GPR17 Is a Negative Regulator of the Cysteinyl Leukotriene 1 Receptor Response to Leukotriene D4 Akiko Maekawaa,B, Barbara Balestrieria,B, K

Home , Uridine diphosphate

GPR17 is a negative regulator of the cysteinyl leukotriene 1 receptor response to leukotriene D4 Akiko Maekawaa,b, Barbara Balestrieria,b, K. Frank Austena,b,1, and Yoshihide Kanaokaa,b,1

aDepartment of Medicine, Harvard Medical School, Boston, MA 02115; and bDivision of Rheumatology, Immunology, and Allergy, Brigham and Women’s Hospital, One Jimmy Fund Way, Boston, MA 02115

Contributed by K. Frank Austen, May 20, 2009 (sent for review May 5, 2009)

The cysteinyl leukotrienes (cys-LTs) are proinflammatory lipid me- CysLT2RtobeLTD4 Ͼ LTC4 Ͼ LTE4 and LTC4 ϭ LTD4 Ͼ diators acting on the type 1 cys-LT receptor (CysLT1R) to mediate LTE4, respectively. The findings that these receptors are ex- smooth muscle constriction and vascular permeability. GPR17, a G pressed not only on human smooth muscle but also on bone protein-coupled orphan receptor with homology to the P2Y and marrow-derived cells of the innate and adaptive immune systems cys-LT receptors, failed to mediate calcium flux in response to revealed a potential for involvement of the cys-LT/CysLTR leukotriene (LT) D4 with stable transfectants. However, in stable pathway in the infiltrating cells of the inflammatory response cotransfections of 6؋His-tagged GPR17 with Myc-tagged CysLT1R, (18, 19). We and others subsequently reported that the mouse the robust CysLT1R-mediated calcium response to LTD4 was abol- CysLT1R can function as a receptor for LTD4 in transfected cells ished. The membrane expression of the CysLT1R analyzed by FACS with a ligand preference similar to that of the human CysLT1R with anti-Myc Ab was not reduced by the cotransfection, yet both (20, 21) and that the mouse CysLT2R exhibits a ligand profile of LTD4-elicited ERK phosphorylation and the specific binding of LTC4 Ն LTD4 Ͼ LTE4 (21, 22). Targeted disruption of LTC4 3 [ H]LTD4 to microsomal membranes were fully inhibited. CysLT1R synthase, CysLT1R, and CysLT2R in mice confirmed the func- and GPR17 expressed in transfected cells were coimmunoprecipi- tion of the cys-LTs in vascular smooth muscle and in the cellular tated and identified by Western blots, and confocal immunofluo- aspects of allergic or chronic fibrotic pulmonary injury (23–27). rescence microscopy revealed that GPR17 and CysLT1R colocalize An additional receptor in porcine pulmonary arterial rings that on the cell surface of human peripheral blood monocytes. Lenti- is responsive to LTC4 and LTD4, but not LTE4, has been viral knockdown of GPR17 in mouse bone marrow-derived macro- recognized by differential pharmacologic responses or resistance phages (BMM⌽s) increased both the membrane expression of to available receptor antagonists on various smooth muscle CysLT1R protein by FACS analysis and the LTD4-elicited calcium flux preparations (28, 29). Presumptive evidence for the existence of ⌽ in a dose-dependent manner as compared with control BMM s, a particular LTE4 receptor was suggested by pharmacologic indicating a negative regulatory function of GPR17 for CysLT1Rin studies with guinea pig tissues (30) and has recently been a primary cell. In IgE-dependent passive cutaneous anaphylaxis, supported by the demonstration of LTE4-mediated vascular GPR17-deficient mice showed a marked and significant increase in permeability in mice lacking both the CysLT1R and the CysLT2R vascular permeability as compared with WT littermates, and this through targeted disruption and crossbreeding (31). vascular leak was significantly blocked by pretreatment of the mice A G protein-coupled orphan receptor, GPR17 (originally with the CysLT1R antagonist, MK-571. Taken together, our findings called ‘‘R12’’), was cloned by homologous screening in human suggest that GPR17 is a ligand-independent, constitutive negative genomic DNA with the chemokine IL-8 receptor (32). Subse- regulator for the CysLT1R that suppresses CysLT1R-mediated func- quently, R12 and its variant form of cDNA were cloned by tion at the cell membrane. homologous screening in a human hippocampus cDNA library with the nucleotide chicken P2Y1 and murine P2Y2 receptors inflammation ͉ knockout mice ͉ lipid mediator ͉ macrophage (33). Phylogenic analysis for P2Y-related receptors revealed that

