Constitutively active Gq/11-coupled Receptors Enable Signaling by Co-expressed Gi/o-coupled Receptors*

Co-expression of guanine nucleotide-binding regulatory (G) protein-coupled receptors (GPCRs), such as the Gi/o-coupled human 5-hydroxytryptamine receptor 1B (5-HT1BR), with the Gq/11-coupled human histamine 1 receptor (H1R) results in an overall increase in agonist-independent signaling, which can be augmented by 5-HT1BR agonists and inhibited by a selective inverse 5-HT1BR agonist. Interestingly, inverse H1R agonists inhibit constitutively H1R-mediated as well as 5-HT1BR agonist-induced signaling in cells co-expressing both receptors. This phenomenon is not solely characteristic of 5-HT1BR; it is also evident with muscarinic M2 and adenosine A1 receptors and is mimicked by mastoparan-7, an activator of Gi/o proteins, or by over-expression of Gβγ subunits. Likewise, expression of the Gq/11-coupled human cytomegalovirus (HCMV)-encoded chemokine receptor US28 unmasks a functional coupling of Gi/o-coupled CCR1 receptors that is mediated via the constitutive activity of receptor US28. Consequently, constitutively active Gq/11-coupled receptors, such as the H1R and HCMV-encoded chemokine receptor US28, constitute a regulatory switch for signal transduction by Gi/o-coupled receptors, which may have profound implications in understanding the role of both constitutive GPCR activity and GPCR cross-talk in physiology as well as in the observed pathophysiology upon HCMV infection.

GPCRs, 1 which can be activated by a diverse array of stimuli, represent the largest group of integral membrane proteins involved in signal transduction. As such, GPCRs are the primary therapeutic target for many of today's drugs (1). The heterotrimeric G proteins mediate signaling from a large number of diverse GPCRs to a variety of intracellular effectors (see Refs. 2 and 3). A large body of work investigating the mechanisms underlying receptor-G protein interactions supports a network of interactions between signaling pathways that converge and diverge at multiple levels, enabling cells to coordinate responses to diverse environmental stimuli (4).
The hitherto existing knowledge of GPCR signal transduction pathways are founded largely upon experimental data obtained by the individual stimulation of the receptor of interest by specific ligands, either in heterologous expression systems or native tissues. Yet, under physiological conditions cells are permanently co-stimulated by various agonists. Investigations using receptor co-stimulation with agonists have recently shown cooperative effects of G i/o and G s (5), G i/o and G q/11 (6), and G s and G q/11 (7), the ability of G i/o -coupled receptors to activate G q -coupled receptors to transduce their signals by G␤␥ exchange (8), as well as G␤␥-mediated glucocorticoid receptor transactivation (9). Such synergistic signaling mechanisms have been suggested to play an important role in signal adaptation (8) and may have a significant role in both physiological and pathophysiological processes (5).
Investigations of GPCRs expressed in surrogate cell systems have greatly modified our understanding of the pharmacological properties of GPCRs and consequently of the mechanistic drug-receptor models used to simulate drug action (10,11). One of the prominent recent additions to our understanding of GPCR action is the occurrence of constitutive, agonist-independent GPCR activity (12)(13)(14)(15). Currently, all models consider agonist-independent activity secondary to spontaneous isomerization of the receptors between the inactive and active receptor state(s), which couple(s) to the G protein; agonists are considered to preferentially bind to the active receptor state, whereas inverse agonists preferentially bind to the inactive receptor state (11). Despite initial concerns, the physiological relevance of constitutive GPCR activity is now accepted, and it appears not only to be a common property of GPCRs but, in various instances, also to be involved in the initiation or progression of disease (15). Various polymorphic GPCR variants have been shown to be highly constitutively active (15,16), but also certain wild-type GPCRs, such as the histamine H 3 receptor, exhibit high constitutive activity in vivo (17). Moreover, viral infection of cells may also result in the expression of virally encoded GPCRs that exhibit high levels of constitutive activity (18,19).
To date, the impact of constitutive GPCR activity on signaling properties on co-expressed GPCRs has not been studied in full detail, despite the potential (patho)physiological relevance. In the present study, we investigated the potential cross-talk between a variety of G i/o -and constitutively active G q/11 -coupled receptors using a heterologous expression system. Our data indicate that both activated histamine H 1 R and HCMVencoded receptor US28 can result in the propagation of G i/ocoupled receptor dependent signaling. Therefore, active G q/11 -coupled receptors constitute a regulatory role in the regulation of G i/o -coupled receptor signaling events. As a consequence, the synergistic activation of signaling cascades that is observed upon co-expression of constitutively active receptors may be modulated by ligands acting at either G i/o -or G q/11 -coupled receptors.
