Molecular Pharmacological Phenotyping of EBI2

Epstein-Barr virus (EBV)-induced receptor 2 (EBI2) is an orphan seven-transmembrane (7TM) receptor originally identified as the most up-regulated gene (>200-fold) in EBV-infected cells. Here we show that EBI2 signals with constitutive activity through Gαi as determined by a receptor-mediated inhibition of forskolin-induced cAMP production and an induction of the serum response element-driven transcriptional activity in a pertussis toxin-sensitive manner. Gαs and Gαq were not activated constitutively as determined by the lack of cAMP production, the lack of inositol phosphate turnover, and the lack of activities of the transcription factors: cAMP response element-binding protein and nuclear factor-κB. Immunohistochemistry and confocal microscopy of FLAG- and green fluorescent protein-tagged EBI2 revealed cell-surface expression. A putative N-terminal truncated version of EBI2, Δ4-EBI2, showed similar expression and signaling through Gαi as full-length EBI2. By using a 32P-labeled EBI2 probe we found a very high expression in lymphoid tissue (spleen and lymph node) and peripheral blood mononuclear cells and a high expression in lung tissue. Real-time PCR of EBV-infected cells showed high expression of EBI2 during latent and lytic infection, in contrast to the EBV-encoded 7TM receptor BILF1, which was induced during lytic infection. EBI2 clustered with the orphan GPR18 by alignment analysis as well as by close proximity in the chromosomal region 13q32.3. Based on the constitutive signaling and cellular expression pattern of EBI2, it is suggested that it may function in conjunction with BILF1 in the reprogramming of the cell during EBV infection.

The orphan EBI2 2 receptor belongs to the superfamily of rhodopsinlike 7TM receptors (seven-transmembrane segment receptors), also known as G-protein-coupled receptors. The sequencing of the human genome has identified around 700 7TM receptors, and approximately half of these are believed to encode sensory receptors. The remaining receptors are divided into different classes of which class A (rhodopsin-like) includes the vast majority of the receptors. The cognate ligands have been identified for ϳ200 non-sensory 7TM receptors, whereas the rest are still orphan receptors (1).
EBI2 was cloned in 1993 as one out of nine up-regulated genes in Epstein-Barr virus (EBV)-infected Burkitt lymphoma cells (2). These nine genes were up-regulated from 4-to Ͼ100-fold upon EBV infection, and two 7TM receptors were identified among the up-regulated genes (Epstein-Barr-induced receptors 1 and 2, EBI1 and -2). EBI1 was later deorphanized as the receptor for the chemokines CCL19/ELC and CCL21/SLC and was consequently renamed CCR7 (3), whereas the ligand(s) for EBI2 still remains to be identified. EBI2 displayed the highest up-regulation (Ͼ200-fold) among the nine EBV-induced genes compared with for example 21-fold for CCR7 (2). Initial expression analyses of the nine genes uncovered an expression of EBI2 in peripheral blood mononuclear cells (PBMCs), tonsils, spleen, and lung tissue (2).
CCR7 and EBI2 are not the only 7TM receptors being regulated by EBV. In fact, up-and down-regulation of endogenous 7TM receptors and ligands are important events following a virus infection, and interestingly, many of these genes belong to the chemokine system (4 -7). Molecular mimicry is another approach used by many herpes-and poxviruses to manipulate the host immune system, and again the chemokine system is "over-represented" compared with other 7TM receptor systems (7,8). Thus, a growing number of herpes-and poxvirus-encoded receptors, and ligands have been identified within the last decade, and most of these receptors bind a broad spectrum of chemokines and signal in a constitutive manner. Of these, the ␥2-herpesvirus-encoded ORF74 receptors are the most abundantly characterized receptors (9 -11). The first ␥1-herpesvirus-encoded 7TM receptor (vGPCR) was identified in 2002 in region A5/BILF1 of the porcine lymphotropic herpesvirus-1 by Goltz et al. (12). The same paper described the presence of BILF1 homologs in other ␥1-herpesviruses, for instance EBV. The EBVencoded BILF1 receptor was recently cloned and characterized as a highly constitutively active receptor (13,14).
The knowledge about the orphan EBV-induced EBI2 receptor is limited to the original cloning paper (2) and to a study showing a lack of simian immunodeficiency virus cell entry cofactor function for EBI2 (15). In the present study we compare the primary structure of EBI2 with other herpesvirus-related 7TM receptors (i.e. the herpesvirusencoded receptors) and a selection of endogenous 7TM receptors. We find that EBI2 does not resemble any of the herpesvirus-encoded 7TM receptors but is closest related to the orphan 7TM GPR18 and the two lipid receptors cysteinyl leukotriene receptor 1 and 2 (CysL1 and -2). In addition, we find that EBI2 clusters with GPR18 in region q32.3 at chromosome 13. We characterize EBI2 in terms of signaling activities at the level of G-proteins as well as at the level of the transcriptional activity.
