Differential Regulation of Formyl Peptide and Platelet-activating Factor Receptors

Formylated peptides (e.g. n-formyl-Met-Leu-Phe (fMLP)) and platelet-activating factor (PAF) mediate chemotactic and cytotoxic responses in leukocytes through receptors coupled to G proteins that activate phospholipase C (PLC). In RBL-2H3 cells, fMLP utilizes a pertussis toxin (ptx)-sensitive G protein to activate PLC, whereas PAF utilizes a ptx-insensitive G protein. Here we demonstrate that fMLP, but not PAF, enhanced intracellular cAMP levels via a ptx-sensitive mechanism. Protein kinase A (PKA) inhibition by H-89 enhanced inositol phosphate formation stimulated by fMLP but not PAF. Furthermore, a membrane-permeable cAMP analog 8-(4-chlorophenylthio)-cAMP (cpt-cAMP) inhibited phosphoinositide hydrolysis and secretion stimulated by fMLP but not PAF. Both cpt-cAMP and fMLP stimulated PLCβ3phosphorylation in intact RBL cells. The purified catalytic subunit of PKA phosphorylated PLCβ3 immunoprecipitated from RBL cell lysate. Pretreatment of intact cells with cpt-cAMP and fMLP, but not PAF, resulted in an inhibition of subsequent PLCβ3phosphorylation by PKA in vitro. These data demonstrate that fMLP receptor, which couples to a ptx-sensitive G protein, activates both PLC and cAMP production. The resulting PKA activation phosphorylates PLCβ3 and appears to block the ability of Gβγ to activate PLC. Thus, both fMLP and PAF generate stimulatory signals for PLCβ3, but only fMLP produces a PKA-dependent inhibitory signal. This suggests a novel mechanism for the bidirectional regulation of receptors which activate PLC by ptx-sensitive G proteins.

Many extracellular signaling molecules including neurotransmitters, hormones, and chemoattractants mediate their biological responses via the activation of G protein-coupled receptors through stimulation of adenylyl cyclase, phospholipase C (PLC), 1 and ion channels (1). Continuous agonist stim-ulation leads to waning of the biological response by a process termed desensitization (2). Receptor phosphorylation by G protein-coupled receptor kinases as well as by second messengeractivated kinases, such as protein kinase A (PKA) and protein kinase C, are important in receptor desensitization (2,3). Additionally, chemoattractant responses are regulated at the level of PLC (4). Chemoattractants such as formyl peptides (e.g., n-formyl-Met-Leu-Phe (fMLP)), the anaphylatoxin C5a, and interleukin-8 activate PLC␤ by releasing ␤␥ subunits (G ␤␥ ) of a pertussis toxin (ptx)-sensitive G protein, likely G i␣2 (5,6). The chemoattractant receptor for PAF couples to both ptxsensitive and -insensitive G proteins. The latter, G ␣q , likely activates PLC␤ by a different mechanism (7,8).
fMLP and PAF receptors have been shown to display differences in susceptibility to desensitization (9). This laboratory has developed methodology to study the regulation of chemoattractant receptors in the leukocyte-like RBL-2H3 (RBL) cell line (10 -13). Using this model, it was found that a membrane permeable cAMP analog caused inhibition of Ca 2ϩ mobilization stimulated by fMLP but not PAF (4). This difference could be potentially related to the distinct G protein usage of these receptors. The present study characterizes this observation and demonstrates that fMLP causes an increase in cAMP production both in neutrophils and transfected RBL cells and that the resulting PKA activation leads to inhibition of a biological response, secretion. In addition, the data show that PLC␤ 3 is a direct substrate for phosphorylation by PKA and that fMLP receptor-stimulated PLC␤ 3 phosphorylation by PKA provides a previously unrecognized mechanism for the counter regulation of cellular activation. (24.4 Ci/mmol), [␥-32 P]ATP (6000 Ci/mmol), and [␥-32 P]GTP (6000 Ci/mmol) were purchased from NEN. fMLP, PAF (1-O-hexadecyl-2-acetyl-sn-glycero-3-phosphocholine), and the protein kinase A inhibitor H-89 were purchased from Calbiochem. Recombinant C5a and cpt-cAMP were purchased from Sigma. Affinity-purified polyclonal antibody against PLC␤ 3 was obtained from Santa Cruz Biotechnology. Pertussis toxin and all tissue culture reagents were purchased from Life Technologies, Inc. The catalytic subunit of PKA was obtained from Promega. The Radioimmunoassay kit for cAMP measurement was purchased from Amersham Corp.