Cell Culture and Transfection-COS-7 African green monkey kidney cells were maintained at 37°C in a humidified 5% CO 2 , 95% air atmosphere in Dulbecco's modified Eagle's medium containing 2 mM Lglutamine, 50 IU/ml penicillin, 50 g/ml streptomycin, and 5% (v/v) fetal calf serum. COS-7 cells were transiently transfected using the DEAE-dextran method as described previously (23,24). The total amount of DNA transfected was maintained constant by the addition of either pcDEF 3 or pcDNA 3 .
[ 3 H]Inositol Phosphate Formation-Cells were seeded in 24-well plates and labeled overnight in inositol-free culture medium supplemented with 1 Ci/ml myo-[2-3 H]inositol 24 h after transfection. Subsequently, the medium was aspirated, and cells were incubated with drugs for 1 h at 37°C in Dulbecco's modified Eagle's medium containing 25 mM Hepes (pH 7.4) and 20 mM LiCl. Incubations were stopped by aspiration of the culture medium and the addition of cold 10 mM formic acid. After 90 min of incubation at 4°C, [ 3 H]inositol phosphates were isolated by anion exchange chromatography as we described previously (23,24) and counted by liquid scintillation.
H 1 R Binding Studies-COS-7 cells used for radioligand binding studies were harvested 48 h after transfection and homogenized in ice-cold H 1 -binding buffer. The COS-7 cell homogenates were incubated for 30 min at 25°C in 50 mM Na 2 /K-phosphate buffer (pH ϭ 7.4) in 400 l with 1 nM [ 3 H]mepyramine. The nonspecific binding was determined in the presence of 1 M mianserin. The incubations were stopped by rapid dilution with 3 ml of ice-cold 50 mM Na 2 /K-phosphate buffer (pH ϭ 7.4). The bound radioactivity was separated by filtration through Whatman GF/C filters that had been treated with 0.3% polyethyleneimine. Filters were washed twice with 3 ml of buffer, and radioactivity retained on the filters was measured by liquid scintillation counting.
5-HT 1B R Binding Studies-COS-7 cells used for 5-HT 1B R binding studies were harvested 48h after transfection and homogenized in ice-cold 5-HT 1B R-binding buffer (50 mM Tris-HCl (pH 7.4), containing 4 mM CaCl 2 , 100 M ascorbic acid, and 10 M pargyline). 5-HT 1B R binding studies were performed using ϳ7 nM [ 3 H]5-HT. The COS-7 cell homogenates were incubated for 30 min at 30°C in 5-HT 1B R-binding buffer in 400 l with 7 nM [ 3 H]5-HT. The nonspecific binding was determined in the presence of 1 M GR-127935. The incubations were stopped by rapid dilution with 3 ml of ice-cold 5-HT 1B R-binding buffer. The bound radioactivity was separated by filtration through Whatman GF/C filters that had been treated with 0.3% polyethyleneimine. Filters were washed twice with 3 ml of buffer, and radioactivity retained on the filters was measured by liquid scintillation counting. A K D of 6.8 nM was subsequently used to calculate the expression levels of the 5-HT 1B R (25).
US28 Receptor Binding Studies-The transfected COS-7 cells used for radioligand binding studies were seeded in 24-well plates; 48 h after transfection, binding was performed on whole cells for 3 h at 4°C using [ 125 I]CCL5 (RANTES) in binding buffer (50 mM HEPES, pH 7.4, 1 mM CaCl 2 , 5 mM MgCl 2 , and 0.5% bovine serum albumin). After incubation, cells were washed four times at 4°C with binding buffer supplemented with 0.5 M NaCl. Nonspecific binding was determined in the presence of 0.1 M unlabeled CCL5.
ELISA-48 h after transfection, receptor expression in COS-7 cells was measured using an ELISA as described previously (18). A mouse anti-HA monoclonal antibody was used as primary antibody and a goat anti-mouse-horseradish peroxidase conjugate as secondary antibody. The 3,3Ј,5,5Ј-tetramethylbenzidine liquid substrate system for ELISA was used as substrate, and the optical density was measured in a Victor 2 at 450 nm.
