Protein Kinase C ζ Phosphorylates a Subset of Selective Sites of the NADPH Oxidase Component p47phox and Participates in Formyl Peptide-Mediated Neutrophil Respiratory Burst

Generation of superoxide anion by the multiprotein complex NADPH phagocyte oxidase is accompanied by extensive phosphorylation of its 47-kDa protein component, p47phox, a major cytosolic component of this oxidase. Protein kinase C ζ (PKC ζ), an atypical PKC isoform expressed abundantly in human polymorphonuclear leukocytes (PMN), translocates to the PMN plasma membrane upon stimulation by the chemoattractant fMLP. We investigated the role of PKC ζ in p47phox phosphorylation and in superoxide anion production by human PMN. In vitro incubation of recombinant p47phox with recombinant PKC ζ induced a time- and concentration-dependent phosphorylation of p47phox with an apparent Km value of 2 μM. Phosphopeptide mapping analysis of p47phox showed that PKC ζ phosphorylated fewer selective sites in comparison to “conventional” PKCs. Serine 303/304 and serine 315 were identified as targets of PKC ζ by site-directed mutagenesis. Stimulation of PMN by fMLP induced a rapid and sustained plasma membrane translocation of PKC ζ that correlated to that of p47phox. A cell-permeant-specific peptide antagonist of PKC ζ inhibited both fMLP-induced phosphorylation of p47phox and its membrane translocation. The antagonist also inhibited the fMLP-induced production of oxidant (IC50 of 10 μM), but not that induced by PMA. The inhibition of PKC ζ expression in HL-60 neutrophil-like cells using antisense oligonucleotides (5 and 10 μM) inhibited fMLP-promoted oxidant production (27 and 50%, respectively), but not that induced by PMA. In conclusion, p47phox is a substrate for PKC ζ and participates in the signaling cascade between fMLP receptors and NADPH oxidase activation.

P olymorphonuclear leukocytes (PMN) 2 play a vital role in the first line of cellular host defense against microorganisms. This function relies in part on the ability of PMN to generate large amounts of superoxide anion (O 2 . ) and related reactive oxygen species, a phenomenon known as the respiratory burst (1). The superoxide-generating enzyme, NADPH oxidase, is a multiprotein complex that comprises a membrane-bound flavocytochrome b composed of two subunits (a 91-kDa glycoprotein and a 22-kDa protein) and four main cytosolic factors (p47 phox , p67 phox , p40 phox , and a small G protein, Rac2) (2)(3)(4)(5)(6). NADPH oxidase activation can be triggered by various stimuli such as chemotactic peptides, phorbol esters, and opsonized particles. Upon activation of NADPH oxidase, p47 phox becomes phosphorylated on several sites (7)(8)(9) and translocates to the plasma membrane (8, 10 -12), where it interacts with cytochrome b (13). Various protein kinases have been involved in the regulation of NADPH oxidase activity (14 -17), among which the protein kinase C (PKC) family appears to play a major role (18 -20). PKC comprises a family of isoenzymes that play a key role in signaling events and cell functions (21,22). PKC has been classified into the following three subgroups on the basis of their molecular structure and mode of activation: 1) "conventional" PKCs (␣, ␤1, ␤2, and ␥), which require phosphatidylserine (PS) and are activated by calcium and diacylglycerol (DAG); 2) "novel" PKCs (␦, ⑀, , , and ), which require PS and DAG but not calcium for activation; and 3) "atypical" PKCs (,), which are calcium-independent and are not activated by DAG (21,22). Among the atypical PKC, PKC can be stimulated by phospholipids, including phosphatidic acid (PA), phosphatidylinositol trisphosphates, and ceramides. Activation of PKCs in intact cells is generally associated with the translocation of the enzyme from the cytosol to particulate compartments (18,(23)(24)(25).
Immunochemical studies have shown that human PMN express five PKC isoforms: ␣, ␤1, ␤II, ␦, and (23, 24, 26 -29). The respective contribution of each PKC isoform in regulating PMN respiratory burst has not yet been elucidated, although it has been shown that PKC ␤ may modulate the respiratory burst of both electropermeabilized PMN (15) and HL-60 cells differentiated into PMN-like cells (30). PKC , which like PKC ␤ is expressed at high levels in human PMN (23), was shown to regulate PMN adhesion and chemotaxis (31). The contribution of PKC to the respiratory burst has not been documented so far, although it has been shown that PKC translocates to the plasma membrane and that its activity is increased in fMLP-stimulated PMN (31)(32)(33).
In the present work we have studied the role of PKC in the signaling pathway between stimulated fMLP receptors and NADPH oxidase activation. The results indicate that PKC phosphorylates selective residues of p47 phox and plays a predominant role in the stimulation of superoxide production in fMLP-stimulated human PMN.

