The CXC chemokine receptor encoded by herpesvirus saimiri, ECRF3, shows ligand-regulated signaling through Gi, Gq, and G12/13 proteins but constitutive signaling only through Gi and G12/13 proteins.

Open reading frame 74 (ORF74) of many gamma(2)-herpesviruses encodes a CXC chemokine receptor. The molecular pharmacological profile of ORF74 from herpesvirus saimiri, ECRF3, is characterized here and compared with that of the well known ORF74 from human herpesvirus 8 (HHV8). The ECRF3 receptor bound the so-called ELR (Glu-Leu-Arg) CXC chemokines (125)I-CXCL1/GRO alpha, (125)I-CXCL6/GCP-2, and (125)I-CXCL8/interleukin-8 with high affinity; but in contrast to ORF74 from HHV8, it did not bind the non-ELR CXC chemokine (125)I-CXCL10/IP10. Interestingly, the B(max) value for CXCL6/GCP-2 was 3-fold higher than the capacity for maximal binding of CXCL1/GRO alpha to ECRF3 and 85-fold higher than that of CXCL8/interleukin-8, despite similar affinities. Like ORF74 from HHV8, ECRF3 activated a broad range of pathways (G(q), G(i), and G(12/13) as well as the cAMP response element-binding protein, NF-kappa B, NFAT, and serum response element transcription factors) in a ligand-regulated manner, with CXCL6/GCP-2 being the most potent and efficacious agonist. ECRF3 signaled constitutively through G(i) and G(12/13), but surprisingly not through G(q). At the level of transcription factor activation, the serum response element was activated constitutively by ECRF3, whereas cAMP response element-binding protein, NFAT, and NF-kappa B were only ligand-regulated. The maximal signaling capacities were similar for the two receptors; however, the ligand-regulated signaling was responsible for the major part of the total ECRF3 signaling and only for a minor part of the total HHV8 ORF74 signaling. The activation pattern of ECRF3 with constitutive activation of some (but not all) of the employed pathways has not been seen before in endogenous or virus-encoded chemokine receptors. The results suggest that the unique ligand selectivity of ECRF3 among ORF74 receptors could reflect differences in the cellular tropism of the gamma(2)-herpesviruses.

The general knowledge of 7TM receptor signal transduction mechanisms and knowledge of the properties of virus-encoded chemokine receptors have developed considerably since the first description of ECRF3 from HVS in 1993 (9). In this study, we characterized ECRF3 on this basis and surprisingly found that it has a very interesting molecular pharmacological phenotype in being constitutively active through G i and G 12/13 (the latter shown by studying the serum response element (SRE) transcription factor in the presence of pathway-specific inhibitors), whereas it activates a broader spectrum of signaling pathways in a ligand-regulated manner, with CXCL6/GCP-2 being the most potent agonist.
Iodination of CXCL6/GCP-2 and CXCL10/IP10 -The Bolton-Hunter reagent was dried under a gentle stream of nitrogen for 30 -60 min. 5-10 g of chemokine was incubated on ice with 1.5 mCi of Bolton-Hunter reagent in a total volume of 50 l of 0.1 mM borate buffer (pH 8.5) for 1 h. The reaction was terminated by addition of 0.25 ml of H 2 O supplemented with 0.1% (v/v) trifluoroacetic acid. The labeled chemokines were purified by reverse-phase high pressure liquid chromatography.
Transfections and Tissue Culture-COS-7 cells were grown at 10% CO 2 and 37°C in Dulbecco's modified Eagle's medium with Glutamax (catalog no. 21885-025, Invitrogen) adjusted with 10% fetal bovine serum, 180 units/ml penicillin, and 45 g/ml streptomycin. HEK293 cells were grown in Dulbecco's modified Eagle's medium adjusted to contain 4500 mg/liter glucose with 10% fetal bovine serum, 180 units/ml penicillin, and 45 g/ml streptomycin at 10% CO 2 and 37°C. The HEK293 cell medium was modified to contain heat-inactivated fetal bovine serum and no penicillin and streptomycin during luciferase-based assays. Transfection of the COS-7 cells was performed by the calcium phosphate precipitation method (8) for the competition binding and PI turnover experiments and by the cationic lipid reagent method with LipofectAMINE TM 2000 reagent in serum-free Opti-MEM I according to the manufacturer's recommendations for the luciferase-based transcription factor experiments. HEK293 cells were transfected by the calcium phosphate precipitation method (8) for the adenylate cyclase inhibition experiments and with Lipofect-AMINE TM 2000 reagent in serum-free Opti-MEM I according to the manufacturer's recommendations for the luciferase-based transcription factor experiments.
