Desensitization of N-Formylpeptide Receptor-mediated Activation Is Dependent upon Receptor Phosphorylation*

The human N-formylpeptide receptor (FPR) represents one of the most thoroughly studied leukocyte chemoattractant receptors. Despite this, little is known about the molecular mechanisms involved in the activation and desensitization of this receptor. To assess the role of phosphorylation in receptor function, U937 promonocytic cells were stably transfected to express the recombinant human FPR. Three mutant forms of the FPR lacking specific serine and threonine residues in the receptor C terminus were studied with respect to activation and desensitization. Replacement of all 11 serine and threonine residues within the C terminus by alanine and glycine residues (ΔST) resulted in a receptor capable of ligand binding and G protein activation similar to the wild-type receptor. However, whereas the wild-type FPR was phosphorylated on both serine and threonine residues upon exposure to agonist and displayed a significantly reduced ability to stimulate G protein-mediated GTP hydrolysis upon subsequent exposure to agonist, ΔST demonstrated a complete lack of phosphorylation and displayed little alteration in its ability to stimulate G protein-mediated GTP hydrolysis upon a subsequent exposure to agonist. In addition to desensitization of G protein-mediated GTP hydrolysis, calcium mobilization was assayed to test whether desensitization occurred at a site distal to G protein activation. However, as observed with G protein activation, ΔST underwent no desensitization of the calcium mobilization response upon a second exposure to agonist. To define more precisely the role of specific serine and threonine residues, two additional mutants were analyzed. Replacement either of Ser328, Thr329, Thr331, and Ser332 (mutant A) or of Thr334, Thr336, Ser338, and Thr339 (mutant B) resulted in functional receptors that exhibited ∼50% the level of phosphorylation following stimulation. Whereas mutant A, like ΔST, could not be significantly desensitized by exposure to agonist, mutant B exhibited partial desensitization. These results indicate that phosphorylation of the FPR is a necessary and sufficient step in cellular desensitization, that multiple phosphorylation sites are involved, and that redundant desensitization does not occur downstream of G protein activation in the signaling cascade.

Neutrophils possess a large number of cell-surface G proteincoupled receptors that respond to structurally diverse ligands such as N-formylpeptides, complement components C5a and C3a, platelet-activating factor, and chemokines such as IL-8 1 (1). Receptor activation results in the stimulation of phospholipases, the mobilization of intracellular calcium, and the activation of a multitude of protein kinases culminating in functions such as chemotaxis, phagocytosis, superoxide production, and degranulation. Following an initial exposure to ligand, resulting in transient cell activation, neutrophils rapidly become unresponsive to continued or subsequent stimulation. This process of cellular acquiescence in the presence of agonist is termed desensitization and has been characterized for many hormonal (2) and neurotransmitter (3) receptors. Although the complex mechanisms involved in this process are poorly characterized, one of the early events has been suggested to involve receptor phosphorylation (4).
G protein-coupled receptor kinases are a family of protein kinases that rapidly phosphorylate seven transmembrane receptors in a ligand-dependent manner (5)(6)(7). Following phosphorylation and the possible association with accessory proteins, such as arrestin, receptors are no longer capable of effectively activating G proteins, a process termed homologous desensitization (8). G protein-coupled receptor kinases are implicated in the light-stimulated phosphorylation of rhodopsin (9) and the agonist-dependent phosphorylation of the ␤ 2 -adrenergic receptor (5). Recently, the FPR as well as additional chemoattractant receptors, such as the C5a and IL-8 receptors, have been shown to become rapidly phosphorylated in a ligand dose-dependent manner (10,11). Phosphorylation of the FPR was demonstrated in the human cell line HL-60 (10), which expresses endogenous FPR after terminal differentiation with agents such as dibutyryl cAMP, as well as in an FPR-transfected rat basophilic cell line, RBL-2H3 (12). Pretreatment of the cells with staurosporine, a protein kinase C inhibitor, failed to prevent the ligand-stimulated phosphorylation of the FPR (10), but resulted in partial inhibition of the phosphorylation of other chemoattractant receptors (12). This indicated that a kinase other than protein kinase C was responsible for the phosphorylation of the FPR. Using a fusion protein containing the carboxyl-terminal cytoplasmic domain of the FPR, it was demonstrated that this region of the FPR is specifically phosphorylated by a neutrophil cytosolic kinase with properties similar to those of the G protein-coupled receptor kinase GRK2 (13). Purified GRK2 and, to a lesser extent, GRK3 were shown to phosphorylate the FPR carboxyl terminus, whereas GRK5 and GRK6 had no activity. Site-directed mutagenesis of numerous regions of the FPR carboxyl terminus suggested a hierarchical mechanism in which phosphorylation of amino-terminal Ser and Thr residues appeared to be required for the subsequent phosphorylation of carboxyl-terminal residues (13).
