FcgammaRI coupling to phospholipase D initiates sphingosine kinase-mediated calcium mobilization and vesicular trafficking.

Aggregation of receptors specific for the constant region of immunoglobulin G activates a repertoire of monocyte responses that can lead ultimately to targeted cell killing via antibody-directed cellular cytotoxicity. The high affinity receptor, FcgammaRI, contains no recognized signaling motif in its cytoplasmic tail but rather utilizes the gamma-chain of FcepsilonRI as an accessory molecule to recruit tyrosine kinases for signal transduction. We show here that, in a human monocytic cell line primed with interferon-gamma, FcgammaRI mobilizes intracellular calcium stores using a novel pathway that involves tyrosine kinase coupling to phospholipase D and resultant downstream activation of sphingosine kinase. Moreover, FcgammaRI is not coupled to phospholipase C; hence, calcium release from intracellular stores occurred in the absence of any measurable rise in inositol triphosphate. Finally, as this novel activation pathway is also shown to be responsible for mediating the vesicular trafficking of internalized immune complexes for degradation, it is likely to play a key role in controlling intracellular events triggered by FcgammaRI.

The macrophage-specific receptor (Fc␥RI) 1 for the constant region (Fc) of IgG plays a central role in the clearance of immune complexes (1,2). Fc␥RI belongs to a family of receptors for IgG that are distinguished by the affinity for ligand. While Fc␥RI is a high affinity IgG receptor, Fc␥RII and Fc␥RIII are both low affinity IgG receptors (reviewed in Refs. 1 and 2). Aggregation of Fc␥RI activates macrophages to undergo a repertoire of responses that can ultimately lead to cell killing through the process of antibody-directed cellular cytotoxicity, a critically important feature in the body's defense against virusinfected cells and in cancer surveillance (3,4). Immune complex aggregation of Fc␥RI initiates signal transduction events, which include protein tyrosine phosphorylation (5,6) and tyrosine kinase-dependent calcium transients (7,8). However, the cDNA for Fc␥RI predicts an integral type I glycoprotein in which, unlike Fc␥RIIa, the cytoplasmic tail contains no recognized signaling motifs (9). Fc␥RI has been shown to associate noncovalently with the signal-transducing ␥-chain (10), which contains an immunoreceptor tyrosine activation motif (11,12) in its cytoplasmic tail, and this association is thought to allow aggregated Fc␥RI to recruit and activate soluble tyrosine kinases (13). The ␥ chain was originally identified in mast cells as a component of the high affinity IgE receptor, Fc⑀RI, but has subsequently been found in macrophages in the absence of the ␣-chain of Fc⑀RI (14). Thus, although expressed in different cell types, the ligand recognition subunits (␣-chains) of Fc␥RI and Fc⑀RI are able to use the same signal-transducing molecule. Recently, Fc⑀RI has been shown to mobilize calcium transients in a mast cell line through the activation of a novel pathway involving sphingosine kinase (15). However, the precise details of the signaling pathway and its relationship to tyrosine kinase activation are as yet unclear.
In this study, we demonstrate that Fc␥RI mobilizes calcium from intracellular stores by activating sphingosine kinase in the absence of phospholipase C activation and resultant generation of inositol 1,4,5-triphosphate (InsP 3 ). We also show that Fc␥RI-stimulated activation of sphingosine kinase is downstream of phospholipase D activation and that both these enzymes are dependent on tyrosine kinase activation. Moreover, activation of this pathway is necessary and sufficient to account for intracellular calcium mobilization after Fc␥RI aggregation in cytokine-primed U937 cells and for efficient vesicular trafficking of internalized immune complexes for degradation.

MATERIALS AND METHODS
Receptor Aggregation-U937 cells, a human monocyte cell line (16), treated with 200 ng/ml interferon-␥ for 18 h were used for all experiments (8,17). For the biochemical assays, approximately 3 ϫ 10 6 cells were harvested and incubated with 1 M human monomeric IgG (Serotec) to occupy surface Fc␥RI. Unbound IgG was removed by dilution and centrifugation of the cells. The cells were resuspended in ice-cold Hepesbuffered saline (HBS), and cross-linking antibody (goat anti-human IgG, 1:100 dilution) was added. The cells were then warmed to 37°C and harvested at specified times for biochemical assay. Where the low affinity receptor was specifically aggregated using anti-Fc␥RIIa, the cells were loaded with the monoclonal antibody 2e1 (1 g) (Serotec) in the presence of saturating concentrations (3 M) of human IgG4 (to block binding of the Fc portion of 2e1 to Fc␥RI). After removal of excess antibody, anti-Fc␥RIIa was aggregated by the addition of goat antimouse IgG F(ab) (1:100 dilution).
