INTRODUCTION

Phosphatidylinositol 3-kinase (PI3K)¹ has been shown to be an important mediator of intracellular signaling in mammalian cells (reviewed in Refs. 1, 2, 3). The major product of PI3K, PIP₃, is generated by phosphorylation of phosphatidylinositol 4,5-bisphosphate at the 3 position of the inositol ring. Formation of PIP₃ has been correlated with both cytoskeletal regulation and mitogenic signaling by growth factors (1, 2, 3). In human leukocytes, studies using PI3K inhibitors indicate that PIP₃ formation is a critical component of the signaling pathway leading from chemoattractant receptors to oxidant production by the NADPH oxidase (4, 5, 6, 7, 8, 9, 10). Formation of PIP₃ via PI3K was originally described in chemoattractant-stimulated human neutrophils (11), and this has remained one of the best characterized cellular systems in terms of PI3K activation by G protein-coupled receptors.

Two types of PIP₃-generating enzymes have been described in neutrophils. The ``classical'' form of the enzyme consists of a p110 catalytic subunit and a p85 regulatory subunit (12, 13, 14). This ubiquitous PI3K can be activated by a number of mechanisms in various cell types (1, 2, 3, 8). These mechanisms include binding of the enzyme to tyrosine-phosphorylated motifs of growth factor receptors via SH2 domains on the p85 subunit (1, 2, 15, 16, 17) and activation of the enzyme by binding of Src-family kinase SH3 domains to proline-rich domains on the p85 subunit (18). Additionally, small GTPases of the Ras and Rho families can stimulate enzyme activity (19, 20, 21, 22). Recently, a novel form of PI3K has been described in myeloid cells whose activity is directly regulated by G protein subunits (23, 24). This form of PI3K has now been cloned and shown to consist of a unique p110 catalytic subunit that lacks the p85 binding domain and therefore does not associate with the p85 subunit (25).

The N-formyl peptide chemoattractant receptor couples to cell activation via heterotrimeric pertussis toxin-sensitive G_i proteins (26). As a result of G_i activation by receptor, G subunits are released and can regulate a number of enzymatic activities, including the formation of IP₃ and diacylglycerol via activation of PLC- isoforms (27, 28, 29, 30). The apparently abundant (23) G-regulated PI3K would presumably be activated as well. An additional signaling pathway activated by this receptor utilizes the Lyn tyrosine kinase (7, 31). In N-formyl peptide-stimulated cells, Lyn physically and temporally associates with the classical p85/p110 form of PI3K (31). A similar situation has been observed with the B cell antigen receptor, where binding of Lyn through its SH3 domain to the p85 subunit directly activates PI3K (18). An important question that has not yet been resolved is the contribution of the Lyn-regulated PI3K pathway versus that of the G-regulated enzyme to the overall formation of PIP₃ in chemoattractant-stimulated neutrophils. This question is of importance for understanding the actual signaling mechanisms that couple these receptors to generation of active oxidants for the purpose of bacterial killing. In this study, we present data that indicate that a tyrosine kinase-dependent pathway presumably mediated via Lyn accounts for a majority of the PIP₃ formed in response to N-formyl peptide receptor activation.

EXPERIMENTAL PROCEDURES

Freshly isolated (31, 32) human neutrophils (>95% purity) were incubated in the presence or the absence of 30-200 ng/ml radicicol for 2.5 h or 100 µM genistein for 20 min at 37 °C. Cell viability was greater than 90% under these conditions (33). Lyn tyrosine kinase activity was determined after immunoprecipitation from lysates of cells stimulated for the indicated time with 1 µM fMetLeuPhe (fMLP) as described in Ref. 31. Shc phosphorylation was determined using the 4G10 anti-phosphotyrosine monoclonal antibody (Upstate Biotechnology, Inc.) as in Ref. 31.

