Priming of Human Neutrophil Respiratory Burst by Granulocyte/Macrophage Colony-stimulating Factor (GM-CSF) Involves Partial Phosphorylation of p47 phox *

Neutrophil superoxide production can be potentiated by prior exposure to “priming” agents such as granulocyte/macrophage colony stimulating factor (GM-CSF). Because the mechanism underlying GM-CSF-dependent priming is not understood, we investigated the effects of GM-CSF on the phosphorylation of the cytosolic NADPH oxidase components p47 phox and p67 phox . Preincubation of neutrophils with GM-CSF alone increased the phosphorylation of p47 phox but not that of p67 phox . Addition of formyl-methionyl-leucyl-phenylalanine (fMLP) to GM-CSF-pretreated neutrophils resulted in more intense phosphorylation of p47 phox than with GM-CSF alone and fMLP alone. GM-CSF-induced p47 phox phosphorylation was time- and concentration-dependent and ran parallel to the priming effect of GM-CSF on superoxide production. Two-dimensional tryptic peptide mapping of p47 phox showed that GM-CSF induced phosphorylation of one major peptide. fMLP alone induced phosphorylation of several peptides, an effect enhanced by GM-CSF pretreatment. In contrast to fMLP and phorbol 12-myristate 13-acetate, GM-CSF-induced phosphorylation of p47 phox was not inhibited by the protein kinase C inhibitor GF109203X. The protein-tyrosine kinase inhibitor genistein and the phosphatidylinositol 3-kinase inhibitor wortmannin inhibited the phosphorylation of p47 phox induced by GM-CSF and by fMLP but not that induced by phorbol 12-myristate 13-acetate. GM-CSF alone did not induce p47 phox or p67 phox translocation to the membrane, but neutrophils treated consecutively with GM-CSF and fMLP showed an increase (compared with fMLP alone) in membrane translocation of p47 phox and p67 phox . Taken together, these results show that the priming action of GM-CSF on the neutrophil respiratory burst involves partial phosphorylation of p47 phox on specific serines and suggest the involvement of a priming pathway regulated by protein-tyrosine kinase and phosphatidylinositol 3-kinase.

Human polymorphonuclear neutrophils play a key role in host defenses against invading microorganisms. In response to a variety of stimuli, neutrophils release large quantities of superoxide anion (O 2 . ) in a phenomenon known as the respiratory burst. O 2 . is the precursor of potent oxidants, which are essential for bacterial killing and also potentiate inflammatory reactions (1). Neutrophil production of O 2 . is dependent on the respiratory burst oxidase, or NADPH oxidase, a multicomponent enzyme system that catalyzes NADPH-dependent reduction of oxygen to O 2 . (2,3). NADPH oxidase is activated and regulated by various neutrophil stimuli at infectious or inflammatory sites. Proinflammatory cytokines such as GM-CSF, 1 tumor necrosis factor, and interleukin-8 modulate NADPH oxidase activity through a priming phenomenon (4 -6). These cytokines induce a very weak oxidative response by neutrophils but strongly enhance neutrophil release of reactive oxygen species on exposure to a secondary applied stimulus such as bacterial N-formyl peptides (7,8). The mechanisms underlying the priming process are poorly understood, although some studies have suggested that priming with various agonists is regulated at the receptor and post-receptor levels (7,9,10). DeLeo et al. (11) recently reported that priming of the respiratory burst by lipopolysaccharide resulted in limited phosphorylation of p47 phox as well as redistribution of oxidase components. However, phosphorylation of NADPH-oxidase components during GM-CSF priming process has not been studied.
In resting cells, NADPH oxidase is inactive, and its components are distributed between the cytosol and membranes. When cells are activated, the cytosolic components (p47 phox , p67 phox , p40 phox , and Rac2) migrate to the membranes, where they associate with the membrane-bound component (flavocytochrome b 558) to assemble the catalytically active oxidase (12)(13)(14)(15)(16). Upon oxidase activation, p47 phox and p67 phox become phosphorylated (17)(18)(19). p47 phox phosphorylation on several serines plays a pivotal role in oxidase activation in intact cells (20 -24). Different kinases have been shown to phosphorylate p47 phox in vitro, but the regulatory pathways involved in different conditions of stimulation in vivo are unknown.
To further define the mechanisms involved in proinflammatory cytokine-induced priming of the neutrophil respiratory burst triggered by fMLP, we analyzed the effects of GM-CSF alone and combined with fMLP on the phosphorylation of p47 phox and p67 phox .

