Identification of the Intracellular Region of the Leukotriene B4 Receptor Type 1 That Is Specifically Involved in Gi Activation*

Many G-protein-coupled receptors can activate more than one G-protein subfamily member. Leukotriene B4 receptor type 1 (BLT1) is a high affinity G-protein-coupled receptors for leukotriene B4 functioning in host defense, inflammation, and immunity. Previous studies have shown that BLT1 utilizes different G-proteins (the Gi family and G16 G-proteins) in mediating diverse cellular events and that truncation of the cytoplasmic tail of BLT1 does not impair activation of Gi and G16 proteins. To determine responsive regions of BLT1 for G-protein coupling, we performed an extensive mutagenesis study of its intracellular loops. Three intracellular loops (i1, i2, and i3) of BLT1 were found to be important for both Gi and G16 coupling, as judged by Gi-dependent guanosine 5′-(γ-thio) triphosphate (GTPγS) binding and G16-dependent inositol phosphate accumulation assays. The i3-1 mutant, with a mutation at the i3 amino terminus, exhibited greatly reduced GTPγS binding but intact inositol phosphate accumulation triggered by leukotriene B4 stimulation. These results suggest that the i3-1 region is required only for Gi activation. Moreover, in the i3-1 mutant, the deficiency in Gi activation was accompanied by a loss of the high affinity leukotriene B4 binding state seen with the wild type receptor. A three-dimensional model of BLT1 constructed based on the structure of bovine rhodopsin suggests that the i3-1 region may consist of the cytoplasmic end of the transmembrane helix V, which protrudes the helix into the cytoplasm. From mutational studies and three-dimensional modeling, we propose that the extended cytoplasmic helix connected to the transmembrane helix V of BLT1 might be a key region for selective activation of Gi proteins.

sional model of BLT1 suggested that the i3-1 region might localize to the cytoplasmic end of the transmembrane helix V. This region is suggested to form an extended helical structure proximate to the membrane, which projects toward the cytoplasm in order to interact with G i .
Cell Culture, Transfection, and Flow Cytometry-Human embryonic kidney 293 (HEK293) and COS-7 cells were maintained in Dulbecco's modified Eagle's medium (Sigma), and Chinese hamster ovary (CHO) cells were maintained in Ham's F-12 medium (Sigma), each containing 10% fatal calf serum, 100 IU/ml penicillin and 100 g/ml streptomycin at 37°C in 5% CO 2 . These cells were transfected with a wild type (WT) or mutated BLT1 expression vector using Lipofectamine PLUS reagent (Invitrogen) according to the manufacturer's protocol. For staining, cells were incubated with anti-HA antibody (3F10; 2 g/ml) in phosphate-buffered saline containing 2% goat serum at 4°C for 30 min, followed by staining with 10 g/ml Alexa-Fluor 488-conjugated anti-rat IgG antibody (Molecular Probes, Inc., Eugene, OR) at 4°C for 30 min. Cells highly expressing WT and mutant BLT1 were collected as polyclonal populations by cell sorting using an EPICS ALTRA (Beckman Coulter, Miami, FL) and maintained in 0.3 mg/ml G418 (Wako, Osaka, Japan). Monoclonal cell lines were established by the limiting dilution method. EPICS XL (Beckman Coulter) was used for flow cytometry.
Membrane Preparation and Western Blotting-Two or 3 days after transfection or after cells were grown to subconfluence, cells were harvested with phosphate-buffered saline containing 2 mM EDTA. In some experiments, cells were treated with 100 ng/ml PTX (List Biological Laboratories, Campbell, CA) for 12 h. Cells were disrupted in ice-cold sonication buffer (20 mM Tris-HCl, pH 7.5, 0.25 M sucrose, 10 mM MgCl 2 , 2 mM EDTA) plus a protease inhibitor mixture (Sigma) and centrifuged at 10,000 ϫ g for 10 min at 4°C. Supernatants were further centrifuged at 100,000 ϫ g for 1 h at 4°C, and the resulting pellets were resuspended in sonication buffer and used as the membrane preparation. The protein concentration in each sample was determined by the Bradford method (Protein Assay Kit) (Bio-Rad). Protein samples were separated on 12% polyacrylamide-SDS gels and transferred to a nitrocellulose membrane. After blocking with 5% skim milk in TBS-T (20 mM Tris-buffered saline, pH 7.4, 0.1% Tween 20), blots were probed with the primary antibody for 1 h. The membrane was washed with TBS-T and incubated with horseradish peroxidase-conjugated antibody for 1 h, and the signal was visualized using an ECL chemiluminescence detection system (Amersham Biosciences).
