Human 5-Hydroxytryptamine_5A Receptors Activate Coexpressed G_i and G_o Proteins inSpodoptera frugiperda 9 Cells

Experimental Procedures

Materials.

Sf9 insect cells were obtained from Invitrogen (Groningen, The Netherlands). The baculovirus transfer vector pAcGP67A and the BaculoGold DNA were purchased from PharMingen (San Diego, CA). The transfer vector pBacPAK9 was obtained from Clontech Laboratories (Palo Alto, CA). [³H]5-Carboxamidotryptamine (5-CT; 50–100 Ci/mmol), [³⁵S]GTPγS (>1000 Ci/mmol), and the chemiluminescent Western detection kit (ECL-Plus) were purchased from Amersham Pharmacia Biotech (Little Chalfont, UK). 5-HT, 5-methoxytryptamine (5-MT), and dihydroergotamine (DHE) were purchased from Acros Organics (Geel, Belgium). Lysergic acid diethylamide (LSD) was obtained from Kenija Industriji (Yugoslavia). 5-CT was obtained from Research Biochemicals Inc. (Natick, MA). Methiothepin was purchased from Hoffman-La Roche (Basel, Switzerland). Pargyline was purchased from Sigma-Aldrich (St. Louis, MO). Grace's supplemented insect cell culture medium, Sf-900 II serum-free insect cell culture medium, and antibiotic/antimycotic solution were obtained from Life Technologies (Gaithersburg, MD). Fetal bovine serum was purchased from BioWhittaker (Walkersville, MD). The protein assay kit and the protein molecular weight marker were obtained from Bio-Rad Laboratories (Hercules, CA). Guanosine-5′-(β,γ-imido)triphosphate (Gpp(NH)p) and GDP were obtained from Boehringer-Mannheim (Mannheim, Germany). The anti-Gα_i/o/t/z/s rabbit antiserum was purchased from Calbiochem (La Jolla, CA). The rabbit antisera for Gα_q/11, Gα₁₂, and Gα₁₃ were obtained from Chemicon International (Temecula, CA). The goat antiserum for Gα₁₆ was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The peroxidase-conjugated anti-rabbit and anti-goat secondary antibodies were obtained from Jackson ImmunoResearch Laboratories (West Grove, PE).

5-HT, 5-CT, and 5-MT were dissolved and diluted in assay buffer. DHE, LSD, and methiothepin were dissolved and diluted in DMSO; the last 20-fold dilution step was performed in assay buffer. The dilution in the assay mixture was 10-fold. In all control assays, DMSO was added to a final concentration of 0.5%.

Baculoviruses containing cDNA-encoding rat Gα_i1, Gα_i2, Gα_i3, and Gα_o subunits were gifts from Dr. J. Garrison (University of Virginia) (Graber et al., 1992). The baculovirus for human Gα_z was a gift from Dr. D. Manning (University of Pennsylvania) (Butkerait et al., 1995). Baculoviruses for bovine Gα_s-short-B, mouse Gα_q, mouse Gα₁₁, and human Gα₁₆ were gifts from Dr. A. Gilman and Dr. T. Kozasa (University of Texas) (Hepler et al., 1993; Kozasa et al., 1993;Linder et al., 1993). Baculoviruses for mouse Gα₁₂ and Gα₁₃ were gifts from Dr. D. Dhanasekaran (Temple University, PA). The bovine Gβ₁γ₂ transfer vector was a gift from Dr. T. Haga (University of Tokyo, Japan) (Nakamura et al., 1995).

Cloning of h5-ht_5A Receptor cDNA.

The coding region of the human 5-ht_5A receptor was amplified from a QuickScreen cDNA library (Clontech) by polymerase chain reaction using primers 5′-GCGATATGGACCCAGAGATGGATTTACCAGTGAACC-3′ and 5′-GCCTCGAGCCTCAGTGTTGCCTAGAAAAGAAGTTCTTG-3′. The inclusion of restriction sites (EcoRV and XhoI) within the oligonucleotide primers allowed cloning of the polymerase chain reaction fragment into the pcDNA3 vector (Invitrogen). The sequence of the insert was identical to that reported by Hurley et al. (1998) and contained a single silent mutation (T to C at nucleotide 300, counting from the A of the start codon), compared with the sequence deposited in the GenBank/EMBL database (accession numbers X81411 and X81412) by Rees et al. (1994).

Construction of Recombinant Transfer Vector.