EXPERIMENTAL PROCEDURES
Reagents-fMLP, proteases and phosphatases inhibitors were from Sigma. [ 32 P]Orthophosphoric acid was from NEN Life Science Products. Kinases inhibitors were from Calbiochem. Injection-grade water and 0.9% NaCl were endotoxin-free (Ͻ0.4 pg/ml) in the limulus test * This work was supported by a grant from Recherches et Partages. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ Contributed equally to this work. § To whom correspondence should be addressed. Neutrophil Preparation-Human neutrophils were obtained in lipopolysaccharide-free conditions by means of dextran sedimentation and Ficoll centrifugation as described previously (19), except that diisopropylfluorophosphate treatment was omitted, as we found that diisopropylfluorophosphate alone from some lots can increase p47 phox phosphorylation. 2 Measurement of O 2 . Production-Neutrophils were treated with GM-CSF at 37°C at various concentrations and incubation times in Hanks' buffered saline solution plus 0.025% endotoxin-free bovine serum albumin. Thereafter, O 2 . production was measured in response to fMLP (10 Ϫ7 M) in terms of superoxide dismutase-inhibitable ferricytochrome c reduction as described previously (16). 32 P Labeling of Neutrophils, Stimulation, and Fractionation-Cells were incubated in phosphate-free buffer containing 0.5 mCi of [ 32 P]orthophosphoric acid/10 8 cells/ml for 60 min at 30°C, as previously reported (19). Neutrophils were then treated with GM-CSF, as described above, and stimulated or not with 10 Ϫ7 M fMLP for a further 2 min. The reaction was stopped by adding ice-cold buffer and centrifugation at 400 ϫ g for 6 min at 4°C. The cells were lysed by resuspending them in lysis buffer, as previously described (19). The suspension was sonicated on ice for 3 ϫ 15 s. The lysate was centrifuged at 100,000 ϫ g for 30 min at 4°C in a TL100 Ultracentrifuge (Beckman).
Electrophoresis and Blotting-The samples were subjected to SDSpolyacrylamide gel electrophoresis (PAGE) in 10% polyacrylamide gels, using standard techniques (25,26). The separated proteins were transferred to nitrocellulose and detected as described elsewhere (16,19).
Two-dimensional Tryptic Phosphopeptide Mapping and Phosphoamino Acid Analysis-Tryptic digestion of p47 phox on nitrocellulose, thin-layer electrophoresis, and thin-layer chromatography were performed as described previously (23,24). The nitrocellulose area containing 32 P-labeled p47 phox was incubated for 30 min at 37°C with polyvinylpyrrolidone, washed, then incubated overnight with trypsin (50 g/ ml) in carbonate buffer. Released peptides were washed three times in a Speed-Vac, redissolved in electrophoresis buffer (17 volumes of water/3 volumes of 88% formic acid), and applied to one corner of a cellulose thin-layer plate (Merck). After electrophoresis (1000 V for 20 min), chromatography was performed as described previously (23,24). The plates were autoradiographed for 1 week at Ϫ75°C. Phosphoamino acid analysis was performed as described previously (19).
Statistical Analysis-All results are expressed as means Ϯ S.E. Significant differences were detected using the Student's t test.

RESULTS
GM-CSF Induces Phosphorylation of p47 phox but Not of p67 phox - Fig. 1 (lower panel) shows that GM-CSF alone did not induce superoxide production; in contrast, it potentiated the respiratory burst induced by 10 Ϫ7 M fMLP, in accordance with the literature (7)(8)(9)(10). To determine the effect of GM-CSF on p47 phox and p67 phox phosphorylation, 32 P-Loaded neutrophils were treated with GM-CSF (500 pM for 20 min) and then stimulated with 10 Ϫ7 M fMLP. p47 phox and p67 phox were immu-noprecipitated with specific antibodies and then analyzed by SDS-PAGE, Western blot, and autoradiography. As shown in Fig. 1 (upper panel), p47 phox phosphorylation in GM-CSFtreated neutrophils was clearly higher than in untreated cells (202.7 Ϯ 8.5% of control untreated cells (n ϭ 4, p Ͻ 0.01) as determined by densitometry analysis); fMLP (10 Ϫ7 M) also stimulated the phosphorylation of p47 phox (210.5 Ϯ 15% of control (n ϭ 4, p Ͻ 0.01). Pretreatment of neutrophils with GM-CSF and subsequent stimulation with fMLP resulted in stronger phosphorylation than that induced by each agent alone (410.5 Ϯ 17.7% of control (n ϭ 4, p Ͻ 0.01). Western blot analysis (middle panel) showed that the same amount of p47 phox was immunoprecipitated from each sample. p47 phox was not detected when irrelevant control antibodies were used (data not shown). Because p67 phox is also phosphorylated during NADPH oxidase activation (18,19), we analyzed the effect of GM-CSF on its state of phosphorylation. p67 phox phosphorylation was not increased by GM-CSF alone (500 pM) relative to resting cells. In addition, GM-CSF did not increase the p67 phox phosphorylation induced by fMLP (10 Ϫ7 M) (data not shown).
