Leukotriene Binding, Signaling, and Analysis of HIV Coreceptor Function in Mouse and Human Leukotriene B4Receptor-transfected Cells*

The mouse leukotriene B4receptor (m-BLTR) gene was cloned. Membrane fractions of human embryonic kidney 293 cells stably expressing m-BLTR demonstrated a high affinity and specific binding for leukotriene B4(LTB4, K d = 0.24 ± 0.03 nm). In competition binding experiments, LTB4was the most potent competitor (K i = 0.23 ± 0.05 nm) followed by 20-hydroxy-LTB4(K i = 1.1 ± 0.2 nm) and by 6-trans-12-epi-LTB4 and LTD4(K i > 1 μm). In stably transfected Chinese hamster ovary cells, LTB4 inhibited forskolin-activated cAMP production and induced an increase of intracellular calcium, suggesting that this receptor is coupled to Gi- and Go-like proteins. In Xenopus laevis melanophores transiently expressing m-BLTR, LTB4 induced the aggregation of pigment granules, confirming the inhibition of cAMP production induced by LTB4. BLT receptors share significant sequence homology with chemokine receptors (CCR5 and CXCR4) that act as human immunodeficiency virus (HIV) coreceptors. However, among the 16 HIV/SIV strains tested, the human BLT receptor did not act as a coreceptor for virus entry into CD4-expressing cells based on infection and cell-cell fusion assays. In 5-lipoxygenase-deficient mice, the absence of leukotriene B4 biosynthesis did not detectably alter m-BLT receptor binding in membranes obtained from glycogen-elicited neutrophils. Isolation of the m-BLTR gene will form the basis of future experiments to elucidate the selective role of LTB4, as opposed to cysteinyl-leukotrienes, in murine models of inflammation.

(LTC 4 , LTD 4 , and LTE 4 ), derived from arachidonic acid metabolism, are synthesized sequentially by 5-lipoxygenase and then by either LTA 4 hydrolase or LTC 4 synthase, respectively (1). The biological actions of the cysteinyl-leukotrienes are mediated through at least two G protein-coupled receptors referred to as CysLT 1 and CysLT 2 whose molecular identities remain uncharacterized (2). LTB 4 mediates its effects through a membrane G-protein-coupled receptor termed BLTR (3). Additionally, LTB 4 was shown to bind to the intracellular transcription factor peroxisome proliferator-activated receptor (4). This facet of binding has been proposed to be part of a negative feedback mechanism to limit the inflammatory actions of LTB 4 . However the binding to peroxisome proliferator-activated receptor has been questioned in recent experiments (5).
LTB 4 , a dihydroxy fatty acid, is one of the most potent known chemoattractant mediators, acting mainly on neutrophils but also on related myeloid cells, mast cells, and endothelial cells (6,7). LTB 4 induces chemotaxis, chemokinesis, and aggregation, causing the migration of neutrophils to sites of inflammation where the cells degranulate, resulting in the release of lysosomal enzymes in addition to other antibacterial and antimicrobicidal agents (8). LTB 4 also promotes the adherence of neutrophils to vascular endothelial cells and their transmigration, which amplifies the inflammatory response. LTB 4 has been implicated in the pathophysiology of various diseases like arthritis, inflammatory bowel disease, and psoriasis. The exact role of LTB 4 in the etiology of these disorders has been debated vigorously. Inhibitors of 5-lipoxygenase and the 5-lipoxygenase-activating protein have been used efficiently in models of ulcerative colitis (9,10), endotoxic shock (11), and induced asthma (12)(13)(14). The development of specific and highly potent BLTR antagonists has lagged behind cysteinyl receptor antagonists, which are currently available in the clinic for treatment of asthmatic inflammatory symptoms. One BLTR antagonist has shown encouraging results in a murine model of collagen-induced arthritis (15).
There has been considerable controversy about the molecular identification of the BLTR. In 1996 two independent research groups (16,17) cloned identical orphan receptor genes believed to encode members of the chemotactic factor subfamily of G protein-coupled seven transmembrane receptors. Initially, one group indicated that the receptor was unable to bind LTB 4 (16) but later retracted this finding to indicate specific binding (18). A third group cloned the identical receptor sequence, which was classified as a purinergic P2Y 7 receptor on the basis of its affinity for binding ATP (19). However, Yokomizo et al. (3) challenged this identification and provided convincing proof that the human BLTR (h-BLTR) had been cloned.