Measurement of Sphingosine Kinase-Sphingosine kinase was assayed as described by Olivera et al. (18). Briefly, reactions were terminated at the times specified in the figures by the addition of ice-cold phosphate-buffered saline (PBS). After centrifugation, the cells were resuspended in ice-cold 0.1 M phosphate buffer (pH 7.4) containing 20% glycerol, 1 mM mercaptoethanol, 1 mM EDTA, phosphatase inhibitors (20 mM ZnCl 2 , 1 mM sodium orthovanadate, and 15 mM sodium fluoride), protease inhibitors (10 g/ml leupeptin, 10 g/ml aprotinin, and 1 mM PMSF), and 0.5 mM 4-deoxypyridoxine. Cells were disrupted by freeze thawing and centrifuged at 105,000 ϫ g for 90 min at 4°C. Supernatants were assayed for sphingosine kinase activity using sphingosine (Sigma) and [␥ 32 P]ATP (2 Ci, 5 mM) as specified by Olivera et al. (18). After incubation, products were separated by TLC on silica gel G60 using chloroform/methanol/acetic acid/water (90:90:15:6) and visualized by autoradiography. The radioactive spots corresponding to sphingosine phosphate were scraped and counted in a scintillation counter.
Measurement of Sphingosine 1-Phosphate-Sphingosine 1-phosphate concentrations were measured as described by Olivera and Spiegel (19). Briefly, cells were preincubated overnight (15 h) in media containing [ 3 H]serine (20 Ci ml) to label cellular sphingolipids and free sphingosine pools. Following labeling, the cells were washed in ice-cold RPMI 1640, 10 mM HEPES, 0.1% bovine serum albumin (RHB medium) and resuspended in ice-cold RHB medium containing 0.1 mM L-canaline and the pyridoxal phosphate analog 4-deoxypyridoxine (0.5 mM) to inhibit the pyridoxal-dependent sphingosine-1-phosphate lyase. Cells were then stimulated by the addition of cross-linking antibody and warming Cross-linking antibody (XL) was added to the cuvette at 300 s, and thapsigargin was added at 475 s. Thapsigargin (250 nM) was added to assess the viability of the stores. A typical trace from 5 separate experiments is shown. B, activation of sphingosine kinase by Fc␥RI aggregation and the effect of tyrosine kinase inhibition. IgG-loaded Fc␥RI was aggregated by the addition of cross-linking antibody (Cross link (XL)), and sphingosine kinase was assayed in cell extracts at specified time points after aggregation. Results were compared with non-cross-linked controls (No Cross-link Control) and to cells pretreated with genistein (0.37 mM) for 30 min prior to the addition of cross-linking antibody to inactivate tyrosine kinases (XL ϩ genistein). Results shown are the mean Ϯ S.D. for triplicate measurements at each time point. The results shown are typical from three separate experiments. C, effect of varying concentrations of genistein on sphingosine kinase activation by Fc␥RI aggregation. Sphingosine kinase activity was measured in cells 30 s after aggregating Fc␥RI in cells pretreated for 30 min with varying concentrations of genistein (0.01, 0.03, 0.1, 0.3, and 1.0 mM) and compared with untreated control cells. Results shown are the mean Ϯ S.D. for triplicate measurements at each concentration. The results shown are typical from three separate experiments. D, increase in sphingosine 1-phosphate concentrations following Fc␥RI aggregation and effect of tyrosine kinse inhibition. Sphingosine 1-phosphate concentrations were measured in cells following aggregation of Fc␥RI (Cross link (XL)) and compared with non-cross-linked control cells (No Cross-link Control) and to cells pretreated with genistein (0.37 mM) for 30 min prior to the addition of cross-linking antibody (XL ϩ genistein). Results shown are the mean Ϯ S.D. for triplicate measurements at each time point.The results shown are typical from three separate experiments.
Fc␥RI Coupling to Phospholipase D and Sphingosine Kinase to 37°C, and the reactions were terminated at specified times. Cells were harvested by centrifugation, and the lipids were extracted and analyzed by TLC on silica gel G60 using chloroform/methanol/acetic acid/water (90:90:15:6). Standard sphingosine 1-phosphate was applied with the samples, and the lipids were visualized using iodine vapors. Bands corresponding to sphingosine 1-phosphate were excised from the plate and counted by liquid scintillation spectrometry. Results were calculated as a percentage of the total radioactivity incorporated in the lipids. Data presented are the mean Ϯ S.D. of triplicate measurements, and the results shown are representative of three different experiments.