In addition to the homologous desensitization of chemoattractant receptors, where a ligand desensitizes only its own receptor, heterologous desensitization, where an activated receptor desensitizes one or more other inactive receptors, has also been demonstrated to occur between chemoattractant receptors (14,15). This latter form of desensitization is at least in part mediated by second messenger-activated kinases, such as protein kinase C (11). The existence of a novel intermediate form of desensitization has been suggested by observations of the desensitization of FPR-mediated inositol 1,4,5-trisphosphate generation and calcium mobilization by C5a and IL-8 in the absence of FPR phosphorylation (16). C5a and IL-8 treatment, however, did not result in desensitization of GTP␥S binding, a measure of receptor-G protein coupling. This suggested that the C5a-or IL-8-mediated desensitization of FPRmediated signaling occurred at the level of either G protein effector coupling or phospholipase C activation. In addition, since fMLF, C5a, and IL-8 all appear to utilize similar signal transduction pathways, stimulation by any one of these three ligands should engage the described downstream desensitization mechanism(s) (16). Thus, it was predicted that fMLF stimulation would result in FPR desensitization independent of receptor phosphorylation.
To investigate the mechanisms of cellular desensitization in response to fMLF, mutant forms of the FPR were generated lacking some or all of the potential phosphorylation sites contained within the carboxyl terminus and expressed in human myeloid U937 cells (17). The results presented here demonstrate that phosphorylation at multiple sites within the carboxyl terminus of the FPR is a necessary step in the desensitization of the receptor and that without it cells remain completely responsive to subsequent challenges with ligand. Thus, the FPR appears not to activate desensitization of a downstream component(s) activated in response to fMLF.

EXPERIMENTAL PROCEDURES
Materials-The cDNA encoding the FPR was obtained from a human HL-60 granulocyte library as described previously (18). fMLF was purchased from Sigma. N-Formyl-Nle-Leu-Phe-Nle-Tyr-Lys-fluorescein and indo-1/AM were obtained from Molecular Probes, Inc. Carrier-and acid-free [ 32 P]orthophosphate was from Amersham Corp. Protein A-Sepharose CL-4B beads were obtained from Pharmacia Biotech Inc. RPMI 1640 medium was from Whittaker Bioproducts; fetal bovine serum was from Hyclone Laboratories.
Construction and Expression of Site-directed Mutants in U937 Cells-The FPR gene was subcloned into the EcoRI site in the polylinker of M13mp18 and mutagenized as described (19). U937 cells were grown in RPMI 1640 medium supplemented with 2 mM L-glutamine, 100 units/ml penicillin, 100 g/ml streptomycin, 10 mM HEPES (pH 7.4), and 10% heat-inactivated fetal bovine serum. For transfection, ϳ4 ϫ 10 6 cells were harvested and resuspended in 400 l of RPMI 1640 medium containing 10 mM glucose and 0.1 mM dithiothreitol (20). Linearized DNA (10 g in a volume of 10 l) was added to the cells and preincubated for 5 min at room temperature. The cells were then subjected to a 240-V pulse from a 960-microfarad capacitor (resulting in a pulse time constant of ϳ30 ms) and immediately returned to 5-10 ml of culture medium. The following day, G418 was added to a final active concentration of 1 mg/ml. As the selection proceeded, the cells were centrifuged and resuspended in fresh medium (containing G418) at 4 -6-day intervals. Cells were cultured at 37°C in a humidified atmosphere of 6% CO 2 and 94% air.