Intense interest in the role of chemokine receptors for facilitation of HIV entry into CD4-positive cells is evident from recent surveys of the literature (20,21). The h-BLTR is structurally related to chemokine receptors (e.g. CCR5) and is expressed in various immune cells. Recent evidence has suggested that the h-BLTR may function as a coreceptor for entry of primary HIV-1 isolates into CD4-positive cells (22). If true, this finding would add a significant new dimension to the interplay of leukotrienes, inflammation, and AIDS pathogenesis.
We report here the cloning and characterization of the m-BLTR, the signaling pathways for this G protein-coupled receptor, and a detailed analysis as to whether the h-BLTR can function as an HIV coreceptor. Screening of Genomic Library-A 1.1-kb fragment (NheI-PstI) from the cDNA, described in Ref. 19 and labeled with [␣-32 P]dCTP, was used as probe to screen the genomic mouse library by standard procedures. Putative positive clones were taken through two additional rounds of screening until plaque-purified. Phage DNA was purified and subjected to restriction enzyme digestion and Southern blot analysis.

Materials-[␣-
Plasmid Constructs-A 2.2-kb XbaI fragment that hybridized to the probe was gel extracted and inserted into the XbaI site of pBluescript II KS. The insert was sequenced entirely on both strands using automated sequencing (Applied Biosystems Big Dye Terminator, Ready Reaction Kit Reagents; ABI 373 sequencer) at the Department of Genetics, University of Pennsylvania.
Cell Culture and Transfection-Human embryonic kidney 293 (HEK 293) and Chinese hamster ovary (CHO) cells from the American Type Culture Collection (ATCC) were maintained in Dulbecco's modified Eagle's medium and Ham's F-12, respectively, containing 100 units/ml of penicillin and 100 g/ml streptomycin, supplemented with 10% (v/v) fetal bovine serum in a 5% CO 2 incubator at 37°C.
Cells were seeded at a density of 1-1.5 ϫ 10 6 cells/100-mm dish. Plasmid DNA and LipofectAMINE were mixed with Opti-MEM and added for 12 h on cells, at 70 -80% confluence for transient transfection or at 40 -50% confluence to select stable transfected cell lines. Cells were selected 24 h after transfection with G418. Isolation of stable clones was achieved by serial dilution or pipette lifting colonies followed by trypsinization and replating colonies in medium containing 2 mg/ml G418. When single clones where isolated and passed several times, the G418 concentration was reduced gradually to 0.5 mg/ml. Experiments were performed on two cell lines transfected with pCR3.1-m-BLTR (HEK-m-BLTR and CHO-m-BLTR).
Northern Blot and Reverse Transcriptase-PCR Analysis-Total RNA was prepared from different murine tissues using TRIzol reagent (Life Technologies, Inc.). RNA blot analysis was carried out with 15 g of total RNA from different tissues. The full-length m-BLTR DNA was used as probe, labeled with [␣-32 P]dCTP. Blots were prehybridized 20 min at 68°C in QuickHyb solution (Stratagene) and hybridized at 68°C for 1 h. The final washing conditions were 0.1 ϫ SSC, 0.1% SDS at 60°C. cDNA was synthesized from 5 g of total RNA with random primers (for reverse transcriptase-PCR). PCR primers CDF370 and CDF367 were used to amplify m-BLTR cDNA.
Binding Experiments-Cells were washed twice with cold phosphatebuffered saline without calcium and magnesium, harvested, and homogenized with a hand-held Polytron on ice in 5 ml of buffer A (10 mM HEPES, 2 mM EDTA, pH 7.4, and a mixture of protease inhibitors (Boehringer Mannheim)). The homogenate was centrifuged 5 min at 140 ϫ g, and the supernatant was then centrifuged at 100,000 g for 1 h. The pellet was resuspended in buffer A to 2 mg/ml of protein for [ 3

H]LTB 4 binding experiments performed within several hours.