The h5-ht_5A cDNA clone in pcDNA3 was digested withPstI, blunt-ended with Klenow DNA polymerase, and digested with XbaI, yielding a 1145-bp fragment encoding the h5-ht_5A receptor. This fragment was subcloned into the BamHI (filled in with Klenow DNA polymerase) andXbaI positions of the multiple cloning site of the baculovirus transfer vector pAcGP67A, such that the gp67 signal sequence was fused in frame to the N terminus of the h5-ht_5A coding sequence via a nine-amino acid linker sequence (gp67-ADRCDMDPE-h5-ht_5A). For the pBacPAK9-based transfer vector, the h5-ht_5A cDNA was excised from the pcDNA3 clone by digestion with EcoRI and XhoI. The 1142-bp fragment encoding the h5-ht_5A receptor was subcloned into the multiple cloning site of the transfer vector pBacPAK9 that was digested with the same restriction enzymes. Protein expression was under control of the polyhedrin promoter in both transfer vectors. The DNA insert sequences were confirmed by sequencing both strands of the double-stranded DNA.

Generation of Recombinant Baculoviruses.

Transfer of the h5-ht_5A receptor cDNA into the wild-typeAutographa californica nuclear polyhedrosis virus genome was accomplished by homologous recombination. Sf9 insect cells were cotransfected with linearized modified A. californicanuclear polyhedrosis virus baculovirus DNA (BaculoGold) and the h5-ht_5A-containing recombinant transfer vector using standard techniques (O'Reilly et al., 1992). Purification of recombinant viruses, amplification of purified virus stocks, and determination of virus titers were performed as described by O'Reilly et al. (1992).

Insect Cell Culture and Baculovirus Infection.

Sf9 cells were grown at 27°C and at an ambient atmosphere in suspension culture using spinner flasks or in monolayers. For viral stock production, Grace's insect cell culture medium was used supplemented with 10% fetal bovine serum, 0.2 mMl-glutamine, and 1% antibiotic/antimycotic solution, whereas Sf-900 II serum-free insect cell culture medium, supplemented with 0.2 mM l-glutamine and 1% antibiotic/antimycotic solution, was used in recombinant protein expression experiments. Cell viability was determined by trypan blue staining. Cells (50–500 ml) at a density of 1 × 10⁶ cells/ml (log phase growth) were infected with a h5-ht_5A receptor-encoding baculovirus at a multiplicity of infection (m.o.i.) of 2 (unless stated otherwise), with a Gβ₁γ₂-encoding virus (m.o.i. = 1) and/or with a Gα-encoding virus (m.o.i. = 2–4). For the expression of single Gα subunits, the m.o.i. was 4 for any Gα baculovirus, whereas for the expression of multiple Gα subunits (Gα_i1, Gα_i2, Gα_i3, and Gα_o, abbreviated as Gα_i/o) the m.o.i. was 2 for each virus. At 48 h postinfection, cells were harvested by centrifugation (10 min at 2000g at 4°C), washed with ice-cold PBS, and stored at −80°C or used directly for membrane preparation.

Membrane Preparation and Determination of Protein Content.

Harvested Sf9 cells were washed with ice-cold 50 mM Tris-HCl buffer, pH 7.4; resuspended in hypotonic 10 mM Tris-HCl buffer, pH 7.4; and homogenized with an UltraTurrax homogenizer (Janke and Kunkel, Staufen, Germany) for 5 s. The homogenate was centrifuged at 30,000g for 20 min at 4°C. The membrane pellet was resuspended in 50 mM Tris-HCl buffer, pH 7.4, containing 10% glycerol and stored in aliquots at −80°C. Protein content in membrane preparations was estimated with the Bradford protein assay (Bradford, 1976), using the Bio-Rad kit. BSA was used as a standard.

Immunoblot Analysis.

Membrane protein (1, 4, or 10 μg) was incubated in 62.5 mM Tris-HCl buffer, pH 6.8, containing 10% glycerol, 5% SDS, and 0.01% bromophenol blue at 37°C for 2 h. Proteins were separated by SDS-polyacrylamide gel electrophoresis and were transferred to polyvinylidene-difluoride membranes, using standard techniques. Immunodetection of Gα subunits was performed with 1:1000 dilutions of the Gα_i/o/t/z/s, Gα_q/11, Gα₁₆, Gα₁₂, and Gα₁₃antisera. The peroxidase-conjugated anti-rabbit and anti-goat secondary antibodies were diluted 1:5000. Bands were visualized by chemiluminescence using the ECL-Plus detection kit.

Radioligand Binding.