The cellular as well as tissue expression of EBI2 is characterized and the expression of EBI2 during the different replication states of EBV is quantified by real-time PCR and compared with the BILF1 receptor expression. We find that the constitutive signaling through G␣ i and the cellsurface localization of EBI2 is similar to BILF1 (13), whereas the expression pattern during virus infection varies, because EBI2 is expressed during latent as well as lytic replication, whereas BILF1 is induced during lytic infection.
Site-directed Mutagenesis-The human EBI2 was inserted into a set of different eukaryotic expression vectors: the pTEJ8 vector (16), the pcDNA3 vector, the pEGFP-N1 vector, and a modified pcDNA3 vector with a FLAG tag inserted upstream of the polylinker region kindly provided by Kate Hansen (7TM-Pharma, Denmark). N-terminal truncation of EBI2 (⌬4) was made by PCR using the wild-type/ full-length EBI2 as template. All reactions were carried out using the Pfu polymerase (Stratagene). The wild-type and truncated EBI2 DNA were verified by DNA sequencing on an ABI 310 sequencer from PerkinElmer Life Sciences.
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 (Invitrogen, cat. no 21885-025) adjusted with 10% fetal bovine serum (FBS), 180 units/ml penicillin, and 45 g/ml streptomycin (PenStrep). HEK293 cells were grown in Dulbecco's modified Eagle's medium adjusted to contain 4500 mg/liter glucose (cat. no. 31966-021) with the same amount FBS and PenStrep as the COS-7 cells at 10% CO 2 and 37°C. The HEK293 cell media was modified to hold heat-inactivated FBS and no PenStrep during luciferase based assays. Stable transfected and naïve L12 cells were grown in RPMI supplemented with 10% FBS, PenStrep, and the selection marker G418 for the EBI2-expressing cell lines. Transfection of the COS-7 cells was performed by using the calcium phosphate precipitation method (11). HEK293 cells were transfected by the LipofectAMINE TM 2000 Reagent in the serum-free media Opti-MEM 1 according to the manufacturer's description for all experiments. L12 cells were stably transfected by electroporation. Briefly, EBI2-GFP DNA was linearized and transferred into a 0.4-cm cuvette, and a single electroporation pulse at 250 V and 960 microfarads was applied. The electroporated cells were incubated for 10 min at room temperature and transferred to culture at 37°C in RPMI supplemented with 10% fetal calf serum. Geneticin (G418) was added to a final concentration of 800 mg/ml 48 h after transfection, and the cells were plated in 96-well plates (25,000 cells/well) (17). After 3 weeks, five G418-resistant clones were selected for further FACS sorting according to EBI2-GFP expression (see below).

FACS Sorting and Fluorescence Analysis of Stably Transfected L12
Cells-Five different EBI2-GFP-expressing L12 cell clones were each FACS-sorted by a BD Biosciences FACS Vantage SE, Diva into a fraction with no EBI2 expression, a fraction with low EBI2 expression, and a fraction with high EBI2 expression. Among the EBI2-expressing fractions, four fractions were chosen representing the spectrum from low to high EBI2 expression. These four fractions were analyzed in respect of adenylate cyclase inhibition (cAMP-assay, see below), and the signaling activities were compared with the EBI2 expression level.
Purification and Fractionation of PBMCs in B-lymphocytes, T-lymphocytes, Monocytes, and NK Cells-Buffy-coats from four healthy blood donors were obtained 16 h after blood collection. The reminiscent erythrocytes were lysed by a 1:1 mixture of the buffy-coats with Hoffman's reagent (0.16 NH 4 Cl, 0.10 mM Na 2 EDTA, and 10 mM NaHCO 3 ), and the granulocytes were separated from the PBMCs by density centrifugation using the reagent Lymphoprep (AXIS-SHIELD, Oslo, Norway) according to the manufacturer's recommendations. The PBMCs were further fractionated into T-lymphocytes, B-lymphocytes, monocytes, and NK cells by CD3-, CD19-, CD14-, and CD56-covered magnetic beads, respectively (Micro-beads from Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). Briefly, the PBMCs were mixed with the antibody-covered Micro-beads (2 l of Micro-beads per 10 7 cells), and the mixtures were subsequently applied to MS-columns (Miltenyi Biotec GmbH) that were attached to a MiniMac separator (Miltenyi Biotec GmbH). The purification efficiencies of T-lymphocytes, B-lymphocytes, monocytes, and NK cells were determined by FACS analysis using perCpCy5 (Peridinin chlorophyll protein cyanin), phycoerythrin, or fluorescein isothiocyanate-labeled CD3, CD19, CD14, and CD56 antibodies (DAKO, Denmark) and found to be 85-95%.