To determine the specific binding on membrane fractions isolated from transiently transfected cells, cells were isolated 36 -40 h after transfection.
For  (23). Measurements of change in Ca 2ϩ levels in stirred cell suspensions were made using a Perkin-Elmer model LS50B luminescence spectrometer and were expressed as ratios of fluorescence emitted at 510 nm in response to excitation at 340 and 380 nm (data sampling interval, 0.5 s). Calcium concentrations were calculated from these ratios after determining the maximum and minimum ratios of fluorescence in the presence and absence of saturating levels of Ca 2ϩ , respectively, according to the ratiometric method described previously (24).
Functional Bioassay-Xenopus laevis melanophores were maintained in culture and used as described previously (25)(26)(27)(28). Briefly, transient expression of pCR3.1-m-BLTR in melanophores was achieved by electroporation. Melanophores were plated (15,000/well) in 96-well tissue culture plates and cultured for 2 days. Before the addition of agonist, cells were washed, incubated with 0.7 ϫ L-15, 0.1% bovine serum albumin as described (25), and then exposed to room light for 1 h. This exposure causes the cells to disperse their pigment granules and darken. The plates were incubated for 1 h in room light and base-line absorbance (A 0 ) reading obtained at 620 nm using a Molecular Devices V max kinetic microtiter plate reader. The agonist LTB 4 was added to the microtiter wells in 20-l aliquots at 10 ϫ their final concentration. Dose-response data were obtained 1 h later (A f ). Data were plotted with y ϭ 1 Ϫ (A f /A 0 ). Data are presented as mean Ϯ S.E.
cAMP Assay-CHO-m-BLTR were plated in 12-well plates at a density of 200,000 cells/well. 2 days later, the cultured medium was removed and replaced with 0.5 ml of culture medium with 25 M forskolin. After 15 min, different concentrations of LTB 4 were added. The medium was removed 10 min after the addition of agonist, and the cAMP produced was extracted by adding 0.5 ml of ethanol to each well. The supernatant was evaporated to dryness, and the pellet was dissolved in Tris (0.05 M) EDTA (4 mM) buffer, pH ϭ 7.5. The cAMP content was measured using a [ 3 H]cAMP radioimmunoassay kit from Amersham Pharmacia Biotech.
BLT Receptor Binding Assays to Glycogen-elicited Neutrophil Membranes-5-Lipoxygenase-deficient mice (29) and wild-type controls (five mice each) from our colony were injected intraperitoneally with glycogen and the neutrophils harvested 5 h later as described previously (29). Binding was carried out on membranes as mentioned above.
Cell-Cell Fusion and Virus Infection Assays-For cell-cell fusion assays, quail QT6 target cells were transfected with plasmids expressing CD4, the desired coreceptor, and luciferase under control of the T7 promoter. The h-BLTR, previously characterized as the P2Y 7 receptor (gift from S. Kunapuli; Ref. 19), was used for these studies (see "Results"). The next day, QT6 effector cells expressing the desired Env protein by transfection and T7 polymerase as a consequence of infection by recombinant vaccinia viruses were added. In this assay, cell-cell fusion results in cytoplasmic mixing and luciferase production, which can be easily quantified. Additional details can be found in previous papers (30,31). For infection assays, we used luciferase reporter viruses. Human 293T cells were transfected with plasmids expressing the desired Env (in a pcDNA3 or psv7d background) and with the NL4 -3 luciferase virus backbone (pNL-Luc-E Ϫ R Ϫ ) (32,33). Virus was harvested from the media the next day and used to infect feline CCCS cells (for HXBch) or 293T cells (for all other Evs) expressing the indicated CD4/coreceptor combinations. Infection was quantified by measuring luciferase activity 3-days postinfection. Data Analysis-Multiple analysis of variance tests followed by Bonferroni analysis were performed on cAMP data. *p Ͻ 0.05; ***p Ͻ 0.001.