Measurement of Inositol Phosphate and DAG Generation-Inositol phosphates were assayed essentially as described by Harnett and Harnett (20). Briefly, U937 cells were labeled with myo-[ 3 H]inositol (1 Ci/10 6 cells) for 16 h at 37°C. The cells were washed three times and resuspended (at 1-3 ϫ 10 7 cells/ml) in RHB medium, pH 7.4, at 4°C. Following stimulation, the cells were harvested, resuspended in 100 l of HBS, transferred to glass trident vials, and extracted by the addition of 0.94 ml of chloroform/methanol (1:2) on ice for 10 min. A Bligh-Dyer phase separation was achieved by the addition of 0.31 ml of chloroform and 0.31 ml of water, vortexing, and centrifugation at 270 ϫ g for 5 min. Levels of [ 3 H]InsP 3 or total [ 3 H]inositol phosphates (reaction mixture containing 10 mM LiCl) were determined by liquid scintillation counting of fractions eluted following Dowex (formate form) ion exchange chromatography of aliquots of the aqueous phase. Results were calculated as a percentage of the total radioactivity incorporated in the lipids. Data presented are the mean Ϯ S.D. of triplicate measurements, and the results shown are representative of three different experiments.
DAG Assay-Mass DAG was measured as described by Briscoe et al. (21). The lower organic phase of Bligh-Dyer extractions was dried in vacuo, and the lipids were solubilized in a Triton X-100/phosphatidylserine mixture. Briefly, phosphatidylserine (30 l; supplied as 25 mM stock from Lipid Products) was dried under nitrogen and then probesonicated in 2.5 ml of 10 mM imidazole buffer, pH 6.6, containing 0.6% (w/v) Triton X-100, until the solution was optically clear. Aliquots (50 l) were added to the lipid samples, which were then sonicated in a bath for 30 min. Once sonicated, 20 l of 250 mM imidazole buffer, pH 6.6, containing 250 mM NaCl, 62.5 mM MgCl 2 , 5 mM EGTA, and 10 l of freshly prepared 100 mM dithiothreitol was added to the solubilized lipid. Escherichia coli diacylglycerol kinase (Calbiochem) was added to a final concentration of 50 milliunits/ml, and the reaction was started by the addition of 10 l of 5 mM ATP containing 1 Ci of [␥-32 P]ATP made up in 100 mM imidazole, pH 6.6; this results in a final ATP concentration of 0.5 mM in a final reaction volume of 100 l. The tubes were incubated at 30°C for 30 min. The reaction was stopped by the addition of 1 ml of chloroform/methanol/HCl (150:300:2). After 10 min, 300 l of chloroform and 400 l of H 2 O are added. The tubes are vortexed and centrifuged at 270 ϫ g for 5 min to promote phase splitting and washed once with 1 ml of a synthetic upper phase. The samples were then dried in vacuo and solubilized in 40 l of chloroform/methanol (19:1), and 20 l was spotted onto silica TLC plate (Whatman catalog no. 4861720, 10 ϫ 20 cm K6F). The plates were developed in chloroform/methanol/acetic acid (38:9:4.5), and radiolabeled bands were located by autoradiography or phosphor imaging. The phosphatidic acid (PtdOH) band (relative to standards) was scraped into scintillation vials, scintillant was added, and the associated radioactivity was determined by liquid scintillation counting.
Measurement of Phospholipase D (PLD) Activity-PLD activity was measured by the transphosphatidylation assay (21). Briefly, U937 cells were labeled (10 6 cells/ml) with [ 3 H]palmitic acid (5 Ci/ml) in RPMI 1640 medium containing 10% (v/v) fetal calf serum for 16 h. Following labeling, the cells were washed in ice-cold RHB medium, resuspended at 2 ϫ 10 6 cells/ml, and incubated at 37°C for 15 min in RHB medium containing butan-1-ol (0.3% final). Specific Fc receptors were crosslinked as described above, and after the times indicated, cells were extracted by Bligh-Dyer phase separation. An aliquot of the lower organic phase was removed and dried down under vacuum (Jouan, catalog number RC1022), and the samples were redissolved in 25 l of chloroform/methanol (19:1, v/v), containing 40 g of unlabeled phosphatidylbutanol (Lipid Products, South Nutfield, Surrey, UK) as standard, and applied to prerun, heat-activated TLC plates (20 ϫ 20 cm, Silica gel 150A grooved plates, Whatman). The plates were developed in the organic phase of the solvent, ethyl acetate/2,2,4-trimethylpentane/ acetic acid/water (11:5:2:10) for approximately 90 min, and the position of the phosphatidylbutanol product was detected using iodine vapor.
[ 3 H]PtdBut-containing silica indicated by the phosphatidylbutanol standard was then scraped into scintillation fluid and counted. Results were calculated as a percentage of the total radioactivity incorporated in the lipids. Data presented are the mean Ϯ S.D. of triplicate measurements, and the results shown are representative of three different experiments.

Measurement of Tyrosine Phosphorylation by Western Blot-U937 cells were loaded with human IgG and cross-linked as described earlier.