Real-time PCR Analysis of EBI2 Expression in T-lymphocytes, B-lymphocytes, Monocytes, and NK Cells-Total RNA was extracted from isolated T-lymphocytes, B-lymphocytes, monocytes, and NK cells using a QIAamp RNA Blood Mini kit (Qiagen, Valencia, CA). Subsequently, cDNA was synthesized from 1 g of total RNA using the ImProm II Reverse Transcriptase kit (Promega, Madison, WI). Quantification of EBI2 transcript levels in the PBMC subsets was performed using the SYBR Premix Ex TaqRT-PCR kit (Takara Bio Inc., Shiga, Japan) and the Mx3000P system (Stratagene). Expression of ␤-actin was used for copy number normalization. A calibrator consisting of a mixture of cDNA from all samples was included in each run to facilitate comparative analysis. Each sample was assessed three times in triplicates. The primer pairs used were as follows: EBI2 forward: 5Ј-GAATCGGAGATGCCT-TGTGT-3Ј; EBI2 reverse: 5Ј-GCCTCCTGCTTTGACATAGG-3Ј; ␤actin forward: 5Ј-CGTCTTCCCCTCCATCGT-3Ј; ␤-actin reverse: 5Ј-CGCCCACATAGGAATCCTTC-3Ј. Amplification efficiency of the primer pair was validated using serial dilutions of the calibrator sample being satisfactory in both cases.
Phosphatidylinositol Assay (PI Turnover)-COS-7 cells were transfected according to the procedure mentioned above. Briefly, 2 ϫ 10 6 cells were transfected with increasing amounts of receptor or vector control cDNA (from 0 to 10 g/flask) with or without 30 g of the promiscuous chimeric G-protein, Gqi4myr, which turns the G␣ i -coupled signal into the G␣ q pathway (phospholipase C activation measured as PI turnover), or the Gq s5 , which turns the G␣ s -coupled signal into the G␣ q pathway (18). One day after transfection, COS-7 cells (2.5 ϫ 10 4 cells/well) were incubated for 24 h with 2 Ci of myo-[ 3 H]inositol in 0.4 ml of growth medium per well. Cells were washed twice in 20 mM Hepes, pH 7.4, supplemented with 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 the cells were incubated in 0.4 ml of buffer supplemented with 10 mM LiCl at 37°C for 90 min. Cells were extracted by addition of 1 ml of 10 mM formic acid to each well followed by incubation on ice for 30 -60 min. The generated [ 3 H]inositol phosphates were purified on AG 1-X8 anion-exchange resin (19). Determinations were made in duplicates.
Adenylate Cyclase Inhibition Assay (cAMP-assay)-HEK293 cells were seeded at 35,000 cells/well in culture plates 1 day prior to transfection and were transfected with 50 ng/well EBI2 or the pcDNA3 vector as control. 24 h after the transfection, the cells were washed twice in HBS buffer (25 mM Hepes, pH 7.2, supplemented with 0.75 mM NaH 2 PO 4 , 140 mM NaCl, and 0.05% (w/v) bovine serum albumin and incubated for 30 min at 37°C in HBS buffer supplemented with 1 mM of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine and various concentrations of the adenylate cyclase activator forskolin. After incubation, the cells were placed on ice, the medium was removed, and the cAMP concentration was measured using the cAMP-EIA kit (Amersham Biosciences). The effect of 100 and 500 ng/ml Ptx was tested by adding it to the cells immediately following transfection. Naïve and stably transfected L12 cell lines were seeded at 50,000 cells/well 24 h before cAMP determination performed as described for the HEK293 cells above. Determinations were made in duplicates.
Constitutive CREB, NF-B, and SRE cis-Reporting Luciferase Assay-Cells were seeded at 35,000 cells/well in culture plates 24 h prior to transfection and were transfected with 50 ng/well of the (cis)-reporter plasmid (pCREB-Luc, pSRE-LUC, and pNF-B-Luc) and concentrations from 0 to 50 ng/well of receptor plasmid. Immediately following transfection, the cells were given new media with variations in the concentrations of FBS (10% for the NF-B and 0.1% for the CREB and the SRE reporter system) and incubated for 18 h in the presence or absence of Ptx. One day after transfection, the cells containing the CREB-luciferase receptor (and EBI2 or pcDNA3 vector) were given new media with or without forskolin and incubated for 5 h. All cells were washed twice in Dulbecco's phosphate-buffered saline (0.9 mM CaCl 2 , 2.7 mM KCl, 1.5 mM KH 2 PO 4 , 0.5 mM MgCl 2 , 137 mM NaCl, and 8.1 mM Na 2 HPO 4 ), and the luminescence was measured in a micro plate scintillation and luminescence counter (Top-counter, Packard) after 10 min incubation in 100 l of Dulbecco's phosphate-buffered saline together with 100 l of LucLite substrate. Determinations were made in quadruples.
Surface Enzyme-Linked Immunosorbent Assay-HEK293 cells were transfected with the N-terminal FLAG-tagged variants of EBI2 wt and ⌬4-EBI2 for the enzyme-linked immunosorbent assay in parallel with cells used for the CREB assay. The cells were washed once in TBS (35 mM Tris-HCl, 140 mM NaCl, pH 7.4), fixed in 4% glutaraldehyde for 10 min, and incubated in blocking solution (2% bovine serum albumin in TBS) for 30 min at room temperature. The cells were subsequently kept at room temperature and incubated 2 h with anti-FLAG (M1) antibody (2 g/ml) in TBS supplemented with 1% bovine serum albumin and 1 mM CaCl 2 . After three washes in TBS with 1 mM CaCl 2 the cells were incubated with goat anti-mouse horseradish peroxidase-conjugated antibody in the same buffer as the anti-FLAG antibody for 1 h. After three washes in TBS with 1 mM CaCl 2 the immune reactivity was revealed by the addition of horseradish peroxidase substrate according to manufacture's instruction.
