h5-HT1B receptor-mediated constitutive Gαi3-protein activation in stably transfected Chinese hamster ovary cells: an antibody capture assay reveals protean efficacy of 5-HT


  • Adrian Newman-Tancredi,

    1. Department of Psychopharmacology, Institut de Recherches Servier, 125, Chemin de Ronde, Croissy-sur-Seine, Paris 78290, France
    Search for more papers by this author
    • 2

      Current address; Neurobiology 2 Division, Institut de Recherche Pierre Fabre, 17, avenue Jean Moulin, 81106 Castres, France.

  • Didier Cussac,

    Corresponding author
    1. Department of Psychopharmacology, Institut de Recherches Servier, 125, Chemin de Ronde, Croissy-sur-Seine, Paris 78290, France
    Search for more papers by this author
  • Laetitia Marini,

    1. Department of Psychopharmacology, Institut de Recherches Servier, 125, Chemin de Ronde, Croissy-sur-Seine, Paris 78290, France
    Search for more papers by this author
  • Manuelle Touzard,

    1. Department of Psychopharmacology, Institut de Recherches Servier, 125, Chemin de Ronde, Croissy-sur-Seine, Paris 78290, France
    Search for more papers by this author
  • Mark J Millan

    1. Department of Psychopharmacology, Institut de Recherches Servier, 125, Chemin de Ronde, Croissy-sur-Seine, Paris 78290, France
    Search for more papers by this author

Department of Psychopharmacology, Institut de Recherches Servier, 125, Chemin de Ronde, Croissy-sur-Seine, Paris 78290, France. E-mail: didier.cussac@fr.netgrs.com


  • Serotonin 5-HT1B receptors couple to G-proteins of the Gi/o family. However, their activation of specific G-protein subtypes is poorly characterised. Using an innovative antibody capture/guanosine-5′-0-(3-[35S]thio)-triphosphate ([35S]GTPγS) binding strategy, we characterised Gαi3 subunit activation by h5-HT1B receptors stably expressed in Chinese hamster ovary (CHO) cells.

  • The agonists, 5-HT, alniditan and BMS181,101, stimulated Gαi3, whereas methiothepin and SB224,289 behaved as inverse agonists. The selective 5-HT1B receptor ligand, S18127, modestly stimulated Gαi3 and reversed the actions of both 5-HT and methiothepin. S18127 (1 μM) also produced parallel, dextral shifts of the 5-HT and methiothepin isotherms.

  • Isotopic dilution experiments ([35S]GTPγS versus GTPγS) revealed high-affinity [35S]GTPγS binding to Gαi3 subunits in the absence of receptor ligands indicating constitutive activity. High-affinity [35S]GTPγS binding was increased 2.8-fold by 5-HT with an increase in the affinity of GTPγS for Gαi3 subunits. In contrast, methiothepin halved the number of high-affinity binding sites and decreased their affinity.

  • h5-HT1B receptor-mediated Gαi3 subunit activation was dependent on the concentration of NaCl. At 300 mM, 5-HT stimulated [35S]GTPγS binding, basal Gαi3 activation was low and methiothepin was inactive. In contrast, at 10 mM NaCl, basal activity was enhanced and the inverse agonist activity of methiothepin was accentuated. Under these conditions, 5-HT decreased Gαi3 activation.

  • In conclusion, at h5-HT1B receptors expressed in CHO cells: (i) inverse agonist induced inhibition of Gαi3, and its reversal by S18127, reveals constitutive activation of this Gα subunit; (ii) constitutive Gαi3 activation can be quantified by isotopic dilution [35S]GTPγS binding and (iii) decreasing NaCl concentrations enhances Gαi3 activation and leads to protean agonist properties of 5-HT: that is a switch to inhibition of Gαi3.

British Journal of Pharmacology (2003) 138, 1077–1084. doi:10.1038/sj.bjp.0705140



CHO cells

Chinese hamster ovary cells


scintillation proximity assay


5-HT1B receptors are involved in the control of mood, motor function and cognition, and 5-HT1B autoreceptors on serotonergic terminals modulate serotonin release (Engel et al., 1986; Barnes & Sharp, 1999; Meneses, 1999; Millan et al., 1999; Sari et al., 1999). Consequently, an understanding of the mechanisms of signal transduction at 5-HT1B receptors is relevant to the etiology and treatment of affective, neurological and cardiovascular disorders and to the management of migraine.