Analytical Methods-All data shown are expressed as means Ϯ S.E. The data from radioligand-binding and functional assays data were evaluated by a nonlinear, least squares curve-fitting procedure using Graphpad Prism® (GraphPad Software, Inc., San Diego, CA).

Constitutive H 1 R Activity and Inverse Agonistic Properties of
Antihistamines-Transient expression of human H 1 R in COS-7 cells resulted in a high affinity binding site for the H 1 R radioligand [ 3 H]mepyramine (K D ϭ 1.7 Ϯ 0.2 nM, B max ϭ 4.6 Ϯ 0.1 pmol/mg protein, data not shown). H 1 R activates both PLCand NF-B-mediated enhancement of gene transcription as measured by the accumulation of [ 3 H]inositol phosphates and NF-B-driven reporter gene assays, respectively, in an agonistdependent and -independent manner (Fig. 1). Moreover, the H 1 R-selective antagonist mepyramine acts as an inverse H 1 R agonist for the constitutive H 1 R-mediated activation of these signal transduction pathways. These data are in agreement with previous results indicating the constitutive H 1 R activity and inverse H 1 R agonistic properties of various antihistamines (24,26). The H 1 R-mediated effects on PLC and NF-B activation are PTX-insensitive (see also Fig. 2A), indicating that the G i/o family of G proteins is not involved in these H 1 R-mediated responses (24,26). Moreover, we have shown that constitutive H 1 R-mediated activation of these pathways can be enhanced selectively via co-expression of G proteins that belong to the G q/11 family of G proteins (24), confirming that the H 1 R is coupled to G␣ q/11 proteins for modulation of these signaling events in these cells.
Constitutive Activation of NF-B by the h5-HT 1B R-We evaluated the human 5-HT 1B R for constitutive activity upon transient expression of this receptor in COS-7 cells. Expression of the receptor was verified by [ 3 H]5-HT radioligand binding studies, indicating 5-HT 1B R expression levels of ϳ310 fmol/mg protein (data not shown). We also assessed the capacity of 5-HT 1B Rs to mediate activation of NF-B. In 5-HT 1B R-expressing cells the selective 5-HT 1B agonist CGS-12066A (CGS) (27) and 5-HT stimulate NF-B activation only ϳ0.6-fold (Fig. 1D). Despite the functional 5-HT 1B R expression in these cells, we did not detect constitutive 5-HT 1B R activation in our reporter gene assay, as the selective inverse 5-HT 1B R agonist SB-224289 (28) was without effect on basal NF-B activation in 5-HT 1B R-expressing cells (data not shown). Consistent with the notion of a G i/o -coupled 5-HT 1B R receptor, 5-HT 1B R-mediated NF-B activation was PTX-sensitive ( Fig. 1D), indicating the involvement of G i/o proteins in 5-HT 1B R-mediated NF-B activation.
5-HT 1B R Signaling in Cells Co-expressing H 1 R-In agreement with previous reports on agonist-mediated signaling (25,29), stimulation of COS-7 cells co-expressing H 1 R and 5-HT 1B R with 10 M CGS, a selective 5-HT 1B R-agonist (27), results in a marked elevation of the inositol phosphate accumulation (Fig.  1C). Surprisingly, however, treatment of cells co-expressing both H 1 R and 5-HT 1B R with 10 M CGS resulted in a robust PLC and NF-B activation, yielding almost the same efficacy as histamine (His), which was completely sensitive to the inverse H 1 R agonist mepyramine (10 M, Fig. 1 As seen in Figs. 2A and 3, A and C, co-expression of both GPCRs not only results in CGS-induced 5-HT 1B R-mediated NF-B activation but also in a 2-3-fold increase in basal NF-B activation. The fold over basal activation of NF-B by histamine is reduced from 4.8 Ϯ 0.4-fold (n ϭ 76), when H 1 R is expressed alone, to 3.3 Ϯ 0.5-fold (n ϭ 9)-fold, upon co-expression of the 5-HT 1B R. These data do not reflect a suppression of the response to histamine when H 1 R is co-expressed with 5-HT 1B R; however, because of the increase in the level of basal signaling, it does result in a reduced fold of stimulation. The constitutive NF-B activation observed upon co-expression of both GPCRs is partially PTX-sensitive but was completely inhibited by mepyramine ( Fig. 2A). PTX treatment of the cells had no effect on mepyramine-treated cells ( Fig. 2A). H 1 R is known to exhibit stereospecificity toward the enantiomers of the partial inverse H 1 R agonist cetirizine (24,26). Indeed, the inhibition of CGS-induced NF-B activation by the enantiomers of cetirizine is found to be stereospecific in cells coexpressing H 1 R and 5-HT 1B R (Fig. 3B).