Binding Experiments-COS-7 cells were transferred to culture plates 1 day after transfection. The number of cells seeded per well was determined by the apparent expression efficiency of the receptors, with the goal of obtaining 5-10% specific binding of the added radioactive ligand. 3-5 ϫ 10 5 cells/well were used to test specific binding. 2 days after transfection, cells were assayed by competition binding for 3 h at 4°C using 12 pM 125 I-CXCL8/IL-8, 125 I-CXCL1/GRO␣, 125 I-CXCL6/ GCP-2, or 125 I-CXCL10/IP10 plus unlabeled ligand in 0.5 ml of 50 mM Hepes (pH 7.4) supplemented with 1 mM CaCl 2 , 5 mM MgCl 2 , and 0.5% (w/v) bovine serum albumin. After incubation, cells were washed quickly four times with 4°C binding buffer supplemented with 0.5 M NaCl. Nonspecific binding was determined as the binding in the presence of 0.1 M unlabeled chemokine. Determinations were made in duplicates.
Phosphatidylinositol Assay (PI Turnover)-1 day after transfection, COS-7 cells (2.5 ϫ 10 4 /well) were incubated for 24 h with 5 Ci/ml FIG. 1. Dendrogram of herpesvirus-encoded chemokine receptors. The phylogenetic tree was constructed using ClustalW Version 1.5 alignment of the whole sequence, followed by analysis using the Distance program of the GCG Wisconsin package (kindly provided by Francois Talabot, Ares-Serono). Constitutively active receptors are marked in red. Solid lines indicate a more peripheral cluster association. CMV, cytomegalovirus; EHV2, equine herpesvirus 2; AtHV, atheles herpesvirus. myo-[ 3 H]inositol in 0.3 ml/well growth medium. Cells were washed twice with 20 mM Hepes (pH 7.4), 140 mM NaCl, 5 mM KCl, 1 mM MgSO 4 , 1 mM CaCl 2 , 10 mM glucose, and 0.05% (w/v) bovine serum albumin and were incubated in 0.3 ml of the same buffer supplemented with 10 mM LiCl at 37°C for 90 min in the presence or absence of chemokines. PTx (100 ng/ml) were added 18 h prior to the experiment. Cells were extracted by addition of 1 ml/well 10 mM formic acid, followed by a 30-min incubation on ice. The generated [ 3 H]inositol phosphates were purified on AG 1-X8 anion-exchange resin (22). Determinations were made in duplicates.
Adenylate Cyclase Inhibition Assay (cAMP Assay)-1 day after transfection, HEK293 cells (2.5 ϫ 10 4 /well) were incubated for 24 h with 2 Ci/ml [ 3 H]adenine in 0.5 ml/well growth medium. Cells were washed twice with 25 mM Hepes (pH 7.2), 0.75 mM NaH 2 PO 4 , 140 mM NaCl, and 0.05% (w/v) bovine serum albumin, and 0.5 ml of the same buffer supplemented with the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (1 mM) was added together with the adenylate cyclase activator forskolin (15 M) to the cells. Ligands were then added, and the reaction was terminated after a 15-min incubation at 37°C. After incubation, the cells were placed on ice; the medium was removed; and the cells were lysed in 1 ml of 5% (w/v) trichloroacetic acid supplemented with 0.1 mM cAMP and 0.1 mM ATP for 30 min. The lysis mixtures were loaded onto Dowex columns, which were washed with 2 ml of water, placed on the top of alumina columns, and washed again with 10 ml of water. The alumina columns were eluted with 6 ml of 0.1 M imidazole into 15 ml of scintillation fluid (Highsafe III). Columns were reused up to 15 times. Dowex columns were regenerated by adding 10 ml of 2 N HCl followed by 10 ml of water; the alumina columns were regenerated by adding 2 ml of 1 M imidazole, 10 ml of 0.1 M imidazole, and finally 5 ml of water. The effect of 100 ng/ml PTx was tested by adding it to the cells 18 h before ligand addition. Determinations were made in triplicates.
Constitutive SRE, NFAT, and NF-B cis-Reporting and CREB trans-Reporting Luciferase Assays-Cells were seeded at 35,000 cells/well in culture plates 24 h prior to transfection and were transfected with the cis-reporter plasmid (pSRE-Luc, pNFAT-Luc, or pNF-B-Luc; 50 ng/ well) and the receptor plasmid (0 -50 ng/well). For the trans-reporting system, 50 ng/well trans-activator plasmid (pFR-Luc) was added together with 6 ng/well trans-reporter plasmid (pFA2-CREB) and receptor plasmids. 24 h after transfection, the cells were washed twice with Dulbecco's phosphate-buffered saline, and luminescence was measured in a microplate scintillation and luminescence counter (Top-counter, PerkinElmer Life Sciences) after a 10-min incubation in 100 l of Dulbecco's phosphate-buffered saline together with 100 l of LucLite substrate. All determinations were made in quadruplicates.
Ligand and Inhibitor Modification of Transcription Factor Activity-The cells were transfected with 10 ng/well receptor cDNA and the above-mentioned concentrations of reporter and activator plasmids. Chemokine ligands were added at concentrations ranging from 10 Ϫ11 to 10 Ϫ7 M for the dose-response curves. Alternatively, a constant concentration (100 nM) was added to the receptor gene dose experiments (see above). In both cases, as well as upon addition of the inhibitors PTx (100 ng/ml) and U-73122 (10 M) to the CREB, NF-B, and NFAT transcription factors and of PTx (100 ng/ml), C3 exoenzyme (10 g/ml), and Y-27623 (10 M) to the SRE transcription factor, the chemokines/inhibitors were added immediately following transfection, and luminescence was measured 24 h post-transfection as described above. All determinations were made in quadruplicates.