[³H]5-CT binding experiments were performed essentially as described previously (Francken et al., 1998). Briefly, 6 μg of membrane protein was diluted in 50 mM Tris-HCl buffer, pH 7.4, containing 10 mM MgCl₂, 1 mM EGTA, and 10 μM pargyline and incubated with [³H]5-CT for 1 h at 25°C in a volume of 0.5 ml. Nonspecific binding was estimated in the presence of 10 μM methiothepin. Reactions were terminated by rapid filtration through glass fiber (GF/B) filters (Whatman, Kent, UK) presoaked in 0.1% polyethyleneimine using a Brandel (Gaithersburg, MD) 96-sample harvester. Filters were washed twice, and filter-bound radioactivity was counted in a liquid scintillation spectrometer (Tricarb) using scintillation fluid (Ultima Gold MV; Packard Instrument Company, Meriden, CT). For radioligand concentration-binding isotherms, 12 concentrations of [³H]5-CT, in a range of 0.1 to 25 nM, were used. Competition binding experiments were performed using 2 nM [³H]5-CT; compounds were added at 7 to 12 concentrations.

[³⁵S]GTPγS Binding.

[³⁵S]GTPγS binding experiments were performed as previously described (Francken et al., 1998). Briefly, 12 μg of membrane protein was diluted in 50 mM Tris-HCl buffer, pH 7.4, containing 50 mM NaCl, 10 mM MgCl₂, 1 mM EGTA, 0.1 mM dithiothreitol, 10 μM pargyline, and 1 μM GDP and preincubated with compound for 30 min at 30°C in a volume of 0.45 ml. Then, 50 μl of [³⁵S]GTPγS in assay buffer was added to a final concentration of 0.2 nM, and the assay mixtures were further incubated for 30 min at 30°C. Reactions were terminated by rapid filtration through GF/B filters, presoaked in assay buffer, using a 40-well manual filtration manifold or a Brandel 48-sample harvester. Filters were washed twice, and filter-bound radioactivity was counted in a liquid scintillation spectrometer. Basal [³⁵S]GTPγS binding was measured in the absence of compound. Compounds were added at 9 to 11 concentrations. Nonspecific [³⁵S]GTPγS binding, as measured in the presence of 100 μM GTPγS, did not exceed 10% of basal binding and was never subtracted from experimental data.

Data Analysis.

Radioligand concentration-binding isotherms (rectangular hyperbola) were calculated by nonlinear regression analysis according to algorithms described by Oestreicher and Pinto (1987), and sigmoidal inhibition curves were calculated by nonlinear regression using the Prism program (GraphPad Software, San Diego, CA).B_max and K_dvalues of the radioligand and IC₅₀ values of inhibitors were derived from the curve fitting.

Stimulation of [³⁵S]GTPγS binding was calculated as 100 times the difference between stimulated and basal binding (in cpm) divided by the amount of basal binding (in cpm). Agonist concentration-response curves and antagonist inhibition curves were analyzed by nonlinear regression using GraphPad Prism. EC₅₀ and IC₅₀ values were derived from the curves. IC₅₀ values were corrected as follows: corrected IC₅₀(IC₅₀-corr) = IC₅₀/{1 + [5-HT]/EC₅₀(5-HT)}. Relative maximum stimulation (E_max) values were calculated as percentage of the maximum stimulation obtained with 10 μM 5-HT, and relative maximum inhibition (I_max) values were calculated as percentage of the inhibition from maximum 5-HT (10 μM)-stimulated [³⁵S]GTPγS binding to basal level.

Statistical F tests and Student's t tests were performed, and all figures were prepared using GraphPad Prism.

Results

Expression of h5-ht_5A Receptors and G Protein Subunits in Sf9 Insect Cells.

The h5-ht_5A receptor coding sequence was cloned from a cDNA library, and recombinant baculoviruses were generated and used to infect Sf9 cells. In preliminary [³H]5-CT concentration-binding experiments on membranes of Sf9 cells infected at an m.o.i. of 3, higher expression levels were found for virus generated with the pAcGP67A transfer vector (B_max = 63 ± 11 pmol/mg protein, K_d = 10.1 ± 2.0 nM, mean ± S.D., n = 5) than for pBacPAK9-based virus (B_max = 23 ± 2 pmol/mg protein,K_d = 5.6 ± 1.0 nM, n= 3), probably due to the presence of the gp67 signal sequence. No specific [³H]5-CT binding could be detected to membranes of uninfected or wild-type baculovirus-infected cells (data not shown). Further expression experiments were performed with the pAcGP67A-based virus at an m.o.i. of 2.

