity) accounting for 12% of the [3H]SCH23390 binding and with a KI value of 7576 ± 574 nM (low affinity). Treatment with 10 µM GppNHp abolished high-affinity binding. Now, the data () fitted to a one-site binding model with a KI value of 6237 ± 448 nM. The data represent mean ± S.E.M (n = 3).

View this table: [in this window] [in a new window]   TABLE 1 Comparison of binding and functional properties of serotonergic ligands at the human 5-HT7A receptor heterologously expressed in HEK293 cells Competition binding experiments using 2 nM [3H]mesulergine were carried out with several serotonergic ligands at various concentrations in HEK293 cell membranes expressing 5-HT7A receptors. Only displacement data for 5-HT and 5-CT were fitted to a two-site binding model; the rest were fitted to a one-site binding model. The values in parentheses represent the relative fraction of high- and low-affinity sites. The IC50 values from dose-response profiles were converted to KI (Cheng and Prusoff, 1973). Intrinsic efficacy for ligands was estimated from an increase in cellular cAMP in intact cells or in 2 nM GTP35S binding in isolated membranes and normalized for that obtained with 10 µM serotonin. The data represent mean ± S.E.M. (n = 3).

We also examined agonist-induced GTP³⁵S binding to G subunits. At 5-HT_7A, 10 µM 5-HT-induced GTP³⁵S binding increased as a function of GTP³⁵S concentrations (0.5 to 20 nM), with an EC₅₀ value of 2.2 ± 0.8 nM and maximal binding of 270 ± 32 fmol/mg of protein (Fig. 2). Similar values were observed with dopamine-activated D1 dopamine receptors (2.6 ± 0.3 nM and 330 ± 53 fmol/mg of protein) (Fig. 2). Methiothepin (100 µM) by itself produced no effect on the basal level, but blocked 1 µM 5-HT-induced GTP³⁵S binding at 5-HT_7A (24 ± 4% over the basal) whereas 100 µM SCH23390 blocked dopamine-induced GTP³⁵S binding at D1 (Fig. 2). Moreover, at 5-HT_7A, the methiothepin-sensitive GTP³⁵S binding was primarily associated with G_s, as detected with the immobilization method with various G-specific antibodies (Okamoto et al., 1992). The relative levels of GTP³⁵S association with the antibodies specific for G_s, G_i, G_o, G_q, and G₁₃ were 24 ± 3, 5 ± 2, 1 ± 1, 4 ± 4, and 0 ± 1%, respectively, above the basal binding (no 5-HT in anti-G_s-coated wells) (Fig. 2).

View larger version (12K): [in this window] [in a new window]   Fig. 2.   Agonist-induced GTP35S binding in HEK293 cell membranes expressing the human 5-HT7A or dopamine D1 receptors, and immobilization of GTP35S-bound G subunits with various G-selective antibodies. A. Enhancement of GTP35S binding by 5-HT or dopamine at 10 µM (a saturating concentration) was measured as a function of the concentration of GTP35S ranging from 1 to 20 nM. Solid lines represent data fitted to a single hyperbolic rectangular, using Sigma Plot (See Text). B. The 5-HT (1 µM)-induced GTP35S (2 nM) binding at 5-HT7A was blocked by its antagonist, methiothepin (100 µM), and was not affected by NEM, a selective inhibitor of Gi/Go subtypes of G proteins. NEM alone decreased the basal binding by 21%. The dopamine-induced GTP35S (2 nM) binding at D1 was blocked by its antagonist, SCH23390 (100 µM). C. Agonist-induced association of GTP35S with G subunits was monitored upon immobilization with various G-specific antibodies. Membranes were treated with 5-HT (10 µM) in the presence of GTP35S at 2 nM for 30 min, and then solubilized with CHAPS at 0.3%. Solubilized G subunits were immobilzed with various G-specific antibodies. The data presented as percent increase from the basal value observed without agonist treatment, and are the mean ± S.E.M (n = 3).

Generally, G_i/G_o subtypes are the most abundant and active cellular G proteins contributing to GTP³⁵S binding. To evaluate independently the potential involvements of G_i/G_o subtypes at 5-HT_7A, membranes were treated with 100 µM NEM, an established, selective inhibitor of G_i/G_o subtypes (Winslow et al., 1987; Nakajima et al., 1990; Alberts et al., 1999). NEM treatment decreased the basal GTP³⁵S (2 nM) binding by 21 ± 3, but with no appreciable effect on the net 5-HT-induced GTP³⁵S binding (Fig. 2). This further supports the primary association of 5-HT_7A with G_s but not G_i/G_o subtypes.