Ligand Binding Assay-The membrane preparation (10 or 5 g of protein) was incubated in 100 l of LTB 4 binding buffer (50 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , and 10 mM NaCl) containing 0.05% bovine serum albumin (BSA; fatty acid-free) and various concentrations of [ 3 H]LTB 4 . To determine nonspecific binding, concentrations of unlabeled LTB 4 at least 1,000 times higher than labeled were used. Mixtures were incubated for 1 h at room temperature, followed by rapid filtration through GF/C filters and washing with ϳ5 ml of LTB 4 binding buffer. Radioactivity was measured using a Top Count scintillation counter (Packard). GTP␥S Binding Assay-The membrane preparation (10 g of protein) was incubated in 100 l of GTP␥S binding buffer (20 mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, 5 M GDP, and 0.1% BSA) containing 0.5 nM [ 35 S]GTP␥S with or without various concentrations of LTB 4 for 30 min at 30°C. To determine nonspecific binding, unlabeled GTP␥S was added to the binding mixture to a final concentration of 10 M. Bound [ 35 S]GTP␥S was separated from free by rapid filtration through GF/C filters and washed with ϳ2 ml of ice-cold TMN buffer (10 mM Tris-HCl, pH 7.5, 25 mM MgCl 2 , and 100 mM NaCl). Radioactivity was measured as described above.
cAMP Assay-CHO stable cells were plated at 3.2 ϫ 10 4 cells/ well in a 96-well plate. On the following day, cells were washed with buffer A (Hanks' balanced salt solution; Invitrogen) containing 0.1% BSA), pretreated with buffer A containing 0.5 M 3-isobutyl-1-methylxanthine for 15 min at room temperature, and then stimulated with a ligand solution (Buffer A, 0.5 M 3-isobutyl-1-methylxanthine and 50 M forskolin) containing various concentrations of LTB 4 for 30 min at room temperature. The reaction was terminated by the addition of lysis buffer (Hanks' buffered salt solution containing 1% Tween 20). cAMP concentrations in the lysate were determined by an Alpha-Screen cAMP assay kit (PerkinElmer Life Sciences).
Inositol Phosphate (IP) Accumulation Assay-Cells were seeded at 1 ϫ 10 5 cells/well in a 12-well plate and transfected with BLT1 and/or G-protein ␣ subunit on the following day. Twenty-four hours later, culture medium was replaced with fresh medium containing 1 Ci/ml myo-[ 3 H]inositol (Amersham Biosciences), followed by incubation for another 18 -24 h. Cells were washed twice and preincubated for 10 min in Dulbecco's modified Eagle's medium containing 0.1% BSA and 10 mM LiCl, followed by stimulation with various concentrations of LTB 4 or vehicle for 30 min. Stimulation was terminated by replacing the medium with ice-cold 5% HClO 4 and incubating on ice for 2 h. Accumulated IP was purified by ion exchange chromatography (AG1-X8 resin; Bio-Rad), and radioactivity was measured by scintillation counting (20,21).
Chemotaxis Assay-Polycarbonate filters with 8-m pores (Neuro Probe Inc., Gaithersburg, MD) were coated with 10 g/ml fibronectin in phosphate-buffered saline for 60 min. The filter was placed in a 96-well Boyden chamber (Neuro Probe) containing LTB 4 solution in the lower wells. CHO cells stably expressing WT or mutated BLT1 were placed at 6 ϫ 10 4 cells/ well in the upper wells. After incubating at 37°C in 5% CO 2 for 4 h, cells on the filter were fixed with methanol and stained using a Diff-Quik staining kit (International Reagents Corp., Kobe, Japan). The upper side of the filter was scraped free of cells. Then the number of cells migrating out to the lower side was determined by measuring optical densities at 595 nm using a 96-well microplate reader (Bio-Rad).