Isolation of m-BLT Receptor
Gene-To clarify the confusion surrounding the molecular identity of the h-BLTR and to advance the study of the role of LTB 4 in murine inflammation models, we sought to clone and characterize the m-BLTR. A mouse genomic library was screened using a 1.1-kb fragment from the cDNA identified previously as encoding the purinergic P2Y 7 receptor (19). Among the positive clones that hybridized to the probe, one genomic clone with a 2.2-kb XbaI fragment was found to display an open reading frame of 1.056 kb, encoding a 351-amino acid receptor with seven putative transmembrane domains, two potential glycosylation sites, and several phosphorylation sites (GenBank accession number AF077673). This open reading frame revealed a deduced amino acid sequence with 77% identity to the h-BLTR (3). The third intracellular loop of the m-BLTR showed 100% identity to the human receptor, with two protein kinase C phosphorylation sites. The N-glycosylation sites and several of the phosphorylation sites are conserved between both species. The putative promoter region (Ϫ1 to Ϫ618) contained a CAAT-like box, a 36nucleotide poly(A) tract, and several other conserved sequences for the putative binding of GATA-1, PEA-3, c-Myc, c-Myb, Sp-1, and NF-IL6 transcription factors. While this manuscript was under review, Huang et al. (34) published a m-BLTR sequence cloned from a murine eosinophil cDNA library that was identical to the m-BLTR gene cloned here.
LTB 4 Binding-Because the genomic clone appeared to be intronless, we proceeded directly to expression studies. m-BLTR was subcloned in the expression vector pCR3.1 and used to transfect HEK 293 cells. Transient transfected HEK cells displayed specific binding for LTB 4 (Fig. 1A), as did cells transfected with the original human P2Y 7 receptor clone (not shown), whereas nontransfected (Ct) and mock transfected cells did not. Membrane fractions of HEK-m-BLTR cells showed a reversible, saturable and high affinity binding for LTB 4 with a K d ϭ 0.24 Ϯ 0.03 nM and B max ϭ 743 Ϯ 168 fmol/mg of protein (Fig. 1B). Displacement curves of [ 3 H]LTB 4 binding indicated that the binding site was specific for LTB 4 with a K i of 0.23 Ϯ 0.05 nM (n ϭ 4), followed by 20-hydroxy-LTB 4 , a metabolite of LTB 4 (K i ϭ 1.1 Ϯ 0.2 nM, n ϭ 4), and by the nonenzymatic breakdown product of LTA 4 , 6-trans-12-epi-LTB 4 , and LTD 4 (K i Ͼ 1 M) (Fig. 1C).

Functional Expression of the m-BLT Receptor-
The tissue distribution of the m-BLTR was investigated by Northern blot analysis of total RNA using two different probes. No signal was detected in the tissues tested (spleen, lung, kidney, liver, pancreas, uterus, testis, heart, aorta, brain). However, using reverse transcriptase-PCR, the m-BLTR mRNA was detected in all tissues, but not aorta (data not shown). The mouse heart and lung cDNAs were sequenced and found to be identical with the positive clone identified by mouse genomic library screening.
X. laevis melanophores provide a rapid, functional and visual readout of receptor activation. These cells disperse their pigment granules upon stimulation of receptors that are coupled to G s , G o , and G q and lead to accumulation of second messengers and thus appear dark. In contrast, stimulation of receptors that are coupled to G i cause a decrease in second messenger levels and result in pigment granule aggregation, and the cells appear light (27). A long term culture of melanophores was used to evaluate the functional activation of the m-BLTR. Stim- ulation of the receptor caused pigment aggregation in a concentration-dependent fashion with an EC 50 ϭ 0.13 nM (Fig. 2) consistent with coupling to G i .
With the same cell line, LTB 4 induced a concentration-dependent increase in the intracellular calcium levels (EC 50 ϭ 4.4 nM; Fig. 3B). With 1 M thapsigargin, an inhibitor of the endoplasmic reticulum calcium ATPase pump, the intracellular increase of calcium induced by 100 nM LTB 4 was inhibited by 89.5 Ϯ 1.4% (n ϭ 4).
BLT Receptor Binding in Leukotriene-deficient Mice-Previously, we (29) and another group (35) developed mice with disruptions in the 5-lipoxygenase gene. These mice were unable to synthesize cysteinyl-leukotrienes or LTB 4 in various inflammatory cell types. To test whether the absence of ligand influences receptor expression, we tested 5-lipoxygenase-deficient and control mice for alterations in LTB 4 binding using membranes from glycogen-elicited neutrophils. Although we obtained specific and competitive binding in these membranes, the lack of ability to synthesize LTB 4 in 5-lipoxygenase deficient mice did not substantially alter BLTR binding (Fig. 4).