We examined the effects of lipid metabolites and similar amphipathic compounds on the 5-HT_7A receptor, because such agents have been reported to allosterically modulate various membrane receptors (Koenig and Martin, 1992; Hedlund et al., 1999). FF-BSA, a scavenger of lipid metabolites, produced a noticeable, concentration-dependent reduction in [³H]5-HT binding, with little effect on [³H]mesulergine (antagonist) binding to 5-HT_7A (Fig. 3). At 3%, FF-BSA decreased [³H]5-HT (2 nM) binding by 46 ± 4%. Scatchard analysis showed a decrease in maximal 5-HT binding from 1.7 ± 0.2 to 1.1 ± 0.1 pmol/mg of protein, with no appreciable change in its K_D value (1.2 ± 0.2 nM; Table 2 and Fig. 4). Among fatty acids tested, only oleic acid at concentrations less than 15 µM increased 2 nM [³H]5-HT binding by 24 ± 5% without affecting [³H]mesulergine binding. Oleic acid at 15 µM increased maximal [³H]5-HT binding from 1.7 ± 0.2 to 2.2 ± 0.2 pmol/mg of protein but showed no effect on its K_D value (1.3 ± 0.2 nM) (Fig. 4). At higher concentrations, however, oleic acid gradually inhibited both [³H]5-HT and [³H]mesulergine (4 nM) binding. This could arise from its ability to disturb ligand binding sites as an amphipathic compound. Stearic acid at 30 µM showed no effects, but at higher concentrations, it decreased both [³H]5-HT and [³H]mesulergine binding (Fig. 3). Similar profiles were observed with palmitic, arachidonic, and myristic acids (data not shown). It seems that FF-BSA could decrease [³H]5-HT binding by two ways, scavenging lipid metabolites such as oleic acid or possible protein-protein interactions with receptors. When fatty acid binding sites at FF-BSA were saturated with innocuous stearic acid (10 µM), FF-BSA lost its action on [³H]5-HT binding (Fig. 3). This indicates that the FF-BSA action may arise from its scavenging of stimulatory lipid metabolites. Furthermore, oleic acid and FF-BSA produced no appreciable effects on [³H]5-HT binding to analogous 5-HT receptors (i.e., gorilla 5-HT_1D and human 5-HT_2C receptors) when expressed heterologously in HEK293 cells. In the presence of oleic acid at 15 µM, for example, [³H]5-HT binding to 5-HT_1D and 5-HT_2C was 104 ± 4 and 100 ± 4% of control, respectively; with FF-BSA at 3%, binding was 97 ± 2 and 104 ± 5%, respectively.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Effects of FF-BSA and various lipid amphipathic compounds on [3H]5-HT and [3H]mesulergine binding to the 5-HT7A receptor expressed in HEK293 cells. Binding of 2 nM [3H]5-HT and 4 nM [3H]mesulergine to the 5-HT7A receptor in isolated membranes was measured in the presence or absence of the indicated amphipathic compounds at the concentration range from 1 to 100 µM, using filtration techniques as described under Materials and Methods. Nonspecific binding was measured in the presence of excess mesulergine (10 µM), and used to compute specific binding, which constitutes more than 90% of the total binding. Relative changes in the specific binding of [3H]5-HT and [3H]mesulergine were obtained upon normalization to control values. A, effects of FF-BSA (solid lines) and 10 µM stearate-saturated FF-BSA (dotted lines); B, oleic acid; C, stearic acid; D, oleamide; E, CHAPS; F, lysophosphatidyl choline; G, lysophosphatidyl glycerol; H, lysophosphatidyl ethanolamine. The data represent the mean ± S.E.M. (n = 3).

View this table: [in this window] [in a new window]   TABLE 2 Effect of various amphipathic compounds on [3H]5-HT binding, cAMP production, and GTP35S Binding at human 5-HT7A receptors expressed in HEK293 cells Cellular cAMP changes were monitored using a FlashPlate assay kit from PerkinElmer in HEK 293 cells in a 96-well plate. A standard curve with various concentrations of cAMP was constructed, following the vendor's protocol. The basal levels of cAMP were less than 0.2 pmol/well and were not appreciably affected by addition of the listed compounds. The 5-HT-induced GTP35S binding was measured in isolated HEK 293 cell membranes in the presence or absence of 10 µM 5-HT and 2 nM GTP35S. Methiothepin at 100 µM blocked 5-HT-induced GTP35S binding and by itself showed no effect. The data represent the mean ± S.E.M. (n = 3).