Statistical Analysis-Data were analyzed for statistical significance using analysis of variance using Prism 4 software (GraphPad Software, Inc., San Diego, CA). Differences were considered significant at p Ͻ 0.05, 0.01, or 0.001, as indicated.
BLT1 Structure Modeling Based on Rhodopsin Structure-A human BLT1 atomic model was constructed based on that of bovine rhodopsin with homology modeling as a rhodopsin subfamily member of GPCRs (22). The amino acid sequences of seven transmembrane proteins were aligned according to the transmembrane sequence alignment in the GPCR data base (23). O software (24) was used to replace amino acid residues according to the sequence alignment from bovine rhodopsin to BLT1 on the atomic model of the bovine rhodopsin (Protein Data Bank code 1F88) (22). Finally, structural idealization with molecular dynamics using CNS software (25) was applied to the BLT1 model. The BLT1 model constructed was checked by Procheck software. All figures were prepared using PyMOL (DeLano Scientific LLC, South San Francisco, CA).

RESULTS
Mutagenesis-To determine regions of human BLT1 responsible for G i and G q protein activation, we performed extensive mutagenesis focusing on the intracellular loops. Three to four tandem amino acids in three regions (i1, i2, and i3) were replaced with alanines by PCR using HA-tagged human BLT1 as a template. We initially constructed 13 mutants designated i1-1 to i1-3, i2-1 to i2-5, and i3-1 to i3-5 ( Fig. 1, A and B). Expression levels of these receptors on the cell surface were semiquantitatively analyzed by flow cytometry as follows. Cells transfected with WT BLT1 or mutants were stained with anti-HA antibody specific to the amino-terminal HA tag and Alexa-Fluor 488-conjugated secondary antibody. WT BLT1 and mutant receptors were expressed on the cell surface, judging from the increased means of fluorescence intensity of the WT BLT1 and mutants compared with the mock transfectant ( Fig. 1C).
BLT1/G␣ Coupling-To determine which G i protein subtypes couple to BLT1, we examined LTB 4 -induced GDP-GTP exchange of G␣ by a GTP␥S binding assay using membrane fractions of HEK293 cells transiently expressing BLT1 and various G-protein ␣ subunits. Transfection of BLT1 without G-proteins caused only a slight increase above basal levels in GTP␥S binding upon application of LTB 4 ( Fig. 2A). When BLT1 was co-transfected with members of the G i family of G-protein ␣ subunits, GTP␥S binding activities were enhanced, whereas co-transfection of G␣ 16 did not show such an effect (data not shown). Co-transfected G␣ i2 showed a significant increase in LTB 4 -induced GTP␥S binding. Moreover, LTB 4induced GTP␥S binding was completely abolished by pretreating the cells with PTX prior to membrane preparation (see Fig.  4A). These results suggest that measuring LTB 4 -induced GTP␥S binding is an appropriate way to evaluate activation of PTX-sensitive G i proteins by BLT1.
The alanine mutants described above were expressed in HEK293 cells along with G␣ i2 and evaluated for GDP-GTP exchange of G␣ by the GTP␥S binding assay. All mutants showed lower GTP␥S binding than WT, but the i1-3, i2-4, i3-1, and i3-2 mutants exhibited significantly reduced GTP␥S binding (Fig. 2B). These results indicate that all three regions, i1, i2, and i3, are crucial for G i coupling.
Next, to determine whether activation of the G q family of G-proteins by mutant receptors is altered, we examined LTB 4induced total IP accumulation in COS-7 cells expressing both BLT1 and various G␣ subunits. As reported by Gaudreau et al. (20), COS-7 cells expressing both BLT1 and G␣ 16 show significantly higher IP accumulation than cells expressing BLT1 alone (Fig. 3A). BLT1 mutants were co-expressed in COS-7 cells with G␣ 16 and tested for their ability to activate G␣ 16 by the IP accumulation assay Among the mutants that showed impaired GTP␥S incorporation, only the i3-1 mutant displayed normal IP accumulation after ligand application (Fig. 3B). The i3-4 and i3-5 mutants exhibited higher IP accumulation than WT BLT1.