The effect of coexpression of various G protein subunits (m.o.i. = 1–4) on the affinity of [³H]5-CT for the h5-ht_5A receptor was determined using [³H]5-CT concentration-binding experiments. Examples of [³H]5-CT saturation curves are presented in Fig. 1. Table1 summarizes the meanK_d and B_maxvalues and shows the binding data for h5-ht_5A-HEK 293 cell membranes for comparison (Francken et al., 1998). All [³H]5-CT concentration-binding isotherms were best fitted to a one-binding-site model compared with a two-binding-site model, using nonlinear regression (F test,P > .05). Receptors expressed alone in Sf9 cells yielded a B_max value of 39 ± 12 pmol/mg protein and a K_d value of 7.8 ± 0.9 nM, which is a K_d value similar to that of the low-affinity form of the receptor in h5-ht_5A-HEK 293 cells. Coexpression of receptors with Gβ₁ and Gγ₂(Gβ₁γ₂ baculovirus, m.o.i. = 1) resulted in a slight, but statistically significant, increase in [³H]5-CT affinity (Student'st test, P < .05) (Fig. 1A, Table 1), suggesting improved coupling of the recombinant receptors to endogenous G proteins. Coexpression of h5-ht_5A receptors with Gα_i1, Gα_i2, or Gα_i3 (m.o.i. = 4) or with a mixture of Gα_i1, Gα_i2, Gα_i3, and Gα_o (further designated as Gα_i/o; m.o.i. = 2 for each virus) also significantly increased [³H]5-CT affinity (Student's t test, P < .05) (Fig. 1B), whereas no effect was observed with Gα_o, Gα_z, Gα_s, Gα₁₁, Gα₁₆, Gα₁₂, or Gα₁₃ subunits (Table 1). The small, but statistically significant, increase in [³H]5-CT affinity that was observed on coexpression with Gα_q is considered as a spurious finding, considering the lack of a Gpp(NH)p effect on agonist binding (Table 1, see Fig. 4). When Gβ₁γ₂ subunits (m.o.i. = 1) were expressed in addition to Gα subunits and receptors, [³H]5-CT affinities further increased for Gα_i/o, Gα_i1, Gα_i2, Gα_i3, and Gα_o (Student's t test,P < .05) (Fig. 1B) but not for Gα_z, Gα_s, Gα_q, Gα₁₁, Gα₁₆, Gα₁₂, or Gα₁₃ (Table 1).

The expression of recombinant Gα subunits was verified by immunoblot analysis using commercially available antibodies. Figure2 shows immunoblots for the membranes ofSf9 cells expressing receptor and mammalian G protein trimers. For the different Gα proteins, immunoreactivity was demonstrated in the respective experimental membranes. It should be noted that using the same antiserum directed against a common peptide sequence, the immunoreactivity for Gα_o was much stronger than that for the Gα_i subunits, suggesting higher Gα protein expression levels.

In the presence of the individual G_i or G_o trimers, the affinity of [³H]5-CT was intermediate to that of the high- and low-affinity forms of the receptor in stably transfected HEK 293 cells (Francken et al., 1998). Only with the simultaneous expression of a mixture of G_i and G_oproteins (G_i/o) did the affinity of [³H]5-CT equal that for the high-affinity state of the receptor, probably due to the occurrence of a lower receptor-to-G protein ratio. Indeed, the infection of Sf9 cells with baculoviruses encoding each of the four G protein subtypes resulted in a relatively low level of receptor binding sites (Table 1), and the high m.o.i. for the Gα protein-encoding baculoviruses implicates an increased overall number of G proteins (Fig. 2). Alternatively, coexpression with the mixture of G_i/o proteins might mimic a more natural situation, in which the receptor is able to interact with all of the used G protein subtypes. It should be noted that a decrease in receptor number was not systematically observed on coexpression with G protein subunits. For example, coexpression with Gα_qand Gβ₁γ₂ resulted in aB_max value that was higher than when the receptor was expressed alone (Table 1). Differences in receptor expression levels between similar experiments were also observed by other groups (Butkerait et al., 1995) and are difficult to explain.

Pharmacological Characterization of h5-ht_5A Receptors Expressed Alone or Coexpressed with G Protein Subunits.

Various 5-HT receptor ligands were used to inhibit [³H]5-CT binding to membranes of baculovirus-infected Sf9 insect cells. pIC₅₀ values were derived from inhibition curves and are summarized in Table 2. The pharmacological profile of h5-ht_5A receptors expressed alone in Sf9 cells was different from that in stably transfected h5-ht_5A-HEK 293 cells; the agonists 5-CT, 5-HT, and 5-MT had 3.2- to 3.6-fold lower affinities (Student's t test, P < .05), whereas DHE, LSD, and methiothepin had slightly higher affinities for the receptor expressed alone in Sf9 cells. The rank order of potency of the tested compounds was LSD > methiothepin > 5-CT > DHE > 5-HT > 5-MT. This profile did not change on coexpression with Gβ₁γ₂subunits, although agonist affinities were slightly, but never significantly, increased.