Calculations-IC 50 and EC 50 values were determined by nonlinear regression, and B max values were calculated using the GraphPad Prism Version 2 software.
ECRF3 Signals via G q in a Ligand-regulated (but Not Constitutive) Manner, with CXCL6/GCP-2 being the Most Potent and Efficacious Agonist-PI accumulation assays performed in transiently transfected COS-7 cells demonstrated that ECRF3 activated G q upon ligand binding. As expected from the initial study of ECRF3 identifying CXCL1/GRO␣ as an agonist in Ca 2ϩ release (9), this chemokine stimulated ECRF3 with a potency (IC 50 ) of 5.5 nM (Fig. 3) and an efficacy of 2.4-fold stimulation above the basal activity, a phenotype resembling the previously shown agonism (IC 50 ϳ 1 nM) of CXCL1/GRO␣ toward ORF74-HHV8 (see Fig. 5B) (8). The neutral ligands for ORF74-HHV8, CXCL8/IL-8 and CXCL7/NAP-2, did not influence the G q activity of ECRF3 at the maximal chosen concentration of 100 nM (Fig. 3). Surprisingly, an inverse agonist for ORF74-HHV8 (13), CXCL6/GCP-2, turned out to be the most potent and efficacious agonist with ECRF3, with an EC 50 of 0.42 nM and an efficacy of 5.4-fold stimulation above the basal activity ( Fig. 3). Two other inverse agonists for ORF74-HHV8, CXCL10/IP10 and vCCL2/vMIP-II, had no effect on the basal activity of ECRF3; however, with respect to antagonizing the agonist-stimulated receptor, vCCL2/vMIP-II (but not CXCL10/ IP10) inhibited the CXCL6/GCP-2 (10 nM)-induced activity with a potency of 77 nM (data not shown).
Gene dose experiments for ECRF3 performed in parallel with ORF74-HHV8 demonstrated that ECRF3, in contrast to ORF74-HHV8, did not activate G q in a ligand-independent (constitutive) manner (Fig. 4A). However, a comparison of the maximal ligand-stimulated activities in the gene dose setup indicated that ECRF3 could be stimulated to the same level as ORF74-HHV8 at equivalent DNA concentrations (Fig. 4B). Thus, 100 nM CXCL6/GCP-2 stimulated ECRF3 to the same level to which 100 nM CXCL1/GRO␣ stimulated ORF74-HHV8 (at equivalent DNA concentrations), corresponding to the almost similar B max values for CXCL6/GCP-2 binding to ECRF3 and for CXCL1/GRO␣ binding to ORF74-HHV8. In contrast, 100 nM CXCL1/GRO␣ stimulated ECRF3 with lower efficacy, corresponding to the lower B max value for CXCL1/GRO␣ binding to ECRF3.
Phospholipase C (PLC) is activated mainly through an interaction with the ␣-subunit from G q (23); however, the ␤␥-dimer released from activated G i is yet another way to activate PLC (24). The ␤␥-dimer has the highest affinity for the PLC␤2 isoform compared with PLC␤1 and PLC␤3; and because COS-7 cells have a very low content of PLC␤2, PLC activation in these cells is believed to be through interaction with G q (25). We performed PI turnover experiments in the presence of the G i inhibitor PTx (100 ng/ml) and observed no inhibition of the efficacy of CXCL6/GCP-2-induced ECRF3 activation or of CXCL1/GRO␣-induced ORF74-HHV8 activation in the presence of PTx (Fig. 5, A and B, respectively). This supports G q activation of PLC. In fact, we observed an increase in the efficacies as well as in the basal constitutive activity of ORF74-HHV8, but not in the basal non-constitutive activity of ECRF3 (Fig. 5, A and B). PTx had no effect on cells transfected with the empty expression vector (data not shown).
ECRF3 Activates a Broad Range of Downstream Transcription Factors in a Ligand-regulated (but Not Constitutive) Manner-Many virus-encoded chemokine receptors have been shown to signal constitutively through a broad range of transcription factors (26 -31). In a gene dose setup using luciferasebased reporter systems for CREB, NF-B, and NFAT (three transcription factors with a known dependence on G q activation) and receptors transiently expressed in HEK293 cells (Fig.  6), none of the transcription factors were constitutively activated by ECRF3. In contrast, all three were activated by increasing doses of ORF74-HHV8 in parallel with ECRF3 (Fig. 6, A-C, for CREB, NF-B, and NFAT, respectively) until a plateau was reached. This plateau was probably due to limitations in the content of intracellular signaling mediators from the cell surface to nucleus. The PI turnover experiments are not de- Ϫ1.07 Ϯ 0.19 (5) FIG . 3. ECRF3 activates G q in a ligand-regulated manner. Transiently transfected COS-7 cells were used for the PI turnover assays. pendent on the same, many steps in the signaling cascade (PLC␤2 is closer to the receptor) and did not reach this plateau (Fig. 4). Moreover, receptor expression was higher in the luciferase-based transcription factor assays due to a more efficient transfection procedure. Upon addition of CXCL6/GCP-2 to the ECRF3-expressing cells, all three pathways were activated with remarkable high and almost identical potencies (EC 50 ) of 0.90, 0.70, and 0.42 nM for CREB, NF-B, and NFAT activities, respectively (Fig. 6, D-F). CXCL6/GCP-2 did not activate these transcription factors in cells transfected with the empty expression vector (data not shown). The same scenario of ligand regulation and lack of constitutive activity for ECRF3 (but maintained for ORF74-HHV8) was observed in transiently transfected COS-7 cells (data not shown).