Since these results suggested that the i3-1 region is responsible for G i2 activation, we constructed another mutant in the more membrane-proximal region of the i3 loop (i3-0; Fig. 1B). As with the i3-1 mutant, the i3-0 mutant BLT1 showed normal IP accumulation but significantly reduced GTP␥S binding after LTB 4 application (Figs. 2B and 3B). These results show that the amino terminus of the i3 loop is important for G i2 activation.
Effects of the i3-1 Mutation on G i2 and G 16 Coupling-LTB 4 stimulated an increase in GTP␥S binding in a dose-dependent manner in HEK293 cells transfected with WT BLT1 and G␣ i2 , but no significant increases were seen with the i3-1 mutant (Fig.  4A). Increases in GTP␥S binding mediated by LTB 4 completely disappeared following pretreatment of cells with 100 ng/ml PTX. These results suggest that the i3-1 region functions in G i2 activation. However, in LTB 4 -induced IP production, dose-re-sponse curves were similar between WT BLT1 and the i3-1 mutant in the presence of G␣ 16 in COS-7 cells, and PTX pretreatment did not affect IP accumulation (Fig. 4B). These results further support the idea that the i3-1 region of BLT1 leads to impairment in PTX-sensitive G i2 activation without affecting PTX-resistant G 16 signaling.
To gain further insight into the role of the i3-1 region in signal transduction, we established polyclonal CHO cells stably FIGURE 1. Targeting of potential G-protein coupling domains of the human BLT1 using alanine substitution. A and B, a secondary structure diagram of the proposed topology of BLT1 (A) and the amino acid sequences of the three intracellular regions (i1, i2, and i3) of WT BLT1 and mutants (B) are shown. Three to four tandem amino acids in the intracellular regions of BLT1 were replaced with alanines. We initially constructed 13 mutants designated i1-1 to -3, i2-1 to -5, and i3-1 to -5. After initial screening, we also constructed another mutant and designated it as i3-0. C, expression levels of the receptors on the cell surface were semiquantitatively analyzed by flow cytometry, shown as the means of fluorescence intensity. HEK293 or COS-7 cells transiently transfected with WT BLT1 and mutants were stained with anti-HA antibody specific to the amino-terminal HA tag, followed by staining with Alexa-Fluor 488-conjugated secondary antibody. Data shown are representative of three separate experiments with similar results.
expressing WT BLT1 or the i3-1 mutant. Cells highly expressing WT or the i3-1 mutant receptor were collected by cell sorting. The expression level of i3-1 mutant was similar to that of WT BLT1, as judged from flow cytometric and Western blotting analyses (Fig. 5, A and B). In CHO stable cells, the maximum response of GTP␥S binding with the i3-1 mutant was only one quarter (23%) that of WT BLT1, although WT and the i3-1 mutant showed LTB 4 -induced GTP␥S binding in a dose-dependent manner (Fig. 5C). The dose-response curve of LTB 4 with the i3-1 mutant was shifted to the right from that seen with WT BLT1 (EC 50 ; 34 versus 5.0 nM) (Fig. 5, C and D). IC 50 values of LTB 4 -induced adenylyl cyclase inhibition in cells expressing the i3-1 mutant and WT BLT1 receptors were 6.0 and 0.25 nM, respectively (Fig. 5E). It has been reported that GPCR-dependent chemotaxis requires activation of G i and a resulting release of G␤␥ dimers (26) and that LTB 4 -induced chemotaxis is completely PTX-sensitive (7). Although both WT and the i3-1 mutant BLT1 cells showed chemotactic responses to LTB 4 with bell-shaped dose-response curves, cells expressing the i3-1 mutant showed impaired migration to low concentrations of LTB 4 (Fig. 5F). However, IP accumulation in WT and the i3-1 mutant BLT1 cells transfected with G␣ 16 were comparable (Fig.  5G). These results confirm the presence of both impaired G i activation and normal G 16 activation by the i3-1 mutant in the same CHO stable cells.