EXPERIMENTAL PROCEDURES
Cell Culture and Assays-RBL cells stably expressing epitope-tagged fMLP and PAF receptors were used throughout this study (4,10,11,14). Cell culture, neutrophil purification, GTPase activity, phosphoinositide hydrolysis, Ca 2ϩ mobilization, secretion, and in vivo PLC␤ 3 phosphorylation were performed exactly as described by us previously (4,9,10). HL-60 cells were differentiated with 1.3% dimethyl sulfoxide for 5-6 days. In vitro phosphorylation of PLC␤ 3 was performed essentially as described for PLC␤ 2 (15). Briefly, cells (5 ϫ 10 6 ) were treated with various agents or buffer for 5 min in the presence of isobutyl-methylxanthine (IBMX) (400 M), lysed, and immunoprecipitated with anti-PLC␤ 3 antibody. The immune complex was washed with a buffer containing 40 mM Tris-HCl, (pH 7.4), MgOAc (20 mM), ATP (20 M) and resuspended in the same buffer (50 l) supplemented with 2 Ci of [␥ 32 P-ATP]. Phosphorylation was started via the addition of 1 l of purified PKA. The reaction was stopped by adding 1 ml of ice-cold buffer, and the immune complex was washed three times. The proteins were resolved on a 6% SDS-polyacrylamide gel electrophoresis and visualized by autoradiography. For cAMP assay, cells (0.5-1.0 ϫ 10 6 / ml) were preincubated for 10 min with 400 M IBMX and stimulated with fMLP, C5a, or PAF. The reactions were quenched, and cAMP measurements were carried out as described in the cAMP kit manual.

Differential Regulation of fMLP-and PAF-mediated Phosphoinositide Hydrolysis and Secretion by cAMP-RBL cells
were preincubated with or without a membrane-permeable cAMP analog, cpt-cAMP (1 mM; 5 min) and dose responses of fMLP-and PAF-stimulated phosphoinositide hydrolysis and degranulation were determined. As shown in Fig. 1A, cpt-cAMP caused a substantial inhibition of fMLP-stimulated generation of inositol phosphates. In contrast, PAF-mediated phosphoinositide hydrolysis was inhibited by only ϳ30% (Fig. 1B). Furthermore, cpt-cAMP substantially inhibited secretion stimulated by fMLP but had no effect on the response to PAF (Fig.  1, C and D). The half-maximal and maximal concentrations of cpt-cAMP for inhibition of fMLP-mediated responses were ϳ0.1 mM and ϳ1 mM, respectively (Fig. 1, E and F). Cpt-cAMP also caused a substantial inhibition of intracellular Ca 2ϩ mobilization stimulated by fMLP but not PAF (4). To test whether cAMP also inhibited responses to other chemoattractant receptors that activate PLC via a ptx-sensitive G protein, its effect on C5a-stimulated Ca 2ϩ mobilization in RBL cells was tested. In the absence of cpt-cAMP, stimulation with C5a (1 nM) re-sulted an increase of intracellular mobilization of 162 Ϯ 5.6 nM (n ϭ 4) over basal. In the presence of cpt-cAMP (1 mM, 5 min), this response was reduced to 27 Ϯ 2.2 nM (83% inhibition).