human GPR17 is homologous to the CysLTRs (34), with an IMMUNOLOGY

he cysteinyl leukotrienes (cys-LTs), leukotriene (LT) C4, amino acid sequence that is 31% and 36% identical to that of the human CysLT R and CysLT R, respectively. Human GPR17 TLTD4, and LTE4, are proinflammatory mediators generated 1 2 by the 5-lipoxygenase (5-LO) pathway after activation of par- also is 90.3% identical to both mouse and rat orthologs in amino ticular bone marrow-derived cells to release arachidonic acid acid sequence (34). Human, rat, and mouse GPR17 have recently from the phospholipids of the outer nuclear membrane. In the been identified as dual receptors for uracil nucleotides and the 35 ␥ presence of the 5-LO-activating protein (1, 2), 5-LO converts cys-LTs, LTC4, and LTD4, based on [ S]GTP S binding assays arachidonic acid to LTA4 (3), which can be conjugated to with transfectants (34, 35). The original report (34) also showed reduced glutathione to form LTC4 by an integral trimeric nuclear by single-cell calcium imaging that 1321N1 cells and COS-7 cells membrane enzyme, LTC synthase (4–6). After energy- expressing human GPR17 could respond to 100 nM of LTD4 and 4 ␮ dependent export of LTC4, glutamic acid and glycine are se- to 100 M of UDP-glucose. quentially cleaved by ␥-glutamyl transpeptidase (7) or ␥-glu- We began a study of mouse GPR17 seeking a novel receptor tamyl leukotrienase (8) and dipeptidases (9, 10) to form LTD for LTE4. However, we could not obtain significant calcium flux 4 ␮ and LTE4, respectively. The cys-LTs are implicated in human to 1 M of LTC4, LTD4,orLTE4 in stable human and mouse bronchial asthma by their pharmacologic actions to constrict GPR17 transfectants using 1321N1, CHO, and HEK-293T cells airway and vascular smooth muscle (11–13) and by the clinical as hosts. In addition, 1321N1 cell clones stably expressing mouse efficacy of agents that block 5-LO or the type 1 receptor for the cys-LTs (CysLT1R) (14, 15). Author contributions: A.M., K.F.A., and Y.K. designed research; A.M., B.B., and Y.K. Two types of human receptors for the cys-LTs, designated performed research; A.M., B.B., K.F.A., and Y.K. analyzed data; and A.M., K.F.A., and Y.K. CysLT1R (16) and CysLT2R (17), which belong to the 7-trans- wrote the paper. membrane, G protein-coupled receptor family, were cloned and The authors declare no conﬂict of interest. shown to be 38% homologous in their amino acid sequences. The 1To whom correspondence may be addressed. E-mail: [email protected] or radio-labeled ligand-binding assay using microsomal membranes [email protected]. from CysLT1R and CysLT2R transfectants revealed the rank This article contains supporting information online at www.pnas.org/cgi/content/full/ order of affinities of the cys-LTs for the CysLT1R and the 0905364106/DCSupplemental.

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0905364106 PNAS ͉ July 14, 2009 ͉ vol. 106 ͉ no. 28 ͉ 11685–11690 Downloaded by guest on September 30, 2021 GPR17 did not respond to 100 ␮M of UDP-glucose. Since heterodimerization of G protein-coupled receptors modulates expression and/or function either negatively (36, 37) or positively (38, 39) in various transfectants, we considered the possibility that GPR17 may associate with CysLT1R to control its calcium signaling function. Here we show that GPR17 may function as a negative regulator for the CysLT1R response to LTD4 not only in cotransfection of transformed cells but also constitutively in primary cells in which its knockdown resulted in increased membrane expression and LTD4-mediated function of CysLT1R. We provide physiologic evidence for this regulatory role of GPR17 by demonstrating that the vascular leak following IgE- dependent, mast cell-mediated passive cutaneous anaphylaxis (PCA) is significantly increased in GPR17-deficient mice and that this response is blocked by administration of a CysLT1R antagonist.

Results and Discussion Cotransfection with GPR17 Inhibits the Function but Not the Mem- brane Expression of the CysLT1R. After failing to find that LTD4- initiated calcium flux in a variety of cell lines transfected with 6ϫHis-GPR17 and expressing membrane protein detected by FACS analysis with mouse monoclonal anti-6ϫHis Ab (data not shown), we considered that GPR17 might influence CysLT1R expression and function. We established 12 independent 1321N1 clones stably expressing Myc-tagged mouse CysLT1R with or without mouse 6ϫHis-tagged GPR17. FACS analysis showed that 6ϫHis-GPR17 protein was expressed in all clones as assessed by staining with anti-His Ab in nonpermeabilized cells [supporting information (SI) Fig. S1A] and that the GPR17

