Chemoattractant Receptors Activate Distinct Pathways for Chemotaxis and Secretion

Human leukocyte chemoattractant receptors activate chemotactic and cytotoxic pathways to varying degrees and also activate different G-proteins depending on the receptor and the cell-type. To determine the relationship between G-protein usage and the biological and biochemical responses activated, receptors for the chemoattractants formyl peptides (FR), platelet-activating factor (PAFR), and leukotriene B4 (BLTR) were transfected into RBL-2H3 cells. Pertussis toxin (Ptx) served as a Gαiinhibitor. These receptors were chosen to represent the spectrum of Gi usage as Ptx had differential effects on their ability to induce calcium mobilization, phosphoinositide hydrolysis, and exocytosis with complete inhibition of all responses by FR, intermediate effects on BLTR, and little effect on PAFR. Ptx did not affect ligand-induced phosphorylation of PAFR and BLTR but inhibited phosphorylation of FR. In contrast, chemotaxis to formylmethionylleucylphenylalanine, leukotriene B4, and platelet-activating factor was completely blocked by Ptx. Wortmannin, a phosphotidylinositol 3-kinase inhibitor, also completely blocked ligand-induced chemotaxis by all receptors but did not affect calcium mobilization or phosphoinositide hydrolysis; however, it partially blocked the exocytosis response to formylmethionylleucylphenylalanine and the platelet-activating factor. Membrane ruffling and pseudopod extension via the BLTR was also completely inhibited by both Ptx and wortmannin. These data suggest that of the chemoattractant receptors studied, G-protein usage varies with FR being totally dependent on Gi, whereas BLTR and PAFR utilize both Gi and a Ptx-insensitive G-protein. Both Ptx-sensitive and -insensitive G-protein usage can mediate the activation of phospholipase C, mobilization of intracellular calcium, and exocytosis by chemoattractant receptors. Chemotaxis, however, had an absolute requirement for a Gi-mediated pathway.

Migration of leukocytes to sites of inflammation is mediated via the activation of G-protein-coupled chemoattractant receptors (1,2). Chemoattractants at low concentrations elicit shape change, pseudopod extension, and chemotaxis, and at higher doses; many of them also trigger degranulation and generation of superoxide anions (1,3). Pathways leading to these activities have been shown to have different dose requirements, kinetics and regulation (4,5), but the role of G-protein usage remains unknown.
Formylpeptides (fMLP), 1 platelet-activating factor (PAF), and leukotriene B 4 (LTB 4 ) are potent chemoattractants for neutrophils and to varying degrees also activate exocytosis and generation of superoxide anions (3,4,6). These activities are mediated through G-protein-coupled receptors (FR, PAFR, and BLTR) (2,7). G-protein usage of chemoattractant receptors was known to be different depending on cell types (1, 8 -10). Previous studies in RBL cells indicated that FR activated G i , whereas PAFR utilized both G i and a Ptx-insensitive G-protein to activate phosphoinositide hydrolysis, calcium mobilization, and exocytosis (11,12). However, the G-protein usage requirements for activating chemotaxis versus exocytosis in these cells were unknown. Therefore, we sought to determine the relationship between G-protein usage versus the subsequent responses activated by these receptors in a single cell line viz. RBL-2H3 cells. Epitope-tagged BLTR, FR, and PAFR were expressed in RBL-2H3 cells, and inhibitors of signaling through G i proteins (Ptx) and the PI3 kinase pathway (wortmannin) were used to determine the role of these pathways in pseudopod extension, chemotaxis, phospholipase C activation, calcium mobilization, and exocytosis. These data demonstrate distinct G-protein usage among chemoattractant receptors and suggest that a G imediated pathway, presumably involving ␤␥ and PI3 kinase, is required for motility-related functions. On the other hand, stimulation of phospholipase C activity, calcium mobilization, and exocytosis can be mediated through activation of a Ptxinsensitive G-protein as well as through G i . All other materials were obtained from sources previously described (12).