Comparison between GM-CSF-induced p47 phox Phosphorylation and the GM-CSF Priming Effect on Superoxide Production-To attempt to establish a relationship between p47 phox phosphorylation and the priming effect of GM-CSF on superoxide production, we performed parallel kinetics and concentration-effect studies on these two processes. As shown in Fig.  2, p47 phox phosphorylation increased with the GM-CSF incubation time. The time course of p47 phox phosphorylation ran closely parallel to the kinetics of GM-CSF-primed superoxide production in response to fMLP (10 Ϫ7 M). In addition, a concen- tration-dependent effect analysis showed a parallel increase in p47 phox phosphorylation and the priming effect on superoxide production, reaching a plateau at 500 pM (Fig. 3).
GM-CSF-induced Phosphorylation of p47 phox Affects One Major Peptide-p47 phox is phosphorylated on several serines in the carboxyl-terminal portion of the protein (23). Tryptic peptide mapping of p47 phox showed that these phosphorylated serines were located on several peptides on the map: six major peptides with strong phosphorylation were regularly present, and a few minor peptides were less strongly phosphorylated and more irregular. As shown in Fig. 4, GM-CSF (500 pM) induced the phosphorylation of only one major p47 phox peptide, whereas fMLP (10 Ϫ7 M) induced weak phosphorylation of all the peptides. Furthermore, pretreatment of neutrophils with GM-CSF followed by stimulation with fMLP (10 Ϫ7 M) resulted in an increase in the phosphorylation of all the peptides. This suggested that GM-CSF induced partial phosphorylation of p47 phox on only certain serines that could potentiate the phosphorylation of the other serines on secondary fMLP stimulation.
Effects of Protein Kinase Inhibitors on GM-CSF-induced Phosphorylation of p47 phox -We then analyzed the pathways involved in p47 phox phosphorylation by GM-CSF relative to fMLP and PMA (Fig. 5). Preincubation of neutrophils for 15 min with the protein kinase C inhibitor GF109203X (5 M) did not inhibit p47 phox phosphorylation induced by GM-CSF (Fig.  5A) but inhibited that induced by fMLP (10 Ϫ6 M) and PMA (0.5 g/ml) (Fig. 5B). Pretreatment of neutrophils with the tyrosine kinase inhibitor genistein or with the PI3K inhibitor wortmannin inhibited the phosphorylation induced by GM-CSF and fMLP but not that induced by PMA (Fig. 5). These results were confirmed using other protein-tyrosine kinase inhibitors, erbstatin (100 M) and lavendustin (25 M); wortmannin (50 nM) and LY294002 (100 M) also inhibited this phosphorylation (data not shown). Taken together, these results suggest that the pathway controlling p47 phox phosphorylation in response to GM-CSF is totally different from that underlying the action of PMA and might belong to the multiple pathways used by fMLP.
Characterization of Phosphorylated Amino Acids on p47 phox -Because p47 phox phosphorylation was inhibited by the tyrosine kinase inhibitor genistein, we analyzed the phosphorylated amino acids of p47 phox isolated from GM-CSF-treated neutrophils. Fig. 6 shows that p47 phox was phosphorylated only on serines. In addition, no phosphorylated tyrosines were detected by Western blotting using an anti-phosphotyrosine antibody (data not shown). This result suggests that a genisteinsensitive protein-tyrosine kinase is an upstream regulator of a serine/threonine kinase that phosphorylates p47 phox .