expression did not alter the cell surface expression of the Fig. 1. Stable cotransfection of 1321N1 cells with GPR17 inhibits CysLT1R- Myc-tagged CysLT1R in replicate stable transfectants (Fig. mediated calcium ﬂux to LTD4.(A) Intracellular calcium mobilization by LTD4 in these transfectants labeled with fura-2. Arrows indicate point of injection S1A). The clones that expressed Myc-tagged CysLT1R alone and were labeled with fura-2 responded to 0.001, 0.01, 0.1, and 1 ␮M of LTD4 or histamine. Experiments were performed 3 times with 12 individual clones for each combination, and data from 1 representative clone are shown. of LTD in a dose-dependent manner (Fig. 1A Left), as previ- 4 (B) Peak relative ﬂuorescence of intracellular calcium mobilization by LTD4 in ously reported (20). In contrast, as depicted for the highest dose 1321N1 cells stably expressing CysLT1R (closed circles), GPR17 (closed squares), of LTD4 (1 ␮M), there was no CysLT1R-mediated calcium signal and cells coexpressing CysLT1R and GPR17 (open triangles). Data represent the in clones coexpressing GPR17, although they responded to 100 mean Ϯ SE from 5 independent experiments. *, P Ͻ 0.005 vs. cells coexpressing ␮M of histamine used as a positive control for function of a CysLT1R and GPR17. constitutive receptor (40) (Fig. 1A Right). Inhibition of the CysLT1R response to LTD4 by coexpression of GPR17 was Cotransfection with GPR17 Inhibits CysLT1R-Mediated ERK Phosphor- significant (P Ͻ 0.005) at all doses of LTD4 as compared to ylation in Response to LTD4. To further assess the effect of CysLT1R alone (Fig. 1B). Repetition of these transfection cotransfection with GPR17 on CysLT1R functions, we per- protocols 3 to 5 times in CHO or HEK-293T cells again showed formed ERK phosphorylation assays (41) in CHO cell stable a robust dose-dependent response to LTD4 (0.001, 0.01, 0.1, and transfectants expressing CysLT R with or without GPR17. The ␮ 1 1 M) with CysLT1R alone, and no response to these doses of bands corresponding to phospho-p44/42 ERK were detected in LTD4 with GPR17 cotransfection or GPR17 alone (data not the cells expressing CysLT1R alone after 10-min stimulation with shown). We also examined the effect of coexpression of GPR17 1 ␮M (Fig. 2A) or 0.01 ␮M (data not shown) of LTD4, but not with CysLT R in stable 1321N1 transfectants on another ligand, 1 in the cells coexpressing CysLT1R and GPR17. These findings LTC . We again observed loss of the CysLT R-mediated calcium 4 1 indicate that GPR17 expression inhibits the CysLT1R response flux function (Fig. S1B). to LTD4 for not only the calcium signaling but also the ERK To exclude the possibility that only LTD4 nonresponsive, signaling pathway, which leads to induction of cytokines/ stable cotransfectants had selectively survived, we performed chemokines (41). It should be noted that we did not detect any the same experiments in 1321N1 cells transiently expressing LTD -mediated ERK phosphorylation in LTD -stimulated cells ϫ 4 4 Myc-tagged mouse CysLT1R, 6 His-tagged mouse GPR17, or expressing GPR17 alone. both together. We observed that the CysLT1R/GPR17 coexpres- 3 sion did not alter the cell surface expression of the Myc-tagged Cotransfection with GPR17 Suppresses [ H]LTD4 Binding by Mem- ϫ 3 CysLT1Rorofthe6 His-tagged GPR17 as compared to single branes from Cells Expressing CysLT1R. We performed [ H]LTD4 transient transfectants but that it completely inhibited the LTD4- binding assays using microsomal membranes prepared from induced, CysLT1R-mediated calcium signal (Fig. S2 A and B). CHO cell clones stably expressing mouse CysLT1R with or The control response to histamine-induced calcium flux was without cotransfection with mouse GPR17. Microsomal mem- intact (Fig. S2B). These findings indicate that GPR17 expression branes from CHO cell clones expressing CysLT1R alone showed 3 inhibits the CysLT1R response to LTD4 without affecting cell increased binding with increasing inputs of [ H]LTD4 alone (Fig. surface expression of CysLT1R or the function of a constitutive 2B, Left). This binding was prevented by inclusion of excess cold receptor. ligand, thereby indicating its specificity as we had previously

11686 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0905364106 Maekawa et al. Downloaded by guest on September 30, 2021 A A C