Tissue Expression Analysis-This was performed using Clontech's Multiple Tissue Expression Array according to the manufacturer's recommendation. Briefly, a PCR fragment corresponding to the EBI2 sequence from bp 160 -403 was obtained using the primers GTCGT-CATTGTTCAAAACAGG (forward) and GCACCACAGCAATG-AAGC (reverse). The fragment was 32 P-labeled and used as probe for hybridization to the Multiple Tissue Expression Array. The data were visualized by a Molecular Imager FX (Bio-Rad) and analyzed using the Quantity One Program (Bio-Rad).
Stimulation of EBV-infected Cells-Akataϩ and B95.8 were grown in RPMI 1640 medium (Invitrogen, 21875-034) supplemented with 10% FBS and 100 units penicillin and streptavidin per milliliter. The cells were incubated at 37°C, and 5% CO 2 . Approximately 10 7 cells were used for each assay. Prior to stimulation cells were pelleted by centrifugation and resuspended in media (10 6 cells/ml). The Akataϩ cells were stimulated with 0.5% v/v IgG antibody (DAKO, A042410), or IgG antibody plus 300 g/ml PAA (phosphonoacetic acid) to inhibit viral DNA replication for 3 h, whereas the B95.8 cells were stimulated with 6 mM butyrate and 10 ng/ml phorbol 12-tetradecanoate 13-acetate. Cells were resuspended in media with or without PAA and grown for 48 h.
EBI2 Is Expressed at the Cell Surface and Signals Constitutively through G␣ i in a Ptx-sensitive Manner-The cellular expression was analyzed in two different cell lines, the laboratory fibroblast HEK293 cell line used for binding and signaling characterization of 7TM receptors in general, and the pre-B-cell lymphocyte cell line L12, chosen due to the original cloning of EBI2 from B-lymphocytes infected with EBV. As shown in Fig. 2, EBI2 fused to green-fluorescent protein (EBI2-GFP) localized predominantly to the cell-membrane in both cell lines, as shown for the majority of endogenous 7TM receptors (1).
To study whether EBI2 is a functional receptor, we tested for putative signaling through a variety of G-proteins. The G␣ i coupling of EBI2 was initially analyzed by means of cAMP formation in transiently transfected HEK293 in the absence and presence of forskolin (a direct activator of adenylate cyclase). As shown in Fig. 3, EBI2 expression resulted in a lower cAMP-formation for two different forskolin doses (15 and 50 M) compared with control cells. Even in the absence of forskolin, EBI2 expression resulted in a lower cAMP-formation compared with control cells, indicating that also the basal G␣ i activity was activated by EBI2. The effect of EBI2 was abolished by Ptx indicating a constitutive activity through G␣ i/o .
EBI2 Is Not Constitutively Active through G␣ s or G␣ q -Gene-dosage experiments in transiently transfected COS-7 and HEK293 cells showed, that EBI2 had no constitutive activity via coupling to G␣ q since no increase in phosphatidylinositol turnover (PI turnover as a result of phospholipase C activation) was observed upon increasing concentrations of EBI2 (Fig. 4A). However, in support of the constitutive G␣ i activity a gene-dose-dependent increase in PI turnover was observed  upon co-transfection with the chimeric G-protein Gqi4myr (Fig. 4A). This increase was not observed upon co-transfection with wt G␣ q (Fig.  4A). In contrast, co-transfection of CCR7 with Gqi4myr did not result in any constitutive activity, yet in the presence of CCL19 or -21, the cells co-transfected with Gqi4myr elicited an increase in PI turnover (Fig.  4B), as expected from the G␣ i coupling for endogenous chemokine receptors (24). To test for a putative G␣ s coupling of EBI2, the receptor was co-transfected with the chimeric Gqs5, a G-protein ␣-subunit that is recognized as G␣ s by the receptors, but transduces a G␣ q signal (25). However, no constitutive PI turnover was observed under these settings, indicating that EBI2 is not constitutively active through G␣ s (Fig.  4C). The gastric inhibitory polypeptide receptor was used as a positive control of Gqs5 and showed a ligand-dependent, but not ligandindependent increase in PI turnover in the presence of Gqs5 (Fig. 4C).