The ability of fMLP and PAF to produce cAMP was also determined. As shown in Fig. 2A, fMLP caused a ϳ2.5-fold increase in cAMP over basal, whereas PAF produced no response. Treatment of cells with ptx (100 ng/ml, overnight) resulted in a complete inhibition of fMLP-stimulated cAMP generation. To determine whether the cAMP increase caused by chemoattractants has a regulatory effect on their cellular responses, cells were preincubated with the PKA inhibitor H-89 and its effect on fMLP and PAF-stimulated generation of inositol phosphates were tested. H-89 pretreatment resulted in a 2.5-fold increase in fMLP-stimulated generation of inositol phosphates (Fig. 2B). In contrast, the response to PAF was enhanced only by ϳ20% and this effect was lost in cells treated with ptx. It was determined whether the findings in transfected RBL cells occurred in human neutrophils and in neutrophil-like HL-60 cells. As shown in Fig. 3A, fMLP caused a significant increase in cAMP generation in neutrophils and in dimethyl sulfoxide differentiated HL-60 cells. C5a also stimulated cAMP production in neutrophils (Fig. 3A). PAF did not stimulate cAMP formation in either neutrophils or HL-60 cells. Cpt-cAMP caused a substantial inhibition of fMLP but not PAF-stimulated Ca 2ϩ mobilization in both cell types (Fig. 3B). Cpt-cAMP also caused a substantial inhibition of C5a-stimulated Ca 2ϩ mobilization in neutrophils (Fig. 3B).
Effect of cAMP on fMLP-and PAF-mediated GTPase Activity-To determine the effect of cpt-cAMP on G protein activation, RBL cells were treated with buffer or cpt-cAMP, then membranes were prepared, and the ability of fMLP and PAF to stimulate GTPase activity was measured. Both fMLP and PAF stimulated GTPase activity in a dose-dependent manner in membranes from buffer or cpt-cAMP-treated cells showing that cpt-cAMP pretreatment had no effect on this PAF-or fMLPstimulated response (data not shown).
In Vivo and in Vitro Phosphorylation of PLC␤ 3 -Using antibodies that specifically recognize different PLC␤ isoforms, it was shown that of the known PLC␤ isoforms only PLC␤ 3 is expressed in RBL cells (4). To determine if other PLC␤ isoforms are expressed in RBL cells at levels below the detection of antibodies, specific oligonucleotide primers for different PLC␤ isoforms were used for reverse transcriptase-polymerase chain reaction on RNA from RBL cells. Rat brain RNA was used as a control. PLC␤ 3 was the only PLC␤ isoform detected in RBL cells (data not shown).
fMLP and cpt-cAMP caused a dose-dependent phosphorylation of PLC␤ 3 (Fig. 4, A and B). To determine whether PLC␤ 3 was a substrate for PKA, cell lysates were immunoprecipitated with anti-PLC␤ 3 antibody and the ability of purified catalytic subunit of PKA to phosphorylate PLC␤ 3 was tested in the presence of [␥-32 P]ATP. As shown in Fig. 4C, PLC␤ 3 phosphorylation by PKA was detected within 1 min and remained elevated for 15 min. In vitro PLC␤ 3 phosphorylation by PKA was blocked in immunoprecipitates prepared from cells treated with cpt-cAMP (Fig. 4D). To determine whether fMLP-stimulated PLC␤ 3 phosphorylation was in part activated by PKA, whole cells were incubated with either fMLP or PAF, PLC␤ 3 was immunoprecipitated, and the ability of PKA to phosphorylate the enzyme was determined. Treatment of cells with fMLP but not PAF resulted in a substantial inhibition of PLC␤ 3 phosphorylation by PKA (Fig. 4D). DISCUSSION Chemotactic, microbiocidal, and cytotoxic effects of phagocytic leukocytes are stimulated by chemoattractants such as formylated peptides and PAF via the G protein-coupled receptor activation of PLC (16). The ability of fMLP to produce a transient increase in cAMP production in neutrophils is well documented (17)(18)(19)(20); however, the physiological effects of this phenomenon were not known. A recent study did demonstrate that a PKA inhibitor enhanced superoxide production stimulated by fMLP in human neutrophils (21), but the mechanism of this effect was also unknown. It also remained to be determined whether PAF stimulated cAMP production in neutrophils and whether inhibition of PKA resulted in the regulation of PAF-mediated biological responses as well. The present work utilizing RBL cells stably expressing fMLP and PAF receptors demonstrated that fMLP, but not PAF, caused an increase in cAMP formation and that preincubation of cells with a membrane permeable cAMP analog resulted in inhibition of both phosphoinositide hydrolysis and exocytotic release of granules stimulated by fMLP. Furthermore, the PKA inhibitor H-89 enhanced fMLP-stimulated phosphoinositide hydrolysis. These data suggest that cAMP produced by fMLP provides a mechanism for counter regulation of an fMLP-stimulated biological response, secretion via the inhibition of PLC activation. This phenomenon appears to be specific for fMLP versus PAF, as the latter did not cause cAMP generation nor did exogenously added cAMP inhibit PAF-induced phosphoinositide hydrolysis or secretion. This difference in the regulation of fMLP-and PAF-mediated responses in transfected RBL cells is likely to be physiologically relevant as similar differences in the generation of cAMP and the regulation of cellular functions by cAMP were observed in the present study in human neutrophils and neutrophil-like HL-60 cells. These differences in the functional regulation of chemoattractant receptors are likely a consequence of G protein usage as suggested by the distinct ptx sensitivity of fMLP versus PAF receptors, with the former being sensitive and the latter is at least partially resistant (8,10,11).