After washing in PBS, the cells were lysed with ice-cold radioimmune precipitation lysis buffer containing 1 mM phenylmethylsulfonyl fluoride, 1 g/ml CLAP (1 mg/ml each of chymostatin, leupeptin, antipain, and pepstatin), 1 mM sodium orthophosphate, and 1 mM sodium fluoride for 30 min. Cellular debris was removed by centrifugation at 13,000 rpm for 15 min, and the cell lysates were incubated with an agaroseconjugated anti-phosphotyrosine monoclonal antibody (clone 4G 10; Upstate Biotechnology, Inc.) at 4°C overnight. Phosphotyrosine proteins were then harvested by centrifugation of the agarose beads and were then dissociated from the beads by boiling in sample buffer (22) containing 50 mM dithiothreitol for 15 min. Samples were run in a 10% SDS-polyacrylamide gel (23). After electrophoresis, the proteins were transferred to a nitrocellulose membrane (0.2-m pore size) as described by Towbi et al. (24). The presence of tyrosine-phosphorylated proteins was then detected by Western blotting with a monoclonal anti-phosphotyrosine antibody (clone 4G 10; Upstate Biotechnology). Western blots were developed using the ECL system (Amersham Pharmacia Biotech).
Measurement of Cytosolic Calcium-Cytosolic calcium was measured in cell populations at 37°C using a Cairn Research Spectrophotometer as described previously (8). Cells were loaded with Fura2 and human monomeric IgG in HBS supplemented with 1 mM Ca 2ϩ to prevent depletion of calcium stores. After dilution and centrifugation to remove excess dye and antibody, the cells were resuspended in a small volume of HBS, 1 mM Ca 2ϩ to give a final density of 10 6 cells/100 l. From this, cells were added to stirred cuvettes containing 1.4 ml of nominally Ca 2ϩ -free HBS (at 37°C) in a Cairn Spectrophotometer system (Cairn Research Ltd.). Excitation wavelengths of 340, 360, and 380 nm were provided by a filter wheel rotating at 35 Hz in the light path. Emitted light was filtered by a 485-nm-long pass filter, and samples were averaged to give a data point every 500 ms. The background-corrected 340/380 ratio was calibrated using the method of Grynkiewicz et al. (25). Following each experiment, cells were lysed by the addition of 50 M digitonin in the presence of external 2 mM Ca 2ϩ to give an R max value. R min was subsequently determined by the addition of 20 mM EGTA (pH 7.4) in the presence of an equimolar concentration of Tris base.
Measurement of Endocytosis and Rate of Trafficking for Degradation-Interferon-␥ (IFN-␥)-treated cells were harvested and washed in phosphate-buffered saline (PBS), 1% bovine serum albumin. The cells were then loaded with 125 I-labeled IgG as described previously (17). After removal of nonbound radiolabel by dilution and centrifugation, cross-linking antibody was added, and the cells were warmed to 37°C for the given times.
Endocytosis-The rate of endocytosis was assessed by measuring the rate of internalization of radiolabeled surface immune complexes. At time 0, triplicate aliquots of cells were harvested into ice-cold PBS (pH 7.4), and this was counted in a Packard ␥-counter to provide a measure of the total counts bound to the cell surface. To measure the proportion of radiolabeled immune complexes internalized after incubation at 37°C, any surface-bound radiolabeled immune complexes can be stripped from the cell by incubating the cells in ice-cold acidified PBS (pH 2.0) (17). Radiolabeled immune complexes that have been internalized remain trapped inside the cell and cannot be released by this acid wash. Thus, to assess the rate of internalization, aliquots of cells were transferred at given times into ice-cold acidified PBS (pH 2.0) for 5 min to strip off cell surface radiolabeled immune complexes (17). The cells were then centrifuged, and the pellets were counted in a Packard ␥-counter to yield the counts that had been internalized, or the cellassociated counts. The cell-associated counts for each time point were then expressed as the percentage of total counts bound at time 0 to provide a measure of the rate of internalization of the immune complexes.
Degradation-After warming the cells to 37°C for long time intervals, the proportion of cell-associated counts was observed to fall. To determine whether this reduction in cell-associated counts represented degradation of the immune complexes, the supernatant following the cell incubation was examined for the presence of trichloroacetic acidsoluble radiolabel indicating that the radiolabeled IgG had been degraded. Thus, cells were also harvested at the same time points to measure the rate of degradation of the internalized counts. Cells were centrifuged, the supernatants were harvested, and trichloroacetic acid was then added to these supernatants. After incubation on ice for 60 FIG. 2. Phospholipase D and not phospholipase C is activated by aggregation of Fc␥RI in IFN-␥-primed cells. A, total InsPs accumulation over 20 min following aggregation of Fc␥ receptors. Inositol phosphate accumulation over 20 min was measured following the min, the samples were centrifuged at 12,000 ϫ g at 4°C, and the supernatants were counted to provide a measure of the trichloroacetic acid-soluble counts in the supernatant. The results were expressed as a percentage of the initial cell-associated counts at time 0.