Transcription Factor Analyzes Confirm the Ptx-sensitive G␣ i Coupling of EBI2-A selection of transcription factors were used to further analyze the signaling pattern of EBI2 in transiently transfected HEK293 cells. The CREB (cAMP response element binding protein) transcription factor is regulated by the level of cAMP (through cAMP-regulated kinases) and therefore maps the G␣ s respective G␣ i coupling of a certain receptor; however, also kinases downstream of G␣ q (such as Ca 2ϩ /calmodulin kinase IV or protein kinase C) have been reported to mediate CREB activation (26,27). We found, that EBI2 in a gene-dose-dependent manner inhibited the forskolin-induced CREB activity and that the presence of Ptx abolished this effect (Fig. 5A). In contrast, the basal CREB activity was unaffected by increasing concentrations of EBI2 (Fig.  5B). In the presence of Gqi4myr, EBI2 increased the CREB activity in a gene-dose-dependent manner, whereas this was not observed in the FIGURE 4. EBI2 is constitutively active through G␣ i but not G␣ q or G␣ s . A, gene-dose experiments with measurement of PI turnover for increasing concentrations of EBI2 in the presence of Gqi4myr (F), the chimeric G␣-protein that turns the G␣ i -coupled signal into the G␣ q pathway, or G␣ q wt (f). B, gene-dose experiments with measurement of PI turnover for increasing concentrations of CCR7 in the presence of Gqi4myr (E) or wt G␣ q (Ⅺ). CCL19 (10 nM) was added to CCR7 (6 g/flask) in the presence (OE) or absence (ࡗ) of Gqi4myr. C, Gqs5, the chimeric G␣-protein that turns the G␣ s -coupled signal into the G␣ q pathway, was co-transfected with EBI2 (gray columns), the gastric inhibitory polypeptide-receptor (black columns), or the pcNDA3 vector (negative control, white columns), and the PI turnover was measured in the presence and absence of the gastric inhibitory polypeptide ligand (n ϭ 3-6). presence of G␣ q wt or for the pcDNA3 vector in the presence of Gqi4myr (Fig. 5C). Thus, the CREB activation pattern clearly supports the constitutive activity through G␣ i (Figs. 3 and 4). For further analysis, we included the NF-B, which has been reported to be regulated by several G-proteins (however, mainly G␣ q ) depending upon cell type (26, 28 -32). We did not observe any increase in NF-B activity upon increasing concentrations of EBI2, whereas the ORF74, as a positive control, resulted in a gene-dose-dependent increase in activity (Fig. 5D). ORF74-HHV8 activates a broad spectrum of signaling pathways, and the NF-B activity induced by this receptor in HEK293 cells has been shown to be mediated by G␣ i , G␣ q , and G␣ 13 (32,33). We therefore included extra positive controls for the G␣ q contribution of the NF-B activation in HEK293 cells using the constitutively active ghrelin receptor and the constitutively active neurotensin receptor 2 for which no direct evidence of constitutive activity through G␣ i or G␣ 13 have been shown in HEK293 (or other) cells (34). Both of these receptors activated NF-B in a constitutively manner, and the phospholipase C inhibitor U73122 inhibited these activities to 9 Ϯ 12% for the ghrelin receptor and 12 Ϯ 11% for neurotensin receptor 2 (Fig. 5E). Thus, the lack of EBI2 activity through NF-B clearly supports the lack of G␣ q coupling observed in Fig. 4. We also included the SRE (serum response element)transcription factor that has been reported to be regulated by G␣ i , ␤␥ as well as G␣ 13 (35)(36)(37). Consistent with the signaling pattern of EBI2 through G␣ i , we observed a Ptx-sensitive increase in the SRE activity for increasing concentrations of EBI2 (Fig. 5F).
Characterization of Two Putative Isoforms of EBI2-The N terminus of EBI2 contains two methionines, Met 1 and Met 5 (Fig. 6A). The mouse and rat homologs of EBI2 lack the first Met and, therefore, consist of N-terminal truncated homologs (⌬4) compared with human EBI2. For the human EBI2 gene it is not known whether it is position 1 (EBI2) or position 5 (⌬4-EBI2) that initiates translation. Therefore, we compared the signaling and cellular expression pattern of these two isoforms. As shown in Fig. 6, the EBI2 and ⌬4-EBI2 displayed similar signaling and expression levels in transiently transfected HEK293 cells. We measured the signaling activities for FLAG-tagged (Fig. 6C) as well as non-tagged receptors (Fig. 6B) and used the CREB activation in the presence of Gqi4myr as signaling read-out in both cases (Fig. 6, B and C). The surface expression was quantified by immunohistochemistry of the two N-terminal FLAG-tagged versions of EBI2 and ⌬4-EBI2 (Fig. 6D) and for these receptor constructs, the expression analyses were performed in parallel with the signaling analyses (Fig. 6, C and D).