The PTX-sensitive increase in basal H 1 R-mediated NF-B activation that is observed upon 5-HT 1B R co-expression, but not the constitutive H 1 R activity itself, is potently inhibited by the inverse 5-HT 1B R agonist SB-224289 (28) (Fig. 3, C and D). activity upon co-expression with H 1 R (Fig. 3C). The observed pIC 50 value of 8.0 Ϯ 0.2 is in good agreement with literature data for SB-224289 acting on 5-HT 1B R (28). SB-224289 is without effect on H 1 R-mediated NF-B activation in cells only expressing the H 1 R, indicating that SB-224289 does not exert its effects of inhibition of constitutive NF-B activation in cells co-expressing H 1 and 5-HT 1B receptors by acting at H 1 Rs. 2 Effects of Co-expression of Other G i/o -coupled Receptors on Constitutive H 1 R Activity-To test whether these observations are exclusive for 5-HT 1B R, we co-expressed H 1 R with other G i/o -coupled receptors: the adenosine A 1 and the muscarinic M 2 receptors. Co-expression of H 1 R with the A 1 or M 2 receptor indeed resulted in a similar increase in basal NF-B activation, which was largely PTX-sensitive (Fig. 2).
As observed after co-expression with the 5-HT 1B R, PTXinsensitive NF-B activation is fully inhibited by mepyramine, indicating that the increased NF-B activation depends on constitutive H 1 R signaling. Stimulation of the M 2 receptor with the muscarinic agonist carbachol results in a further increased mepyramine-and PTX-sensitive NF-B activation (Fig. 2B). We also tested the ability of a G i/o -coupled ␣ 2A adrenergic receptor, in which the C terminus is fused to a G␣ o subunit that harbors a mutation that renders the G␣ o insensitive to PTX (␣ 2A -G o C351I) (20), to enhance NF-B activation of co-expressed H 1 Rs. As expected, stimulation of the ␣ 2A receptor with the specific agonist clonidine resulted in a PTX-insensitive increase in NF-B activation that could be fully inhibited by mepyramine, indicating that the PTX-insensitive G␣ o C351I subunit fused to the ␣ 2A receptor mediates NF-B activation via H 1 R (data not shown).
Effects of Co-expression of G i/o -and G q/11 -coupled Chemokine Receptors-To investigate whether the observed responses are H 1 R-specific or more generally relevant, we tested the effect of the co-expression of two chemokine receptors on their signaling properties. We co-expressed the G q/11 -coupled virally encoded chemokine receptor US28, a chemokine receptor homolog encoded by HCMV, which binds with high affinity the CC-chemokine CCL5 (RANTES) Fig. 4, together with the human G i/ocoupled chemokine receptor CCR1. We previously reported the high constitutive activity of US28 in activation of both PLCand NF-B-mediated enhancement of gene transcription as measured by the accumulation of [ 3 H]inositolphosphates and NF-B-driven reporter gene assays, respectively (23) (Fig. 4, A,  B, and E). CCL5 binds with high affinity to receptor US28 but does not increase the level of activation of receptor US28, and therefore CCL5 acts as a neutral antagonist for this receptor (Fig. 4A) (23). In contrast, CCL5 binds to CCR1 with high affinity and acts as a CCR1 agonist, inducing calcium mobilization in a PTX-sensitive manner (see Ref. 30). When tested for activation of NF-B, CCR1 does not exhibit basal NF-B activation, and stimulation of CCR1 with CCL5 results in a small, PTX-sensitive NF-B activation. Upon co-expression of receptor US28 and CCR1, however, CCL5 elicits a robust agonist response that is completely PTX-sensitive, indicating the involvement of G i/o proteins in CCL5-mediated NF-B activation (Fig. 4A). The binding of CCL5 to receptor US28 is not involved in the observed phenomenon as shown by co-expression of CCR1 with an N-terminal truncation mutant of receptor US28 (⌬(2-22)-US28), in which the first 22 residues of receptor US28 are deleted. The N-terminal truncation mutant ⌬(2-22)-US28 receptor does