The results shown are the mean Ϯ S.D. of triplicate measurements and are representative of three different experiments.

RESULTS
Aggregation of Fc␥RI Activates Sphingosine Kinase in a Tyrosine Kinase-dependent Manner-In IFN-␥-primed U937 cells, aggregation of Fc␥RI with surface-bound immune complexes results in calcium transients in the form of a single spike (see Ref. 8 and Fig. 1A). The Fc␥RI-associated accessory transducing molecule, ␥-chain, has recently been reported to mobilize calcium via activation of sphingosine kinase when coupled to the high affinity IgE receptor, Fc⑀RI (15). Thus, to compare the nature of this Fc␥RI-calcium response to that of Fc⑀RI, the effect of DL-threo-dihydrosphingosine (DHS) on the release of calcium from intracellular stores was determined. Pretreatment of cells with 25 M DHS completely abolished the Fc␥RImediated rise in cytosolic calcium, indicating that intracellular calcium stores are mobilized in these cells in a similar fashion to that observed for Fc⑀RI in mast cells (15). The calcium stores were intact in cells treated with DHS, since the subsequent addition of thapsigargin (250 nM) resulted in a prompt increase in cytosolic calcium, thereby demonstrating that the failure to observe a rise in calcium following aggregation of Fc␥RI in cells pretreated with DHS was not secondary to depletion of intracellular calcium stores.
Since DHS acts as a competitive inhibitor of sphingosine kinase, the activity of this enzyme after Fc␥RI aggregation was next assessed. Aggregation of Fc␥RI stimulated a prompt increase in sphingosine kinase activity, which was detectable within 30 s (Fig. 1B). Sphingosine kinase activation by Fc␥RI aggregation in these cytokine-primed cells was dependent on tyrosine kinase activation, since treatment of the cells with genistein (0.37 mM) completely abolished the response (Fig.  1B). Pretreatment with genistein at a lower concentration (0.1 mM) also completely inhibited Fc␥RI activation of sphingosine kinase (Fig. 1C), although concentrations below this only resulted in partial inhibition (Fig. 1C).
In parallel with the activation of sphingosine kinase, Fc␥RI aggregation resulted in a prompt increase in the concentration of sphingosine 1-phosphate in these cells (Fig. 1D). The concentration of sphingosine 1-phosphate peaked 30 s after receptor aggregation, and although levels fell gradually thereafter, concentrations remained elevated above control values 5 min after receptor aggregation. Pretreatment of cells with genistein (0.37 mM) completely abolished the Fc␥RI-mediated increase in sphingosine 1-phosphate generation.

Aggregation of Fc␥RI Activates Phospholipase D and Not
Phospholipase C in a Tyrosine Kinase-dependent Manner-Since immune complex aggregation of Fc␥RI has previously been reported to lead to tyrosine phosphorylation of phospholipase C␥1 (5) with presumed generation of InsP 3 and DAG, the role of this phospholipid signaling pathway in mediating the cytosolic calcium response was also investigated. Surprisingly, no increase in InsP 3 could be detected (data not shown). Since InsP 3 generation can be transient in nature, the accumulation of total inositol phosphates (InsPs) was measured to ensure that any small transient InsPs signals did not go undetected.
No accumulation of total InsPs over 20 min could be detected in IFN-␥-primed U937 cells after aggregation of Fc␥RI ( Fig. 2A). Phospholipase C signaling was, however, functional in these cells, since aggregation of a related immune receptor, the low affinity IgG receptor (Fc␥RIIa), using monoclonal antibodies resulted in an easily measurable accumulation of InsP 3 (data not shown) and total InsPs ( Fig. 2A). Unlike Fc␥RI, the low affinity receptor possesses an integral, albeit unconventional, immunoreceptor tyrosine activation motif in its cytoplasmic tail; the tyrosine residues are separated by an unusually long intervening sequence (26). Taken together, these data indicate that the high affinity receptor, Fc␥RI, mobilizes calcium stores through a novel pathway that, unlike the low affinity receptor (Fc␥RIIa), does not involve InsP 3 . Interestingly, despite the lack of generation of InsPs over 20 min, mass DAG concentrations were elevated following aggregation of Fc␥RI (Fig. 2B). Thus, in an attempt to delineate alternative lipid signaling pathways involved in mediating the response to Fc␥RI, DAG was measured in the presence of 0.3% butan-1-ol to block the generation of DAG derived from PtdOH resulting from PLD activation (21). Under these conditions, the primary alcohol, butan-1-ol, traps the phosphatidyl-moiety generated by PLD-mediated hydrolysis of phosphatidylcholine as PtdBut; PtdBut is not a substrate for the enzyme, phosphatidic acid phosphohydrolase, that converts PtdOH to DAG (21). Fc␥RI-coupled DAG was indeed shown to be derived from Pt-dOH generated by phospholipase D activation, since pretreating cells with 0.3% butan-1-ol completely abolished the receptor-stimulated rise in mass levels of DAG (Fig. 2B).