Phosphorylation of p47 phox in vitro by PKC
Phosphorylation of recombinant p47 phox by recombinant PKC was performed in 40 l of a mixture containing 50 mM glycero-2 phosphate, 0.4 mM EGTA, 10 mM magnesium acetate, 35 mM Tris-HCl buffer (pH 7.5), various concentrations of wild-type or mutant p47 phox , and 0.15 U/ml recombinant PKC (1 U corresponding to 1 nmol phosphate incorporated in peptide ⑀/min) without activators. This PKC preparation has been used previously to phosphorylate peptide ⑀ and myristoyl alanine-rich C kinase substrate peptide (37). Phosphorylation was initiated with 50 M [␥-32 P]ATP (specific radioactivity, 150 -300 Ci/mmol) and conducted at 30°C. After 30 min of incubation, 20 l of the reaction mixture was spotted onto Whatman p81 paper (Tewksbury, MA) followed by three washes with 5% H 3 PO 4 , and the radioactivity was counted. The remaining fraction (20 l) was mixed with 3ϫ Laemmli sample buffer (Bio-Rad, Ivry/Seine, France) and subjected to SDS-PAGE (10% acrylamide gel) (39), then blotted (40) for autoradiography analysis with Amersham Hyperfilm MP (Arlington Heights, IL).

Cyanogen bromide (CNBr) cleavage of p47 phox and twodimensional phosphopeptide mapping analysis
The phosphorylated p47 phox was subjected to SDS/PAGE and transferred to nitrocellulose. The band of interest was excised and incubated in darkness with 12.5 mg/ml CNBr in 70% (v/v) formic acid for 16 h at room temperature. The reaction mixture was quenched with 1 vol of water and lyophilized in a Speed-Vac (Savant, Holbrook, NY). The digested peptides were then separated by Tris-N-[2-hydroxy-1,1bis(hydromethyl)ethyl]glycine (Tricine)-SDS-PAGE as previously described (16,41,42). For peptide mapping, the phosphorylated p47 phox blotted onto a nitrocellulose membrane was cut out and resuspended in 200 l of 50 mM (NH 4 )HCO 3 buffer (pH 7.8) and digested overnight with 25 g trypsin at 37°C. After drying, the resulting phosphopeptides were resolved on thin-layer cellulose plates by electrophoresis at pH 1.9 for 30 min at 1000 V and 4°C in the first dimension, followed by ascending chromatography developed with isobutyric acid buffer (isobutyric acid/butanol/pyridine/acetic acid/water, 25/1/ 2.5/1.8/10 (v/v)) (16,43). 32 P-labeled peptides were detected by autoradiography at Ϫ70°C.

Chemiluminescence measurement
A suspension of 5 ϫ 10 5 cells was incubated in 0.3 ml HBSS at 37°C in the absence (control) or presence of various concentrations of PKC antagonist for 15 min in the presence of 0.025% BSA. Cells were then incubated with 10 M luminol and stimulated with 20 nM fMLP or PMA (100 ng/ml). Results are expressed as the percentage of the peak of chemiluminescence or in cpm.

Fractionation of PMN
PMN were treated in HBSS containing 0.025% BSA in the absence or presence of PKC antagonist for 30 min and stimulated with fMLP or PMA under the conditions described above. Membranes were prepared as described (44). Briefly, the PMN suspensions (10 8 cells/ml) were sonicated three times for 5 s on ice in 1 ml relaxation buffer containing 10 mM PIPES, (pH 7.3), 3 mM MgCl 2 , 100 mM KCl, and 5 mM NaCl, supplemented with 0.5 mM PMSF, 1 mM EGTA, 10 g/ml leupeptin, and 10 g/ml pepstatin. Cytosolic and membrane fractions were prepared by centrifugation (200,000 ϫ g, 30 min) of the homogenates on 15-34% sucrose gradient at 4°C. A membrane preparation equivalent of 5 ϫ 10 6 cells was used for SDS-PAGE.