Neutrophils were suspended at a concentration of 1 × 10⁸/ml in Buffer A (30 mM Hepes, pH 7.2, 110 mM NaCl, 10 mM KCl, 1 mM MgCl₂, 10 mM glucose) and 1 mCi/ml [³²P]orthophosphate (HCl-free, DuPont NEN) was added. The cells were incubated at 37 °C for 90 min and then washed three times with Buffer A. Labeled cells were treated ± 200 ng/ml radicicol for 2.5 h or 100 µM genistein for 20 min at 37 °C, and subsequently stimulated with 1 µM fMLP or buffer for the indicated times. The reaction was stopped by addition of 3 ml chloroform/methanol (1:2, v/v), followed by 4 ml chloroform/2.4 M HCl (1:1, v/v). The resulting organic lower phase was removed and the aqueous layer washed four times with 1 ml of chloroform. The combined organic phases were evaporated to dryness under N₂ and resuspended in 90 µl of chloroform for thin layer chromatography (11). 20 × 20 cm Silica Gel 60 plates (EM Science) impregnated with potassium oxalate were used for analysis of lipids, with development in chloroform/acetone/methanol/acetic acid/water (80:30:26:24:14, v/v/v/v/v). Radioactive spots were detected by autoradiography using Kodak X-Omat film, and total cellular PIP₃ quantified by both densitometry and an AMBIS B scanning system (San Diego, CA), with comparable results.

PIP₃ formed by PI3K associated with Lyn or p85 subunit immunoprecipitates was determined as described in Refs. 4 and 31 in the presence or the absence of inhibitors. The G-regulated PI3K was assayed using sonicated micelles containing 600 µM bovine liver phosphatidylethanolamine and 300 µM bovine liver phosphatidylinositol in a solution containing 40 mM NaHepes, pH 7.4/2 mM EGTA/1 mM dithiothreitol/0.2 mM EDTA/120 mM NaCl/5 mM MgCl₂/1 mM -glycerophosphate/50 µM sodium orthovanadate/1 mg/ml bovine serum albumin. Inhibitors were added in 2 µl of dimethyl sulfoxide. Reactions were initiated by the addition of 10 µM ATP with [-³²P]ATP at 3.0 µCi/assay and by transfer from 4 to 30 °C. Reactions were terminated after 30 min by addition of a 2:1 methanol/chloroform solution. Samples were centrifuged, and the lower organic phase containing the lipids was extracted with a solution containing 48% methanol/3% chloroform/0.5 M HCl/1 mM EDTA/10 mM tetrabutylammonium hydrogen sulfate. The lower organic phase was dried, suspended in 2:1 methanol/chloroform, and spotted onto TLC plates that were developed with methanol/chloroform/ammonia/water (10:7:1.5:2.5) for analysis of product formation. Quantitation was by a PhosphorImager, with results presented in arbitrary units.

Partial Purification of G-regulated PI3K

U937 cells were grown in 10-liter spinner flasks in RPMI 1640 with 1% fetal calf serum (UBI), 5 units/ml of penicillin, 4 µg/ml streptomycin, and 1 × lipid concentrate (Life Technologies, Inc.) at 37 °C to a density of 10⁶ cell/ml. Cells were harvested by centrifugation at 1000 × g for 20 min, suspended in phosphate-buffered saline, pH 7.4, containing 10 mM phenylmethylsulfonyl fluoride, and collected again by centrifugation for 20 min. The cells were suspended in 75 ml of a solution of 50 mM Tris-HCl, pH 7.5/2 mM EGTA/1 mM EDTA/1 mM dithiothreitol/350 mM sucrose plus a protease inhibitor mixture (2 µg/ml leupeptin, 10 µg/ml aprotinin, 10 µg/ml soybean trypsin inhibitor, 10 µM phenylmethylsulfonyl fluoride, 1 µg/ml pepstatin A, 10 µg/ml L-1-tosylamido-2-phenylethyl chloromethyl ketone, and 20 µg/ml 1-chloro-3-tosylamido-7-amino-2-heptanone) that was included in all of the buffers used throughout the preparation. The cells were disrupted in a Parr cell disruption bomb by rapid decompression after equilibration at 4 °C for 45 min with N₂ at 500 p.s.i. Unbroken cells and intact nuclei were removed by centrifugation at 700 × g for 20 min at 4 °C. The membranes were removed by centrifugation at 100,000 × g for 1 h at 4 °C. The cytosolic supernatant was frozen in liquid N₂ and stored at 70 °C until use.