Effects of the i3-1 Mutation on [ 3 H]LTB 4 Binding-To examine the effect of the i3-1 mutation on ligand binding, we performed [ 3 H]LTB 4 binding experiments.
Membrane preparations (10 g of protein) from HEK293 cells transiently expressing WT BLT1 or the i3-1 mutant were incubated with various concentrations of [ 3 H]LTB 4 (Fig. 6A). From Scatchard analyses, WT BLT1 showed high affinity LTB 4 binding with a K d value of 3.5 nM and B max value of 2.1 pmol/mg protein (Fig.  6B). In contrast, the i3-1 mutant showed low affinity LTB 4 binding with a K d of 6.4 nM and a B max value of 1.8 pmol/mg protein. To analyze whether G i coupling is required for high affinity LTB 4 binding, we examined LTB 4 binding of membrane preparations of PTX-treated cells. WT BLT1-expressing cells pretreated with PTX displayed only low affinity LTB 4 binding similar to the i3-1 mutant (Fig. 6B).
Experiments Using Monoclonal Cell Lines-In the above experiments, the B max of HEK293 cells transiently transfected with the i3-1 mutant was lower than that in cells transfected   with WT BLT1 (Fig. 6). To confirm that the defects in G i signaling in the i3-1 mutant were not due to decreased expression of the mutant receptor, we established monoclonal cell lines of WT BLT1 (lines 10 and 13) and the i3-1 mutant (lines 30 and 46) and determined B max values and signaling (Fig. 7). As shown in Fig. 7, A and B, the LTB 4 binding assay revealed that B max values for lines 10, 13, 30, and 46 of 8.1, 4.9, 7.6, and 14.4 pmol/mg protein, respectively. The K d values were 1.9, 1.5, 7.4, and 8.3 nM, respectively. Dose-response curves of the i3-1 mutants (lines 30 and 46) were shifted to the right by 1 order of magnitude compared with WT BLT1 (lines 10 and 13; Fig. 7C). Although the i3-1 mutant 46 was expressed higher than the WT BLT1 (lines 10 and 13), the IC 50 values of G i -dependent adenylyl cyclase inhibition in the i3-1 mutant were higher than that in WT BLT1 (lines 10 and 13). However, a G 16dependent IP accumulation was similar among cells expressing either WT or mutant receptors (Fig.  7D). These results further confirm the impaired G i activation and normal G 16 activation.

DISCUSSION
Using mutagenesis studies, we show that three intracellular loops of BLT1 function in G-protein activation (Figs. 2B and 3B). Among these regions, we identified the i3-1 region as a selective activation site for G i protein. The i3-1 mutant showed impaired G i activation but normal G 16 activation (Figs. 4 and  7). The i3-1 region is located at the amino terminus of the third intracellular (i3) loop of BLT1. Previous reports suggest that the i3 region of several GPCRs functions in G-protein activation (27)(28)(29)(30). In contrast, our findings show that the i3-1 (possibly i3-0) region of BLT1 is specifically required for G i activation.