Is the BLT Receptor a Coreceptor for HIV/SIV Entry into CD4-expressing Cells?-Primate lentiviruses utilize CD4 and a coreceptor (most often the chemokine receptors CCR5 and CXCR4) to enter target cells (20). The importance of CCR5 as a HIV-1 coreceptor was demonstrated by the finding that individuals who lack CCR5 because of a naturally occurring polymorphism are highly resistant to HIV infection (36). In addition to CCR5 and CXCR4, approximately one dozen other chemokine and related orphan receptors have been shown to function as coreceptors for more limited numbers of virus strains in vitro (37). However, the in vivo relevance of these alternative coreceptors is uncertain. Recently, using a PCRbased entry assay, h-BLTR was shown to serve as a coreceptor for some X4 HIV-1 stains (22). To investigate the ability of the BLTR to serve as a coreceptor for SIV and additional HIV-1 strains using a virus infection assay, we expressed CD4 and h-BLTR in human 293T or CCCS cells. The cells were then infected with luciferase reporter viruses bearing various HIV-1 and SIV Env proteins (Fig. 5). None of the viral Env proteins tested could mediate infection of cells expressing CD4 and h-BLTR, although infection was readily observed when cells expressed either CCR5 or CXCR4 (depending on the virus strain). In addition to the results depicted in Fig. 5, SIVmac1A11, smPBj6.6, macB670-clone 3, and sm62B were unable to use h-BLTR to infect cells. In separate experiments using transiently transfected HEK 293 cells the number of specific binding sites was found to be approximately 90 fmol/mg of protein.
The results described above indicate that the h-BLTR does not serve as a coreceptor for the virus strains tested or does so only inefficiently. We therefore employed a more sensitive cellcell fusion assay (Fig. 6) in an attempt to detect h-BLT coreceptor activity. We have found that coreceptors that support virus infection also support Env-mediated cell-cell fusion. However, there are cases when a given Env-coreceptor combination supports cell-cell fusion but does not support virus infection (38,39). Cells expressing the indicated Env proteins (including four primary HIV-1 strains that utilize CXCR4: 89.6, UG024.2, 93ZR001.3, and DH12) were mixed with cells expressing CD4 and either CCR5, CXCR4, or h-BLTR. Cell-cell fusion was quantified 7 h later using a gene-reporter system (30,31). Although fusion was readily observed with the major HIV-1 coreceptors, fusion using this sensitive assay was not observed when CD4 and h-BLTRs were coexpressed. Thus, the h-BLTR did not function as a coreceptor for either virus infection or Env-mediated cell-cell fusion.

DISCUSSION
The high level of sequence identity for the m-BLTR gene cloned here with the human sequences R2 (17), CMKRL1 (16), BLTR (3), and the clone classified as a P2Y 7 purinergic receptor (19) combined with the binding studies confirm the fact that these are mouse and human orthologs of the BLTR. Additional confirmation comes from a recent independent cloning of m-BLTR cDNA from murine eosinophils (34). LTB 4 affinity for m-BLTR was similar to that for h-BLTR (0.24 versus 0.154 nM, respectively). The NH 2 terminus as well as the three putative extracellular loops of the receptor show the lowest percentage identity between species (between 25 and 75% identity), in contrast with the high percent identity for the transmembrane domains (between 86 and 95% identity) and the intracellular loops (between 61 and 100%). The third intracellular loop of the m-BLTR, which shows two putative protein kinase C phosphorylation sites, is identical to the human sequence. These phosphorylation sites could be critical in the process of homologous desensitization of the BLTR that has been established in neutrophils (40), eosinophils (41), and granulocytes (42). The potential protein kinase C as well the casein kinase 2 (a Ser/Thr protein kinase) phosphorylation sites of the COOH-teminal loop could also be involved in the regulation of the activation/ desensitization of the receptor or in the uncoupling of the receptor to G protein.