Detailed Expression Profile of EBI2-The original paper of EBI2 included Northern blot analysis of different tissues (heart, brain, liver, skeletal muscle, kidney, and pancreas), and among these, only lung tissue displayed abundant EBI2 expression (2). We decided to further characterize the tissue expression pattern of EBI2 by using a Multiple Tissue Expression Array with 32 P-labeled EBI2 cDNA as hybridization probe. In this array poly(A)ϩ RNAs from different tissues were normalized to the mRNA expression levels of eight different housekeeping genes (among these ␤-actin, glyceraldehyde-3-phosphate dehydrogenase, ubiquitin, ␣-tubulin, and ribosomal protein S9) and immobilized in separate dots. As shown in Fig. 7, EBI2 was abundantly expressed primarily in lymphoid tissues. Thus, very high expression (Ͼ20-fold higher compared with background expression in negative controls) was found in PBMCs, lymph node, and spleen. Also lung tissue showed a high expression of EBI2 (Ͼ15-fold higher expression compared with background) and was followed by a medium-to-high expression in aorta and right and left atrium (from 6.4-to 11-fold higher expression compared with background (Fig. 7B)). The expression of EBI2 in the different parts of the gastrointestinal tract was medium-to-high (from 4.2-to 11.5-fold higher expression compared with background) with the highest expres-sion in appendix (Fig. 7D). The urogenital system displayed low-tomedium expression (from 1.5-to 7.5-fold higher expression compared with background (Fig. 7C)). The lowest expression was found in the brain (from 1.2 to 2.7 higher expression compared with background, with the exception of a 4.7-fold induction in corpus callosum and medulla oblongata (Fig. 7A)). Thus, this detailed analysis of EBI2 expression corresponded very well to the original expression profile (2), and the extended analysis of the different tissues clearly show that EBI2 is expressed in high levels in lymphoid organs and tissues with a high concentration of leukocytes (e.g. appendix). In contrast, the high expression was not found in the hematopoietic organs (bone marrow and fetal liver). The expression of EBI2 was further analyzed in purified subfractions of PBMCs. Thus, by magnetic beads covered by antibodies to CD3, CD19, CD14, and CD56 (Miltenyi Micro-beads), we purified T-lymphocytes, B-lymphocytes, monocytes, and NK cells, respectively, with a purity of 85-95% (verified by FACS analysis with the PerCpCy5-, phycoerythrin-, and fluorescein isothiocyanate-labeled antibodies). The relative EBI2 expression in these four cellular subsets was analyzed by Real-time PCR and normalized to the housekeeping gene ␤-actin. By this technique it was found that the expression of EBI2 is highest in B-lymphocytes, followed by T-lymphocytes, NK cells, and lowest in monocytes (Fig. 7B, inset). Thus, also this detailed analysis of EBI2 FIGURE 7. Tissue expression pattern of EBI2. The tissue expression pattern was determined by a Multiple Tissue Expression Array from Clontech with 32 P-labeled EBI2 cDNA as hybridization probe. A, different brain areas together with whole brain. B, heart, skeletal muscle, and lung tissue. C, urogenital system and lymphoid organs (inset in C: relative EBI2 expression in B-lymphocytes, T-lymphocytes, NK cells, and monocytes as determined by real-time PCR. D, gastrointestinal (GI) tract and different glands. The data were visualized by a Molecular Imager FX (Bio-Rad) and analyzed using the Quantity One Program (Bio-Rad). The ordinate indicates the expression in arbitrary units, where the average expression (with the background subtracted) in the different brain areas was set to one. MAY 12, 2006 • VOLUME 281 • NUMBER 19 expression in PBMC subsets is consistent with previously published data of high expression of EBI2 in B-and T-lymphocytes (4).

EBI2: Undressing an Orphan 7TM Receptor
EBI2 Is Expressed during Latent and Lytic Infection-EBI2 was originally identified as being highly up-regulated in latently infected Burkitt lymphoma cells compared with non-infected cells (38). This up-regulation has also been shown in other B-lymphocytes (4), and we therefore decided to focus on the expression pattern of EBI2 during latent and lytic infection (Fig. 8A). We used two different EBV-infected cell lines (Akataϩ and B95.8) and induced lytic infection by addition of 0.5% v/v IgG antibody to the Akataϩ cells and 6 mM butyrate and 10 ng/ml phorbol 12-tetradecanoate 13-acetate to the B95.8 cells. The expression pattern of EBI2 was compared with the expression pattern of the recently characterized EBV-encoded 7TM receptor BILF1 (Fig. 8B). The early lytic gene BZLF1 and late lytic gene gp350 were used as controls for the early and late lytic gene expression, and as shown in Fig. 8 (C and D, respectively), these two lytic genes were induced equally well in both cell lines. As expected, the BZLF1 mRNA expression was independent of viral DNA as evident from the insensitivity to inhibition by PAA in both cell lines (Fig. 8C). In contrast, the expression of gp350, as expected, was dependent on viral DNA replication as evident from the sensitivity to inhibition by PAA (Fig. 8D). As previously published, BILF1 was induced during lytic infection (13,14). Interestingly however, the kinetics of BILF1 mRNA expression differed between Akataϩ and B95.8 cells, because BILF1 behaved as a late lytic gene (it was inhibited by PAA) in the Akataϩ cell line, whereas it behaved as an early lytic gene (it was not inhibited by PAA) in the B95.8 cell line (Fig. 8B). This difference could perhaps reflect a constitutive low lytic replication in the B95.8 cell line. The expression of EBI2 was not altered by induction of lytic infection in these two cell lines; however, the expression level was ϳ10-fold higher in the B95.8 than in the Akataϩ cells (Fig. 8A). The experiments were also performed in the absence of reverse transcriptase (minus controls) to test for any false positives based in reminiscent DNA. We found an almost absent DNA contamination, because the average copies were 8,000-fold lower in the absence of reverse transcriptase compared with the presence of reverse transcriptase in Akataϩ cells, and 80,000-fold lower in B95.8 cells (data not shown).