Flow Cytometry-U937 cells were harvested by centrifugation, washed once with phosphate-buffered saline, and resuspended at 10 6 cells/ml in phosphate-buffered saline. Binding was carried out in 0.5 ml with N-formyl-Nle-Leu-Phe-Nle-Tyr-Lys-fluorescein at 10 nM (21). Following incubation for at least 15 min on ice, the cells were analyzed on a FACScan flow cytometer (Becton Dickinson) for fluorescent intensity. Debris and dead cells were excluded with a gate on forward and side scatter. Nonspecific binding was determined in the presence of 1 M fMLF.
Radioligand Binding-Ligand binding assays were performed on membranes prepared by nitrogen cavitation in a final volume of 0.2 ml of binding buffer. Membranes (30 g of protein) were suspended in binding buffer (pH 7.4) consisting of 140 mM NaCl, 1.0 mM KH 2 PO 4 , 5 mM Na 2 HPO 4 , 1.5 mM CaCl 2 , 0.3 mM MgSO 4 , 1 mM MgCl 2 , and 0.2% bovine serum albumin. Binding was initiated by the addition of various amounts of [ 3 H]fMLF. Equilibrium binding was carried out at 23°C for 45 min and terminated by rapid filtration through Whatman GF/C filters followed by three washes with 0.75 ml of ice-cold binding buffer. Specific binding was calculated as total binding minus nonspecific binding, determined in the presence of 50 M unlabeled fMLF. Each determination was done in duplicate. The amount of bound ligand was estimated by scintillation counting, and the binding data were analyzed by fitting to a double rectangular hyperbola with the nonlinear curve fitting program SigmaPlot (Jandel Scientific). In the presence of GTP␥S, only one binding affinity was observed, corresponding to the low affinity state of the receptor. In the absence of GTP␥S, two binding affinities were observed, representing the high and low affinity binding states, corresponding to the G protein-coupled and uncoupled forms of the receptor, respectively.
GTPase Activity-G protein activation was determined directly by measuring the ligand-induced hydrolysis of GTP by G proteins. Membranes (20 g of protein) prepared by nitrogen cavitation were incubated with [␥-32 P]GTP in the following buffer: 25 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 2.5 mM MgCl 2 , 1 mM ATP, 0.5 mM AMP-PNP, 0.25 mM ouabain, 1 mM creatine phosphate, 5 units/ml phosphocreatine kinase, and 1 M GTP (containing 2-4 Ci of [␥-32 P]GTP). The incubation was carried out for 5-10 min at room temperature in the presence or absence of ligand. The assay was stopped with a 9-fold excess of ice-cold 5% (w/v) Norit A charcoal (J. T. Baker Inc.) in an aqueous solution containing 57 mM phosphoric acid. The charcoal and bound guanine nucleotides were sedimented by centrifugation, and the liberated [ 32 P]phosphate in the supernatant was determined by scintillation counting.
Measurement of [Ca 2ϩ ] i -Cells were harvested by centrifugation, washed once with phosphate-buffered saline, and resuspended at 5 ϫ 10 6 cells/ml in Hanks' buffered saline solution. The cells were incubated with 5 M indo-1/AM for 25 min at 37°C, washed once with Hanks' buffered saline solution, and resuspended at ϳ10 6 cells/ml in Hanks' buffered saline solution containing 1.5 mM EGTA (pH 8.0). The elevation of intracellular Ca 2ϩ by various amounts of fMLF was monitored by continuous fluorescent measurement using an SLM 8000 photon-counting spectrofluorometer (SLM-AMINCO) detecting at 400 and 490 nm, as described (18). The concentration of intracellular Ca 2ϩ was calculated as described (22).