5-HT1B receptors couple inhibition of adenylyl cyclase (Schoeffter & Hoyer, 1989; Pauwels & Palmier, 1994), a response which is abolished by pre-treatment with Bordetella pertussis toxin which ADP-ribosylates Gα subunits of the Gi/o family, indicating coupling of 5-HT1B receptors to these G-proteins. Several studies have investigated the influence of serotonergic ligands on G-protein activation by 5-HT1B receptors (Thomas et al., 1995; Pauwels et al., 1997; Gaster et al., 1998; Selkirk et al., 1998; Newman-Tancredi et al., 2000). These studies demonstrated robust activation of G-proteins by agonists, and also revealed that 5-HT1B receptors exhibit marked constitutive activity for G-protein activation. Correspondingly, several inverse agonists have been identified at 5-HT1B receptors, including methiothepin and the selective 5-HT1B ligand, SB224,289 (Gaster et al., 1998; Selkirk et al., 1998; Audinot et al., 2001a,2001b). However, these studies were carried out using techniques which measure “overall” G-protein activation without distinguishing the precise G-protein subtype(s) involved. 5-HT1B receptors couple, in fact, to several Gα subtypes, including members of the Gi/o family and Gα15 subunits (Bae et al., 1997; Clawges et al., 1997; Wurch & Pauwels, 2000) but not to Gαt subunits (Bae et al., 1997). Reconstitution of 5-HT1B receptors expressed in Sf9 cells with different purified Gαi subunits increased the affinity of agonists indicating coupling to several G-protein subtypes (Clawges et al., 1997). Compared with Gαi2 and Gαo, the Gαi3 subtype induced the greatest increase in agonist affinity, suggesting preferential coupling of 5-HT1B receptors to this G-protein subtype.

In view of the above considerations, the present study employed a recently developed antibody-capture technique coupled to SPA detection (De Lapp et al., 1999; Cussac et al., 2002; Newman-Tancredi et al., 2002a) to characterise Gαi3 subunit activation by recombinant human 5-HT1B receptors stably expressed in CHO cells (Newman-Tancredi et al., 2000; Audinot et al., 2001a,2001b). This cell line constitutes a useful model to investigate 5-HT1B receptor coupling since they express high levels of Gαi3 subunits in comparison to only low levels of Gαo, while Gαi1 is undetectable (Gerhardt & Neubig, 1991; Law et al., 1993; Gettys et al., 1994).

Herein, we investigated the influence of several parameters on Gαi3 activation. First, we compared ligand potencies and efficacies for Gαi3 activation with those determined previously in identical membrane preparations (Newman-Tancredi et al., 2000) employing conventional [35S]GTPγS binding, which does not distinguish between G-protein subtypes (Lazareno et al., 1993; Lorenzen et al., 1993). Second, constitutive 5-HT1B receptor-mediated Gαi3 protein activation was quantified employing [35S]GTPγS versus GTPγS homologous inhibition curves. Such binding isotherms allow the detection of high-affinity (HA) and low-affinity (LA) binding components (Breivogel et al., 1998; Selley et al., 1998), and can be used to directly quantify the amount of agonist-independent, constitutive G-protein activation without requiring the use of inverse agonists (Newman-Tancredi et al., 2000; Audinot et al., 2001b; Rouleau et al., 2002). Third, detection of constitutive activity may be dependent on the experimental conditions employed and particularly on the presence of Na+ ions. Thus, NaCl influences receptor conformation by favouring a shift from a G-protein coupled to an uncoupled state, thereby reducing the affinity of agonists (Pert & Snyder, 1974; Horstman et al., 1990). Further, NaCl reduces basal G-protein activation in membranes of cells expressing G-protein coupled receptors (Costa et al., 1990; see de Ligt et al., 2000 for review). However, few studies have examined the influence of NaCl on specific G-protein subtypes (Wenzel-Seifert et al., 1998; Seifert, 2001). Thus, we describe here its influence on the constitutive activity of h5-HT1B receptor-activated Gαi3 subunits, and upon the actions of 5-HT and methiothepin.

A preliminary communication of these data was presented in abstract form (Newman-Tancredi et al., 2002b).