Fig. 2. Cotransfection with GPR17 inhibits ERK phosphorylation and 3 [ H]LTD4 binding in microsomal membranes from cells stably expressing CysLT1R. (A) Western blot with anti-phospho p44/42-ERK (Upper panels) or anti-total p44/42-ERK (Lower panels) in CHO cell stable transfectants stimulated for 10 min with or without 1 ␮MofLTD4. Results are representative from 3 independent experiments. (B) Microsomal membranes from CHO cells transfected with CysLT1R alone (Left) or cotransfected with GPR17 (Right) were 3 incubated with 0–1,000 pM [ H]LTD4 in the presence (open circles) or absence 3 (closed circles) of an excess amount of cold LTD4. Bound [ H]LTD4 was sepa- 3 rated from free [ H]LTD4 by filtration through glass fiber filters. The residual 3 Fig. 3. GPR17 and CysLT R coimmunoprecipitate in CHO cells cotransfected membrane-associated [ H]LTD4 on the filter was determined by liquid scintil- 1 lation counting. The insert shows cell surface staining with isotype control as with both receptors and GPR17 and CysLT1R colocalize at the plasma membrane in human monocytes. (A and B) HEK-293T cells were transfected with a shaded histogram and with anti-CysLT1R Ab as a solid line. Data represent the mean Ϯ SE from 3 independent experiments. *, P Ͻ 0.005 vs. cells pEYFP-N1 vector (YFP) ϩ pCXN vector (lanes 1 and 5), GPR17-YFP ϩ pCXN (lane 2), YFP ϩ CysLT R-Myc (lane 3), or GPR17-YFP ϩ CysLT R-Myc (lanes 4 and 6). coexpressing CysLT1R and GPR17. 1 1 (A) Lysates from each transfectant were immunoprecipitated with rabbit antiserum against GFP, and the precipitates (Left) and total cell lysates (Right) demonstrated (20). In contrast, there was no total or specific were separated by SDS/PAGE and analyzed by Western blot with rabbit [3H]LTD binding in microsomal membranes from clones co- polyclonal anti-Myc Ab. (B) Lysates from each transfectant were immunopre- 4 cipitated with agarose conjugated with mouse anti-Myc mAb and the precip- transfected with GPR17 and CysLT1R (Fig. 2B, Right). The cell itates (Left) and total cell lysates (Right) were separated by SDS/PAGE and surface expression of CysLT1R with and without GPR17 co- analyzed by Western blot with mouse anti-GFP mAb. Molecular weight mark- transfection was similar by FACS analysis with rabbit polyclonal ers are shown at Left. An arrowhead and an asterisk indicate CysLT1R-Myc anti-CysLT1R peptide Ab (RB34) (Fig. 2B, Insets). There was no dimers and GPR17-YFP dimers, respectively. Results are representative from 3 3 [ H]LTD4 specific binding in microsomal membranes from independent experiments. IP, immunoprecipitation; WB, Western blot anal- stable clones expressing GPR17 alone (data not shown). Specific ysis; IgH, Ig heavy chains. (C) Human peripheral blood monocytes were stained IMMUNOLOGY 3 with anti-human GPR17 N-terminal peptide Ab and anti-human CysLT R [ H]LTD4 binding on microsomal membranes expressing 1 C-terminal peptide Ab. Immunoreactivities to GPR17 and CysLT1R were visu- CysLT1R was not inhibited when they were mixed with microsomal membranes from clones expressing only GPR17, indicat- alized with secondary Abs conjugated with Cy3 (red) for GPR17 and with FITC (green) for CysLT1R. Colocalization is visualized in the merged image as ing that GPR17 needed to be coexpressed with CysLT1Rto yellow-white, because of overlap of the colors for Cy3 and FITC. Results are 3 suppress its ligand binding (data not shown). [ H]LTD4 binding representative from 4 independent experiments. was also robust in 1321N1 cells transiently expressing Myc- tagged mouse CysLT1R and absent with co-transfection with 6ϫHis-tagged GPR17 (Fig. S2C). Ϸ120 kDa for GPR17-YFP only in immunoprecipitates prepared from the cells cotransfected with CysLT1R-Myc and CysLT1R and GPR17 Coimmunoprecipitate in CHO Cells Cotransfected GPR17-YFP (Fig. 3 A and B, lane 4). We only detected the bands with both Receptors. To determine if GPR17 can physically corresponding to each monomer when we oversaturated the interact with CysLT1R, we used a fusion protein of mouse signals, suggesting that the major form of both receptors for GPR17 and the YFP (GPR17-YFP) and a Myc-tagged-mouse interaction may be dimers that can progress to oligomers as CysLT1R (CysLT1R-Myc). The GPR17-YFP and CysLT1R-Myc observed in bands with higher molecular weights. The direct constructs were transiently transfected alone or in combination Western blots of these lysates show comparable bands. These into CHO cells, and the cell lysates were immunoprecipitated findings indicate that GPR17 and CysLT1R can form het- with mouse anti-Myc mAb or with rabbit antiserum against GFP, erodimers or heteromultimers. which recognizes denatured recombinant GFP and YFP. The immunoprecipitates were separated by SDS/PAGE and analyzed CysLT1R and GPR17 Colocalize on Human Monocytes. To examine if by Western blot. Mouse anti-GFP mAb (JL-8) was used for GPR17 and CysLT1R colocalize on the plasma cell membrane in Western blots of the immunoprecipitates obtained with anti-Myc primary cells, we stained, fixed, and permeabilized human Ab, and rabbit polyclonal anti-Myc Ab was used for immuno- monocytes with rabbit polyclonal anti-human GPR17 N- precipitates obtained with anti-GFP Ab. We detected bands terminal peptide Ab and goat polyclonal anti-human CysLT1R corresponding to dimers at Ϸ70 kDa for CysLT1R-Myc and C-terminal peptide Ab. Visualization with secondary Abs, Cy3-

Maekawa et al. PNAS ͉ July 14, 2009 ͉ vol. 106 ͉ no. 28 ͉ 11687 Downloaded by guest on September 30, 2021 AB C

D E

Fig. 4. Knockdown of GPR17 in BMM⌽s increases CysLT1R expression and function. (A) RT-PCR analysis for GPR17 and CysLT1R in mouse BMM⌽s. (B) Quantitative RT-PCR for mouse GPR17 in total RNA from BMM⌽s 2 days after infection with lentiviral particles carrying pLKO1 vector (gray) or individual constructs (TRCN124057 and TRCN124056) of pLKO1-mouse GPR17 shRNA (black). (C) Cell surface expression of CysLT1R on mouse BMM⌽s 3 days after infection with lentiviral particles carrying pLKO1 vector (dotted line) or pLKO1-mouse GPR17 shRNA (TRCN124057, thick black line). Isotype control staining is shown in shaded histogram. (D) Dose-dependent intracellular calcium ﬂux response to LTD4 (Upper)or50␮M of ATP (Lower) in mouse BMM⌽s 3 days after infection with lentiviral particles carrying pLKO1 vector or pLKO1-mouse GPR17 shRNA (TRCN124057). Arrows indicate point of injection of LTD4 or ATP. Results are representative from 5 independent experiments. (E) Peak relative ﬂuorescence of intracellular calcium mobilization by LTD4 in mouse BMM⌽s infected with lentiviral particles carrying pLKO1 vector (open triangles) or individual constructs [TRCN124057 (closed circles) and TRCN124056 (closed squares)] of pLKO1-mouse GPR17 shRNA. Data represent the mean Ϯ SE (n ϭ 5–7) from 5 independent experiments. *, P Ͻ 0.01; #, P Ͻ 0.05 vs. control BMM⌽s.