The simultaneous expression of receptor and Gβ₁γ₂ together with Gα_i1, Gα_i2, Gα_i3, Gα_o, or the mixture of Gα_i/o subunits, but not together with G_z or G_s, resulted in an increase in the agonist affinities; the pIC₅₀values were very similar to those for h5-ht_5A-HEK 293 membranes (Student's t test, P > .05) (Table 2). Figure 3 compares the inhibition curves of the tested compounds for Sf9 cells expressing the h5-ht_5A receptor alone and in combination with Gα_i1 and Gβ₁γ₂. In contrast to the agonists, the affinity of methiothepin significantly decreased up to 1 log unit on coexpression of G_i/G_o proteins. Decreases in DHE and LSD affinities were minor on coexpression of individual G_i or G_o proteins but appeared significant on coexpression of the mixture of G_i/o proteins.

Effect of Gpp(NH)p on [³H]5-CT Binding.

The interaction of h5-ht_5A receptors with endogenous or coexpressed G proteins in membranes of baculovirus-infected Sf9 cells was investigated by measuring the sensitivity of agonist binding to the addition of the nonhydrolyzable GTP analog Gpp(NH)p. [³H]5-CT concentration-binding experiments in the presence and absence of 100 μM Gpp(NH)p were performed in parallel, and the K_d values were compared using a paired Student's t test (Table 1). Figure4 visualizes the ratio ofK_d values for [³H]5-CT binding in the presence and absence of Gpp(NH)p. The affinity of [³H]5-CT observed for the h5-ht_5A receptor expressed alone was unaffected by Gpp(NH)p, suggesting the absence of interaction with endogenous G proteins. The small increase in affinity achieved by Gβ₁γ₂ coexpression was completely reversed by Gpp(NH)p. The affinity of [³H]5-CT significantly decreased on Gpp(NH)p addition to membranes of Sf9 cells coexpressing the h5-HT_5A receptor together with Gβ₁γ₂ and either the Gα_i/o mixture, Gα_i1, Gα_i2, Gα_i3, or Gα_o, whereas no significant decrease in affinity was observed for the other G protein coexpressions (Fig. 4).

5-HT-Stimulated [³⁵S]GTPγS Binding.

Receptor-mediated activation of G proteins was examined by concentration-dependent 5-HT-stimulated binding of [³⁵S]GTPγS to membranes of baculovirus-infected Sf9 cells. The activation of G proteins involves stimulation of GDP/GTP exchange at the Gα subunit and can be measured by the incorporation of the nonhydrolyzable GTP analog [³⁵S]GTPγS (Wieland and Jakobs, 1994). Figure 5 depicts the dose-dependent increase in 5-HT-stimulated [³⁵S]GTPγS binding for Sf9 cells expressing h5-ht_5A receptors alone or together with Gβ₁γ₂ and/or the Gα_i/o mixture. In membranes of Sf9 cells expressing only h5-ht_5A receptors without mammalian G protein subunits, stimulation of the receptors with 5-HT resulted in an increase in [³⁵S]GTPγS binding to a maximum of 40% over basal, probably due to the activation of endogenous G proteins. Coexpression of Gβ₁γ₂ resulted in a significant increase of the maximum response (Student's ttest, P < .05), up to 110% over basal. When the receptor was coexpressed with the mixture of Gα_i/o subunits, without or with Gβ₁γ₂, the maximum stimulation was 330 and 570%, respectively. The effect of 5-HT was specific for the h5-ht_5A receptor, because 5-HT did not affect [³⁵S]GTPγS binding to membranes from uninfected or wild-type baculovirus-infectedSf9 cells (data not shown).

Mean pEC₅₀ and maximum stimulation values for 5-HT, tested on a series of 26 receptor/G protein combinations expressed in Sf9 cells, are summarized in Table3. Coexpressions with the individual Gα_i1, Gα_i2, Gα_i3, or Gα_o subunits yielded maximum responses that were comparable with those for the Gα_i/o mixture. For the Gα_i subunits, additional coexpression of Gβ₁γ₂ markedly increased stimulation of [³⁵S]GTPγS binding, as observed for the Gα_i/o mixture (Table 3). For Gα_o, however, coexpression with Gβ₁γ₂ significantly (Student's t test, P < .05) decreased the maximum stimulation of [³⁵S]GTPγS binding. It should be noted that the absolute values for basal [³⁵S]GTPγS binding (in cpm) were 2.6-fold higher for Gα_o/Gβ₁γ₂than for Gα_i/Gβ₁γ₂when coexpressed with h5-ht_5A receptors, in contrast to coexpressions with Gα_i or Gα_o, which showed comparable levels of agonist-independent [³⁵S]GTPγS binding (data not shown). For the coexpressions including Gα_z, Gα_s, Gα_q, Gα₁₁, and Gα₁₆, a small 5-HT-induced stimulation of [³⁵S]GTPγS binding was detected, but the maximum stimulation was never significantly higher than the appropriate control sample. No stimulation was observed for the Gα₁₂ and Gα₁₃coexpressions.