[35S]GTPγS binding by antibody capture and scintillation proximity assay detection

CHO-h5-HT1B cell membranes expressing 8.5 pmol mg−1 of h5-HT1B receptors (Newman-Tancredi et al., 2000) were purchased from Euroscreen (Brussels, Belgium). To specifically detect [35S]GTPγS binding to h5-HT1B receptor-mediated Gαi3 G-protein subunits, an antibody-capture strategy was adopted, coupled to detection by SPA. Procedures were similar to those described by De Lapp et al. (1999) and are detailed in Newman-Tancredi et al. (2002a) and Cussac et al. (2002). Briefly, CHO-h5-HT1B cell membranes were incubated on 96-well plates with agonists and/or antagonists and [35S]GTPγS (0.2 nM) for 1 h at 22°C in a buffer containing: 20 mM HEPES (pH 7.4), 3 μM GDP, 3 mM MgCl2 and 100 mM NaCl unless otherwise indicated. The reaction was stopped by solubilising cell membranes with detergent (NP40 0.3% v/v final) and gently agitating for 30 min. Mouse anti-Gαi1/3 monoclonal antibodies (Biomol, San Diego, CA) were then added (0.1 μg of IgG per well) and plates incubated for a further 2 h to allow antibody-Gα complexes to form. Since CHO cells do not express Gαi1, the assay detects activation of Gαi3 (see Gettys et al., 1994; Newman-Tancredi et al., 2002a and Introduction). The specificity of the anti-Gαi1/3 antibody itself was verified by Western blots against a range of purified Gα subunits indicating an absence of cross-reactivity with Gαi2, Gαo, GαS, Gαq, Gα13 (Cussac et al., 2002). At the end of the incubation period, SPA beads coated with anti-mouse 2nd antibody (Amersham, Les Ulis, France), were added at the manufacturer's recommended concentrations and incubated with gentle agitation overnight before radioactivity counting. All incubation steps were carried out at room temperature. Non-specific binding was defined with 10 μM GTPγS and, unless indicated otherwise, subtracted from observed [35S]-GTPγS binding prior to analysis.

Data analysis

Concentration – response isotherms for the effect of drugs on Gαi3 subunit activation were analysed by nonlinear regression using the program “Prism” (Graphpad Software Inc., San Diego, CA, (U.S.A.)). Isotherms were fitted to a four-parameter, logistic equation to derive values of pseudo-Hill coefficient (nH), potency (pEC50) and maximal stimulation (EMAX). The latter was defined as the amount of specific [35S]GTPγS binding expressed as a percentage of basal (agonist-independent) binding (=100%). Antagonist potency (pKB values) for dextral shift of agonist-induced stimulation or inverse agonist-induced inhibition isotherms of Gαi3 activation was calculated by KB=[Antagonist]/{([EC50′]/[EC50])−1}, where [Antagonist]=antagonist concentration, [EC50]=concentration of agonist/inverse agonist producing half-maximal effect and [EC50′]=concentration of agonist producing half-maximal effect in the presence of antagonist. KB values for inhibition of 5-HT and methiothepin (10 and 100 nM, respectively)-stimulated [35S]GTPγS binding were calculated by KB=IC50/{[(2+(Agonist/EC50)nH)nH−1]−1}, where IC50 is the inhibitory concentration50 of the antagonist, Agonist the 5-HT or methiothepin concentration; EC50 the effective concentration50 of 5-HT/methiothepin alone and nH the Hill coefficient of the 5-HT/methiothepin stimulation isotherm.

Inhibition of [35S]GTPγS binding by unlabelled GTPγS at CHO-h5-HT1B membranes

Isotopic dilution experiments were carried out as described by Newman-Tancredi et al. (2000). Binding of radiolabelled [35S]GTPγS to Gαi3 subunits was inhibited with GTPγS and the resulting isotherms were best fitted by a two-site nonlinear regression analysis, giving inhibitory concentration (IC50) values for HA and LA binding components, respectively. HA binding observed under basal conditions (i.e., not agonist-induced) reflects receptor-dependent G-protein activation and provides a direct measure of constitutive activity (Newman-Tancredi et al., 2000; Audinot et al., 2001b), whereas LA binding reflects endogenous GDP/GTP turnover of CHO cell membrane Gα subunits. Binding data from these experiments expressed in fmol mg−1 of protein were normalised to account for the concentration of [35S]GTPγS present in the assay. Hence, units are denoted fmol of [35S]GTPγS bound mg−1 of protein nM−1 [35S]GTPγS in the incubation.