Materials
Construction of Epitope-tagged BLTR-Nucleotides encoding a nine amino acid hemagluttinin-epitope sequence (YPYDVPDYA) are in-* This work was supported in part by National Institutes of Health Grants AI-43184, AR-39162, DE-03738, HL-54166, HL-57629, and AI-38910. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  serted between the N-terminal initiator methionine and the second amino acid of human BLTR by polymerase chain reaction methods as described previously for other chemoattractant receptors with minor changes in polymerase chain reaction conditions (12,13). Because of high GC content of the DNA encoding the carboxyl-terminal third of the BLTR, cDNA polymerase chain reaction amplification occurred only in reactions containing 10% dimethyl sulfoxide (Me 2 SO). The integrity of the epitope tag as well as the rest of the molecule was verified by dideoxy sequencing after cloning into eukaryotic expression vector PRK-5. In BLTR a single nucleotide change resulted in a conservative substitution of amino acid 87 (Leu 3 Met). Competition ligand binding experiments and calcium mobilization studies indicated an affinity for LTB 4 (K d 4.4 Ϯ 0.63 nM) comparable to that seen in neutrophils and in other BLTR-transfected cell lines (7). In some experiments, i.e. calcium mobilization, RBL cells expressing a clone without this point mutation showed identical results.
Cell Culture, Transfection, Phosphoinositide Hydrolysis, Ca 2ϩ Mobilization, and Phosphorylation-We previously isolated RBL cells stably expressing FR and PAFR, and cells expressing BLTR were prepared following essentially the same procedures (11,12). Methods for RBL-2H3 culture, transfection and clonal selection, phosphoinositide hydrolysis, calcium mobilization, 32 P labeling, and immunoprecipitation of epitope-tagged receptors were essentially as described previously. (12) Chemotaxis-Migration of RBL stable transfectants was measured by a 48-well micro-chemotaxis chamber technique as described (14). Polycarbonate filters (25 ϫ 80 mm 8 pore size Neuroprobe) were coated with 50 g/ml of rat type I collagen (Collaborative Biomedicals) in HEPES-buffered RPMI 1640 medium for 2 h at 37°C. A dry coated filter was placed on a 48-well chamber containing different amounts of LTB 4 , fMLP, or PAF. RBL cells expressing the corresponding receptors (5 ϫ 10 4 /well) were added to the top wells, and the chamber was incubated at 37°C in 5% CO 2 for 4 h. Cells migrating to the underside of the filter were fixed using a leukostat staining kit. Cells from four high power (400X) fields representing at least two independent wells for each concentration of ligand were counted. Chemokinesis, measured with ligand at same concentration in both chambers, was less than 15% of the directed migration.
Micromanipulation-Cell micromanipulation was performed as described elsewhere (15). Briefly, the cells were added to a thermostated chamber, which was 2 mm thick and open from both sides to allow micromanipulation. The experiments were performed at 30°C 15 min after removing the cells from the culture dishes. To prevent nonspecific cell activation, the bottom of the chamber was coated with collagen. After adding the cells to the experimental chamber a single cell was chosen and held with a collagen-coated pipette. A local region of the membrane of the chosen cell was exposed for 5 min to a solution containing 10 nM LTB 4 , which was delivered from another pipette. The cell was observed with an inverted microscope Nikon 200 with 60ϫ oil immersion objective. The microscope images were recorded using a COHU CCD camera.

Functional Expression of LTB 4 Receptor in RBL-2H3 Cells-
Several chemoattractant receptors including FR and PAFR were previously expressed in RBL-2H3 cells for functional and regulation studies (11,12). Native RBL cells did not express any LTB 4 receptors as determined by lack of calcium mobilization to 1.0 M LTB 4 . To determine the signal transduction pathways activated by LTB 4 receptors, they were stably expressed in transfected clonal lines of RBL cells (BLTR). These receptors induced calcium mobilization, inositol phosphate production, exocytosis, and chemotaxis (see below) in RBL cells. Competition ligand binding experiments indicated an affinity for LTB 4 (K d 4.4 Ϯ 0.63 nM) comparable to that seen in neutrophils and in other BLTR-transfected cell lines (7) Phosphorylation of Chemoattractant Receptors-LTB 4 and phorbol 12-myristate 13-acetate stimulated the phosphorylation of BLTR by severalfold (Fig. 1A, lanes 1-3). As the cDNA is predicted to encode a 35-kDa protein, the two phosphoprotein bands at 40 and 60 kDa likely represent distinct glycosylation of the receptor. Previous studies showed that both FR and PAFR also displayed ligand-induced phosphorylation, whereas phosphorylation of PAFR but not FR was stimulated by phorbol 12-myristate 13-acetate (11,12). Pertussis toxin (16), an inhibitor of signaling through G i and G o family of G-proteins, did not have any effect on ligand or phorbol 12myristate 13-acetate-induced phosphorylation of BLTR (Fig.  1A) or PAFR (Fig. 1C). In contrast, fMLP-induced phosphorylation of FR was significantly inhibited by Ptx (Fig. 1B).