not exhibit [ 125 I]CCL5 binding as observed by [ 125 I]CCL5 saturation binding analysis (Fig. 4D), in line with reports that chemokine binding to chemokine receptors is crit-ically dependent on the N terminus of the receptor (31). We used an ELISA to confirm cell surface expression of ⌬(2-22)-US28 (45% of wild-type receptor US28, Fig. 4C). ⌬(2-22)-US28 still activates both NF-B (Fig. 4B) and PLC constitutively (Fig. 4E), and co-expression of (⌬(2-22)-US28 with CCR1 resulted in CCL5-induced CCR1-mediated activation of NF-B (Fig. 4B). In contrast, a mutant US28 receptor (US28-R129A) (22), which is expressed at the cell surface (Fig. 4, C and D), and which binds CCL5 with unchanged affinity in comparison with wild-type receptor US28 (Fig. 4D) but is devoid of constitutive activity (Fig. 4, B and E), does not transduce CCR1-mediated NF-B activation upon co-expression with CCR1 (Fig. 4B); this indicates the importance of the constitutive activity of receptor US28 for the observed response to CCL5.
The effects of CCL5 that are observed upon co-expression of CCR1 and receptor US28 are similar to the effects observed for 5-HT 1B R agonists in cells co-expressing H 1 R and 5-HT 1B R. The co-expression of the G q/11 -coupled receptor US28 with the G i/ocoupled CCR1 results in the alteration of CCR1 signaling capabilities. By co-expressing receptor US28 and CCR1, CCR1 acquires the capability to signal through NF-B upon agonist (CCL5) stimulation. However, it does not appear to result in increased constitutive receptor activity, as we did not observe modulation of basal NF-B activation in cells expressing receptor US28 upon co-expression of CCR1 (Fig. 4A).

Mechanism of Synergistic NF-B Activation by Co-expressed G i/o -and G q/11 -coupled Receptors-Various
GPCRs may activate signal transduction pathways leading to NF-B activation (see Fig. 7). Both the H 1 R and receptor US28 are reported to activate G proteins belonging to both the G i/o and G q/11 families of G proteins (23,24,26,(32)(33)(34)(35). Yet, G␣ i subunits are not implicated in the H 1 R or receptor US28-mediated activation of phospholipase C or NF-B, as PTX treatment of the cells did not alter H 1 R or receptor US28-mediated [ 3 H]inositol phosphate production or NF-B activation (23, 24) (see also Figs. 2A and 4A).
Co-expression of H 1 R or receptor US28 with G␣ q or G␣ 11 resulted in a G␣ expression level-dependent increase in (constitutive) H 1 R-or receptor US28-mediated NF-B activation and inositol 1,4,5-trisphosphate production (23,24) (Fig. 5A). Scavenging of G␤␥ subunits by co-expression with either G␣ t (Fig. 5) or GRK2 (data not shown) effectively reduced constitutive H 1 R and receptor US28-mediated NF-B activation, whereas co-expression with G␤␥ subunits resulted in the elevation of the constitutive receptor activity of both H 1 R and receptor US28 (23,24) (Fig. 5A). In line with published data on GPCR-mediated activation of NF-B (see also Fig. 7) these data implicate both the G␣ q/11 and G␤␥ subunits, which are released upon H 1 R or receptor US28-mediated activation of G q/11 proteins, in the activation of signal transduction pathways resulting in activation of NF-B (23,24).
Activation of H 1 R (36 -38) and G␣ q signaling (39,40), as well as NF-B activation, have been linked to the activation of small G proteins and especially the Rho-like G proteins (41)(42)(43)(44)(45). We therefore investigated the role of Rho signaling in the H 1 receptor-mediated activation NF-B. C. botulinum c 3 exoenzyme, which prevents the coupling of RhoA, -B, and -C to its downstream effectors by irreversible ADP-ribosylation, has often been used to demonstrate the involvement of Rho in GPCR signaling, including NF-B activation (42). Co-expression of the c 3 exoenzyme abrogates all, both constitutive as well as agonist-induced, human H 1 R-mediated NF-B activation (Fig. 5B).