EBI2 Expression in B-lymphocytes Correlates to Basal Signaling Levels-Due to the high expression of EBI2 in B-lymphocytes, we decided to characterize the EBI2 signaling in a more natural cell line (the pre B-lymphocyte cell line L12) as a supplement to the characterized signaling in HEK293 and COS-7 cells (Figs. [3][4][5]. Thus, we created EBI2-GFP stably transfected L12 cell lines and FACS-sorted these according to level of fluorescence as an indication of EBI2-GFP expression. Four of these fractions were selected representing the spectrum from low to high EBI2-GFP expression. The basal (Fig. 9A) and forskolin-induced (Fig. 9B) signaling of these four fractions was clearly correlated to the EBI2-GFP expression measured as average fluorescence intensity in each cell fraction (mean fluorescence intensity multiplied with the percentage of cells with EBI2-GFP expression at the day of signaling determination, Fig. 9C). Thus, as shown in Fig. 9, we found the lowest level of basal and forskolin-induced cAMP in the cell-fraction with highest EBI2 expression (fraction A), whereas the cell fraction with medium expression displayed a higher cAMP level (fraction B), and the two fractions with the lowest EBI2 expression had even higher cAMP levels, although still lower than the cAMP levels in naïve L12 cells.

DISCUSSION
In the present study, we describe constitutive activity through G␣ i for the EBI2 receptor. This activity may have been overlooked previously due to the fact that many 7TM receptors are screened, in the search for ligands, by calcium mobilization assays. We furthermore describe the cellular and tissue expression of EBI2, together with the receptor expression during latent

EBI2: Undressing an Orphan 7TM Receptor
and lytic viral infection. Thus, although the endogenous ligand(s) remains unknown, our studies "undress" EBI2 by providing a functional characterization as well as cellular and tissue expression pattern for this 7TM receptor.
Signaling Pattern of EBI2-The signaling of EBI2 was investigated at the level of G-protein activation and at the level of transcriptional activity. We found that EBI2 signaled constitutively through G␣ i , because it inhibited the forskolin-induced and basal G␣ i activity in a Ptx-sensitive manner in the HEK293 fibroblast cell line (Fig. 3) as well as in the L12 pre-B-lymphocyte cell line (Fig. 9). Furthermore, it activated the transcription factor SRE in a Ptx-sensitive manner. In the PI turnover experiments we observed a constitutive activation of phospholipase C upon co-transfection with the chimeric G-protein, Gqi4myr, but no activation in the presence of G q wt (Fig. 4) indicating constitutive G i , but not G q activation. Consistent with the constitutive G i signaling (Figs. 3 and  4), we observed a gene-dose-dependent increase in CREB activity upon co-transfection with Gqi4myr (Fig. 5B) and an inhibition of the forskolin-induced CREB activity by EBI2 in the absence of Gqi4myr (Fig. 5A). To further delineate the cellular signaling of EBI2 we included two other transcription factors. For the G␣ i -and G␣ 13 -dependent SRE, we observed a gene-dose-and Ptx-sensitive increase in the activity (Fig.  5D), supporting constitutive G␣ i activation. In contrast, no constitutive activity was observed through the G␣ q -dependent NF-B transcription factor (Fig. 5C). In summary, EBI2 signals constitutively through G i (but not G␣ s or G␣ q ) in a Ptx-sensitive manner.
Constitutive activation per se is a rather normal phenomenon among virus-encoded 7TM receptors (23). Some of the virus-encoded receptors display constitutive activity in a rather promiscuous manner through many pathways (e.g. ORF74-HHV8) (39), whereas other display constitutive activity only through G i (e.g. ORF74-EHV2) (40). The constitutive activity for the virus-related (virus-encoded and -induced) receptors is therefore not a unique phenomenon, nor is it unique among endogenous 7TM receptors. Thus, some endogenous naturally occurring 7TM receptors display constitutive activity through one or several pathways, e.g. the ghrelin and melanocortin receptors, and the orphan receptors GPR3, -6, -12, and -39 (41)(42)(43). However, the majority of rhodopsin-like 7TM receptors do not display detectable constitutive activity. In contrast, severe pathology has been ascribed to the presence of constitutive active mutants of 7TM receptors, e.g. mutations in rhodopsin (retinitis pigmentosa), in the MSH (pigmentation defects), and in the TSH receptor (hyperfunctional thyroid adenomas) (44,45).