5-HT creatinine sulphate was purchased from Sigma (Saint Quentin Fallavier, France) and methiothepin maleate from Tocris Cookson (Southampton, England). SB224,289 (1′-methyl-5-[[2′-methyl-4′-(5-methyl-1,2,4-oxadiazol-3-yl)biphenyl-4-yl]-carbonyl]-2,3,6,7-tetrahydrospiro-[furo-[2,3f]-indole-3,4′-pi-peridine]-oxalate) and S18127 (N-[1-(2,3-dihydro[1,4] dioxin-5-yl)piperid-4-yl] indan-2-yl-amine) dichlorohydrate were synthesised by Jean-Louis Peglion, Servier. BMS 181,101 (5-fluoro-3-{3-[4-(5-methoxy-pyrimidin-4-yl)-piperazin-1-yl]-propyl}-1H-indole) di-hydrochloride and alniditan were synthesised by Gilbert Lavielle, Servier. Compounds were dissolved in water at 1 mM or in dimethylsulphoxide at 10 mM and diluted in the appropriate assay buffer to the required experimental concentrations.


Specific Gαi3 G-protein subunit activation

The antibody capture/SPA technique detected h5-HT1B receptor-mediated Gαi3 subunit activation. 5-HT and methiothepin, up to micromolar concentrations, did not modify Gαi3 activation from basal levels in membranes of control (untransfected) CHO cells (not shown) demonstrating that h5-HT1B receptors mediate actions described below. Further, when experiments were carried out in the absence of anti-Gαi3 antibodies, or in the presence of another IgG (monoclonal anti-Extracellular Regulated Kinase; pERK, Nanotools, Germany), no stimulation of [35S]GTPγS binding was detected (Newman-Tancredi et al., 2002a; and data not shown).

Influence of agonists, partial agonists and inverse agonists

In CHO-h5-HT1B cell membranes, 5-HT yielded sigmoidal [35S]GTPγS binding isotherms with a pEC50 value of 8.97±0.20 (Figure 1; Table 1). BMS181,101 and alniditan also efficaciously stimulated [35S]GTPγS binding, whereas the selective 5-HT1B receptor ligand, S18127, only slightly increased Gαi3 activation suggesting weak partial agonist properties. Methiothepin and the selective 5-HT1B receptor ligand, SB224,289, exhibited inverse agonist properties in inhibiting basal Gαi3 activation (Figure 1; Table 1). 5-HT (10 nM)-stimulated Gαi3 activation was concentration-dependently reversed by S18127 (Figure 2) with a pKB of 7.91±0.35. Further, S18127 (1 μM) shifted the 5-HT isotherm to the right with a pKB value of 8.01±0.15. S18127 also reversed methiothepin (100 nM)-inhibited Gαi3 activation, with a pKB of 6.80±0.11. Similarly, S18127 (1 μM) shifted the methiothepin inhibition isotherm to the right with a pKB of 6.86±0.17 (Figure 2).

Figure 1.

Action of agonists and inverse agonists on [35S]GTPγS binding to Gαi3 subunits in CHO-h5-HT1B cell membranes. Panel a: concentration – response isotherms of 5-HT, BMS181,101 and methiothepin. Panel b: concentration – response isotherm of alniditan, S18127 and SB224,287. Data points are means of duplicate determinations from representative experiments repeated on at least three independent occasions. Data from these experiments are summarised in Table 1.

Table 1. Stimulation of i3 G protein subunits at h5-HT1B receptors
 Gαi3 activationTotalG–proteinsAffinity
LigandpEC50EMAX (%)pEC50EMAX (%)pki
  • Activation of Gαi3 subunits in CHO cells stably expressing h5-HT1B receptors was determined employing an antibody-capture/SPA detection technique. Data are expressed as means±s.e.m.'s of three or more determinations performed in duplicate. 5-HT stimulated Gαi3 subunit activation by 1.81±0.16-fold above specific basal binding (see Fig.1). EMAX values in the table are expressed as a percentage of the stimulation induced by a maximally effective concentration of 5-HT (1 μM). For comparative purposes, the table shows previously determined potencies and efficacies for stimulation of ‘total’ G-proteins in identical membrane preparations. Ligand affinities (pKi) determined by competition binding with [3H]GR125,743 are also shown. (Data are from Newman–Tancredi et al., 2000; Audinot et al., 2001a.)