We have previously reported that constitutively activated (Q205L) G␣ i/o proteins do not enhance H 1 R-mediated NF-B activation (24), which indicates that activation of G i/o -coupled receptors most likely results in the release of G␤␥ subunits from activated G␣ i/o proteins, which then may serve to potentiate signaling mediated by G q/11 -couled receptors. To mimic activation of G i/o -coupled receptors we used mastoparan-7 (M7), a relatively stable analogue of mastoparan, to directly activate G␣ i/o proteins and to stimulate both G␣ i/o -and G␤␥mediated signal transduction pathways (46). M7 induced a PTX-sensitive NF-B activation in cells expressing 5-HT 1B R to a similar extent as the 5-HT 1B R agonist CGS (Fig. 6A). M7 also induces NF-B activation in cells expressing H 1 R; this effect is effectively blocked by the inverse H 1 R agonist mepyramine (Fig. 6B), demonstrating that activated G␣ i/o proteins may enhance constitutive H 1 R-mediated NF-B activation. Mepyramine did not affect M7-induced NF-B activation in cells that do not express H 1 R (Fig. 6A). DISCUSSION We co-transfected COS-7 cells with cDNAs encoding various GPCRs and examined agonist-induced as well as constitutive GPCR-mediated stimulation of PLC-and NF-B-regulated gene expression. Expression of either the human histamine H 1 R or the HCMV-encoded receptor US28 in COS-7 cells results in the activation of both PLC-and NF-B-regulated gene expression in an agonist-dependent (H 1 R) and -independent manner (H 1 R and receptor US28) (23,24,26). H 1 R-or receptor US28-mediated activation of PLC and NF-B is PTX-insensitive (23,24) (Figs. 1, 2, and 4), indicating that G i/o proteins are not involved in these H 1 R-and receptor US28-mediated responses in COS-7 cells. By co-expression of a variety of G protein subunits, we have shown earlier that G␣ q/11 as well as G␤␥ subunits are involved in NF-B activation via H 1 R or receptor US28 (23,24) (Fig. 7), whereas for both receptors, the activation of PLC is mediated via the classical G␣ q/11 pathway (23,24). In the present study we have investigated the effects of co-expression of various G i/o -coupled receptors on agonistinduced and constitutive activity of the G q/11 -coupled H 1 R and receptor US28 using a heterologous expression system.
In agreement with reports on synergistic agonist-induced PLC activation in cells co-expressing H 1 R and G i/o -coupled 5-HT 1B R (25,29), stimulation of cells co-expressing both receptors with the selective 5-HT 1B agonist CGS-12066A (27) results in a marked synergistic elevation of inositol phosphate accumulation (Fig. 1). Moreover, upon co-expression of 5-HT 1B R with H 1 R, CGS elicited a robust PTX-sensitive activation of NF-B, whereas CGS had no such effects on NF-B activation in cells lacking 5-HT 1B Rs (Fig. 1).
Unexpectedly, we found that H 1 R-mediated effects of H 1 Rselective inverse H 1 R agonists such as mepyramine and cetirizine include the inhibition of CGS-induced 5-HT 1B R-mediated activation of both PLC and NF-B in cells co-expressing both 5-HT 1B R and H 1 R (Figs. 1, 2, and 4). In contrast, inverse H 1 R agonists had no such effects on 5-HT 1B R-mediated signaling in cells lacking H 1 Rs. Radioligand binding studies confirmed the H 1 R selectivity of the inverse H 1 R agonist, indicating that inverse H 1 R agonists do not modulate 5-HT 1B R signaling via a direct interaction with either G i/o -coupled 5-HT 1B R or with the activated G i/o proteins. We concluded from these results that the interaction of inverse H 1 R agonists with H 1 Rs induces biological responses that negatively interfere with 5-HT 1B R signaling. Because these experiments were performed in the absence of histamine, it seems that the constitutive activity of the H 1 R (24, 26) is a critical determinant for the observed functional 5-HT 1B R-mediated responses to the 5-HT 1B R agonist.