conjugated anti-rabbit IgG (red) and FITC-conjugated anti-goat the GPR17 knockdown BMM⌽s was increased in a dose- IgG (green), showed discrete colors, while analysis by confocal dependent manner over a 1,000-fold range of the ligand con- immunofluorescence microscopy showed a pseudocolored yel- centrations as compared with BMM⌽s infected by control vector low-white signal resulting from the merged immunostained construct (Fig. 4D and Fig. S3A). The sensitivity for a compa- images with red and green signals. Both GPR17 and CysLT1R rable signal was increased by about 1–2 logs for each of the were dominantly expressed on the plasma membrane and they knockdown constructs in BMM⌽s (Fig. 4E). Importantly, ATP- were partially colocalized (Fig. 3C, merged image). elicited calcium flux (24) was similar in both GPR17 knockdown and control BMM⌽s, suggesting that GPR17 does not affect GPR17 Silencing in BMM⌽s Increases CysLT1R Expression and Function. calcium flux mediated by another intrinsic receptor (Fig. 4D, To seek a regulatory function of GPR17 in primary cells, we used Lower panels). We also observed the enhanced CysLT1R re- shRNA interference by the lentivirus system to knock down sponse to a weaker CysLT1R ligand, LTC4, with GPR17 knock- GPR17 expression in mouse bone marrow-derived macrophages down (Fig. S3B). These results reveal that GPR17 can negatively (BMM⌽s). This cell type expressed both GPR17 and CysLT1R regulate the surface expression and function of the CysLT1Rin as assessed by RT-PCR (Fig. 4A). Quantitative RT-PCR re- macrophages. vealed that lentivirus infection with 2 of 5 shRNA constructs In contrast to the enhanced CysLT1R expression in our (TRCN124057 and TRCN124056) effectively suppressed the knockdown studies (Fig. 4C), cooverexpression of GPR17 did GPR17 mRNA expression in BMM⌽s by about 70% and 40%, not affect the cell surface expression of CysLT1R(Fig. S1A and respectively (Fig. 4B). Mouse BMM⌽s were then infected with S2A). Not all potential G protein-coupled receptors expressed lentivirus particles carrying a control vector or the GPR17- on the cell surface are associated with G proteins according to knockdown constructs and cultured for 72 h. FACS analysis with the allosteric ternary complex model (42). As we overexpressed anti-CysLT1R peptide Ab (RB34) revealed that the cell surface CysLT1R under the control of a strong promoter, the chicken expression of CysLT1R was enhanced 1.5- to 2-fold by the ␤-actin promoter, it is likely that most of the CysLT1Risnot GPR17 knockdown with both constructs as compared with associated with G proteins, and thus not functional, but still BMM⌽s infected by control vector construct (Fig. 4C and Fig. immunoreactive. Perhaps GPR17 only interacts with the func- S3A). In a parallel experiment, the LTD4-elicited calcium flux in tional CysLT1R. It is also conceivable that cooverexpression of