In the studies reported here, fMLP stimulated a 40 -50% increase in cAMP over basal levels in neutrophils and HL-60 cells (Fig. 3A). These data are consistent with previous findings from this and other laboratories (17)(18)(19). Furthermore, the ability of a PKA inhibitor to enhance fMLP-stimulated superoxide generation in neutrophils (21) suggests that the small fMLP-stimulated cAMP increase is sufficient to counter regulate the response to fMLP. This contention is supported by the finding that phosphodiesterase inhibitors which cause ϳ50% increase of cellular cAMP also result in a substantial inhibition of fMLP-stimulated superoxide release in neutrophils (22). The mechanism by which fMLP causes an increase in cAMP is not known (18,23). Of the nine adenylyl cyclases identified, activation of types I and III are inhibited by all three forms of G i proteins (24 -26). In contrast, types II and IV are activated by G s and G ␤␥ of ptx-sensitive G proteins in a synergistic manner (23). In guinea pig neutrophils, fMLP greatly potentiates cAMP production stimulated by prostaglandin E 1 receptor and this enhancement is totally inhibited by ptx (27). In HEK 293 cells, which endogenously expresses adenylyl cyclase type III, fMLP causes an inhibition of cAMP production (28,29). This inhibitory effect of fMLP likely results from the interaction of G i␣ with type III adenylyl cyclase. However, in the same cell line transiently expressing adenylyl cyclase type II, fMLP causes the stimulation of cAMP production and this response is inhibited by ptx (30). This indicates that fMLP can either stimulate or inhibit cAMP formation depending on the subtype of adenylyl cyclase expressed. The observation that fMLP caused cAMP formation in RBL cells and that this response was completely inhibited by ptx suggest that G ␤␥ directly interacts with adenylyl cyclases types II or IV to stimulate cAMP production. This contention is supported by the finding that PAF, which utilizes a ptx insensitive mechanism to cause intracellular Ca 2ϩ mobilization in RBL cells, HL-60 cells and neutrophils, did not cause cAMP production in any of these cell types ( Figs. 2A and 3A).
The demonstration that cAMP did not cause phosphorylation of fMLP receptor (11) and had no effect on fMLP-stimulated GTPase activity indicates that its ability to block fMLP-stimulated inositol phosphate generation and secretion is not mediated at the level of the receptor or its coupling to G protein.