Inhibitors of two different enzymes in the early branches of the signal transduction pathways were used to describe the G q versus G i dependence of the transcriptional activity for the two receptors. We tested the effect of 100 ng/ml PTx (G i inhibitor) and 10 M U-73122 (PLC inhibitor) on the CXCL6/GCP-2 doseresponse curves for ECRF3 (Fig. 7, A-C) and on the constitutive activity of ORF74-HHV8 (Fig. 7, D-F). The effect of the inhibitors was remarkably similar for the two receptors. For CREB activation, PTx had only very little effect, in agreement with the G q /G s -dependent nature of this pathway (32), whereas U-73122 resulted in 55-70% inhibition of activity. Concomitant application of the inhibitors increased the inhibition to 85-95% (Fig. 7, A and D). The NF-B activation of ECRF3 by CXCL6/ GCP-2 was totally inhibited by U-73122, whereas the activation of ORF74-HHV8 was inhibited by 70%, consistent with the G q -dependent nature of NF-B (32). PTx resulted in a surprisingly high inhibition of 45-60%, and concomitant application of these two inhibitors totally abolished the NF-B activity of both receptors (Fig. 7, B and E). NFAT is traditionally described as being G q -dependent (32); and consistent with this, we found that U-73122 efficiently inhibited the activities by 40 -50%. However, PTx also had an effect on NFAT activity, with 60% inhibition for ECRF3 and 40% inhibition for ORF74-HHV8. In combination, these inhibitors resulted in almost total inhibition (Fig. 7, C and F). Thus, NF-B and NFAT were substantially influenced by the PLC inhibitor U-73122, in agreement with the known G q dependence of these transcription factors (32). Surprisingly, both transcription factors were also inhibited by PTx, indicating that G i plays a role in the regulation of these pathways. The content of the PLC␤2 isoform in HEK293 cells makes it most likely that the PTx effect on NF-B and NFAT activation is through inhibition of ␤␥mediated PLC activation. In fact, PTx inhibition of constitutive and/or ligand-induced NF-B activation by ORF74-HHV8 has been reported previously by several groups (26,27,31).

ECRF3 Constitutively Activates the G i Pathway, and Agonists and Inverse Agonists Further Regulate This Signaling-
Most (if not all) of the endogenous chemokine receptors characterized so far signal in a PTx-sensitive and ligand-dependent manner through inhibition of adenylate cyclase (3). In addition, ORF74-HHV8 also constitutively activates G i (26,27,31). To study constitutive G i signaling of ECRF3, transiently transfected HEK293 cells were treated with forskolin. Stimulation of ECRF3 with CXCL6/GCP-2 resulted in inhibition of the forskolin-induced adenylate cyclase activity, with an EC 50 of 0.30 nM (Fig. 8A). In contrast, CXCL1/GRO␣ acted as a partial agonist (Fig. 8A), whereas CXCL8/IL-8 had no effect within the chosen concentrations (data not shown). The antagonist vCCL2/vMIP-II acted as an inverse agonist, as it inhibited the constitutive activity with a surprisingly high potency (EC 50 ) of 0.14 nM, i.e. the highest potency ever observed for vCCL2/ vMIP-II with any receptor yet tested (13,21,33). The chemokine modulations could be eliminated by the presence of 100 ng/ml PTx, indicating G i as an involved G-protein (data not shown). In theory, cells expressing constitutively active G icoupled receptors display lower levels of cAMP production upon forskolin-induced adenylate cyclase stimulation compared with control cells (34). Consistent with this, we observed slightly lower levels of forskolin-induced adenylate cyclase stimulation in ECRF3-expressing cells compared with control cells. Importantly, PTx addition eliminated this difference, supporting the constitutive nature of G i activation by ECRF3 (Fig. 8B).