  • a

    Unpub. obs.

  • b

    pIC50 for inverse agonists.

Figure 2.

Antagonism by S18127 of 5-HT-stimulated and methiothepin-inhibited [35S]GTPγS binding to Gαi3 subunits in CHO-h5-HT1B cell membranes. Panel a: concentration – response isotherms of 5-HT and methiothepin alone or in the presence of S18127 (1 μM). Panel b: concentration-dependent reversal by S18127 of 5-HT (10 nM)-induced stimulation or methiothepin (100 nM)-induced inhibition of Gαi3 activation. Data points are the means of duplicate determinations from representative experiments repeated on at least three independent occasions.

Isotopic dilution “saturation binding”

Under basal conditions (absence of receptor ligand), inhibition of [35S]GTPγS binding by GTPγS produced biphasic isotherms (two-site fit statistically superior to single-site fit; P<0.05). This is consistent with spontaneous induction of G-protein activation, that is, constitutive activity (Figure 3). 5-HT (1 μM) increased the number of HA sites by 2.8-fold and increased the pIC50 of the HA component from 8.32±0.09 under basal conditions to 9.34±0.04 with 5-HT (Table 2). Conversely, the number of HA sites was reduced by the inverse agonist, methiothepin (1 μM), which also decreased the pIC50 value to 7.81±0.07 (Table 2).

Figure 3.

Inhibition, by isotopic dilution with GTPγS, of [35S]GTPγS binding to Gαi3 subunits in membranes of CHO cells expressing h5-HT1B receptors. Panel a: [35S]-GTPγS isotopic dilution under basal conditions (no receptor ligands) and in the presence of 5-HT (1 μM) or methiothepin (1 μM). The dotted lines indicate the HA binding component of the isotherms. Isotherms determined by nonlinear regression are biphasic (two-site fit statistically superior to a single site fit; P<0.05, F-test). Panel b: Gαi3 subunit saturation binding isotherms derived from [35S]GTPγS isotopic dilution experiments (as described in Materials and Methods). Isotherms are shown under basal conditions and in the presence of 5-HT (1 μM) or methiothepin (1 μM). Points shown are from representative experiments performed in duplicate and repeated on at least three independent occasions. Data from these experiments are summarised in Table 2.

Table 2. Inhibition by GTPγS of [35S]GTPγS binding to Gαi3 G protein subunits at h5-HT1B receptors
 HA sites+LA sites (fmol mg1nM1)HA sites (fmol mg1nM1)pIC50(HA)LA sites (fmol mg1nM1)pIC50(LA)
  • Activation of Gαi3 G-protein subunits in CHO cells stably expressing h5-HT1B receptors was determined employing an antibody-capture/SPA detection technique. [35S]GTPγS binding was inhibited with unlabelled GTPγS in the presence or absence of 5-HT (1 μM) or methiothepin (1 μM). Isotherms were analysed by nonlinear regression yielding high-affinity (HA) and low-affinity (LA) specific binding components (expressed as fmol of [35S]GTPγS bound mg−1 of protein nM−1 [35S]GTPγS present in the experiment). In all cases a two-site fit was statistically superior to a single site fit (P<0.05, F-test). Data are means±s.e.m.'s of at least three independent experiments. Data in italics within parentheses are the ratios of values in the presence of ligand to those under basal conditions.

  • a

    P<0.05, two-tailed unpaired t-test versus respective basal values.

Basal336 (1.0)244±27 (1.0)8.32±0.0992±95.45±0.10
5-HT792 (2.4)675±84a (2.8)9.34±0.04a117±55.90±0.07a
Methiothepin236 (0.7)129±22a (0.5)7.81±0.07a107±125.29±0.16

In contrast to HA sites, the number of LA binding sites observed in [35S]GTPγS versus GTPγS binding isotherms was not affected by the presence of receptor ligands. However, an increase in the affinity of the low potency component was observed (pIC50 increased from 5.45 under basal conditions to 5.90 with 5-HT), possibly suggesting the presence of additional coupling states of the receptor to Gαi3 subunits (cf. Rouleau et al., 2002, who reported a third affinity component for H3 receptors).