Viral Exploitation of Endogenous 7TM Receptor Systems-One strategy of microbial survival is the exploitation of the cytokine system of the host. This is done either by molecular mimicry or by a manipulation of the expression levels of endogenous genes within the cytokine system. The molecular mimicry has been known for more than 10 years (46,47), and many of the virus-encoded genes belong to the chemokine system. These genes have presumably been obtained by the viruses through an ancient act of molecular piracy and serve important functions in immune system surveillance (and control) (7). One example is the Th2 immune system polarization by a combined action of the antagonistic CC-chemokine vCCL-2 and the agonistic vCCL-1 and vCCL-3, all from HHV8 (48). Another example comes from the CXC-chemokine binding ORF74 receptor family mentioned above, because the ORF74 from murine ␥-herpesvirus 68 has been shown to be important for virus reactivation (49). The EBV-encoded BILF1 was the first 7TM receptor from ␥1-herpesviruses to be characterized from a functional point of view (13,14). BILF1 from EBV shares many functional similarities to EBI2, because both receptors are surface-expressed, and both couple in a constitutive manner to the G i subclass of G-proteins in a Ptx-sensitive manner (see Figs. 2-5 for EBI2). However, one difference is the induction of BILF1 during lytic infection, in contrast to the unchanged EBI2 expression during lytic as well as latent infection (Fig. 8). It is therefore rather tempting to suggest that these two receptors function in conjunction and in synergy in the cellular reprogramming in benefit of EBV replication and survival and, consequently, that drugs targeting one (or both) of these receptors would be efficient antiviral drugs (7). The endogenous cytokine system is orchestrated and manipulated in a sophisticated manner by viruses. Thus, the expression of several chemokines and/or chemokine receptors are either up-or down-regulated upon virus cell-entry, as reviewed in Refs. 6 and 7. One example is the up-regulation of the CCL5/RANTES (regulated on activation normal T cell expressed and secreted) upon infection with herpes simplex virus (5). Another example is the up-and down-regulation of chemokine receptors upon EBV cell entry (2,50). In summary, viruses exploit the endogenous cytokine system by at least two different approaches, molecular mimicry (exemplified by BILF1) and regulation of the expression of endogenous genes related to the immune system (exemplified by EBI2).
EBI2 Is Still an Orphan Receptor-EBI2 was originally described as being closest related to the thrombin receptor (2), yet more recently alignment studies have suggested EBI2 to be related to at least three other subgroups of 7TM receptors. One comprehensive phylogenetic analysis from 2002 comprising 277 human 7TM receptors suggested that EBI2 belongs to a rather broad subgroup of 7TM receptors consisting of the nucleotide receptors (e.g. GPR86, KI01, P2Y10, and -12), the platelet-activating factor receptor, the thrombin-and thrombin-like receptors (THR, PAR2, and -3), and a handful of orphan receptors (GPR17, -18, -20, -34, and -87) (22). Another comprehensive 7TM alignment and expression analysis from 2003 comprising 367 human and 392 mice receptors suggested that EBI2 clusters with GPR18 and the lipid receptors, CysL1 and -2 (21). A third attempt to classify 7TM receptors, this time based on alignment-independent extraction of principal chemical properties of primary amino acid sequences from 2002, suggested that EBI2 belongs to the large and broadly defined group of peptide receptors (20). Thus, the classification of EBI2 is still a rather "open" issue; yet the close relationship with GPR18 seems consistent. The close proximity in region 13q32.3 (Fig. 1B) supports this relationship and suggests that they may have evolved from gene duplication events (as for some of the chemokine receptors). Interestingly, a recently published expression analysis of GPR18 (21) uncovers important similarities to the expression pattern we observed for EBI2. Thus, both receptors display high expression in PBMCs and moderate expression in lung tissue (21). GPR18 in contrast displayed moderate expression in ovary and different areas in the brain (striatum, cerebellum, hypothalamus, and cortex), whereas we found a low expression of EBI2 in the brain. However, the alignment studies, the close genomic proximity, and the similar receptor expression patterns suggest that EBI2 and GPR18 may have similar biologic actions and may constitute a novel family, presumably involved in immune system regulation based on the high expression in PBMCs but also with widespread functions in other organs.
Is There an Endogenous Agonist or Inverse Agonist for EBI2?-The expression profile of EBI2 (Fig. 7) suggests a role in immune system modulation. We furthermore show that EBI2 is constitutively active, also in lymphocytes (L12), and that the basal signaling in these cells can be modulated according to the expression level of EBI2 (Fig. 9). Interestingly, EBI2 is not the first constitutively active endogenous receptor within this system. Thus, for instance the neutrophilic expressed che-moattractantreceptorsforN-formyl-L-methionyl-L-leucyl-L-phenyl-alanine and the complement C5a receptor also display high constitutive activity (51). Interestingly, another complex physiologic system, appetite regulation, also contains several constitutively active 7TM receptors, e.g. the ghrelin and the melanocortin (MC) receptors (52). In fact, the only identified naturally occurring inverse agonist for 7TM receptors, agouti-related peptide, acts via the MC4 receptor (53). Apparently, it must be an advantage for the fine-tuning of important biologic systems (like the immune system and the food intake) to be controlled by ligands acting in two opposite directions, agonists and inverse agonists, on the same receptor, instead of having ligands (agonists) acting through two receptors with opposite biologic functions. On the other hand, there may not at all be any endogenous agonist(s) (or inverse agonist(s)) for EBI2. This receptor could be controlled solely at the level of the promoter region, i.e. expression control, and thereby a regulation of the activity simply due to the constitutive nature of the signaling. Thus, presently the endogenous ligand(s) remains unidentified, and may even not exist. Our studies provide the first functional characterization of EBI. Future transgenic expression studies and/or knock-out experiments, or the identification of endogenous ligands/ development of exogenous ligands, will be needed to further elucidate the biological function of EBI2.