In Vivo Phosphorylation and Immunoprecipitation-FPR-transfected U937 cells were grown to a density of 1.0 -1.5 ϫ 10 6 cells/ml and washed three times with 150 mM NaCl and 10 mM HEPES (pH 7.4) to remove traces of phosphate. Cells were resuspended in phosphate-free RPMI 1640 medium containing 10 mM HEPES (pH 7.4) to a density of 10 7 cells/ml in a volume of ϳ0.5 ml to which was added 1 mCi of carrierand acid-free [ 32 P]orthophosphate (10 mCi/ml). Cells were loaded for 3 h at 37°C in a humidified atmosphere of 6% CO 2 and 94% air. Following loading, cells were stimulated with fMLF as indicated and immediately lysed by the addition of 0.33 volume of 4 ϫ radioimmune precipitation assay buffer (40 mM Tris-HCl (pH 7.5), 600 mM NaCl, 4 mM EDTA, 0.4% SDS, 2% deoxycholate, 4% Triton X-100, 4 mM p-nitrophenyl phosphate, 40 mM sodium phosphate, 40 mM NaF, 20 g/ml soybean trypsin inhibitor, 20 g/ml leupeptin, 2 mM phenylmethylsulfonyl fluoride, 400 ng/ml aprotinin, and 200 g/ml pepstatin A). Following lysis and extraction for 10 min while rotating at 4°C, samples were centrifuged at 15,000 ϫ g for 15 min at 4°C to remove insoluble debris. The supernatant was added to 10 mg of protein A-Sepharose that had been precoated with 15 l of a rabbit antiserum directed against the C-terminal 12 amino acids of the FPR. The use of this antibody for immunoprecipitating the photoaffinity-labeled FPR has been previously described (23). Following binding for 1 h while rotating at 4°C, the beads were washed as follows: once with 1 ml of 50 mM Tris-HCl, 500 mM NaCl, 1% Triton X-100, and 0.2% SDS (pH 8.0); once with 1 ml of 50 mM Tris-HCl, 500 mM NaCl, 1% Triton X-100, and 0.1% SDS (pH 8.0); once with 1 ml of 50 mM Tris-HCl and 500 mM NaCl (pH 8.0); and finally with phosphate-buffered saline. Laemmli sample buffer (2-fold concentrated, 40 l) was added, and the samples were heated at 37°C for 10 min, followed by electrophoresis on a 12.5% SDS-polyacrylamide gel. Gels were dried, and relative determinations of 32 P content were performed with a Molecular Dynamics PhosphorImager.
Phosphoamino Acid Analysis-Phosphoamino acid content was determined as described (24). Briefly, following transfer of the SDS gel to an Immobilon P membrane (Millipore Corp.), the immunoprecipitated band was excised and hydrolyzed in 100 -200 l of 6 M HCl for 1 h at 100 -110°C. The sample was dried in a SpeedVac concentrator; resuspended with 0.5 g each of phosphoserine, phosphothreonine, and phosphotyrosine in a volume of 3-5 l; and spotted onto a cellulose thinlayer plate (100 m; EM Laboratories). Phosphoamino acids were separated by chromatography in 5:3 isobutyric acid/ammonium hydroxide (0.5 M). Unlabeled phosphoamino acid standards were visualized with ninhydrin, whereas 32 P-labeled phosphoamino acids were visualized with a PhosphorImager.
Desensitization-For desensitization of GTPase activity, cells were harvested, resuspended in RPMI 1640 medium containing 10 mM HEPES (pH 7.4) to a density of 10 7 cells/ml, divided into two equal parts, and stimulated with either 1 M fMLF or buffer for 10 min at 37°C. Cells were then added to ice-cold buffer, harvested, and processed for membranes by nitrogen cavitation. Membranes from fMLFtreated and untreated cells were then assayed for GTPase activity as described above. For desensitization of the calcium mobilization response, cells (5 ϫ 10 6 ) were loaded with indo-1/AM as described above for calcium determinations and divided into two parts. One was stimulated with 1 M fMLF for 10 min, whereas the other was treated with only buffer. The cells were then washed three times with Hanks' buffered saline solution at room temperature to remove surface-bound fMLF and resuspended for assay of calcium mobilization as described above.