The isotopic dilution experiments (n=4) were used to derive KD (nM) and BMAX values (fmol mg−1) for Gαi3 subunits as described by Newman-Tancredi et al. (2000). BMAX values did not vary significantly: basal conditions: 900±280 fmol mg−1; with 1 μM 5-HT: 510±30 fmol mg−1; with 1 μM methiothepin: 1180±270 fmol mg−1. In contrast, KD values were significantly modified in the presence of receptor ligands: basal conditions: 3.0±0.6 nM; with 1 μM 5-HT: 1.2±0.2 nM (P<0.05, two-tailed t-test versus basal conditions); with 1 μM methiothepin: 6.5±1.0 nM (P<0.05, two-tailed t-test versus basal conditions).

Modulation of Gαi3 activation by NaCl concentration

The influence of NaCl on Gαi3 activation was examined by investigating the responses to 5-HT and methiothepin in the presence of four different concentrations (10, 30, 100, 300 mM) (Figure 4; Table 3). When the concentration of NaCl was increased to 300 mM, basal Gαi3 activation was reduced to just ∼40% of basal control values (100 mM NaCl; Table 3). Under these conditions, no constitutive activity was detectable as demonstrated by the absence of inhibition by the inverse agonist, methiothepin. Conversely, at 10 mM NaCl, basal Gαi3 activation was high, attaining 237% of control values. Under these conditions, methiothepin exhibited robust inhibition of Gαi3 activation (Table 3). In the case of 5-HT, a striking NaCl-dependent reversal of its actions from stimulation to inhibition of Gαi3 activation was observed. Thus, at 300 and 100 mM NaCl, basal binding was low and 5-HT robustly stimulated Gαi3 activation. In contrast, at 30 or 10 mM NaCl, no stimulation by 5-HT was observed: indeed, 5-HT concentration-dependently inhibited Gαi3 activation. The isotherms of 5-HT in the presence of 10, 30 or 100 mM NaCl converged at a similar level of Gαi3 activation (∼160% above basal control values; Figure 4), suggesting stabilisation of a common conformation of a h5-HT1B receptor-Gαi3 subunit complex.

Figure 4.

Influence of NaCl concentrations on [35S]GTPμS binding to Gαi3 subunits in CHO-h5-HT1B cell membranes. The influence of four concentrations of NaCl on the actions of a full agonist, 5-HT (panel a), and on the actions of the inverse agonist, methiothepin (panel b), are shown. B = basal binding. Data points are the means of duplicate determinations from representative experiments repeated on at least three independent occasions with similar results. Data from these experiments are summarised in Table 3.

Table 3. Influence of NaCl on [35S]GTPγS binding to Gαi3 G protein subunits at h5-HT1B receptors
 10 mMNaCl30 mMNaCl100 mMNaCl300 mMNaCl
  • Activation of Gαi3 in CHO cells stably expressing h5-HT1B receptors was determined employing an antibody-capture/SPA detection technique. Isotherms were analysed by a 4-parameter logistic equation. “Bottom” refers to the lower plateau (normalised to the binding observed under standard basal conditions, that is, 100 mM NaCl without non-specific subtraction). “Top” is the upper plateau, pEC50 is the concentration of ligand inducing a half-maximal effect and nH is the pseudo-Hill coefficient (positive for stimulation of Gαi3, negative for its inhibition). Data are means±s.e.m.'s of at least three independent experiments performed in duplicate. n.c.=not computable. Representative isotherms for these data are shown in Fig.4.

  • a

    Inhibition of basal binding.