We used different cell lines for the different assays in the course of screening mutant receptors. We used HEK293 cells for GTP␥S binding assay, because HEK293 cells FIGURE 5. LTB 4 -induced G i signaling is impaired, but G 16 signaling is normal in polyclonal CHO cells stably expressing the i3-1 mutant. A, polyclonal CHO cells stably expressing WT BLT1 or the i3-1 mutant were stained with an anti-HA antibody and an Alexa-Fluor-488-conjugated secondary antibody and analyzed by flow cytometry. B, membrane preparations of polyclonal CHO cells stably expressing WT BLT1 or the i3-1 mutant were examined for receptor expression by Western blotting using an anti-HA antibody. C and D, a GTP␥S binding assay was performed using membrane preparations of CHO cells stably expressing WT BLT1 or the i3-1 mutant (C, mean Ϯ S.D., n ϭ 3). The same data of C are represented as a percentage of the maximum responses in D. E, CHO cells stably expressing WT BLT1 or the i3-1 mutant were stimulated with 50 M forskolin and various concentrations of LTB 4 , and cAMP concentrations were determined (mean Ϯ S.D., n ϭ 4). F, dose dependence of LTB 4 -induced chemotaxis was measured in CHO cells stably expressing WT or the i3-1 mutant (mean Ϯ S.D., n ϭ 4). G, IP accumulation assay was performed using CHO stable cells transiently expressing G␣ 16 protein (mean Ϯ S.D., n ϭ 3). Data shown are representative of at least three separate experiments with similar results (G was repeated twice). transiently expressing WT BLT1 showed higher LTB 4 -induced GTP␥S binding than COS-7 cells. We used COS-7 cells for the IP accumulation assay, since Gaudreu et al. (20) had reported that COS-7 cells transfected with BLT1 and G␣ 16 showed LTB 4 -induced IP accumulation. CHO stable cells are commonly used for GPCR research, because they are effective in examining intracellular signaling and cellular events, such as calcium mobilization, cAMP assay, and chemotaxis. Using these cells, we confirmed that the i3-1 mutant BLT1 exhibited impaired G i activation but intact G 16 activation. LTB 4 -dependent G i2 activation as assessed by GTP␥S binding assay was impaired in the i3-1 mutant in HEK293 and CHO cells (Figs. 4A and 5, C and D). LTB 4 -dependent inhibition of adenylyl cyclase and chemotaxis was also impaired in the i3-1 mutant in CHO stable cells (Fig. 5, E and F). However, LTB 4dependent IP accumulation in COS-7 (Fig. 4B), HEK293 (data not shown), and CHO stable cells (Fig. 7G) expressing either the WT or i3-1 mutant was similar. These results show that impaired G i activation by the i3-1 mutant is not cell type-dependent.
We also isolated several monoclonal CHO cells expressing the WT or the i3-1 mutant BLT1 and performed similar experiments. Dose-response curves of LTB 4 -induced adenylyl cyclase inhibition were shifted to the right in both high and low expresser of the i3-1 mutant compared with WT BLT1 (Fig.  7C). However, IP accumulations remained intact in the clones (Fig. 7D). We conclude that impaired G i activation by the i3-1 mutant is not dependent on the expression level of the receptor.
The binding studies showed that WT BLT1 exhibited high affinity for LTB 4 (Figs. 6 and 7, A and B). The i3-1 mutants and PTX-pretreated WT BLT1 showed only a low affinity binding state. In agreement with this data, Igarashi et al. reported that specific LTB 4 binding was increased by reconstitution with G␣ proteins (G␣ i/o ) (31). Among them, reconstitution with G i2 increased LTB 4 binding most efficiently. Banères et al. (32) showed that in the presence of G i2 proteins, BLT1 showed a high affinity binding state, using recombinant proteins produced in E. coli. In addition, Masuda et al. (33) reported that G i -coupled BLT1 exhibited high affinity LTB 4 binding, whereas G 16 -coupled BLT1 exhibited low affinity binding in a reconstitution system on baculovirus. These results and ours strongly suggest that G i coupling enables BLT1 to maintain a high affinity state, whereas BLT1 coupled to G 16 mediates a signal but exhibits only a low affinity state. Pharmacological studies suggest the presence of high and low affinity binding sites for LTB 4 in spleen and granulocytes (34 -36). Characterization of this low affinity binding site cannot be explained by the presence of BLT2, which constitutes a low affinity LTB 4 receptor (8 -11), since human granulocytes do not express BLT2 (data not shown). Therefore, BLT1 might be responsible for low affinity binding of LTB 4 in neutrophils. The physiological relevance of the presence of low and high affinity binding states in BLT1 remains to be determined.  . The i3 region is colored (red, i3-0; orange, i3-1; pink, i3-2 to -5). The three-dimensional model suggested that the i3-1 region might localize to the cytoplasmic end of helix V and that side chains of Asp, Ile, and Arg residues in the i3-1 region might interact with G i proteins. Banères and Parello (32) clearly showed that BLT1 was in a low affinity binding state in a reconstituted system completely devoid of any G-protein, whereas the addition of G-protein (G␣ i2 ␤ 1 ␥ 2 ) binding induced a high affinity binding state. They also showed that homodimerized BLT1 associates with G-protein and forms BLT1 2 ⅐G-protein pentamers by using chemical cross-linking. Judging from the profiles in Western blotting analysis (Fig. 5B), it seems that the i3-1 mutant might be able to form a dimer, although more detailed studies will be required to confirm this (work in progress).