RESULTS
To examine the role of receptor phosphorylation in cellular desensitization in response to fMLF, a novel model system employing a human myeloid cell line was used. In addition to expressing the recombinant wild-type FPR, three mutant forms of the receptor were also expressed (Fig. 1). The first mutant, ⌬ST, had alanine or glycine substituted for each of the 11 serine and threonine residues of the carboxyl terminus. The high number of glycine residues used was due to the fact that 9 of 12 amino acids between Ser 328 and Thr 339 would have been alanine in the final receptor sequence. Given the helix-forming propensity of such an alanine-rich sequence, numerous serine and threonine residues were converted to glycine with the intent of eliminating any possible structural aberrations. In addition to ⌬ST, two mutants were generated changing only four Ser and Thr residues at a time (mutants A and B) (Fig. 1). These mutants, previously expressed as glutathione S-transferase fusion proteins containing the carboxyl-terminal 47 amino acids of the FPR, have been shown to represent potential sites of phosphorylation by GRK2 in vitro (13). Their role in desensitization was to be elucidated here.
Mutant forms of the FPR cDNA were generated by sitedirected mutagenesis, subcloned into the mammalian expression vector pSFFV.neo, and introduced into the human myeloid cell line U937 by electroporation. Transfected cells were selected in the presence of G418 and analyzed for cell-surface expression of ligand binding by flow cytometry (Fig. 2). All three mutants demonstrated binding of the fluorescent ligand N-formyl-Nle-Leu-Phe-Nle-Tyr-Lys-fluorescein at levels comparable to those of the wild-type receptor. Expression of the wild-type FPR in U937 cells has previously been shown to yield binding affinities for [ 3 H]fMLF similar to those of the FPR from dibutyryl cAMP-differentiated U937 cells and neutrophils (17). Detailed ligand binding studies performed here using [ 3 H]fMLF demonstrated that the mutant forms of the FPR exhibited both high and low affinity binding sites similar to those of the wild-type FPR (Table I). In the presence of GTP␥S, which irreversibly activates and dissociates G proteins, only one low affinity binding site was observed. These result are consistent with a fraction of the wild-type and mutant receptors being coupled to G proteins in the absence of GTP␥S, suggesting that the mutant receptors might be capable of G protein activation.
To examine the functional coupling of the wild-type FPR as well as the mutant forms of the FPR, monitoring of intracellular calcium fluxes was performed with indo-1. Stimulation of each transfected cell line with 1 M fMLF resulted in a rapid rise in intracellular calcium (Fig. 3). Stimulation of untransfected or vector-transfected cells yielded no such response, indicating that undifferentiated U937 cells do not express any FPR (data not shown). The rise in intracellular calcium was transient, returning to base line within ϳ60 s. Taken together, these results indicate that the Ser and Thr residues of the carboxyl terminus of the FPR do not play a direct role in G protein activation or affect receptor function in such a way as to preclude ligand binding or signal transduction.
Previous data have suggested that the carboxyl terminus of the FPR is capable of being phosphorylated by a neutrophil kinase with properties similar to those of the G protein-coupled receptor kinase GRK2 as well as by purified GRK2 itself (13). To examine this reaction in vivo, transfected U937 cells were loaded with 32 P-labeled inorganic phosphate, stimulated with fMLF, and extracted with detergent for immunoprecipitation of the FPR. Compared with unstimulated FPR-transfected U937 cells (Fig. 4A, lane 1), fMLF-stimulated cells showed a diffuse band with a molecular mass of 50 -70 kDa (lane 2). This diffuse band has previously been shown to represent the glycosylated form of the FPR. Immunoprecipitation of the phosphorylated FPR was blocked by preincubation of the anti-FPR antiserum with the C-terminal peptide used as antigen (Fig. 4A, lane 3), but not by a peptide from the third intracellular loop (lane 4). The phosphorylated FPR was also not immunoprecipitated from FPR-transfected cells by preimmune serum (Fig. 4A, lane   5) or from vector-transfected U937 cells by the anti-FPR antiserum (Fig. 4B, lane 2), again confirming the absence of the FPR in the latter cells.