Constitutive and Ligand-regulated Activation of the Transcription Factor SRE-The SRE transcription factor has been shown to be dependent on G i as well as G 12/13 , the latter via activation of Rho and Rho kinase (32,35,36). A gene dose experiment revealed constitutive signaling for ECRF3 (Fig. 9, A and B) as well as for ORF74-HHV8 (Fig. 9B) expressed in transiently transfected COS-7 cells through the SRE pathway. The constitutive activity of ECRF3 was ϳ2.5-fold higher than the basal cell activity (Fig. 9A), whereas the constitutive activity of ORF74-HHV8 was ϳ10-fold higher than the basal cell activity (Fig. 9B). Furthermore, both receptors where activated further by their respective agonists. Thus, CXCL6/GCP-2 stimulated the ECRF3-mediated SRE activity with a potency of 64 pM (Fig. 9C) 9D). A similar scenario of constitutive and ligand-regulated activities was obtained in transiently transfected HEK293 cells (data not shown). PTx (100 ng/ml) was applied to determine the G i contribution to the constitutive and ligand-regulated SRE activities. For both receptors, PTx was found to inhibit (albeit not eliminate) the constitutive activity. Thus, PTx resulted in a 47-55% reduction of the basal ECRF3 activity (Fig. 9A) and a 60 -65% reduction of the basal ORF74-HHV8 activity (Fig. 9B). With respect to the ligand-mediated activities, PTx (100 ng/ml) resulted in a 50% reduction in the efficacy of CXCL6/GCP-2induced ECRF3 activity, and the potency was concomitantly shifted 0.6-fold to the right. PTx did not reduce the efficacy of CXCL1/GRO␣ for ORF74-HHV8, but the potency was shifted 4-fold to the right. Thus, G i contributed to the SRE activities of both receptors; however, since PTx only reduced (but did not eliminate) the activities, we investigated the G 12/13 contribution by application of inhibitors of the G 12/13 -dependent signal mediators Rho and Rho kinase. Thus, we applied the C3 exoenzyme (C3 transferase) from C. botulinum, which irreversibly ADP-ribosylates and inactivates Rho (37,38), and the Rho kinase inhibitor Y-27632 (39) to the constitutive as well as ligand-mediated activities of ECRF3 and ORF74-HHV8. Both inhibitors resulted in a reduction (but not elimination) of the SRE activities of both receptors (Fig. 9, E and F). Thus, the C3 exoenzyme (10 g/ml) inhibited the constitutive activities by 34 and 46% for ECRF3 and ORF74-HHV8, respectively, whereas the ligand-regulated signaling (10 nM CXCL6/GCP-2 for ECRF3 and 10 nM CXCL1/GRO␣ for ORF74-HHV8) was inhibited by 65-67%. The SRE activities of the two receptors were similarly reduced by Y-27632 (10 M) since the constitutive activities were inhibited by 31-34%, whereas the ligand-regulated activities were inhibited by 34% for ECRF3 and by 71% for ORF74-HHV8. The two inhibitors did not influence basal cellular SRE activity (data not shown). In summary, the constitutive and ligand-regulated SRE activities of both receptors were dependent on G i as well as G 12/13 activation. DISCUSSION In this study, we have characterized the molecular pharmacological profile of ECRF3 in terms of ligand binding and signaling activities and compared it with that of ORF74-HHV8. Different ELR CXC chemokines bound to ECRF3 with high and similar affinities, yet the maximal binding capacities varied dramatically for the three radioligands 125 I-CXCL6/GCP-2, 125 I-CXCL1/GRO␣, and 125 I-CXCL8/IL-8. ECRF3 did not bind the non-ELR CXC chemokine 125 I-CXCL10/IP10, in contrast to ORF74 from HHV8. Signal transduction analysis revealed activation of a broad range of pathways for ECRF3, similar to ORF74-HHV8. However, in contrast to ORF74-HHV8, which promiscuously activated all pathways constitutively, the ECRF3 receptor activated G i and G 12/13 (measured via SRE activation) constitutively, whereas the other pathways (G q and the transcription factors NFAT, CREB, and NF-B) were activated solely in a ligand-regulated fashion.
Ligand Repertoire for ECRF3 Compared with Other ORF74 Receptors-In the initial study of ECRF3, CXCL1/GRO␣ was found to be the most potent agonist, followed by CXCL8/IL-8 and CXCL7/NAP-2 (9). In agreement with this, ORF74-HHV8 was later shown to recognize the GRO peptides (CXCL1-3) as agonists, with CXCL1/GRO␣ being the most potent agonist (6,8,13). Also ORF74 from MHV68 responds to GRO-related ELR motif-containing chemokines since the murine chemokines MIP-2 and KC (homologs of CXCL1/GRO␣) in addition to LIX (homolog of CXCL6/GCP-2 and CXCL5/ENA78) act as agonists for this receptor (10). Our analysis of ECRF3 confirms that CXCL1/GRO␣ indeed is a highly potent agonist; however, another ELR CXC chemokine, CXCL6/GCP-2 was the most potent and efficacious agonist for ECRF3, with a 13 times higher potency and a 2.2 times higher efficacy with respect to G q activation compared with CXCL1/GRO␣. The reason for the recognition of CXCL6/GCP-2 as the most potent agonist for ECRF3 and as an inverse agonist for ORF74-HHV8 (13) remains to be explained. The ELR CXC chemokines are chemoattractants of neutrophils and are angiogenic, whereas the non-ELR CXC chemokines attract lymphocytes and are in general angiostatic (40 -42). Two ELR CXC chemokines (CXCL8/IL-8 and CXCL6/GCP-2) bind to the structurally related CXCR1 and CXCR2 (43,44), whereas the rest, e.g. the GRO peptides (CXCL1-3), are selective for CXCR2 (45,46). A selective nonpeptide antagonist (SB 225002) has been developed for CXCR2 (47), and chemokines have been shown to bind differently to the two receptors (48,49) and to be expressed differently (50). Thus, CXCL1/GRO␣ and CXCL6/GCP-2 (and CXCL8/IL-8) differ with respect to receptor recognition and physiological roles, which could explain the choice of CXCL6/GCP-2 as a full agonist for ECRF3. The lack of agonistic properties of CXCL8/IL-8 for ECRF3 expressed in COS-7 or HEK293 cells is in contrast to the initial agonistic properties for ECRF3 expressed in Xenopus laevis oocytes (9). However, we observed a very low B max for CXCL8/IL-8 compared with the B max values for CXCL6/ GCP-2 and CXCL1/GRO␣, and it could be anticipated that the lack of agonistic properties in our system was due to this diminutive B max value for CXCL8/IL-8.