A portion of the soluble fraction (10 ml at 3 mg/ml) was applied to a 1-ml Mono Q column (Pharmacia) equilibrated with Buffer B (50 mM Tris-HCl, pH 7.5/2 mM EGTA/0.2 mM EDTA/1 mM dithiothreitol/50 mM NaCl/0.05% Tween 20 and protease inhibitors). The column was washed with 60 ml of Buffer B; this was followed by elution with a gradient from 50 to 400 mM NaCl in Buffer B. Fractions were assayed in the presence and the absence of 300 µM bovine brain (G), purified according to Sternweis and Robishaw (34). Fractions containing the highest activity in the presence of were pooled, avoiding fractions that contained activity in the absence of . Peak -stimulated fractions were pooled and frozen in aliquots at 70 °C.

Neutrophils were washed once with Hanks' balanced salt solution, resuspended to 5 × 10⁶/ml in Hanks' balanced salt solution, and incubated with 5 µM indo-1 acetoxymethyl (1 mM stock in dimethyl sulfoxide) at 37 °C for 30 min. Cells were then washed and treated with radicicol, genistein, or buffer only exactly as described for PIP₃ measurements. Following a brief wash with the same medium, neutrophils were resuspended to 15 × 10⁶/ml in Hanks' balanced salt solution and maintained on ice until use. The cells were diluted 10-fold for assay and stimulated with 1 µM fMLP. Continuous fluorescent measurements (excitation wavelength, 340 nm) of calcium-bound and free indo-1 were made using an SLM 8000 photon counting spectrofluorimeter (SLM-Aminco) detecting at 400 and 490 nm, respectively. Intracellular free [Ca²⁺] during the initial phase of IP₃-dependent Ca²⁺ entry was determined according to the equation [Ca²⁺]_i = K_d(F  F_min)/(F_max  F) where K_d is 250 nM, F is the fluorescence ratio 400 nm/490 nm, F_max is the ratio at saturating Ca²⁺, and F_min is the ratio with no Ca²⁺ (i.e. excess EDTA).

RESULTS

It was previously established that human neutrophil N-formyl peptide receptors activate a signaling pathway that involves the Lyn tyrosine kinase (7, 31). Associated with Lyn immunoprecipitates in chemoattractant-stimulated neutrophils is the tyrosine-phosphorylated Shc adapter protein and the p85/p110 form of PI3K (31). The interaction of Lyn and Shc occurred through the SH2 domain of Shc and presumably required the chemoattractant-stimulated autophosphorylation of Lyn. In the present study, we observed that the tyrosine kinase inhibitor genistein at the lowest concentration exerting maximum inhibition of N-formyl peptide-stimulated oxidant production (33) almost totally blocked Lyn activity in response to fMLP (Fig. 1). Concomitantly, the tyrosine phosphorylation of Shc was also blocked by genistein (Fig. 1), consistent with the predicted need for Lyn kinase activity to mediate the interaction with and phosphorylation of Shc.

We determined whether the enhanced PI3K activity, which rapidly becomes associated with Lyn immunoprecipitates during leukocyte activation by chemoattractants (31), could be inhibited by tyrosine kinase inhibitors. As shown in Fig. 1, genistein effectively blocked the increases in PI3K activity observed in Lyn precipitates from stimulated neutrophils. Inhibition was constant over a 10-min time period after N-formyl peptide stimulation (data not shown). Vlahos and Matter (35) have previously established that p85/PI3K itself is not tyrosine-phosphorylated in human neutrophils, nor is it effectively inhibited directly by genistein at the concentration used in these studies (Ref. 36; data not shown). Thus, inhibition of Lyn tyrosine kinase activity is correlated with loss of Lyn-associated PIP₃ formation.