To gain insight into the molecular mechanism underlying how the i3-1 region of BLT1 recognizes G i protein, we constructed a BLT1 three-dimensional model based on the structure of bovine rhodopsin (Fig. 8) (19,22). This model suggests that the i3-1 region might localize at the cytoplasmic end of helix V, not in the i3 loop of conventional seven-transmembrane models based on bacteriorhodopsin (Fig. 1A). The crystal structure of bovine rhodopsin shows that the cytoplasmic ends of helix V and VI are prominently extended below the plasma membrane in two different crystal forms, and these extended regions are highly flexible according to crystallographic temperature factors at more than 100 Å 2 (22,37). In rhodopsin, it has also been suggested that this region of helix V and the membrane-proximal region of helix VI function in activation of transducin via a carboxyl-terminal region of G␣ t (27,38,39).
We propose here that the BLT1 i3-1 region interacting with G i protein may consist of the end of helix V in the model. Notably, alanine substitution in the i3-0 region, which contains three amino acid residues upstream of the i3-1 region, resulted in impaired G i activation but intact G 16 activation, as was seen with the i3-1 mutant (Figs. 2B and 3B). NMR analysis of the undecapeptide corresponding to C terminus of the i3 region of the cannabinoid receptor 1 has recently suggested formation of an ␣-helix in the interaction with G i1 protein, an interaction essential for activation of GTPase activity by the receptor (28). Based on the structure of G-protein-bound mastoparan (40) and artificial GEF peptide-G␣ i1 complex (41,42), short amphiphilic ␣-helices might constitute a common mechanism of protein-protein interaction by which G␣ subunits adopt a conformation suitable for GDP-GTP exchange. In this context, the i3-1 region of BLT1 contains bulky hydrophobic (Ile) and charged (Asp and Arg) residues with an ␣-helix-destabilizing Gly residue interacting specifically with spherical and charged residues suitable for the proposed coupling to G-protein (41). In addition, the i3-0 region also contains two polar (Ser) and bulky hydrophobic (Tyr) residues.
The homology modeling of GPCRs based on bovine rhodopsin have been accepted for predicting and gaining insights into molecular basis of GPCR structure activity, although current homology modeling of GPCRs is not final or best structural solution. Combination of deduced model and functional experimental analysis have elucidated the molecular bases of GPCR structure activity, such as ligand binding and function of helix VIII (18,19,43). The i3 regions of BLT1 and bovine rhodopsin can be aligned automatically using the T-Coffee program (44). The i3 region of BLT1 is shorter than those of bovine rhodopsin. In the nature of homology modeling, the model of deletion region is easier to construct than that of the insertion region. In this context, the three-dimensional model of the i3 region of BLT1 has confidence to some extent, although more detailed experiments are required to prove the above hypothesis in the future.
In summary, our work analyzes specific interactions between GPCRs and G-proteins using BLT1 as a model. These data suggest that BLT1 couples to both G i and G 16 proteins using overlapping intracellular region but that the i3-1 region functions only in G i coupling, not in G 16 coupling. The i3-1 mutant prefers G 16 to G i and maintains a low affinity LTB 4 binding state due to deficiencies in G i coupling. This mutant receptor would be useful in determining specific signal transduction pathways mediated by G 16 proteins via BLT1 and the biological significance of low and high affinity LTB 4 binding to BLT1. We propose that the helical structure of GPCRs acts as a switch for activation of G i proteins. Examination of the roles of the cytoplasmic end of helix V of GPCRs other than BLT1 in determining specificity for coupling to G-proteins should be undertaken.