To determine the identity of the phosphorylated amino acids, acid hydrolysis was performed on the isolated phosphorylated FPR. Chromatographic separation of the phosphoamino acids revealed that both serine and threonine residues were phosphorylated, with no phosphotyrosine detected. Approximately equal amounts of the two phosphoamino acids were detected (55% phosphothreonine and 45% phosphoserine), similar to what was demonstrated for GRK2-mediated in vitro phosphorylation of the isolated FPR carboxyl terminus. Phosphorylation of the wild-type FPR in response to fMLF occurred in a dose-dependent manner (Fig. 5A). The EC 50 for phosphorylation was similar to the EC 50 for calcium mobilization, ϳ2 ϫ 10 Ϫ8 M, suggesting that the ligand-induced active G protein coupling state of the FPR also represented the substrate for the kinase. Phosphorylation was rapid, with a half-time of ϳ1-2 min (Fig. 5B).
Analysis of the site-directed mutants revealed that the ⌬ST mutant was not phosphorylated even at the highest concentra-  Following immunoprecipitation of the 32 P-phosphorylated FPR and separation on a 12% SDS-poylacrylamide gel, the proteins were transferred to an Immobilon P membrane, and the FPR protein band was excised and hydrolyzed as described under "Experimental Procedures." The circles indicate the positions of the unlabeled phosphoamino acid standards as determined with ninhydrin staining. PY, phosphotyrosine; PT, phosphothreonine; PS, phosphoserine.
tions of fMLF tested (Fig. 6). This result demonstrated that the Ser and Thr residues of the carboxyl terminus were indeed the sites of fMLF-dependent FPR phosphorylation. Both mutants A and B demonstrated phosphorylation at a level ϳ50% that of the wild-type receptor (Fig. 6), with an EC 50 identical to that of the wild-type receptor (data not shown). These results suggest that residues within both regions covered by mutants A and B are phosphorylated, consistent with previous results of GRK2mediated phosphorylation of the FPR carboxyl terminus, where it was found that phosphorylation appeared to proceed in a hierarchical manner, with residues within mutant A (site A) being phosphorylated before residues within mutant B (site B) (13).
To determine the possible role of phosphorylation in desensitization, alterations in the profile of ligand-induced G protein activation following desensitization were investigated. Desensitization was accomplished by treating cells with saturating doses of fMLF (1 M) for 10 min at 37°C. This treatment is more than sufficient to achieve maximal phosphorylation of the wild-type receptor (see Fig. 5). Following this, cells were cooled to 0°C, and membranes were prepared as described for ligand binding. Parallel (non-desensitized) controls were prepared that were not treated with fMLF. Determination of the ability of the wild-type FPR to stimulate G protein-mediated GTP hydrolysis in non-desensitized control membranes demonstrated that fMLF stimulation resulted in a 2-fold increase in GTP hydrolysis over the background of unstimulated membranes. This level of stimulation in the presence of fMLF was similar for non-desensitized cells expressing ⌬ST, mutant A, and mutant B (data not shown). However, when membranes from fMLF-treated (i.e. desensitized) cells expressing the wildtype FPR were assayed, the extent of GTP hydrolysis was reduced by ϳ80%, reflecting desensitization of the receptormediated activity (Fig. 7). In contrast, when membranes from fMLF-treated, desensitized cells expressing the ⌬ST mutant were compared with membranes from non-desensitized cells, only a slight reduction in the level of GTPase activity was observed (Fig. 7). This result suggests that phosphorylation of the carboxyl terminus of the FPR is required for desensitization of receptor-G protein coupling. Further analysis of mutants A and B revealed that both of these mutants were also deficient in their ability to undergo desensitization, although not to the same extent as the ⌬ST mutant, suggesting that phosphorylation within both sites A and B is necessary for maximal desensitization.