Modulation of [³⁵S]GTPγS Binding by 5-HT Receptor Ligands.

Several 5-HT receptor ligands were examined for their ability to modulate [³⁵S]GTPγS binding to membranes of Sf9 cells, expressing h5-ht_5A receptors with Gβ₁γ₂ and either Gα_i1, Gα_i2, Gα_i3, or Gα_o. Figure6 shows, as an example, the mean curves for the coexpression of h5-ht_5A receptors with Gα_i1 and Gβ₁γ₂. Table4 summarizes the pEC₅₀, E_max, pIC₅₀-corr, and I_maxvalues from all [³⁵S]GTPγS dose-response and inhibition curves.

For each of the coexpressed combinations tested, 5-CT and 5-MT produced maximum responses similar to 5-HT, confirming their full agonistic properties (Francken et al., 1998). DHE and LSD stimulated [³⁵S]GTPγS binding to about 50% of the 5-HT level for the G_i coexpressions (i.e., behaved as partial agonists), whereas for the G_ocoexpression, maximum stimulation approached the level of 5-HT (i.e., DHE and LSD behaved as full agonists). Methiothepin behaved as an inverse agonist as it inhibited [³⁵S]GTPγS binding to about −10% below its basal level (5-HT level set at 100%) for Gα_i1, Gα_i2, or Gα_i3 and to −24% below its basal level for Gα_o (see Fig. 6 for Gα_i1; data not shown for Gα_i2, Gα_i3, and Gα_o). However, no reproducible curves could be derived from the methiothepin data points. The antagonistic properties of DHE, LSD, and methiothepin were investigated using [³⁵S]GTPγS binding to membranes of the same four coexpressions. DHE and LSD inhibited 5-HT (10 μM)-stimulated [³⁵S]GTPγS binding to the level of their own agonistic effect. Methiothepin behaved again as an inverse agonist, inhibiting [³⁵S]GTPγS binding below the basal level.

Discussion

Little is known about the pharmacological and functional properties of cloned 5-ht₅ receptors. Recently, h5-ht_5A receptors were shown to mediate inhibition of adenylate cyclase activity in transfected HEK 293 cells (Francken et al., 1998; Hurley et al., 1998). High-affinity agonist binding and agonist-stimulated [³⁵S]GTPγS binding to h5-ht_5A-HEK 293 membranes were found to be pertussis toxin-sensitive (Francken et al., 1998), indicating the involvement of G_i/G_oproteins. To provide further insight in its signaling properties, we coexpressed the h5-ht_5A receptor inSf9 insect cells with a series of 11 mammalian G proteins, from each of the four Gα families. Using radioligand and [³⁵S]GTPγS binding assays, we demonstrated selective coupling of the h5-ht_5A receptor to coexpressed G_i and G_oproteins and the absence of coupling to G_z/G_s/G_q/G₁₁/G₁₆/G₁₂and G₁₃ proteins. Hence, the h5-HT_5A receptor does not show promiscuous coupling to various G protein families. Although no clear coupling preference to either of the G_i/G_o subtypes was evident, we have observed differences in the coupling behavior of G_o versus G_i.

The overexpression in Sf9 cells of h5-ht_5A receptors alone resulted in a predominantly uncoupled phenotype, as demonstrated by guanine nucleotide-insensitive, low-affinity agonist binding. Although not evident from the binding data, h5-ht_5A receptors coupled to endogenous G proteins to some extent; 5-HT stimulated [³⁵S]GTPγS binding to 40% over basal. We conclude that a large excess of uncoupled receptors is present in h5-ht_5A-Sf9 membranes. Although the activation of G proteins by a small fraction of coupled receptors can be detected due to the sensitivity of the [³⁵S]GTPγS binding assay, the curve-fitting algorithms for the concentration-binding isotherms cannot reliably detect a high-affinity binding component of less than 10% of theB_max value.