11688 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0905364106 Maekawa et al. Downloaded by guest on September 30, 2021 blocked by pretreatment of the mice with the CysLT1R antagonist, MK-571. This enhanced vascular leak in GPR17-deficient mice was also evident when the dose of monoclonal IgE per ear was reduced by a log so that there was virtually no response in the WT littermates. Again, the vascular leak in GPR17-deficient mice was markedly and significantly suppressed by MK-571 (Fig. S4D). Our findings of a lack of LTC4, LTD4, and LTE4 recognition by GPR17 transfectants are consistent with the previous observation by Heise et al. that there was no ClϪ conductance in response to LTD4 in Xenopus oocytes when the cRNA for GPR17 was microinjected (17). Instead, we found that cotransfection of GPR17 eliminated CysLT1R binding and signaling in Fig. 5. GPR17-deficient mice show markedly increased CysLT1R-mediated response to a full range of LTD4 concentrations without pre- vascular permeability in PCA. GPR17-deficient mice (KO) and their WT litter- venting expression at the membrane, as assessed by FACS mates (WT) were sensitized with 10 ng anti-DNP IgE or saline and challenged analysis. Importantly, knockdown of GPR17 in mouse BMM⌽s with 100 ␮g DNP-human serum albumin with 1% Evans blue dye. These mice increased membrane expression of CysLT1R and increased the had (black columns) or had not (white columns) been pretreated with MK-571. magnitude and sensitivity to LTD -induced calcium flux. Finally, The net Evans blue dye extravasation is expressed as the mean Ϯ SE (n ϭ 8 mice 4 per group) from 3 independent experiments. *, P Ͻ 0.01; #, P Ͻ 0.05. GPR17-deficient mice showed an enhanced vascular permeability in IgE-dependent PCA as compared to their WT littermates, and this enhanced response was MK-571 sensitive. These results GPR17 may stabilize the CysLT1R localization on the cell indicate that GPR17 is a negative regulator of CysLT1R- surface by forming heterodimers. Indeed, cooverexpression of mediated responses in vitro and in vivo. CysLT1R also did not affect the cell surface expression of GPR17 In a recent study of another class of orphan G protein-coupled (Fig. S2A). The knockdown study avoids the problem of induced receptor, Levoye et al. reported that the orphan receptor GPR50 CysLT1R overexpression by decreasing GPR17 transcription in and the melatonin 1 (MT1) receptor formed a heterodimer and the presence of a native expression system for CysLT1R. In this showed that coexpression of GPR50 and MT1 receptor specif- setting, there was both increased membrane expression of ically abolished MT1 receptor function in the transfectants (43). This inhibition was suppressed by deletion of a large portion of CysLT1R and its G protein-coupled calcium flux. the C-terminal tail of GPR50 although the heterodimerization of Viability of a GPR17-Deficient Mouse Strain. To demonstrate that GPR50 and MT1 receptor was retained, thereby suggesting that the C-terminal domain of GPR50 modulates a regulatory por- dampening of the CysLT1R function was a physiological role of GPR17 in vivo, we obtained a GPR17-deficient mouse strain. tion of MT1. These authors consider this finding to be an Although this strain is commercially available (Deltagen), its example of constitutive regulation of a receptor of the neuroen- phenotypic characterization has not been reported. The target- docrine system by a nonligand-binding G protein-coupled or- ing vector was designed so that a lacZ-neo gene casette interrupts phan receptor. Our observation that GPR17, an orphan receptor 76 bp of the coding region on exon 2 of the mouse GPR17 (Fig. with homology to the CysLTRs, negatively regulates the expres- S4A). The genotypes of the pups from intercrossing of heterozy- sion and function of CysLT1R provides the initial example of such nonligand-dependent, constitutive regulation in the 5-LO/ gotes (N7, C57BL/6) determined by PCR with genomic tail DNA LTC synthase pathway. Importantly, we have validated this using GPR17 gene-specific primers and a neo gene primer 4 possibility by showing that in the absence of GPR17 the demonstrated the 440-bp and 299-bp bands of the mutant and CysLT R-mediated component of the PCA response is pro- WT alleles, respectively, as illustrated for 1 litter in Fig. S4B. The 1 foundly aggravated. ratio, homozygotes:heterozygotes:WT, was consistent with the IMMUNOLOGY expected Mendelian frequency, and similar numbers of male and Materials and Methods female homozygotes were produced. The GPR17-deficient mice cDNA Cloning of the Mouse and Human GPR17. Methods to clone the full-length developed normally and exhibited no apparent clinical abnor- GPR17 cDNAs from mouse brain and human placenta cDNA libraries by PCR are malities up to 9 months of age. Because mouse GPR17 is described in SI Materials and Methods. abundantly expressed in the brain (35), we examined whether the lacZ-neo gene casette insertion disrupted expression of the Cell Culture and Stable Expression of the Mouse CysLT1R With or Without Mouse GPR17 mRNA in this tissue. By Northern blot analysis with total GPR17. Stable transfectants of the mouse CysLT1R with or without mouse RNAs, a 6-kb band was detected in the brain of WT mice, GPR17 were established as described in SI Materials and Methods. whereas the band was not detected in the sample from GPR17- deficient mice (Fig. S4C). Intracellular Calcium Mobilization. The 1321N1 or CHO cell transfectants and mouse BMM⌽s were labeled with fura-2, stimulated with LTC4, LTD4,orLTE4, and assayed as described (20). GPR17 Deficiency in Mice Increases CysLT1R-Mediated Vascular Per- meability in PCA. We previously reported that the vascular leak ERK Phosphorylation. After6hofserum starvation, the CHO cell stable accompanying IgE-dependent, mast cell-mediated PCA is highly transfectants were stimulated with vehicle (ethanol) or LTD4 at 1 ␮M and 0.01 dependent on the integrity of LTC4 synthase (23) and the ␮M for 10 min at 37 °C and assayed as described in SI Materials and Methods. CysLT1R (24). To demonstrate a physiologic regulatory role of 3 3 GPR17 in this CysLT1R-mediated inflammatory response in [ H]LTD4 Specific Binding. [ H]LTD4 binding to the microsomal membranes vivo, we repeated such PCA studies in GPR17-deficient mice and from the CHO cell transfectants was assayed as described (20). their WT littermates. After local sensitization of ear mast cells with 10 ng of monoclonal IgE and i.v. challenge with hapten- Flow Cytometry. To confirm the expression of epitope-tagged GPR17 and CysLT1R in 1321N1 cells, Alexa Fluor 647-conjugated mouse Penta-His mAb specific antigen, the extravasation of plasma proteins associated (Qiagen, 5 ␮g/mL) and Alexa Fluor 488-conjugated mouse anti-Myc tag mAb with Evans blue dye was markedly and significantly increased in (4A6, Upstate, 5 ␮g/mL), respectively, was used. Rabbit polyclonal anti-CysLT1R GPR17-deficient as compared to their sufficient littermates (Fig. peptide Ab (RB34, against the conserved sequence DEKNNTKCFEPPQNN of 5). Furthermore, this protein leak was significantly (Ϸ80%) extracellular loop 3 of both mouse and human CysLT1R, 5 ␮g/mL) (37) and