Activation of phosphatidylcholine-specific phospholipase D (PtdCho-PLD) following aggregation of Fc␥RI was demonstrated by the definitive transphosphatidylation assay (21). These experiments showed that aggregation of Fc␥ receptors in IFN-␥-primed, [ 3 H]palmitate-labeled cells stimulated activation of PtdCho-PLD, as evidenced by substantial generation of A, sphingosine kinase activity following Fc␥RI aggregation in cells treated with butan-1-ol. Cells were preincubated with 0.3% butan-1-ol and harvested at given times after aggregation of Fc␥RI is trapped at the lower concentration of primary alcohol. Lower concentrations of butan-1-ol (0.1%) resulted in less measurable [ 3 H]PtdBut. Thus, the optimal concentration of butan-1-ol to trap the phosphatidyl moiety is 0.3% (Fig. 2D). The specificity of the measurement was confirmed using butan-2-ol, which is unable to trap the phosphatidyl moiety generated by PtdCho-PLD. No [ 3 H]PtdBut could be detected in cells preincubated with butan-2-ol even at 1% preincubation.
The increase in PtdCho-PLD activity following receptor aggregation was tyrosine kinase-dependent, since it was completely abolished by treating the cells with genistein. Moreover, the concentration dependence of genistein-mediated inhibition of PtdCho-PLD (Fig. 2E) showed a similar profile to that obtained for sphingosine kinase coupling (Fig. 1C).
Thus, Fc␥RI is coupled through tyrosine kinases to the activation of PtdCho-PLD and sphingosine kinase.
Activation of Sphingosine Kinase Is Downstream of Phospholipase D Activation-To assess the relative relationship of activation of PtdCho-PLD and sphingosine kinase, studies were initially undertaken to explore comparative kinetics of activation and use of selective inhibitors. However, comparison of the relative kinetics was complicated by the difference in the assay characteristics for measuring sphingosine kinase and phospholipase D. Thus, sphingosine kinase is measured as an in vitro kinetic kinase assay, whereas the assay for phospholipase D relies on the accumulation of a nonhydrolyzable product. The difference in assay characteristics, therefore, precluded definitive early comparative time course analysis. The relationship of phospholipase D and sphingosine kinase activation was therefore addressed by examining selective inhibitors of the two enzymes.
To determine whether sphingosine kinase activation was upstream or downstream of phospholipase D, cells were preincubated with butan-1-ol or butan-2-ol for 20 min before aggregation of Fc␥RI, and the resultant sphingosine kinase activity was compared with that of control cells. Pretreating cells with butan-1-ol (0.3%) completely abolished the normal sphingosine kinase response to Fc␥RI aggregation (Fig. 3A). Consistent with these results, the rise in sphingosine 1-phosphate observed after aggregation of Fc␥RI was also blocked by pretreating cells with butan-1-ol (0.3%) (Fig. 3B).
These data suggested that activation of sphingosine kinase is dependent on the activation of phospholipase D and generation of PtdOH. To assess this in detail, the effect on peak (30 s after receptor aggregation) sphingosine kinase activity of preincubating cells with varying concentrations of butan-1-ol previously shown to influence phospholipase D was examined, and the effects of these concentrations were compared with the same concentrations of butan-2-ol, which does not influence phospholipase D (Fig. 2D). Peak activity of sphingosine kinase following Fc␥RI aggregation was abolished by incubating cells with either 0.3 or 1.0% butan-1-ol but was completely unaf-fected by preincubation with butan-2-ol even at the highest concentration (Fig. 3C). Incubation of cells with a lower concentration of butan-1-ol (0.1%) partially inhibited peak sphingosine kinase activity. Of interest, 0.1% butan-1-ol resulted in lower concentrations of [ 3 H]Ptd-But (Fig. 2D), suggesting that, at this concentration, butan-1-ol is only able to trap a proportion of the phosphatidyl moiety generated by phospholipase D and that as a result some phosphatidic acid may be produced.
These data using butan-1-ol indicates that phospholipase D is upstream of sphingosine kinase. Consistent with this observation, DHS, a competitive inhibitor of sphingosine kinase, had no effect whatsoever on phospholipase D activation at all concentrations examined, even up to 100 M (Fig. 3D). The potency of DHS on sphingosine kinase was measured directly; DHS at concentrations of 30 M and above completely abolished the peak sphingosine kinase activity observed after Fc␥RI aggregation; 10 M DHS inhibited peak sphingosine kinase activity by about 75% (Fig. 3E).