In this study, we observed no specific binding of the non-ELR CXC chemokine 125 I-CXCL10/IP10 to ECRF3, and binding analysis with 125 I-CXCL1/GRO␣ revealed a very low affinity for CXCL10/IP10. Thus, the binding and inverse agonistic properties of CXCL10/IP10 observed for ORF74-HHV8 (8,13,14,51) and the antagonistic properties of murine and human IP10 (CXCL10) for ORF74-MHV68 (10) are not shared by ECRF3. This lack of CXCL10/IP10 binding (and lack of modulation and down-regulation of ECRF3 by CXCL10/IP10) is, however, explainable in light of the cellular tropism for HVS since HVS has tropism for T-cells, in contrast to the B-cell tropism of HHV8 and MHV68 (8,51). CXCR3 is highly upregulated and expressed abundantly in activated T-cells, and the CXCR3 ligands (e.g. CXCL10/IP10) are present in high concentrations in T-cell-rich areas (1,52). Thus, an interaction of ECRF3 with CXCL10/IP10 and consequent inhibition would be potentially harmful for HVS.
vCCL2/vMIP-II encoded by HHV8 is an unnatural ligand for ECRF3; and during in vivo settings, this ligand-receptor pair will never meet due to the virus-host discrepancy. Nevertheless, vCCL2/vMIP-II serves important purposes as determined in studies of virus-encoded receptors due to the promiscuous binding to the majority of endogenous and viral CC, CXC, XC, and CX 3 C chemokine receptors, in most cases as an antagonist (8,14,21,33) and more seldom as an agonist (53). We employed vCCL2/vMIP-II as a prototype antagonist and observed inverse agonistic and antagonistic properties of this chemokine for ECRF3.
Constitutive and Non-constitutive Activities of ECRF3-Besides the broad-spectrum ligand binding and the exploitation of several signal transduction pathways, the majority of virusencoded receptors are unique due to the high degree of constitutive activity they usually display through several pathways (Table II) (6, 8, 26, 30, 54 -58), in contrast to the quiescent endogenous chemokine receptors. In general, high constitutive activity is not common for endogenous 7TM receptors (59,60); however, some receptors display high constitutive activity, e.g. the ghrelin and histamine receptors (61, 62), whereas the majority display low to undetectable levels of constitutive activity (60). In this study, we have characterized signaling pathways in close proximity to the receptor (G i and G q activation and G 12/13 indirectly through SRE) and signaling mediators farther downstream (transcription factors CREB, NFAT, NF-B, and SRE). In contrast to the rest of the constitutively active virusencoded receptors, we found, for ECRF3, selectivity in the constitutive activation of two of the three G-proteins investigated (G i and G 12/13 , but not G q ). The constitutive G i activity was measured directly by looking at the inhibition of cAMP production in transiently transfected HEK293 cells, in which vCCL2/vMIP-II functioned as an inverse agonist and CXCL6/ GCP-2 functioned as an agonist (Fig. 8A) and in which PTx reversed the constitutive adenylate cyclase inhibition by ECRF3 seen after forskolin treatment (Fig. 8B). We also measured the constitutive G i activity indirectly from the SRE activity in the presence of pathway-specific inhibitors since SRE has been described as being dependent on G i as well as on G 12/13 (32,35,36). The constitutive SRE activity was 4 -5-fold lower for ECRF3 compared with ORF74-HHV8 (tested in parallel) (Figs. 9B and 10 and Table III). We applied the G i inhibitor PTx and the Rho and Rho kinase inhibitors C. botulinum C3 exoenzyme and Y-27632, respectively (Rho and Rho kinase are downstream mediators of G 12/13 activation) (32,35,36), and observed partial inhibition of the constitutive as well as ligandregulated activities of all three inhibitors, indicating a contribution of G i as well as G 12/13 to the SRE activation of these two receptors. The G i contribution to the basal and ligand-regulated activities of ECRF3 was ϳ50% as judged from the influence of the inhibitors (47-55% inhibition by PTx concomitant with 34 -67% inhibition by Y-27632 and C3 exoenzyme of the constitutive and ligand-regulated activities of ECRF3) (Fig. 9, A, C, and E). For ORF74-HHV8, the G i pathway was most important for the basal SRE activity (60 -65% inhibition by PTx concomitant with 31-46% inhibition by Y-27632 and C3 exoenzyme, respectively) (Fig. 9, B and F), whereas G 12/13 was most important for the ligand-regulated signaling (small effect of PTx and 65-71% inhibition by C3 exoenzyme and Y-27632) (Fig. 9, D and F). Thus, using these inhibitors, it became clear that the G i activity (from SRE activation) was 6 -7-fold lower for ECRF3 than for ORF74-HHV8, whereas the G 12/13 activity was 3-4-fold lower for ECRF3 than for ORF74-HHV8 ( Fig. 10 and Table III). G 12/13 activation by ORF74-HHV8 is consistent with previously published data (31).