View larger version (13K): [in this window] [in a new window]   Fig. 4.   Effects of FF-BSA and CHAPS on 5-HT binding parameters, 5-HT-induced cAMP production and GTP35S binding. A, Scatchard analysis of binding data for [3H]5-HT at various concentrations in cell membranes expressing 5-HT7A receptors without () or with FF-BSA (3%) () or oleic acid (15 µM) (). Binding experiments were carried out with filtration techniques at room temperature. [3H]5-HT concentration varied from 0.25 to 10 nM. Nonspecific binding was measured in the presence of excess mesulergine (10 µM) and was used to compute specific binding. The solid lines represent linear regression analysis using Sigma Plot (see Table 2 for parameters). B, 5-HT-induced cAMP production was measured in intact cells using FlashPlate assay kit from PerkinElmer without () or with FF-BSA (3%) or oleic acid (15 µM). Cells were treated with 5-HT from 0.5 to 500 nM for 30 min in the presence of 3-isobutyl-1-methylxanthine, an inhibitor of phosphodiesterase. The data fitted to the equation of single rectangular hyperbola from Sigma plot (solid lines). Maximal cAMP accumulation reached about 8 to 9 pmol/well in a 96-well plate. C. 5-HT (10 µM)-induced 2 nM GTP35S binding was measured without or with FF-BSA (3%) or oleic acid (15 µM) and normalized to control. The data represent the mean ± S.E.M. (n = 3).

It has been reported that oleamide and oleic acid at 0.1 µM decrease allosterically the affinity of [³H]5-HT for rat 5-HT_7A receptors by 2- to 3-fold in HeLa cells, without affecting maximal binding (Hedlund et al., 1999). Here oleamide and oleic acid at the concentration range 0.01 to 1 µM showed no appreciable effects on [³H]5-HT and [³H]mesulergine binding to human 5-HT_7A receptors in HEK293 cells. At micromolar concentrations, oleic acid (<15 µM) stimulated [³H]5-HT binding in HEK293 cells, and oleamide (>50 µM) blocked gradually both [³H]5-HT and [³H]mesulergine binding (Fig. 3). These differences in receptor sensitivity could be ascribed to various factors, including species variations in rat and human receptors and possibly different surrounding lipid environments in the HeLa and HEK293 cells. In fact, the rat 5-HT_7A in HeLA cells displayed a K_D value for 5-HT (12.68 nM) nearly 10-fold greater than the human counterpart in HEK293 cells (1.5 nM) (Hedlund et al., 1999).

We also studied lysophospholipids, which are also generated from hydrolysis of phospholipids by phospholipases. Various lysophospholipids (oleoyl and stearyl) at 15 µM or less showed no selective effects on [³H]5-HT or mesulergine binding. Even at higher concentrations (50 µM), only lysophosphatidyl choline decreased selectively [³H]5-HT binding (no effect on [³H]mesulergine binding) by reducing maximal [³H]5-HT binding from 1.7 ± 0.2 to 1.2 ± 0.2 pmol/mg of protein, but not its K_D value (1.4 ± 0.1 nM) (Table 2). Lysophosphatidyl glycerol decreased both [³H]5-HT and [³H]mesulergine binding (Fig. 3), and similar profiles were observed with lysophosphatidyl inositol and lysophosphatidyl serine (data not shown). Lysophosphatidyl ethanolamine and lysophosphatidic acid at concentrations up to 100 µM marginally inhibited [³H]5-HT and mesulergine binding, with decreases of less than 20%.

Another nondenaturing zwitterionic detergent, CHAPS at 0.2% or less, markedly increased [³H]5-HT binding, but marginally decreased [³H]mesulergine binding (Fig. 3). In the presence of 0.1% CHAPS, maximal [³H]5-HT binding increased from 1.7 ± 0.2 to 2.6 ± 0.1 pmol/mg of protein, with no appreciable change in its K_D value (1.4 ± 0.1 nM compared with 1.5 ± 0.1 nM). For [³H]mesulergine binding, 0.1% CHAPS showed no effect on maximal binding (2.5 ± 0.2 compared with 2.6 ± 0.1 pmol/mg of protein) but increased its K_D value from 4.3 ± 0.1 to 6.9 ± 0.4 nM. The opposite effects by lysophaphatidyl choline and CHAPS, two similar zwitterions, suggest that their actions arise from specific interactions with receptors, not from general membrane perturbations.

Functionally, these agents also affected 5-HT-induced GTP³⁵S binding. FF-BSA (3%) and 50 µM lysophosphatidyl choline reduced 10 µM 5-HT-induced GTP³⁵S binding to 56 ± 4 and 55 ± 10%, respectively, as normalized to that observed with 10 µM 5-HT, whereas oleic acid and CHAPS increased the GTP³⁵S binding to 137 ± 5 and 133 ± 7%, respectively. In the analogous Gs-coupled D1 receptor, however, 15 µM oleic acid, 3% FF-BSA, 50 µM lysophosphatidyl choline, and 0.1% CHAPS produced no appreciable effects on 10 µM dopamine-induced, SCH-23390-sensitive GTP³⁵S binding (96 ± 28, 95 ± 3, 83 ± 16, and 109 ± 22%, respectively).