Maekawa et al. PNAS ͉ July 14, 2009 ͉ vol. 106 ͉ no. 28 ͉ 11689 Downloaded by guest on September 30, 2021 allophycocyanin-conjugated donkey anti-rabbit IgG were used on CHO cells Conventional and Quantitative RT-PCR. Methods for conventional and quan- stably expressing CysLT1R. Nonspecific rabbit and mouse IgG (Jackson Immuno- titative RT-PCR are described in SI Materials and Methods. chemical) were used as controls. Analyses were performed on a FACSCanto flow cytometer (BD Biosciences), and data were analyzed with the FlowJo software. GPR17-Deficient Mice and Northern Blot Analysis. Detailed information of GPR17-deficient mice and methods for Northern blot analysis are described in Immunoprecipitation and Confocal Immunofluorescence Microscopy. Meth- SI Materials and Methods. ods for immunoprecipitation with transfectants and confocal immunofluorescence microscopy with human monocytes are described in SI Materials PCA. PCA was performed in the ear of GPR17Ϫ/Ϫ and their WT littermates as and Methods. described (24) except that the intradermal dose of mouse monoclonal anti- dinitrophenyl (DNP) IgE (Sigma) was reduced from 25 ng to 10 or 2.5 ng. The Preparation of Lentiviral Particles and Infection. Five shRNA constructs for the details are described in SI Materials and Methods. mouse GPR17 in a pLKO1 vector (TRCN124054–8) were purchased from Open Biosystems. Infectious viral particles were prepared by cotransfection of HEK- Statistical Analysis. Results were expressed as mean Ϯ SE. Student’s t test was 293T cells with each shRNA construct, a packing vector (psPAX2), and an used for the statistical analysis. A value of P Ͻ 0.05 was considered envelope vector (pMD2.G) according to the manufacturer’s protocol. The viral significant. stocks were titrated in HT1080 cells. Mouse BMM⌽s were cultured from ⌽ BALB/c mouse bone marrow cells as described (44). BMM s were incubated ACKNOWLEDGMENTS. We thank Tyler Hickman for technical assistance and with viral particles in culture medium to achieve a multiplicity of infection of Dr. Joshua Boyce for critical discussions of this work. These studies were 10. After 24 h, the medium was replaced and the cells were cultured for an supported by National Institutes of Health grants P01HL36110, R01HL82695, additional 48 h before the functional studies were performed. R01HL090630 (to Y.K.), and K08AI064226 (to B.B.).