h5-HT1B receptor-mediated Gαi3 activation by agonists and inverse agonists

Although, as noted in the Introduction, previous studies have investigated the coupling of 5-HT1B receptors to G-proteins, little is known concerning the activation patterns of specific G-protein subtypes. As shown in Table 1, the pattern of activation of Gαi3 generally resembled that described using classical [35S]GTPγS binding filtration assays (Thomas et al., 1995; Selkirk et al., 1998; Newman-Tancredi et al., 2000; Audinot et al., 2001a). Thus, BMS181,101 and alniditan exhibited high efficacy for Gαi3 activation relative to 5-HT, whereas SB224,289 and methiothepin exhibited inverse agonist properties. However, a detailed comparison of results revealed some interesting differences. First, whereas BMS181,101 behaved as an efficacious but submaximal agonist (EMAX 85%) in conventional [35S]GTPγS binding experiments, its efficacy for Gαi3 activation consistently exceeded that of 5-HT. Second, S18127 behaved as a “neutral” antagonist in [35S]GTPγS binding (Audinot et al., 2001a) but exhibited modest partial agonist properties for Gαi3 activation. Third, potencies of agonists for Gαi3 activation were markedly greater than for “total” G-protein activation. Thus, the profile of action of h5-HT1B receptor ligands for Gαi3 subunit activation may differ from that of other G-potein subtypes. “Classical” (overall) G-protein activation assays mask these differences by yielding a composite response of multiple G-proteins. Future SPA studies of other Gi subtypes, in particular Gαi2 (Clawges et al., 1997), coupled to h5-HT1B receptors must address this question directly. In contrast to agonists, negative efficacies of inverse agonists for inhibition of Gαi3 activation corresponded closely to those reported by conventional methods. This raises the possibility that a major proportion of constitutive activity in CHO-h5-HT1B cell membranes may be mediated by Gαi3 subunits. A further point should be noted: the selective ligand, S18127, reversed both 5-HT-mediated stimulation and methiothepin-mediated inhibition of Gαi3 subunit activation (Figure 2), but pKB values obtained differed (7.9 versus 5-HT compared with 6.8 versus methiothepin). The lower pKB values observed for S18127 against methiothepin are consistent with previous data derived from “total” G-protein assays (Audinot et al., 2001a) and may be due to methiothepin's proposed capacity to stabilise specific conformations of G-protein-coupled receptors, such as 5-HT1A receptors (McLoughlin & Strange, 2000). Whatever the mechanism involved, blockade by S18127 of the actions of inverse agonists provides compelling evidence for constitutive activation of Gαi3 subunits in these CHO-h5-HT1B cell membranes.

Isotopic dilution Gαi3 saturation binding

In previous studies, we (Newman-Tancredi et al., 2000; Audinot et al., 2001b) and others (Rouleau et al., 2002) described a methodology to directly quantify the degree of constitutive G-protein activation without the need for inverse agonists. The strategy was based on isotopic dilution of [35S]GTPγS with increasing concentrations of GTPγS. Thus, high-affinity [35S]GTPγS binding sites observed under basal conditions (absence of receptor ligands) provide a measure of constitutive G-protein activation. In our previous study of h5-HT1B receptors, agonists increased the amount of high-affinity [35S]GTPγS binding. Conversely, inverse agonists decreased the number of high-affinity binding sites. Neither agonists nor inverse agonists markedly influenced low-affinity [35S]GTPγS binding sites (Newman-Tancredi et al., 2000; Audinot et al., 2001b). A similar pattern was observed in the present study of Gαi3. Thus, under basal conditions, [35S]GTPγS binding to Gαi3 was biphasic, indicating the presence of constitutive Gαi3 activation. As for “total” G-protein activation assays, 5-HT increased HA [35S]GTPγS binding, with a marked increase in the affinity of GTPγS for Gαi3 subunits (pIC50 increased from 8.32 to 9.34; Table 2). In contrast, methiothepin decreased the number of high affinity sites with a reduction in the affinity of GTPγS (pIC50 decreased from 8.32 to 7.81; Table 2). It is interesting that methiothepin reduced the number of high affinity sites by about 50% compared with a 70% decrease in conventional [35S]GTPγS binding experiments (Newman-Tancredi et al., 2000). This suggests that methiothepin displays sub-maximal “negative” efficacy at Gαi3. Indeed, just as agonists exhibit “agonist-directed trafficking of receptor signalling” (Kenakin, 1995a), inverse agonists may differentially inhibit signalling of receptors to specific G-proteins (Berg et al., 1999). Indeed, the BMAX of Gαi3 subunits determined from the isotopic dilution experiments (leqslant R: less-than-or-eq, slant1 pmol mg−1), amounted to a third of “total” G-protein activation (∼3 pmol mg−1) in this expression system (Newman-Tancredi et al., 2000) indicating that other populations of G-proteins expressed in CHO cells couple to h5-HT1B receptors. It would be interesting to investigate the influence of alterations in receptor/G-protein stoichiometry on activation of Gαi3 subunits. Changes in the expression levels of Gαi3 may lead to alterations in the equilibrium of coupling of 5-HT1B receptors to other G-protein subtypes. Though Gαi2 is likely to be involved (Clawges et al., 1997), this remains to be directly demonstrated.