When the h5-ht_5A receptor was coexpressed with G_i1/G_i2/G_i3or G_o proteins (Gαβ₁γ₂heterotrimers), the coupled phenotype was achieved, as evident from guanine nucleotide-sensitive, high-affinity agonist binding. In addition, the affinity of methiothepin, which was identified as an inverse agonist at h5-ht_5A-HEK 293 cells (Francken et al., 1998), decreased on coexpression of G_i/G_o proteins. These observations are consistent with two distinct states of the h5-ht_5A receptor, according to the two-state model (Leff, 1995). Receptors are proposed to exist in an active form (R*) that is G protein-coupled and an inactive form (R). Agonists show high affinity for R* and low affinity for R, whereas inverse agonists display the opposite behavior (Milligan et al., 1995). Our binding data suggest that h5-ht_5A receptors expressed inSf9 cells convert to the active, high agonist affinity state (R*) through interaction with coexpressed G_i/G_o proteins. Remarkably, the affinities of DHE and LSD, which were identified as partial agonists at h5-ht_5A-HEK 293 cells, decreased on coexpression of G_i and/or G_o proteins. This observation might indicate that the two-state model of agonist action is not generally applicable to partial agonists.

Evidence for h5-ht_5A receptor-mediated G_i/G_o protein activation was obtained using [³⁵S]GTPγS assays. The maximum level of 5-HT-stimulated [³⁵S]GTPγS binding to coexpressed G_o proteins was similar to that found for h5-ht_5A-HEK 293 cells, whereas G_i1/G_i2/G_i3and the mixture of G_i/o proteins were stimulated by 5-HT with approximately 4-fold higher efficacy. The lower level of 5-HT-mediated stimulation of G_o, compared with G_i, might be explained by the 2.6-fold higher basal [³⁵S]GTPγS binding that was found for coexpressed G_o. This high agonist-independent [³⁵S]GTPγS binding most probably originates from a larger number of G_o proteins in theSf9 membranes compared with G_i, as Gα_o appeared more abundant than the various Gα_i subunits in immunoblot analysis. Alternatively, the h5-ht_5A receptor may exhibit stronger constitutive activation of G_o, compared with G_i. High agonist-independent binding complicates the detection of agonist-induced increases in [³⁵S]GTPγS binding (Wieland and Jakobs, 1994). It could be that the assay conditions (e.g., buffer composition and incubation temperature) optimal for G_oactivation differ from the applied conditions, such that the actual maximum stimulation of G_o by 5-HT might well be higher than reported.

Coexpression of the h5-ht_5A receptor with one of the other G proteins tested (G_z/G_s/G_q/G₁₁/G₁₆/G₁₂or G₁₃) had no effect on agonist binding, and no or only minor 5-HT-induced activation of these G proteins could be detected. The expression of the different heterologous Gα proteins in the Sf9 membranes was confirmed using immunoblotting. All Gα proteins were highly expressed, and only Gα₁₂ showed weak immunoreactivity. Hence, poor subunit expression is not the reason for the absence of effects for the various G proteins, except perhaps for Gα₁₂. It should be noted that in the [³⁵S]GTPγS studies, the assay conditions were not optimized for each individual G protein type. Under the applied conditions, which were optimized for [³⁵S]GTPγS binding to h5-ht_5A-HEK 293 membranes, activation of some G protein types may therefore be suboptimal. As the need to optimize assay conditions for individual G proteins has been reported previously (Wieland and Jacobs, 1994), it would be rash to conclude the absolute absence of h5-ht_5A receptor coupling to G_z/G_s/G_q/G₁₁/G₁₆/G₁₂or G₁₃ proteins based exclusively on the absence of increases in [³⁵S]GTPγS binding. The lack of receptor interaction with these G proteins is only suggested by the fact that coexpression of these G proteins did not induce guanine nucleotide-sensitive, high-affinity agonist binding to the h5-ht_5A receptor. It appears that h5-ht_5A receptors selectively couple to G_i/G_o proteins, which is in agreement with the finding that pertussis toxin pretreatment completely abolished high-affinity agonist binding and 5-HT-stimulated [³⁵S]GTPγS binding to h5-ht_5A-HEK 293 membranes (Francken et al., 1998).