Taken together, these data clearly indicate that activation of sphingosine kinase is secondary to activation of phospholipase D and generation of PtdOH.

Tyrosine Phosphorylation Is Triggered Promptly by Aggregation of Fc␥RI and Is Upstream of both Phospholipase D and
Sphingosine Kinase-Tyrosine phosphorylation events were monitored in these cytokine-primed U937 cells after aggregation of Fc␥RI by immunoprecipitating tyrosine-phosphorylated proteins with a monoclonal antibody to phosphotyrosine. Consistent with other reports (5-7, 27, 28), the addition of crosslinking antibody to form surface immune complexes resulted in the prompt appearance of a large number of tyrosine-phosphorylated proteins. Preincubating cells with either butan-1-ol (0.3%) or DHS (25 M) did not influence the pattern of tyrosine phosphorylation (Fig. 4), results consistent with our findings that both PtdCho-PLD (Fig. 2C) and sphingosine kinase (Fig.  1C) activation are downstream of tyrosine kinase activation.

Activation of Phospholipase D Is Necessary for both Mobilization of Intracellular Calcium and for Trafficking of Immune
Complexes for Degradation-The release of intracellular stores of calcium by aggregation of Fc␥RI was significantly inhibited by pretreating the cells with 0.3% butan-1-ol (Fig. 5A), thus providing further support for the role of this pathway in mobilizing calcium and the concept that PtdCho-PLD is upstream of sphingosine kinase. The possibility that butan-1-ol affected calcium mobilization through nonspecific effects was ruled out, since the subsequent addition of thapsigargin (250 nM) resulted in a prompt response in cytosolic calcium. In addition, butan-1-ol had no effect on the InsP 3 -dependent mobilization of calcium following aggregation of the related low affinity receptor, Fc␥RIIa (Fig. 5A). The difference in release of calcium after thapsigargin is not likely to be significant following manual injection as undertaken here. Although the speed of calcium release by thapsigargin can be influenced by a number of in- tracellular factors such as the amount of calcium in the stores, it is well recognized that there is considerable variability between runs for thapsigargin-mediated calcium release, and the largest influence is the rate of addition of thapsigargin and its mixing in the cuvette (29,30).
As observed previously (17), the internalization of surfacebound immune complexes is very rapid in IFN-␥-treated U937 cells. The cell-associated counts plateau between 15 and 30 min. However, over prolonged incubations, the internalized cell-associated counts were found to diminish gradually in the control cells such that, by 120 min, approximately 50% of the cell-associated counts had been lost. This reduction was entirely matched by the appearance of counts in the culture media of the cells, and these counts were not precipitable by trichloroacetic acid (TCA) (Fig. 5B). The rate of appearance of these trichloroacetic acid-soluble counts in media is an indication of the rate of lysosomal degradation of the radiolabeled immune complexes (31) and is, therefore, a sensitive measure of the rate of intracellular trafficking of internalized immune complexes from endosomes to lysosomes.
Pretreatment of cells with butan-1-ol (0.3%) to inhibit PtdOH generation appeared to reduce to a small extent the initial phase of endocytosis (peak percentage of surface-bound counts internalized for control cells, 93 Ϯ 3%; butan-1-ol-treated cells, 77 Ϯ 2%). Nonspecific effects of the alcohol were eliminated, since butan-2-ol (0.3%) had no effect on the rate of endocytosis (peak percentage of surface bound counts internalized butan- 2-ol-treated cells, 90 Ϯ 3%). Following longer time intervals after internalization of immune complexes, the amount of radiolabel trapped inside the cells gradually decreased in the untreated cells and in cells treated with butan-2-ol (0.3%) such that about 50% of the initial internalized radiolabel had been lost from the cells after 2 h of incubation. This loss of cellassociated counts was entirely matched by the appearance in the cell supernatant of radiolabel in a form that was soluble in trichloroacetic acid. Thus, after 2 h, 47 Ϯ 2% of the initial counts in the control cells appear as trichloroacetic acid-soluble counts within the supernatant; this is a measure of trafficking of immune complexes for lysosomal degradation (31). Treatment of cells with butan-1-ol significantly inhibited trafficking of immune complexes for degradation. Thus, the rate of loss of cell-associated (internalized) counts was significantly slowed in cells treated with butan-1-ol (after a 2-h incubation, the percentage of counts remaining internalized for control cells was 43 Ϯ 5%; for butan-1-ol-treated cells, it was 63 Ϯ 3%). In addition, the rate of appearance of trichloroacetic acid-soluble counts in the media over prolonged incubations was significantly slower for cells pretreated with 0.3% butan-1-ol compared with the control untreated cells or cells treated with butan-2-ol (Fig. 5B). Thus, following 120 min of incubation, only 19 Ϯ 0.5% of counts appeared as trichloroacetic acidsoluble counts in the media of cells treated with butan-1-ol in contrast to approximately 45 Ϯ 1% for the control cells and 40 Ϯ 0.4% for cells treated with butan-2-ol.