GM-CSF Does Not Induce but Enhances fMLP-induced
Translocation of p47 phox and p67 phox -NADPH oxidase assembly at the membrane site is dependent on complete phosphorylation of p47 phox (17,27). We observed that the partial phosphorylation of p47 phox induced by GM-CSF was insufficient to induce p47 phox or p67 phox translocation (Fig. 7). However, neutrophil priming by GM-CSF followed by stimulation with 10 Ϫ7 M fMLP resulted in stronger translocation than that observed with fMLP alone (163.8 Ϯ 13.8% and 177.5 Ϯ 10.6% control (fMLP alone) (n ϭ 4, p Ͻ 0.01) for p47 phox and p67 phox , respectively, as determined by densitometry analysis). This suggested that prephosphorylation of p47 phox induced by GM-CSF may accelerate the completion of p47 phox phosphorylation on other sites and the movement of the cytosolic subunits to membranes. DISCUSSION GM-CSF alone induced partial phosphorylation of the cytosolic oxidase component p47 phox , suggesting a role of this proc-ess in the priming effect of GM-CSF on the respiratory burst. The kinetics and dose-effect studies of GM-CSF-induced p47 phox phosphorylation was parallel to GM-CSF-induced priming of superoxide production by neutrophils in response to fMLP 10 Ϫ7 M. This phosphorylation was distinct from that induced by PMA, which phosphorylates other peptides and the sensitivity of which to protein kinase inhibitors (GF109203X, genistein, and wortmannin) was totally different. In addition, the partial GM-CSF-induced p47 phox phosphorylation differed from fMLP-induced phosphorylation, as it was insensitive to GF109203X, although it was also inhibited by genistein and wortmannin. These results suggest that the GM-CSF priming pathway might share one of the multiple pathways used by fMLP signaling. Furthermore, GM-CSF increased both the degree of phosphorylation of p47 phox peptides and the translocation of p47 phox and p67 phox when fMLP was subsequently added at a low concentration. These results suggest that GM-CSF triggers phosphorylation of only certain serines on p47 phox that represent the first cytosolic step in the phosphorylation cascade, as GM-CSF alone did not induce p47 phox or p67 phox membrane translocation. Stepwise phosphorylation of p47 phox has previously been suggested on the basis of two-dimensional gel analysis in cytochrome b-deficient neutrophils (17) during translocation to the membranes (27) or cytoskeleton (16) and also in a kinase-activating cell-free system (28). Green et al. (29) have reported that priming of NADPH oxidase by interleukin 8, another proinflammatory cytokine, does not induce translocation of cytosolic oxidase components. However, DeLeo et al. (11) recently reported that priming of the respiratory burst by lipopolysaccharide resulted in translocation of cytochrome b and p47 phox but not p67 phox or rac2. The mechanisms underlying lipopolysaccharide priming could be different from those underlying GM-CSF and IL-8 priming.
Our observation that GM-CSF induced partial p47 phox phosphorylation but not p67 phox phosphorylation supports the idea that p47 phox phosphorylation on critical sites plays a major role in the priming of NADPH oxidase in intact cells. The position of the phosphorylated peptides on the phosphopeptide map of p47 phox (24) suggests that candidate targets for GM-CSF-induced phosphorylation are serines 328, 345, 348, 359, and 370, contrary to serines 303, 304, 320, and 315.
The GM-CSF receptor is not coupled to classical G proteins. Engagement of GM-CSF with its receptor activates a number of signal transduction pathways, including protein-tyrosine kinase (30), PI3K (31), phopholipase A 2 (32), and mitogen-activated protein kinase (ERK1/ERK2) (33,34). Our inhibition experiments suggest that a genistein-sensitive protein-tyrosine kinase and PI3K are involved in GM-CSF-induced partial p47 phox phosphorylation and in NADPH oxidase priming. Genistein and wortmannin also inhibited GM-CSF priming of the respiratory burst in response to fMLP in parallel to p47 phox phosphorylation (data not shown). In addition, Kodama et al.  7. Effect of GM-CSF priming on membrane translocation of p47 phox and p67 phox . Neutrophils were incubated with or without GM-CSF (500 pM) for 20 min and challenged with or without fMLP (10 Ϫ7 M) for 3 min. Membranes and cytosols were prepared as described under "Experimental Procedures" and analyzed by SDS-PAGE and Western blot with anti-p47 phox and anti-p67 phox antibodies. Membranes shown on the gel are from 12 ϫ 10 6 cells, and cytosols are from 1 ϫ 10 6 cells. The data are representative of three different experiments. (35) recently described a close correlation between the inhibitory effect of wortmannin on PI3K activity, superoxide production, and p47 phox phosphorylation in GM-CSF-primed and fMLP-treated neutrophils; their results support our hypothesis. The GM-CSF receptor interacts with lyn, a member of the src-family nonreceptor tyrosine kinase family, and PI3K can be directly regulated by lyn (31,36). Because we only detected phosphorylated serines on p47 phox , these data suggest that GM-CSF activates Ser/Thr-protein kinase(s) that phosphorylate(s) p47 phox downstream of lyn and PI3K. In addition, GF109203X-sensitive protein kinase C isoforms are not involved in this process.
In conclusion, the data presented here show that priming of the neutrophil respiratory burst by GM-CSF results in partial p47 phox phosphorylation, which increases the degree of p47 phox phosphorylation and NADPH oxidase assembly in response to a secondary applied stimulus such as fMLP. This GM-CSF priming pathway involves one or more protein-tyrosine kinases and PI3K. Identification of the Ser/Thr kinase involved in p47 phox phosphorylation will require further investigations.