In order to address the question of the predominant source of total PIP₃ formed in response to stimulation of the N-formyl peptide receptor, we examined the effect of tyrosine kinase inhibitors on total PIP₃ formation. Because we detected a direct inhibitory effect of genistein on the G-regulated PI3K (see Table I), we utilized a structurally and mechanistically distinct inhibitor of tyrosine kinases, radicicol, which has no effect on either G-regulated PI3K (Table I) or p85 PI3K (data not shown). We confirmed that radicicol effectively inhibited activation of Lyn after fMLP-stimulation at concentrations from 30 to 200 ng/ml, with complete inhibition seen at 200 ng/ml (data not shown). This is similar to the effective doses reported in the literature for Src and Lyn kinase inhibition (37, 38, 39).

Table I. Effects of tyrosine kinase inhibitors on PI3K activity in vitro Assay Condition   +a Radicicol controlb 1254c 7914 +666 ng/ml radicicol 1426 7860 Genistein control 1728 8652 +100 µM genistein 1277 5514 +300 µM genistein 739 2769 a  600 µM subunits. b  Dimethyl sulfoxide vehicle controls were added. c  Phosphatidylinositol 3-phosphate was quantitated on TLC plates using a PhosphorImager, which quantifies data in arbitrary units. Activity was determined as described under ``Experimental Procedures.'' Assays were performed in duplicate, and the data are representative of two separate experiments each.

Analysis of [³²P]orthophosphate-labeled neutrophils showed the rapid stimulation of PIP₃ formation by fMLP, with peak formation observed at 30 s to 1 min post activation, as reported previously (4, 11). Formation of PIP₃ was effectively blocked at every time point by radicicol pretreatment (Fig. 2). Although fMLP increased PIP₃ levels by an average of 5-fold by 30 s and 4-fold by 1 min in untreated cells, in the presence of 200 ng/ml radicicol these increases were blunted by 64.3 ± 10 and 68.3 ± 3.5% (S.D.) respectively (n = 3). Radicicol had no significant effect on unstimulated cellular levels of PIP₃.

Tyrosine Kinase Inhibitors Do Not Block All G-mediated Signaling Pathways

In order to make certain that the tyrosine kinase inhibitors genistein and radicicol did not exert nonspecific effects on G-mediated signaling that could account for some of the block of PIP₃ formation, we measured fMLP-stimulated Ca²⁺ mobilization. It has been established that increases in intracellular Ca²⁺ by N-formyl peptides results from the mobilization of intracellular Ca² stores by inositol trisphosphate generated as a result of phospholipase C activation (8, 40, 41, 42). In human neutrophils, this occurs solely through G-regulated PLC  isoforms (27, 28, 29). As shown in Fig. 3, neither genistein nor radicicol significantly prevented the increases in intracellular Ca²⁺ induced by fMLP. Thus, these drugs did not exert any direct inhibitory effect on G function.

Genistein and radicicol do not block G-dependent Ca²⁺ mobilization.

Genistein but Not Radicicol Inhibits the G-regulated PI3K

In our initial studies, we had used genistein as a ``specific'' tyrosine kinase inhibitor (43) and observed that it caused a dramatic inhibition of PIP₃ formation in response to fMLP (i.e. greater than 90%, see Fig. 1). When we performed control experiments to examine whether genistein had any direct effect on the -regulated PI3K, we found that genistein attenuated the catalytic activity of the isolated enzyme. As shown in Table I, genistein caused inhibition of both the basal and -stimulated activity of the partially purified U937 cell -sensitive PI3K. At the concentrations used in Fig. 1, 100 µM, inhibition was only partial (~30-40%). Inhibition approached 100% at higher concentrations of the drug. In contrast to genistein, radicicol had no inhibitory effect even at high concentrations, up to 666 ng/ml. Similar results were obtained using a recombinant p110 catalytic subunit.