Consistent with the biochemical data defining the signaling pathway, the inhibition of sphingosine kinase with DHS (25 M) also significantly inhibited trafficking of immune complexes for lysosomal degradation (Fig. 5B). Thus, activation of this intracellular signaling pathway involving phospholipase D and sphingosine kinase is required for the appropriate trafficking of internalized immune complexes along the degradative pathway.

DISCUSSION
Taken together, these data indicate that Fc␥RI in cytokineprimed U937 cells is coupled through tyrosine kinase activation to a novel pathway responsible both for mobilizing calcium transients through an InsP 3 -independent route and for trafficking internalized immune complexes for degradation. This novel pathway involves the activation of PtdCho-PLD, in the absence of measurable activation of phospholipase C, and this is upstream of activation of sphingosine kinase, which generates sphingosine 1-phosphate.
Sphingosine 1-phosphate has been proposed previously to play a role in mobilizing calcium from intracellular stores (32)(33)(34). However, this proposal has proven highly controversial due to the presence of extracellular G protein-coupled receptors for sphingosine 1-phosphate (35,36), which are able to mobilize calcium through conventional InsP 3 receptor-dependent pathways. The recent cloning of the SCaMPER receptor (37) provides additional evidence that sphingoid derivatives are able to engage intracellular receptors and effect calcium release from stores independently of InsP 3 generation. The data presented here provide evidence for specific immune receptor triggering of this pathway in myeloid cells. Thus, aggregation of Fc␥RI resulted in the rapid activation of sphingosine kinase and consequent cellular increases in sphingosine 1-phosphate concentrations. In these same cells, neither product of phospholipase C activation could be detected; no accumulation of total InsPs could be measured even in the presence of lithium chloride to prevent breakdown. Moreover, the observed increase in DAG could be completely blocked by pretreatment of cells with butanol, indicating PtdCho-PLD rather than phospholipase C activation as the source of the DAG. In contrast, aggregation of an alternative immune receptor, Fc␥RIIa, on these cells, resulted in increases in both phospholipase C-dependent DAG and inositol phosphate generation, indicating that this pathway is intact and functional in these cells and that the assays used were potentially able to detect any such receptor-triggered changes.
Taken together, the data presented here suggests that the high affinity receptor, Fc␥RI, mobilizes intracellular calcium through this sphingosine kinase-dependent, InsP 3 -independent pathway. In this respect, Fc␥RI is behaving like the high affinity IgE receptor, Fc⑀RI, in mast cells (15). Of interest, both these receptors use the same signal-transducing molecule (␥chain) (10) to recruit soluble tyrosine kinases to mediate cellular activation. However, the mechanism of coupling of tyrosine kinases to sphingosine kinase activation following Fc⑀RI aggregation in mast cells was unclear (15). Here, we demonstrate that PtdCho-PLD is activated following aggregation of Fc␥RI in myeloid cells and that sphingosine kinase activation is dependent on PtdCho-PLD activation. The immediate product of Ptd-Cho-PLD is phosphatidic acid, and this is subsequently converted to DAG through the action of phosphatidic acid phosphohydrolase. Previous studies have shown that sphingosine kinase is activated by phosphatidic acid (38) and not by DAG (38), a product of both phospholipase D and phospholipase C. Our finding that sphingosine kinase is downstream of Ptd-Cho-PLD is, therefore, consistent with this in vitro work. Moreover, both components of this novel Fc␥RI-coupled intracellular signaling pathway involving the sequential activation of Ptd-Cho-PLD and sphingosine kinase depend on tyrosine kinase activation. This finding is consistent with previous in vitro studies demonstrating that v-Src can activate PLD (39).
Aggregation of Fc␥RI in myeloid cells triggers a number of effector functions. The novel intracellular signaling pathway demonstrated here appears to be functionally interactive/associated with these. Thus, previous studies have implicated phosphatidic acid in modulating neutrophil function, in particular by influencing the respiratory burst/NADPH oxidase cascade (40). In the study reported here, inhibiting this pathway at either the PtdCho-PLD or sphingosine kinase level reduced or abolished the ability of this receptor to mobilize calcium from intracellular stores. In addition, the inhibition of PtdCho-PLD significantly slowed the rate of trafficking of internalized immune complexes for degradation. Of interest, ADP-ribosylation factor plays a major role in regulating vesicular trafficking, and this small molecular weight G protein has also been demonstrated to regulate phospholipase D activity (41). The finding that Fc␥RI is coupled to the release of intracellular calcium stores and vesicular trafficking via a novel pathway that does not use InsP 3 has profound implications for the development of strategies for therapeutic intervention against differential myeloid responses to immune complexes.