Recent results have suggested that desensitization of chemoattractant-mediated signaling can also occur at a site following the activation of G proteins and prior to the activation of phospholipase C. To test this possibility, cells expressing the wild-type and mutant receptors were treated with a saturating  7. Desensitization of fMLF-stimulated, G protein-mediated GTPase activity. Wild-type (WT) or mutant FPR-transfected cells were incubated with either buffer or 1 M fMLF for 10 min at 37°C to generate non-desensitized and desensitized cells, respectively. Membranes were then prepared and assayed for fMLF-stimulated GTP hydrolysis as described under "Experimental Procedures." Basal levels of GTP hydrolysis were identical in both fMLF-treated (i.e. desensitized) and untreated (i.e. non-desensitized) membranes. To determine the degree of desensitization, the amounts of fMLF-induced GTP hydrolysis were compared in the fMLF-treated (i.e. desensitized) and buffer-treated (i.e. non-desensitized) membranes. Desensitization (%) is defined as 100 ϫ ((N Ϫ D)/N), where N represents the fMLF-stimulated GTP hydrolysis of buffer-treated (i.e. non-desensitized) membranes, and D represents the fMLF-stimulated GTP hydrolysis of fMLF-treated (i.e. desensitized) membranes. dose of fMLF for 10 min at 37°C to desensitize them as described above, washed extensively to remove the ligand, and assayed for calcium mobilization, the result of phospholipase C activation. Desensitized cells were then compared with cells that had not been desensitized. When U937 cells transfected with the wild-type FPR were examined, desensitization resulted in a significantly higher concentration (30 -50-fold) of fMLF being required to obtain a response to ligand, which at its maximum was less than half that obtained with the non-desensitized receptor (Fig. 8). When cells expressing the ⌬ST mutant were examined, no such change in the calcium mobilization response was observed. In fact, ⌬ST was completely refractory to desensitization, exhibiting an identical dose-response curve after the "desensitizing" treatment as compared with the untreated mutant cells, which themselves were indistinguishable from the untreated wild-type receptor. These results indicated that, whereas the Ser and Thr residues of the carboxyl terminus of the FPR play no role in the ability of the receptor to activate G protein (Figs. 3 and 8), they are essential to the process of desensitization, presumably through their ability to become phosphorylated. Analysis of mutant A also demonstrated no decrease in responsiveness following treatment with fMLF, indicating that mutation of the Ser and Thr residues within site A is sufficient to reproduce the effects seen with ⌬ST. Mutant B, however, yielded an intermediate profile between the wild-type receptor and the ⌬ST mutant. Although the maximal response following desensitization approached the level of the untreated cells, 5-10-fold higher concentrations of fMLF were required to achieve this level of response. The EC 50 values for untreated mutants A and B were essentially identical to that of the wild-type receptor. DISCUSSION In this report, the relationship between chemoattractant receptor phosphorylation and the desensitization of down-stream signaling was investigated. Although it has long been known that stimulation of the FPR and other chemoattractant receptors results in desensitization of neutrophil functions, the mechanisms responsible for this desensitization are poorly understood (25,26). Studies with membranes from fMLF-and C5a-desensitized leukocytes originally revealed that coupling between the FPR and G proteins was impaired under conditions of desensitization (27,28). More recently, it has been demonstrated that desensitization of the FPR by C5a or IL-8 can occur in the absence of FPR phosphorylation, resulting in decreased inositol 1,4,5-trisphosphate generation and calcium mobilization without desensitization of G protein activation (16). Such results have led to the conclusion that desensitization of peptide chemoattractant receptors occurs downstream of G protein activation, possibly at the level of phospholipase C activation. In addition to this mechanism of desensitization, protein kinase C-mediated receptor phosphorylation, in the case of the C5a and IL-8 receptors, although not the FPR, appears to ameliorate receptor signaling. Since the FPR and the C5a and IL-8 receptors utilize similar if not identical signal transduction pathways, any downstream desensitization initiated by one of these receptors should be similarly initiated by the others. Furthermore, downstream desensitization must also be present under the conditions