The rank order of LTB 4 binding affinity and of related compounds for the m-BLTR is similar to the h-BLTR (3) and consistent with previous experiments using guinea pig lung membranes (43) and porcine spleen membranes (44). 20-Hydroxy-LTB 4 , a metabolite of LTB 4 , was the most potent competitor after LTB 4 . There was a discrepancy in the efficiency of LTB 4 to induce an increase in Ca i 2ϩ and to inhibit cAMP production in CHO cells. There was also a distinction in the ability of LTB 4 to inhibit cAMP production in CHO cells versus melanophores. These differences could be caused by different affinities of the receptor for G o -and G i -like proteins or by a difference in the type of G protein coupled to the m-BLTR in these two cell types.
The 5Ј-upstream sequence of the m-BLTR gene did not contain a TATA box but revealed the presence of putative consensus sequences for binding various transcription factors, including myeloid-specific factors, which could be implicated in the regulation of transcription of the m-BLTR. There was also a repetitive poly(A) sequence whose role remains to be defined.
The alteration of LTB 4 synthesis in vivo could modify m-BLTR expression. However, we found that LTB 4 binding to neutrophil membranes from 5-lipoxygenase knockout mice, who do not synthesize any leukotrienes, and from wild type mice were similar. In some pathological situations (45, 46), a modification in the level of BLTR binding has been demonstrated. The functional characterization of the promoter region, combined with mutation and deletion studies of this region, would help to elucidate transcription factors involved in the regulation of transcription of m-BLTR, and potentially link the results to inflammatory pathologies where LTB 4 is involved. The recent discovery of high expression of m-BLTR in eosinophils (34) will be important to explore in the context of airway inflammation models.
In addition to CCR5 and CXCR4, a host of other chemokine and orphan seven-transmembrane domain receptors have been shown to support infection by smaller numbers of HIV or SIV strains (for review, see Ref. 37). Recently, Owman et al. (22) reported that the h-BLTR functions as an efficient coreceptor for a number of virus strains, especially primary virus isolates that utilize CXCR4. However, we were unable to detect coreceptor activity for any of the 16 SIV and HIV-1 Env proteins examined using either virus infection or a sensitive cell-cell fusion assay with h-BLTR. There are a number of possible explanations for these discrepancies. First, it is simply possible that h-BLTR may function as a coreceptor for the strains tested by Owman et al. (22) but does not function as a coreceptor for the strains tested here. Second, h-BLT coreceptor function may be dependent upon the cell type or assay used. Owman et al. (22) used a PCR-based entry assay, which may be overly sensitive, whereas we used a reporter virus infection and cell-cell fusion assays. It will be important to demonstrate h-BLT coreceptor function in stable cell lines using more standard methods of viral quantification, such as p24 measurements. Third, coreceptor function can be highly dependent upon surface expression levels (47,48). Reagents are not yet available which make it possible to measure BLTR surface expression accurately. It is possible that we expressed suboptimal amounts of plasma membrane BLTR in our system, although this receptor did express well for binding studies in transiently transfected HEK 293 cells. Regardless of the explanation, the importance of h-BLTR as a viral coreceptor will rest upon the number and types of virus strains that can use it for infection, the levels at which it is expressed in vivo, and the CD4-positive cell types in which it is expressed. Only additional work will clarify these points.
FIG. 5. Infection assay using luciferase reporter viruses to test for h-BLT coreceptor function. 293T or feline CCCS cells were transfected with the indicated coreceptors and CD4 and infected with luciferase reporter viruses pseudotyped with the SIV and HIV-1 Env proteins listed. Data are presented from a representative experiment. Infections were performed two or three times depending on the Env protein.
FIG. 6. h-BLTR does not function as an HIV coreceptor in cell-cell fusion assays. Quail QT6 target cells were transfected with plasmids expressing CD4, a "coreceptor" (CXCR4, CCR5 or h-BLTR), and luciferase under control of the T7 promoter. The next day, QT6 effector cells expressing the desired Env protein by transfection and T7 polymerase as a consequence of infection by recombinant vaccinia viruses were added. Cell-cell fusion results in cytoplasmic mixing and luciferase production, measured as relative light units (RLU). Results depict one representative experiment (n ϭ 2) performed in triplicate.