We observed no constitutive activity with respect to PI turnover (G q ), CREB, NFAT, and NF-B for ECRF3 as opposed to a high constitutive activity for ORF74-HHV8 through all these pathways (Figs. 4 and 7), again tested in the same cell type and at the same time as ECRF3. The lack of constitutive activation of CREB, NF-B, and NFAT (Fig. 7) was expected from the lack of constitutive PLC activation (Fig. 4) and the general G q dependence of these transcription factors (32). However, the pathway inhibitors revealed that only the CREB activation was solely G q -dependent, whereas the NF-B and NFAT activation was G i -and G q -dependent for both receptors (Fig. 7). Thus, despite the direct and indirect evidence of constitutive G i activation for ECRF3 (Figs. 8 and 9), there was no detectable constitutive activation of NF-B and NFAT. However, this could be due to the fact that the constitutive G i activity for ECRF3 was poor compared with that for ORF74-HHV8 (6 -7fold lower) ( Fig. 10 and Table III) and therefore may be undetectable in the NF-B and NFAT activation. The G i dependence of NF-B activation by ORF74-HHV8 has previously been shown in HeLa, COS-7, and lung endothelial cells (26,27,31). The lack of constitutive G q activation was observed in terms of both PI turnover (Fig. 3) and CREB activation (Fig. 6); and for both assays, we could not detect any increase in basal activities, not even low activation corresponding to one-seventh of the ORF74-HHV8 activation. From our studies of the ORF74 receptor encoded by MHV68, we have been able to detect even very low increases in activation, as low as one-seventh to onetenth of the corresponding activities of ORF74-HHV8. Despite this, we cannot rule out the possibility that ECRF3 possesses very low constitutive activity through G q , below detectable levels. However, compared with ORF74-HHV8, the basal G q activity of ECRF3 is negligible. We tested the activity of all four transcription factors in COS-7 and HEK293 cells (in parallel for the two receptors) and found no differences in the relationship of ligand-mediated and constitutive activities in the two cell lines.
Ligand-regulated Activities of ECRF3-Despite the lack of constitutive activity for G q , CREB, NF-B, and NFAT, we observed a brilliant ligand-regulated signaling through all pathways for ECRF3. Thus, CXCL6/GCP-2 induced a 5.4-fold increase above the basal non-constitutive G q activity of ECRF3, resulting in a maximal stimulation similar to the maximal CXCL1/GRO␣-mediated stimulation of ORF74-HHV8 (Fig. 4), tested in parallel for the two receptors. The PI turnover was not inhibited by the presence of PTx, indicating that the PLC activity is indeed a result of G q activity in COS-7 cells and not mediated by the ␤␥-subunits released from activated G i . In fact, an increase in efficacy for both receptors (as well as in the constitutive activity of ORF74-HHV8) occurred in the presence of PTx (Fig. 5B), indicating that inhibition of G i potentiates PLC activation. This phenomenon has been observed before for ORF74-HHV8 expressed in endothelial cells (27) and is explainable by the law of mass action since we found that ECRF3 couples to several G-proteins, (G i , G q , and G 12/13 ); the same has been shown for ORF74-HHV8 (6,27,29,31,63). A receptor in the active state couples only to one G-protein at a given time (64), and elimination of active G i by PTx could raise the probability for an interaction with another G-protein and consequently result in an increase in PI turnover, as we observed for both receptors.
Quantitative Comparison of ECRF3 and ORF74-HHV8 Signaling Properties-Not only did the two receptors express very similarly as determined from the B max values for their full agonists ( 125 I-CXCL6/GCP-2 binding to ECRF3 with a B max of 34 Ϯ 6.7 fmol/10 5 cells and 125 I-CXCL1/GRO␣ binding to ORF74-HHV8 with a B max of 33 Ϯ 5.9 fmol/10 5 cells), the total signaling capacities or efficiencies of ECRF3 and ORF74-HHV8 were also very similar. Thus, both in terms of G q (Fig. 4) and G i and G 12/13 coupling, based on SRE activation in the absence/ presence of inhibitors (Fig. 9), we observed a total stimulation in the same range for these two receptors (Fig. 10, full bars; and Table III). However, whereas ORF74-HHV8 displayed very high constitutive signaling (between 50 and 80% of the maximal signaling capacity) (Fig. 10, black bars; and Table III), in all signaling pathways that it had the ability to signal through, ECRF3 displayed clear constitutive activity (14 -20% of the maximal signaling capacity) only through the G i , G 12/13 , and SRE pathways (Fig. 10, black bars; and Table III). In contrast to ORF74-HHV8, no sign of constitutive signaling was observed for ECRF3 through G q or through the other tested transcription factors, despite the fact that the full agonist for ECRF3 stimulated this receptor to the same maximal signaling capacity as that observed for ORF74-HHV8 ( Fig. 10 and Table III).