Distinct Pathways for Chemotaxis and Exocytosis-The ef-  4 , fMLP, or PAF as indicated. Total inositol phosphate generation was determined as described previously (12). The experiment was repeated three times with similar results. Ligand-induced secretion of ␤-hexoseaminidase (D, E, and F) was determined in cells left untreated (open circles), pretreated overnight with 100 ng/ml Ptx (closed squares), or pretreated for 3 h with 100 nM wortmannin (closed triangles) as described previously (12). The data are expressed as percentage of total ␤-hexoseaminidase present in cells. Data are mean Ϯ S.E. of a single representative of three experiments performed in triplicate. fects of Ptx and the PI3-kinase inhibitor wortmannin (17) on ligand-induced calcium mobilization (Fig. 2), phosphoinositide hydrolysis, exocytosis (Fig. 3), and chemotaxis (Fig. 4) were measured to determine the differential regulation of these responses. For BLTR, Ptx did not have a significant effect on the initial spike of calcium release but the length of the response was diminished (Fig. 2A). Whereas FR-induced calcium mobilization was completely inhibited by Ptx, it did not affect the initial spike or sustained calcium mobilization by PAFR (Fig. 2,  B and C). A similar differential effect of Ptx was also seen in phosphotidylinositol hydrolysis and exocytosis responses of these three receptors. Whereas the responses of FR were completely inhibited, BLTR showed some resistance to inhibition by Ptx and about 50% of the PAFR responses were Ptx-resistant (Fig. 3). Wortmannin had no effect on calcium mobilization by BLTR, FR, or PAFR (Fig. 2) and also did not inhibit phosphotidylinositol hydrolysis (Fig. 3). Wortmannin did not affect the exocytosis by BLTR but partially inhibited these responses to FR and to low concentrations of PAF. In contrast to these differences, chemotaxis to fMLP, PAF, and LTB4 was completely inhibited by Ptx (Fig. 4). Wortmannin also inhibited chemotaxis to a similar extent (ϳ90%) for all three receptors (Fig. 4).
Micromanipulation of RBL-BLTR Cells-Cytoskeletal alterations and pseudopod formation in the direction of the source of chemoattractant are essential events in leukocyte migration (18). To better define the role of distinct signaling events in leukocyte chemotaxis, ligand-induced membrane ruffling and pseudopod formation were studied using the micromanipulation assay (Fig. 5) (15, 19). A pipette held a single cell, and a region of its body was exposed to 10 nM LTB 4 blown continuously from a second pipette. The local exposure of the cell surface to chemoattractant provided a condition for the formation of a single pseudopod-like structure in suspension (Fig.  5a). Incubation of cells with Ptx (Fig. 5b) or wortmannin (Fig.  5c) resulted in complete inhibition of LTB 4 -induced pseudopod formation. However, Ptx-treated cells spread on culture dishes when replated (data not shown) indicating that they retain their ability to extend lamellipodia during physical attachment but were unable to respond to LTB 4 activation. They also displayed spontaneous activation and membrane ruffles (6 of 10 cells) at other sites on the cell body distinct from the area of stimulation by the second pipette. Cells treated with wortmannin also failed to extend a pseudopod-like structure. However, many of these cells showed a partial response and extended several membrane ruffles at the site of stimulation (7 of 10 cells) as shown in Fig. 5c. DISCUSSION Expression of chemoattractant receptors that activate different G-proteins in a single cell line allowed the determination of the role of specific G-proteins in activating distinct biological responses. The data show that chemoattractant receptors may activate phospholipase C, calcium mobilization, and exocytosis through multiple G-proteins, but ligand-stimulated pseudopod extension and chemotaxis require the activation of a Ptx-sensitive G-protein and PI3 kinase.