Influence of NaCl on h5-HT1B receptor-mediated Gαi3 constitutive activation

Previous studies have demonstrated the importance of NaCl in influencing receptor/G-protein coupling (see Introduction). Indeed, as for “total” G-protein assays at other G-protein coupled receptors (De Ligt et al., 2000), NaCl concentrations herein were inversely related to basal Gαi3 activation (Figure 4) and the inverse agonist, methiothepin, failed to inhibit [35S]GTPγS binding in the presence of 300 mM NaCl when constitutive activity was entirely suppressed. These data are reminiscent of studies at 5-HT1A receptors, where inverse agonist properties of certain ligands could only be observed in the absence of NaCl (Cosi & Koek, 2000). However, a striking result was obtained concerning the influence of 5-HT on Gαi3 activation. Whereas, under standard conditions (100 mM NaCl), 5-HT exhibited robust stimulation of [35S]GTPγS binding to Gαi3 subunits, at low NaCl concentrations (10 mM) it inhibited Gαi3 activation. Hence, under the latter condition, 5-HT behaved as an inverse agonist, reversing basal [35S]GTPγS binding to Gαi3 subunits in a manner similar to that of methiothepin (Figure 4; Table 3). This observation is consistent with the concept of a protean agonist (Kenakin, 1995b), by which ligands can exhibit either agonist or inverse agonist properties depending on “receptor tone”. When the latter is low (i.e., little constitutive G-protein activation), agonist properties of the ligand (in this case, 5-HT) are revealed. In contrast, when receptor tone is pronounced (i.e., a high level of constitutive activity), the ligand stabilises the receptor in a conformation which is less able to activate G-proteins than the free, non-ligand-occupied, receptor. It is noteworthy that, over a wide range of NaCl concentrations (10–100 mM), saturating levels of 5-HT induced a similar degree of Gαi3 labelling, either by stimulating or inhibiting [35S]GTPγS binding from basal levels. At an intermediate NaCl concentration (about 40 mM), 5-HT would be expected to have no effect on Gαi3 activation, thus behaving as a neutral antagonist. It would be interesting to examine other 5-HT1B agonists to determine whether they display similar sensitivity to NaCl.

Previous reports have described protean agonist properties of secretin at constitutively active mutants of secretin receptors (Ganguli et al., 1998), and of medetomidine and the dexefaroxan analogue, RX831003, at α2A-adrenoceptors (Jansson et al., 1998; Pauwels et al., 2002). However, to our knowledge, this is the first demonstration that 5-HT can exhibit protean agonism at non-mutant receptors. The physiological implications of these data are unclear, but Na+/H+ exchange in CHO cells is known to be regulated by Gαi3 subunits (Garnovskaya et al., 1997) and modulation of Na+ currents is involved in multiple physiological functions (Urenjak & Obrenovitch, 1996; Cantrell & Catterall, 2001). Indeed, whereas plasma levels of sodium are about 120 mM, G-proteins located on the intracellular surface of plasma membranes are exposed to sodium concentrations of 5–10 mM, except under neuronal depolarisation conditions, when extracellular sodium enters the cells. Thus, it may be speculated that the response to 5-HT by 5-HT1B receptor-activated Gαi3 subunits may shift from inhibition to stimulation depending on polarisation state. Further, it would be of interest to investigate whether these actions of NaCl are due to an influence upon Gαi3 subunits and/or its allosteric regulation of a conserved aspartate residue at the intracellular terminal of the putative second transmembrane segment of the receptor. Mutation of this residue in α2-adrenoceptors abolishes the ability of Na+ to modulate receptor–ligand interactions (Horstman et al., 1990).


The present data demonstrate that: (i) h5-HT1B receptor activation of Gαi3 subunits can be increased by agonists and decreased by inverse agonists. The reversal of the inhibitory actions of methiothepin by the selective ligand, S18127, demonstrates the presence of constitutive activity; (ii) quantification of Gαi3 constitutive activity by homologous [35S]GTPγS versus GTPγS inhibition isotherms indicates that methiothepin inhibits about half of the high affinity binding sites (i.e. constitutive activity) suggesting that it exhibits submaximal negative efficacy for Gαi3 subunit activation and (iii) modulation of NaCl concentrations reveals the protean agonist properties of 5-HT for activation of Gαi3 subunits. It would be interesting to investigate whether similar signalling patterns are observed for other Gα subunits and signal transduction responses, such as adenylyl cyclase inhibition.