Since the maximal signaling capacities in the various pathways were rather similar for the two receptors, this means that the ligand-mediated signaling was responsible for the major part of the signaling capacity for ECRF3 (ligand-regulated/constitutive signaling ratio (L/C) Ͼ 1) and only a minor part for ORF74-HHV8 (L/C Ͻ 1), as most clearly seen in Fig. 10 (gray versus black bars) and in Table III. The lowest ligand contribution compared with constitutive activity was found for the G i coupling of ORF74-HHV8 (L/C ϭ 0.2). The G i activity of ECRF3 (assessed by the cAMP inhibition experiments in HEK293 cells) (Fig. 8) resulted in a surprisingly low ligand regulation compared with the constitutive activity (L/C ϭ 0.8). Thus, for this measurement (but not for the G i component of SRE), the constitutive signaling slightly exceeded the ligand-regulated signaling (assuming that CXCL6/GCP-2 and vCCL2/vMIP-II are full agonist and full inverse agonist, respectively). This discrepancy in the ligand-regulated versus constitutive component of G i signaling could be a reflection of the very different steps at which we measured in the G i signaling cascade. Adenylate cyclase inhibition is close to the G-protein, whereas SRE is far away and influenced by several regulators from the G-protein to the nucleus. In addition, adenylate cyclase inhibition is dependent on pre-activation with forskolin. Since forskolin activates adenylate cyclase in all cells, the actual window of G i activity is therefore dependent on transfection efficiency. In summary, the ligand-regulated signaling compared with the constitutive signaling was greater for ECRF3 than for ORF74-HHV8 in all cases. The ligand-regulated activities were also greater for ECRF3, whereas the constitutive activities were greater for ORF74-HHV8 ( Fig. 10 and Table III).
What Are the Roles of Virus-encoded 7TM Chemokine Receptors?-Experimental evidence from receptor knockout studies indicates a certain importance of the chemokine receptors in the virus-host interplay. For instance, it has been shown that knockout of the murine cytomegalovirus-encoded receptor M33 results in a virus strain with decreased virulence and with less severe host infection (65). Similarly, deletion mutants of ORF74 from MHV68 point to a receptor function in the early reprogramming of the virus-infected cells and upon virus reactivation from latency (66,67). Transgenic expression of ORF74-HHV8 in mice has shown that this receptor is important for Kaposi's sarcoma development (19,20) as a side effect of high constitutive activity since abrogation of the constitutive activity decreases the appearance of Kaposi's sarcoma-like lesions (68).
In conclusion, virus-encoded chemokine receptors serve important functions, although they are not mandatory for virus survival. From a molecular pharmacology point of view, these receptors constitute a unique opportunity to study basic principles of ligand recognition, the activation mechanism of 7TM receptors, internalization, and recycling pathways as examples of targeted evolution, i.e. the receptors obtained from the host and through "combinatorial chemistry" structurally and functionally varied by random mutagenesis. The receptor phenotypes are highly interesting, with most of them being constitutively active. This study of ECRF3 addresses differences among virus-encoded chemokine receptors characterized so far with respect to ligand repertoire and signaling properties. The signaling pathways were determined in transiently transfected COS-7 cells (in parallel for the two receptors). The ratios were calculated from the data in Fig. 4 (G q ) and in Fig. 9 (SRE). The ratios for G i and G 12/13 were calculated from the SRE experiments (SRE included in last column) in the presence or absence of pathway-specific inhibitors (Fig. 9). NDCA, no detectable constitutive activity.
FIG. 10. Quantitative comparison of ECRF3 and ORF74-HHV8 signaling properties. A, the ligand-regulated (gray bars) and constitutive (black bars) activities of ECRF3 and ORF74-HHV8 (expressed in parallel in transiently transfected COS-7 cells) are shown here for the three different G-proteins (G q , G i , and G 12/13 ) activated by both receptors. 100% equals the total activity (ligand-regulated ϩ constitutive) of ORF74-HHV8 for each pathway. B, the signaling pattern with respect to SRE activation is also provided since the analysis of this transcription factor provided the data for the G i and G 12/13 activities (by addition of pathway-specific inhibitors to SRE activity). The constitutive ORF74/ ECRF3 activity ratio for each pathway is given by the height of the black bars; the ligand-regulated ORF74-HHV8/ECRF3 activity ratio is given by the height of the gray bars; and the ligand-regulated/constitutive activity ratio for each receptor in each pathway is given by the height of the gray versus black bars. (Table III summarizes these  ratios.)