DISCUSSION

We previously demonstrated that the Shc adapter protein was phosphorylated on tyrosine in response to chemoattractant receptor activation and that this phosphorylated Shc was physically associated with Lyn (31). In the present study, we provide further evidence that Shc is tyrosine-phosphorylated by Lyn by demonstrating that blockade of Lyn tyrosine kinase activity inhibits Shc phosphorylation (Fig. 1). This is consistent with our results showing that the interaction of Lyn and Shc occurs via a Shc SH2 domain (31).

The data presented here strongly indicate that the Lyn-regulated PI3K pathway of PIP₃ formation is of primary importance quantitatively during chemoattractant-mediated leukocyte activation. We have previously shown that the enhancement of PI3K activity associated with Lyn immunoprecipitates is rapid and parallels both PIP₃ formation and cell activation (4, 31). This is in contrast to the data of Stephens et al. (7), who reported Lyn-associated PIP₃ formation was much slower. The Src tyrosine kinase inhibitor radicicol substantially attenuated the rise in PIP₃ formation stimulated by the chemoattractant fMetLeuPhe at a concentration that totally blocked stimulated Lyn tyrosine kinase activity (Fig. 2). This inhibition was not due to a direct inhibitory effect on either the p85/p110 PI3K or the -regulated enzyme (Table I), nor was there any nonspecific inhibition of -mediated signaling in general (Fig. 3). Because there is no involvement of nor requirement for a tyrosine kinase in activating the -regulated enzyme (23, 24, 25), our data strongly point to the Lyn pathway as a quantitatively important source of PIP₃ during early neutrophil activation. Radicicol inhibited the increase in cellular PIP₃ levels by nearly 70%, indicating that at least two-thirds of the PIP₃ formed originated from a tyrosine kinase-regulated PI3K pathway rather than from the -regulated enzyme. Indeed, if the decrease in actual mass of PIP₃ product formed is considered, the level of inhibition is even greater. An alternative explanation for our results would have to invoke the need for a tyrosine kinase for in vivo activation of the -regulated PI3K, a hypothesis for which there is no supporting evidence.

The predominant importance of the tyrosine kinase-requiring PI3K pathway is also evidenced by the data we obtained with genistein. Although not as clearcut as the radicicol data, we observed that genistein almost completely blocked (>90%) chemoattractant-stimulated PIP₃ formation while not decreasing unstimulated levels of PIP₃. Even taking into account the inhibition that could occur due to direct effects on PI3K itself (30-40% in vitro), the majority of the inhibition appears to be as a result of tyrosine kinase blockade. The observation that genistein was a direct inhibitor of the -sensitive PI3K (p110) at higher concentrations may have bearing on the widespread use of this compound as a specific inhibitor of tyrosine kinases (43). It is clear that at higher concentrations genistein can interact with the p110 and the p85-associated p110 isoforms (, ) as well (36). This is likely to be due to the ability of genistein to compete with ATP for binding to the enzyme (43). Because radicicol is structurally distinct from genistein and inhibits tyrosine kinases in a manner noncompetitive with ATP (37, 38), it has no effect on either form of PI3K.

In summary, we have provided evidence that a tyrosine kinase-requiring pathway likely to be the Lyn pathway is quantitatively predominant for formation of the lipid mediator PIP₃ during early leukocyte activation by chemoattractants. Thus, the regulation of neutrophil functions that depend upon PIP₃ formation, such as oxidant formation via the NADPH oxidase (9, 10), is likely to occur via the tyrosine kinase-initiated pathway. The Rac GTPase is a critical regulator of the NADPH oxidase (44), and we have previously shown that Rac translocation from cytosolic complex to membrane oxidase requires the activity of a tyrosine kinase (33). PI3K activity has also been implicated in Rac activation (45). The data presented here reconcile these observations by demonstrating the primary importance of the tyrosine kinase mechanism for generating the lipid mediator PIP₃ in chemoattractant-stimulated human neutrophils.