1. Dixon RA, et al. (1990) Requirement of a 5-lipoxygenase-activating protein for leuko- 24. Maekawa A, Austen KF, Kanaoka Y (2002) Targeted gene disruption reveals the role of triene synthesis. Nature 343:282–284. cysteinyl leukotriene 1 receptor in the enhanced vascular permeability of mice under- 2. Ferguson AD, et al. (2007) Crystal structure of inhibitor-bound human 5-lipoxygenase- going acute inflammatory responses. J Biol Chem 277:20820–20824. activating protein. Science 317:510–512. 25. Beller TC, et al. (2004) Cysteinyl leukotriene 1 receptor controls the severity of chronic 3. Rouzer CA, Matsumoto T, Samuelsson B (1986) Single protein from human leukocytes pulmonary inflammation and fibrosis. Proc Natl Acad Sci USA 101:3047–3052. possesses 5-lipoxygenase and leukotriene A4 synthase activities. Proc Natl Acad Sci USA 26. Beller TC, Maekawa A, Friend DS, Austen KF, Kanaoka Y (2004) Targeted gene 83:857–861. disruption reveals the role of the cysteinyl leukotriene 2 receptor in increased vascular 4. Yoshimoto T, Soberman RJ, Spur B, Austen KF (1988) Properties of highly purified permeability and in bleomycin-induced pulmonary fibrosis in mice. J Biol Chem leukotriene C4 synthase of guinea pig lung. J Clin Invest 81:866–871. 279:46129–46134. 5. Ago H, et al. (2007) Crystal structure of a human membrane protein involved in 27. Kim DC, et al. (2006) Cysteinyl leukotrienes regulate Th2 cell-dependent pulmonary cysteinyl leukotriene biosynthesis. Nature 448:609–612. inflammation. J Immunol 176:4440–4448. 6. Martinez Molina D, et al. (2007) Structural basis for synthesis of inflammatory medi- 28. Back M, et al. (2000) Antagonist resistant contractions of the porcine pulmonary artery ators by human leukotriene C4 synthase. Nature 448:613–616. by cysteinyl-leukotrienes. Eur J Pharmacol 401:381–388. 7. Anderson ME, Allison RD, Meister A (1982) Interconversion of leukotrienes catalyzed 29. Back M, et al. (2000) Prostacyclin modulation of contractions of the human pulmonary by purified gamma-glutamyl transpeptidase: Concomitant formation of leukotriene artery by cysteinyl-leukotrienes. Eur J Pharmacol 401:389–395. D4 and gamma-glutamyl amino acids. Proc Natl Acad Sci USA 79:1088–1091. 30. Lee TH, Austen KF, Corey EJ, Drazen JM (1984) Leukotriene E4-induced airway hyper- 8. Shi ZZ, Han B, Habib GM, Matzuk MM, Lieberman MW (2001) Disruption of gamma- responsiveness of guinea pig tracheal smooth muscle to histamine and evidence for glutamyl leukotrienase results in disruption of leukotriene D4 synthesis in vivo and three separate sulfidopeptide leukotriene receptors. Proc Natl Acad Sci USA 81:4922– attenuation of the acute inflammatory response. Mol Cell Biol 21:5389–5395. 4925. 9. Lee CW, Lewis RA, Corey EJ, Austen KF (1983) Conversion of leukotriene D4 to 31. Maekawa A, Kanaoka Y, Xing W, Austen KF (2008) Functional recognition of a distinct leukotriene E4 by a dipeptidase released from the specific granule of human polymor- receptor preferential for leukotriene E4 in mice lacking the cysteinyl leukotriene 1 and phonuclear leucocytes. Immunology 48:27–35. 2 receptors. Proc Natl Acad Sci USA 105:16695–16700. 10. Habib GM, Shi ZZ, Cuevas AA, Lieberman MW (2003) Identification of two additional 32. Raport CJ, et al. (1996) New members of the chemokine receptor gene family. J Leukoc members of the membrane-bound dipeptidase family. FASEB J 17:1313–1315. Biol 59:18–23. 11. Dahlen SE, Hedqvist P, Hammarstrom S, Samuelsson B (1980) Leukotrienes are potent 33. Blasius R, Weber RG, Lichter P, Ogilvie A (1998) A novel orphan G protein-coupled constrictors of human bronchi. Nature 288:484–486. receptor primarily expressed in the brain is localized on human chromosomal band 12. Weiss JW, et al. (1982) Bronchoconstrictor effects of leukotriene C in humans. Science 2q21. J Neurochem 70:1357–1365. 216:196–198. 34. Ciana P, et al. (2006) The orphan receptor GPR17 identified as a new dual uracil 13. Soter NA, Lewis RA, Corey EJ, Austen KF (1983) Local effects of synthetic leukotrienes nucleotides/cysteinyl-leukotrienes receptor. EMBO J 25:4615–4627. (LTC4, LTD4, LTE4, and LTB4) in human skin. J Invest Dermatol 80:115–119. 35. Lecca D, et al. (2008) The recently identified P2Y-like receptor GPR17 is a sensor of brain 14. Israel E, et al. (1990) The effects of a 5-lipoxygenase inhibitor on asthma induced by damage and a new target for brain repair. PLoS ONE 3:e3579. cold, dry air. N Engl J Med 323:1740–1744. 36. McGraw DW, et al. (2006) Airway smooth muscle prostaglandin-EP1 receptors directly 15. Manning PJ, et al. (1990) Inhibition of exercise-induced bronchoconstriction by MK- modulate beta2-adrenergic receptors within a unique heterodimeric complex. J Clin 571, a potent leukotriene D4-receptor antagonist. N Engl J Med 323:1736–1739. Invest 116:1400–1409. 16. Lynch KR, et al. (1999) Characterization of the human cysteinyl leukotriene CysLT1 37. Jiang Y, Borrelli LA, Kanaoka Y, Bacskai BJ, Boyce JA (2007) CysLT2 receptors interact receptor. Nature 399:789–793. with CysLT1 receptors and down-modulate cysteinyl leukotriene dependent mitogenic 17. Heise CE, et al. (2000) Characterization of the human cysteinyl leukotriene 2 receptor. responses of mast cells. Blood 110:3263–3270. J Biol Chem 275:30531–30536. 38. White JH, et al. (1998) Heterodimerization is required for the formation of a functional 18. Figueroa DJ, et al. (2001) Expression of the cysteinyl leukotriene 1 receptor in normal GABA(B) receptor. Nature 396:679–682. human lung and peripheral blood leukocytes. Am J Respir Crit Care Med 163:226–233. 39. Xu H, et al. (2004) Different functional roles of T1R subunits in the heteromeric taste 19. Figueroa DJ, et al. (2003) Expression of cysteinyl leukotriene synthetic and signalling receptors. Proc Natl Acad Sci USA 101:14258–14263. proteins in inflammatory cells in active seasonal allergic rhinitis. Clin Exp Allergy 40. Yanai K, Maeyama K, Nakahata N, Nakanishi H, Watanabe T (1992) Calcium mobili- 33:1380–1388. zation and its desensitization induced by endothelins and sarafotoxin in human 20. Maekawa A, Kanaoka Y, Lam BK, Austen KF (2001) Identification in mice of two astrocytoma cells (1321N1): Comparison of histamine-induced calcium mobilization. isoforms of the cysteinyl leukotriene 1 receptor that result from alternative splicing. Naunyn Schmiedebergs Arch Pharmacol 346:51–56. Proc Natl Acad Sci USA 98:2256–2261. 41. Mellor EA, Austen KF, Boyce JA (2002) Cysteinyl leukotrienes and uridine diphosphate 21. Ogasawara H, et al. (2002) Characterization of mouse cysteinyl leukotriene receptors induce cytokine generation by human mast cells through an interleukin 4-regulated mCysLT1 and mCysLT2: Differential pharmacological properties and tissue distribution. pathway that is inhibited by leukotriene receptor antagonists. J Exp Med 195:583–592. J Biol Chem 277:18763–18768. 42. Lefkowitz RJ, Cotecchia S, Samama P, Costa T (1993) Constitutive activity of receptors 22. Hui Y, et al. (2001) The murine cysteinyl leukotriene 2 (CysLT2) receptor. cDNA and coupled to guanine nucleotide regulatory proteins. Trends Pharmacol Sci 14:303–307. genomic cloning, alternative splicing, and in vitro characterization. J Biol Chem 43. Levoye A, et al. (2006) The orphan GPR50 receptor specifically inhibits MT1 melatonin 276:47489–47495. receptor function through heterodimerization. EMBO J 25:3012–3023. 23. Kanaoka Y, Maekawa A, Penrose JF, Austen KF, Lam BK (2001) Attenuated zymosan- 44. Hume DA, Gordon S (1983) Optimal conditions for proliferation of bone marrow- induced peritoneal vascular permeability and IgE-dependent passive cutaneous ana- derived mouse macrophages in culture: The roles of CSF-1, serum, Ca2ϩ, and adher- phylaxis in mice lacking leukotriene C4 synthase. J Biol Chem 276:22608–22613. ence. J Cell Physiol 117:189–194.

11690 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0905364106 Maekawa et al. Downloaded by guest on September 30, 2021