We also examined how these agents affect cAMP accumulation at 5-HT_7A in the HEK293 cell. Typically, 5-HT increased cAMP production in a concentration-dependent manner with an EC₅₀ value of 8.1 ± 0.7 nM and maximal stimulation of 9 ± 1 pmol/well in a 96-well plate (Fig. 4). With cholera toxin treatment (5 µg/ml culture for overnight), the basal level of cAMP increased from 0.1 to 4 pmol/well, but 5-HT produced no responses; the covalent modification of G_s subtypes by cholera toxin (ADP-ribosylation) rendered them into permanently activated states (Milligan, 1988). FF-BSA at 3% and lysophosphatidyl choline at 50 µM showed no effects on the basal cAMP production (<0.2 pmol/well) but increased the 5-HT concentration (EC₅₀) from 8.1 to 13.9 ± 1.2 and 15.8 ± 1.6 nM, respectively, without affecting maximal cAMP production (9.1 ± 0.8 and 8.3 ± 0.3 pmol/well, respectively; Fig. 4 and Table 2). On the other hand, 15 µM oleic acid and 0.1% CHAPS decreased the EC₅₀ value from 8.1 ± 0.7 to 3.7 ± 0.4 and 4 ± 1 nM, respectively, but with little effect on maximal production (8.4 ± 0.3 and 9 ±1 pmol/well, respectively) and on basal cAMP production.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The human 5-HT_7A receptor is a G_s-coupled receptor; when expressed heterologously in HEK293 cells, more than 60% of its population displayed high-affinity [³H]5-HT binding. We discovered here that the high-affinity agonist binding decreased in the presence of FF-BSA and increased in the presence of oleic acid without appreciable effects on antagonist binding. The actions of FF-BSA and oleic acid seem to be highly selective for 5-HT_7A, judging from the following observations. 1) FF-BSA and oleic acid showed no appreciable effects on [³H]5-HT binding to analogous 5-HT receptors (5-HT_1D and 5-HT_2C). 2) Structural analogs of oleic acid, such as palmitic, stearic, and arachidonic acids, were not effective. 3) FF-BSA lost its action when preincubated with inactive 10 µM stearic acid, supporting its role in scavenging endogenous lipid metabolites rather than protein-protein interaction. 4) FF-BSA and oleic acid reduced maximal binding for 5-HT, not the K_D value, indicating their selective action at receptors. 5) Oleic acid was effective at concentrations (<15 µM) too low to form micellar structures, and its action was not mimicked by lysophospholipids, another class of amphipathic lipid metabolites. 6) In general, the human 5-HT_7A receptor seems to be susceptible to allosteric modulations by amphipathic agents. For example, lysophosphatidyl choline decreased maximal 5-HT binding for 5-HT (not antagonist binding), and CHAPS, a similar zwitterion, increased it. From these results, we propose that endogenous amphipathic compounds, probably including oleic acid and others that remain unidentified, may allosterically modulate 5-HT_7A to increase high-affinity 5-HT binding.

In addition, the amphipathic compounds at concentrations that selectively affected [³H]5-HT binding also influenced 5-HT-induced GTP³⁵S binding at 5-HT_7A. Oleic acid and CHAPS increased and FF-BSA and lysophosphatidyl choline decreased 5-HT-induced GTP³⁵S binding. No similar actions by these agents at D1 (a prototypic G_s-coupled receptor) also support the view of their specific interactions with 5-HT_7A, but not with G proteins in the signaling pathways.

In this study, we also observed that GppNHp produced no appreciable effect on the high-affinity agonist binding to 5-HT_7A but abolished high-affinity dopamine binding to D1. Two possibilities can be cited: 1) At 5-HT_7A, a certain portion could be sensitive to GppNHp but not detectable because the overwhelming majority of high-affinity binding was insensitive to the guanine nucleotide. 2) 5-HT_7A receptors could form a very tight complex with Gs that could not be destabilized by GppNHp. A similar situation has been proposed for Mel_1a-metatonin receptor and G_i (Roka et al., 1999). Further study will be needed.

In summary, endogenous amphipathic lipid metabolites that could be scavenged by FF-BSA, such as oleic acid, seem to allosterically modulate 5-HT_7A to enhance its high-affinity agonist binding, at least in part, and to confer more efficient coupling with G_s subtypes of G proteins. Their interactions with 5-HT_7A seem to be specific, judging from the narrow structural requirement for modulators (oleic acid and CHAPS) and their inactivity at analogous 5-HT and dopamine receptors.