of homologous desensitization if the activating peptide chemoattractant receptors cannot be "distinguished between." The results presented in this paper demonstrate that the mechanisms of desensitization are not this simple. By completely preventing phosphorylation of the FPR with the ⌬ST mutant, not only was homologous desensitization of G protein activation abolished, but so was the downstream calcium mobilization response. This result demonstrates that activation by the FPR does not lead to the activation of a redundant downstream desensitization mechanism prior to the site of calcium mobilization (16). Two possible reasons exist to explain this result. First, it may be that the FPR does not activate the downstream desensitization machinery, as do the IL-8 and C5a receptors, or second, that it can circumvent the effects of this process under the conditions of homologous desensitization. If the former were true, then a model in which FPR signaling is different from IL-8 and C5a signaling would need to be invoked. To date, there is no evidence to substantiate this idea. If the latter were true, then a model in which heterologous desensitization can be distinguished from homologous desensitization would need to be created. In such a model, only when the second stimulus came from a receptor different from the first would the cell exhibit desensitization.
Further details of the mechanism of desensitization were revealed by the characteristics of the two partially phosphorylation-defective mutants, A and B. Mutants A and B both exhibited significant reductions (ϳ50%) in the level of phosphorylation following fMLF stimulation. The discrepancy between the levels of phosphorylation of mutant A in vivo (ϳ50%) and in vitro (ϳ20%) may be a result of the influence of other regions of the FPR or of accessory proteins, such as arrestin-related proteins. Although phosphorylation of purified rhodopsin has been shown to proceed to a stoichiometry as high as 7-9 mol/mol receptor (29), arrestin was found to limit phosphorylation to 1-3 mol/mol rhodopsin (30,31). Arrestin has also been found to promote the initial phosphorylation of rhodopsin (30). Thus, it is unclear exactly what effect on phosphorylation would be expected after substitution of Ser and Thr residues. In fact, if an arrestin homologue were to bind preferentially to the FPR when phosphorylated at site A as opposed to site B, it might be predicted that site B would become hyperphosphorylated in mutant A when compared with the wild-type FPR or mutant B. The effects of the mutations on desensitization were clear, however. Mutant A as well as mutant B, to a slightly lesser extent, exhibited a decrease in the level of desensitization of G protein activation, suggesting that phosphorylation within both these sites is necessary for complete desensitization. However, when calcium mobilization was evaluated, mutant A demonstrated a complete lack of desensitization, identical to that seen with ⌬ST, whereas mutant B exhibited only a partial defect in the ability to undergo desensitization. These results indicate that a difference exists in the ability of the two sites to promote desensitization. Phosphorylation of site A appears to be critical in desensitizing the FPR since although mutant A demonstrated ϳ50% the level of phosphorylation as compared with the wild-type receptor, phosphorylation of site B resulted in no desensitization. Phosphorylation of only site A can result in partial desensitization of the FPR, as observed with mutant B. This desensitization was incomplete, however, indicating that for complete desensitization, residues within both sites A and B must be phosphorylated. These results are consistent with a sequential model of phosphorylation previously proposed, where residues within site A are phosphorylated, first resulting in partial desensitization with subsequent phosphorylation of residues within site B, leading to complete desensitization.
The results presented here demonstrate that the recombinant FPR expressed in U937 cells undergoes ligand-stimulated phosphorylation and desensitization as it does in the neutrophil. Furthermore, it was demonstrated that receptor phosphorylation is required for desensitization of immediate responses such as G protein activation as well as downstream responses such as calcium mobilization. The proposed model, in which chemoattractant-induced desensitization of calcium mobilization occurs in the absence of receptor phosphorylation, appears inconsistent with the results presented here. Further studies will be necessary to determine if perhaps homologous and heterologous desensitization mechanisms involve different pathways.