Inhibition of membrane inositol phospholipid resynthesis and thus a reduction in the availability of substrate for PLC has been postulated as a mechanism by which cAMP inhibits fMLP-stimulated generation of inositol phosphates in human neutrophils (31). This mechanism is unlikely because cAMP did not inhibit phosphoinositide hydrolysis stimulated by PAF in RBL cells, and it had no effect on PAF-induced Ca 2ϩ mobilization in human neutrophils and HL-60 cells. The selective inhibition of fMLP response by cAMP is therefore likely to be mediated via the modification of PLC. Using reverse transcriptase-polymerase chain reaction (this study) and Western blotting with PLC␤ isoform-specific antibodies (4) it was shown that PLC␤ 3 is the only known PLC␤ isozyme expressed in RBL cells. Furthermore, both fMLP and cpt-cAMP caused phosphorylation of PLC␤ 3 in this cell line. In addition, purified catalytic subunit of PKA phosphorylated PLC␤ 3 immunoprecipitated from an RBL cell lysate. The observation that preincubation of cells with cpt-cAMP blocked subsequent in vitro PLC␤ 3 phosphorylation by PKA suggests that PLC␤ 3 is a direct substrate for PKA. The ability of cpt-cAMP to inhibit fMLP-induced phosphoinositide hydrolysis and secretion is likely mediated via the phosphorylation of PLC␤ 3 by PKA. Importantly, the finding that fMLP stimulated the formation of cAMP and that pretreatment of cells with fMLP resulted in a partial inhibition of PKA-stimulated PLC␤ 3 phosphorylation in vitro indicate that fMLP-stimulated PLC␤ 3 phosphorylation is mediated, at least in part, by PKA. The ability of the PKA inhibitor H-89 to enhance fMLP-stimulated inositol phosphates generation suggests that counter regulation of fMLP-stimulated biological responses is likely mediated via the PKA-induced phosphorylation of PLC␤ 3 . This form of inhibition appears to be specific for fMLP versus PAF, which did not stimulate cAMP formation and did not block PLC␤ 3 phosphorylation by PKA in vitro. Furthermore, the PKA inhibitor H-89 had no effect on the ptx-insensitive component of PAF-mediated generation of inositol phosphates.
The data presented herein revealed that fMLP activated both PLC␤ 3 and cAMP production via a ptx-sensitive pathway. Interestingly, PKA activated by this mechanism phosphorylated PLC␤ 3 and blocked the subsequent activation of PLC by fMLP. It is likely that fMLP activates PLC␤ 3 through the release of G␤␥ (8,10,11). Thus the PKA-mediated phosphorylation of PLC␤ 3 can be hypothesized to selectively block activation by G ␤␥ as opposed to G ␣q , which is activated by PAF. This suggests a novel selective counter regulatory pathway for certain G protein-coupled receptors such as those for chemoattractants that activate a ptx-sensitive signaling pathway. Other receptors such as the ␣2-adrenergic, dopamine D 2 and adenosine A 1 receptors also activate both adenylyl cyclase and PLC (23). Given that PLC␤ 3 is expressed abundantly in many cells and tissues such as platelets, leukocytes, brain, testes, and lung (32-34), phosphorylation of PLC␤ 3 by PKA may be a novel mechanism for the counter regulation of some but not other G protein-coupled receptors.
Receptors that couple to G s and cause an elevation of intracellular cAMP levels are known to inhibit PLC-mediated responses to other receptors that couple to ptx-sensitive G protein (15). The data herein suggest that the mechanism for this phenomenon may be through the phosphorylation of PLC␤ 3 by PKA at a site which blocks G ␤␥ -mediated activation. PLC␤ other than PLC␤ 3 may also be regulated by PKA. For example, phosphorylation of PLC␤ 2 by PKA has been suggested as a mechanism by which cAMP inhibits fMLP-stimulated phosphoinositide hydrolysis in differentiated HL-60 cells (15). It is, however, important to note that the neutrophil-like HL-60 cells express both PLC␤ 2 and PLC␤ 3 (4). Furthermore, cAMP causes phosphorylation of PLC␤ 3 in the human monocyte-like U937 cells and the murine macrophage-like J774.1 cells (4). The finding that fMLP-stimulated phosphoinositide hydrolysis and Ca 2ϩ mobilization were not completely blocked in neutrophils isolated from mice deficient in PLC␤ 2 suggests that both PLC␤ 2 and PLC␤ 3 are activated by fMLP (35). Unlike PLC␤ 2 , which is expressed only in certain cells of hematopoietic origin, PLC␤ 3 is expressed in many cell types and tissues (32,33). Therefore, cross-talk between the adenylyl cyclase and PLC pathways is likely to be mediated via the phosphorylation of PLC␤ 2 , PLC␤ 3 , or both depending on the cell type on which the receptors are expressed.