The Gβγ complex has already been shown to be required for optimal receptor-G protein interaction (Fung, 1983; Butkerait et al., 1995). We have used the Gβ₁γ₂dimer to enhance G protein coupling to the h5-ht_5A receptor, because this dimer was reported to interact with members of the four Gα families (Barr et al., 1997). However, the subunit composition of Gβγ affects receptor-G protein coupling specificity (Kisselev and Gautam, 1993; Kleuss et al., 1993;Richardson and Robishaw, 1999), such that other Gβγ subunit compositions may yield different receptor coupling profiles. Therefore, we also investigated h5-ht_5A receptor-G protein coupling in the absence of the mammalian Gβ₁γ₂ complex. The interaction of receptor with Gα_i1, Gα_i2, and Gα_i3 could still be detected in agonist binding and [³⁵S]GTPγS assays, but it was indeed less effective than that in the presence of Gβ₁γ₂. Remarkably, coexpression with Gα_o did not induce high-affinity agonist binding in the absence of Gβ₁γ₂. Previously,Jockers et al. (1994) found similar results for adenosine A₁ receptors expressed in Escherichia coli; reconstitution of high-affinity agonist binding by purified G proteins was poor in the absence of Gβγ for G_o, but not for G_i, whereas in the presence of Gβγ, their maximum responses were similar. Despite the lack of effect on agonist affinity of Gα_o, the activation of h5-ht_5A receptors produced a maximum stimulation of [³⁵S]GTPγS binding similar to Gα_i subunits. Coexpression of h5-ht_5A receptors and either Gα_z/Gα_s/Gα_q/Gα₁₁/Gα₁₆/Gα₁₂or Gα₁₃ without Gβ₁γ₂ did not result in the coupled phenotype, as expected from the lack of effect when Gαβ₁γ₂ heterotrimers were expressed. We conclude that the Gβ₁γ₂ complex greatly facilitates coupling of G_i/o proteins to the h5-ht_5A receptor when coexpressed inSf9 cells.

Coexpression of h5-ht_5A receptors and Gβ₁γ₂ without mammalian Gα subunits revealed that Gβ₁γ₂ enhances interaction of heterologous receptor with insect G proteins. Similar results were reported for the serotonin 5-HT_1Aand the dopamine D_2S receptor (Butkerait et al., 1995; Boundy et al., 1996). Considering this finding, one should note that an improved interaction of recombinant receptors with endogenous G proteins due to coexpressed Gβ₁γ₂ subunits may confuse the interpretation of receptor-G protein interaction specificity. Regardless, it is clear that the overexpression of specifically interacting G proteins should yield effects that exceed these observed for the appropriate controls.

For some receptors that couple to pertussis toxin-sensitive G proteins, preferential interaction with one of the G_i/G_o subtypes has been demonstrated (Senogles et al., 1990; Parker et al., 1991; Rubinstein et al., 1991; Grünewald et al., 1996; Clawges et al., 1997; Lorenzen et al., 1998). Our data indicate that the heterotrimeric G_i1, G_i2, G_i3, or G_o proteins interacted equally well with the h5-ht_5A receptor to induce its high-affinity conformation, and no significant differences in the affinities of the tested compounds were observed. However, in contrast to Gα_i, Gα_o did not induce high-affinity agonist binding in the absence of Gβ₁γ₂, suggesting diminished receptor interaction. Furthermore, some striking differences between G_o and G_i proteins appeared from the [³⁵S]GTPγS experiments. Maximum stimulation of [³⁵S]GTPγS binding by 5-HT was significantly lower at G_o than at G_i, possibly due to the high agonist-independent [³⁵S]GTPγS binding to G_o. In addition, the relative efficacies of DHE and LSD were dependent on the G protein type expressed. Both compounds were full agonists at the h5-ht_5A receptor when coexpressed with G_o, whereas coexpression with G_i proteins resulted in partial agonistic behavior, which was also found at the h5-HT_5A-HEK 293 membranes (Francken et al., 1998). These data might be explained by a difference in receptor/G protein stoichiometry, which can influence both agonist potency and efficacy (Hermans et al., 1999). Alternatively, the efficacy of compounds may be determined by the type of G protein interacting with the receptor. In this respect, Yang and Lanier (1999) have reported that recombinant expression of Gα_o, but not Gα_i1, increased the relative efficacy of clonidine in NIH-3T3 cells cotransfected with α₂-adrenergic receptor and Gα subunit, an effect that was not an issue of G protein or receptor levels. Although we cannot exclude that differences in h5-ht_5A receptor-to-G protein ratio cause the distinct behavior of G_o and G_i proteins, it is tempting to speculate that structural differences exist in their interaction with the h5-HT_5A receptor. However, differences in the nucleotide binding properties of the G protein types themselves should also be taken into account; as such, G_o may be easier to activate by receptors than G_i.

In summary, the h5-HT_5A receptor selectively coupled to mammalian G_i1/G_i2/G_i3and G_o but not to G_z/G_s/G_q/11/16or G_12/13 proteins, when coexpressed inSf9 insect cells. Although G_odisplayed different receptor coupling characteristics than G_i proteins, no clear coupling preference was evident.