The co-expression of G q/11 -and G i/o -coupled receptors may not only give rise to agonist induced signaling of G i/o -coupled receptors but may also affect constitutive signaling. In fact, co-expression of H 1 R and 5-HT 1B R allows the detection of otherwise undetectable constitutive 5-HT 1B R activity. Co-expression of 5-HT 1B R together with H 1 R resulted in a significantly increased constitutive activation of NF-B (Figs. 2 and 3), which was completely inhibited by the inverse H 1 R agonist mepyramine but also, partially, by the inverse 5-HT 1B R agonist SB-224289 (28). This inverse 5-HT 1B R agonist inhibited only the increased constitutive NF-B activation that is observed upon co-expression of the 5-HT 1B R but not the constitutive H 1 R-mediated NF-B activation (Fig. 3). Based on these find-ings we conclude that the otherwise minimal 5-HT 1B R agonistinduced or undetectable constitutive 5-HT 1B R responses are enhanced through constitutive H 1 Rs.
We subsequently investigated whether the observed phenomena are receptor-specific by testing the effects of co-expression of other G q/11 -and G i/o -coupled receptors on NF-B activation. To test whether these phenomena were specific for 5-HT 1B R, we co-expressed several other G i/o -coupled receptors with H 1 R. Stimulation of cells co-expressing the muscarinic M 2 receptor (M 2 R) with H 1 R with the M 2 R agonist carbachol resulted in a PTX-sensitive activation of NF-B. Similar to our findings upon H 1 R/5-HT 1B R co-expression, the inverse H 1 R agonist mepyramine potently inhibited carbachol-induced M 2 R signaling in cells co-expressing H 1 Rs (Fig. 2). As observed for co-expression of 5-HT 1B R, we saw an increased constitutive activation of NF-B upon co-expression of M 2 R with H 1 R. Because similar observations were made with the G i/o -coupled adenosine A 1 (Fig. 2) and the adrenergic ␣ 2 receptors, constitutively active G q/11 -coupled H 1 R seems to act as a regulatory switch for important cellular functions of G i/o -coupled GPCRs, including the activation of PLC-and NF-B-mediated gene transcription.
To assess whether the observed phenomenon is limited to H 1 R, we tested an unrelated GPCR, the virally encoded chemokine receptor US28, which is highly constitutively active (23), for its capacity to transduce signaling events mediated through activation of G i/o -coupled receptors. As observed in the experiments with cells co-expressing the 5-HT 1B , M 2 , or ␣ 2 receptors together with H 1 R, activation of the human G i/ocoupled chemokine receptor CCR1 with the CCR1 agonist CCL5 induces a robust PTX-sensitive NF-B activation in cells co-expressing the constitutively active G q/11 -coupled receptor US28 (Fig. 4). In contrast, CCL5 does not affect NF-B-medi- ated gene transcription in cells that express CCR1 but not receptor US28. Moreover, CCL5 has no effect on cells that express receptor US28, for which CCL5 is a neutral antagonist (3). However, the binding of CCL5 to US28 is not involved in the observed phenomenon, as CCL5 also induces a robust PTXsensitive NF-B activation in cells co-expressing CCR1 together with the constitutively active ⌬22-US28 receptor, which is an N-terminal deletion mutant receptor US28 that does not bind CCL5. Co-expression of CCR1 with a mutant receptor US28 (US28-R129A) (22) that is devoid of constitutive receptor activity confirmed that the constitutive activity of receptor US28 is crucial for CCL5-mediated activation of NF-B by co-expressed CCR1 receptors (Fig. 4). These data show that in COS-7 cells the active state of the G q/11 -coupled receptors, such as H 1 R and receptor US28, is crucial for signaling by various G i/o -coupled receptors. Apparently, constitutively active G q/11coupled receptors act as regulators of G i/o -coupled receptormediated signal transduction cascades.
GPCRs activate NF-B through a wide variety of signaling mechanisms (Fig. 7). The signal transduction pathways of G q/11 -and G i/o -coupled receptors provide ample opportunities for cross-talk, ultimately converging at the level of NF-B activation. Our previous data implement both G␣ q/11 and the released G␤␥ subunits in the activation of NF-B by H 1 R and receptor US28 (23,24). We have shown by co-expression or scavenging of G␤␥ subunits that G␤␥ subunits are essential but not exclusive signaling moieties for NF-B activation. Recent evidence indicates that active G␣ q/11 proteins activate RhoA (40), which can lead to activation of NF-B via PLD (47,48). Previous studies have also reported RhoA as a signaling partner for H 1 R (49,50). Rho-mediated signaling has been implicated in, for example, H 1 R-mediated airway hyper-responsiveness (38) and activation of PLD (43). Co-expression of C. botulinum c 3 exoenzyme, which is known to inactivate the small G protein Rho (42), completely abolishes H 1 R-mediated NF-B activation (Fig. 5), implying that Rho is required in the H 1 R-mediated activation of NF-B.