Comparison of biochemical responses activated by FR, BLTR, and PAFR, all of which activate G i , but couple to varying degrees to a Ptx-insensitive G-protein, allowed clear distinctions to be made for the role of G-proteins in different activities. Previous studies have indicated that receptor occupancy, but not signaling through G-proteins, is essential for receptor phosphorylation (20,21). Whereas the PAFR and BLTR were phosphorylated to similar levels in Ptx-treated cells, FR phosphorylation was inhibited. These data suggest that FR is likely phosphorylated by G-protein-coupled receptor kinases GRK2 or GRK3, because both require free ␤␥ for translocation to membrane and activation (20). Studies on in vitro phosphorylation of the C terminus of FR also support this contention (22). Lack of any affect, by Ptx, on ligand-induced receptor phosphorylation of PAFR and BLTR indicates that signal transduction through the Ptx-sensitive G-protein is not required for this event.
Previous studies in neutrophils and in transfected cell lines indicated that all FR responses are completely Ptx-sensitive (1,11,23). In contrast, both PAFR and BLTR can activate Ptxinsensitive G-proteins as well as G i (7,23,24). Our previous observations in RBL cells showed that calcium mobilization was only marginally affected and phosphotidylinositol hydrolysis was partially inhibited by Ptx, which suggests that PAFR activate a Ptx-insensitive G-protein in RBL cells (12). PAFR activated sufficient levels of Ptx-insensitive G-protein in these cells to cause exocytosis. The current studies show that like PAFR, BLTR expressed in RBL cells could utilize a Ptx-insensitive G-protein to activate phosphotidylinositol hydrolysis and mobilize intracellular calcium. However, the most significant finding of the current study is that despite activating different G-proteins all three receptors displayed an absolute requirement for G i to activate chemotaxis. This suggests that signal transduction pathways activated through Ptx-insensitive Gproteins are not sufficient, and the activation of a G i -mediated event is essential for chemotaxis. Recent studies also indicate that ␤␥-mediated responses are essential for cell migration (25)(26)(27). The data presented herein indicates that this ␤␥ must come from the activation of a Ptx-sensitive G-protein, presumably G i . The ␤␥ released from a Ptx-insensitive G-protein by the same chemoattractant receptor was not sufficient for chemotaxis.
The single cell manipulation assay has provided a means to correlate biochemical and motility-related cellular responses. The ability of RBL cells to extend a pseudopod similar to the one produced by leukocytes suggests molecular events in the migration of leukocytes and transfected RBL cells are related (19). Whereas cell lines of non-lymphoid origin like Chinese hamster ovary cells and HEK 293 cells also undergo chemotaxis (7,25), it is not known whether they use similar pathways or cytoskeletal processes as leukocytes for migration. In wortmannin-treated RBL cells, the low level of chemotactic responsiveness was associated with weak membrane ruffling and no pseudopod extension (Fig. 5). These results again underscore the dissociation of chemotaxis from the calcium release and exocytosis, which remain unaffected by wortmannin.
The unique biochemical pathways activated by the ␤␥ subunits of G i proteins that mediate chemotaxis are at present unknown. The ␤␥ induced PI3 kinase activity, and its related signaling pathway could be one of these events (28 -30). Other studies also indicated that GTP␥S could activate actin polymerization in soluble cell extracts of leukocytes in the presence of the low molecular weight G-protein CDC42 but not Rho or Rac (31,32). Therefore it is possible that ␤␥ dimers released from G i , but not the G q family of G-proteins, activate one or more low molecular weight G-proteins that may participate in leukocyte migration. In any case, the data clearly demonstrate that chemoattractant receptors may couple to individual or multiple G-proteins. To stimulate chemotaxis however, they must activate G i . Moreover, distinct signaling pathways among G-proteins are likely to determine other cellular responses initiated by chemoattractant receptors. These data may explain numerous observations in neutrophils and other cells that different chemoattractant receptors display different abilities to trigger nonmotility-related functions (i.e. respiratory burst, mitogenesis) despite exhibiting similar chemotactic responses (1,23). The availability of a model system in which migration responses may be measured both at population and single cell levels will allow further investigation of the divergent pathways for motility and cytotoxic functions.