Although the activation of G i/o -coupled receptors may result in a limited degree of NF-B activation, the expression of activated G␣ i proteins in COS-7 cells does not result in NF-B (23,24) or Rho activation (51). It has been suggested that G i/o proteins are not sufficient or necessary for GPCR-mediated activation of Rho (52). To obtain evidence that the G␤␥ subunits of G i/o proteins have a role in G i/o -coupled receptor-mediated NF-B activation, we stimulated COS-7 cells with mastoparan-7. Activation of heterotrimeric G␣ i/o proteins using M7 results in the release of both the activated G␣ i/o and G␤␥ subunits (46). M7 induced PTX-sensitive NF-B activation in COS-7 cells expressing 5-HT 1B R to a similar extent as the 5-HT 1B R agonist CGS (Fig. 6). Moreover, M7 stimulated NF-B activation in cells expressing the H 1 R to a similar extent as achieved by co-expression of G␤␥ subunits (Figs. 5 and 6); this response was completely sensitive to the inverse H 1 R agonist mepyramine. Activation of G i/o -coupled receptors may therefore provide additional free G␤␥ subunits that synergize with constitutive H 1 R-mediated NF-B activation.
While this manuscript was in preparation, constitutively active G i/o -coupled cannabinoid CB1 receptors (CB1R) were shown to sensitize MAPK activation by the G q/11 -coupled orexin 1 receptor (OX1R), upon co-expression in Chinese hamster ovary cells (53). In this study orexin-mediated MAPK activation is shown to be sensitive to an inverse CB1R agonist (53). The authors explain the observed synergism by heterodimerization of the two GPCRs (53). In view of our findings with M7 and the apparent lack of specificity for GPCR combinations, we do not consider heterodimerization a major determinant for our observations. However, our study does underscore the importance of constitutive GPCR activity in the modulation of cell signaling.
Our findings are consistent with the reported necessity for preactivation of various G q/11 -coupled receptors for G i/o -coupled receptor-mediated signaling both in vitro and ex vivo (54). Pretreatment of blood vessels with histamine, for example, is known to yield an enhanced contractile response to 5-HT, which is mediated by previously "silent" 5-HT receptors (55). The conditional synergy might have important (patho)physiological significance, because constitutive activity of G q/11 -coupled receptors might be prominent in (patho)physiology (33). Multiple mechanisms have previously been implicated in the NF-B activation that is observed upon cytomegalovirus infection (56). HCMV infection of smooth muscle cells may result not only in the expression of constitutively active US28 receptors (18) but also in G q/11 -mediated (31) as well as PTX-sensitive NF-B activation (57). Our present data suggest that expression of receptor US28 upon cytomegalovirus infection may be implicated in both G q/11 -mediated and PTX-sensitive NF-B activation through the unmasking of signaling responses of G i/o -coupled receptors. Therefore, these results provide a new mechanism by which the expression of constitutively active GPCRs, which may be constitutively active under physiological conditions (17) because of receptor mutation (15,16) or viral infection (18,19), may regulate signaling events through sensitization of cellular communication that may ultimately result in the initiation and/or progression of disease. These findings may also attribute potentially new roles to naturally occurring inverse agonists (see Ref. 58). Future studies will be required to address these issues in more detail.
In conclusion, our findings reveal a heretofore unrecognized role for constitutively active G q/11 -coupled receptors in the signaling events initiated by G i/o -coupled receptors. Constitutively active G q/11 -coupled receptors constitute a regulatory switch for signal transduction by G i/o -coupled receptors to unmask signaling events. The observed conditional synergy may be a key to the identification of the constitutive activity of native G q/11coupled GPCRs in vivo and could be useful as a sensitive screening strategy in drug discovery (e.g. for "de-orphanizing" receptors). Our findings suggest that endogenous levels of constitutive GPCR tone can be considerably higher than antici-pated on cellular receptor number and expression levels of G proteins or effector molecules. Also, our data indicate that selective inverse agonists may affect signaling events that are induced upon activation of unrelated GPCRs. Although the implications of the present observations remain to be fully ascertained, these data clearly demonstrate the importance of cellular environment for GPCR function.