In vitro phosphorylation of PMN endogenous p47 phox induced by PA
Phosphorylation of endogenous substrates was performed with 20 l of cytosol fraction from unlabeled PMN (2 ϫ 10 6 cell equivalent) in a final volume of 100 l containing 50 mM glycero-2 phosphate, 0.4 mM EGTA, 10 mM magnesium acetate, and 35 mM Tris-HCl buffer (pH 7.5) to which PA (100 M) or DAG/PS/calcium (0, 2/0, 1/1, 2 mM) was added. In some experiments, cytosolic fractions were pretreated in the presence or absence of nonmyristoylated PKC peptide antagonist. The reaction was initiated with [␥-32 P]ATP at a final concentration of 50 M (specific radioactivity, 150 -300 Ci/mmol). Incubation was conducted at 30°C for 30 min, and the reactions were stopped with 3ϫ Laemmli buffer. Proteins were heat-denatured, subjected to SDS-PAGE (10% acrylamide gel), and blotted for autoradiography and Western blot analysis.

Culture of HL-60 cells and oligonucleotide treatment
Cells were grown in suspension in RPMI 1640 medium (Life Technologies, Cergy Pontoise, France) supplemented with 10% (v/v) heat-inactivated FCS, 10 mM HEPES, 2 mM L-glutamine, streptomycin (50 g/ml), and penicillin (50 U/ml) in a 5% CO 2 atmosphere. To induce myeloid differentiation, cells were seeded at a density of 5 ϫ 10 5 -1.0 ϫ 10 6 /ml and treated with 1.3% DMSO for 4 -6 days. For antisense treatment, we adapted a procedure described by Xu et al. (38), using an oligonucleotide complementary to the 5Ј end of the PKC messenger, starting at the translation start codon. Three days after cell treatment with DMSO, HL-60 cells were washed, resuspended in DMEM, and treated with 5 and 10 M PKC oligonucleotides (antisense or sense) in the presence of 15 g/ml lipofectin for 8 h at 37°in 5% CO 2 following the manufacturer's instructions. The medium was then replaced by fresh culture medium containing oligonucleotides with 1.3% DMSO without lipofectin. After 48 h, cells were washed and suspended in HBSS, and superoxide anion production was continuously recorded by monitoring superoxide dismutase-inhibitable reduction of cytochrome c, as described in Ref. 45, with minor modifications. Aliquots of 5 ϫ 10 5 cells in 500 l of HBSS were used to generate superoxide anion induced by fMLP (100 nM) or PMA (160 nM). Cells were also lysed in Laemmli sample buffer to analyze the expression of PKC isoforms and p47 phox in immunoblotting experiments as described above.
In control experiments, lipofectine per se did not induce a marked cytotoxic effect (14 Ϯ 2%) relative to untreated cells (6 Ϯ 2%), as assessed by the trypan blue exclusion test. The functional activity of differentiated HL-60 was also decreased by ϳ15%, determined by measurement of PMA-or fMLP-induced superoxide production.

Statistical analysis
Each experiment was performed in duplicate and repeated at least three times. Unless otherwise indicated, data represent means Ϯ SEM. Statistically significant differences between means were identified using Student's paired t test with a threshold of p Ͻ 0.05.

In vitro phosphorylation of p47 phox by PKC
The phosphorylation of p47 phox on multiple serines is required for NADPH oxidase activation. Previous studies have reported that p47 phox is a good substrate for PKCs. However, the contribution of each isoform to the multiphosphorylation process of p47 phox is not known. We and others have reported that PKC is expressed in human neutrophils and translocates to the plasma membranes upon PMN activation (23,24,31,32). To determine whether p47 phox is a potential substrate for PKC , we analyzed the phosphorylation of recombinant p47 phox in vitro by a constitutively active recombinant PKC . The results show (Fig. 1, right) that PKC phosphorylated p47 phox as a function of the incubation time. In this experiment, which was conducted at 30°C in the absence of enzyme activators, phosphorylation of p47 phox began to be visible after 3 min and was maximal at 30 min. In the absence of p47 phox , a weakly phosphorylated band of 80 kDa was detected (Fig. 1,  left), which probably corresponded to autophosphorylated PKC because the band was recognized by an anti-PKC Ab in Western blotting experiments (results not shown). Phosphorylation of p47 phox by PKC was also dependent on the concentration of the protein ( Fig.  2A). The apparent K m value of PKC for p47 phox , calculated from a Lineweaver-Burk representation of the data, was ϳ2.0 M (Fig.  2B), indicating that p47 phox is a good substrate for PKC .

Mapping of p47 phox phosphorylation targets for PKC in vitro
In PMN stimulated with PMA or fMLP, p47 phox is phosphorylated on multiple serine residues (42). Phosphorylated residues were located in the carboxyl-terminal portion of the protein resulting from CNBr cleavage of p47 phox (42), and previous studies identified serine residues that were specifically phosphorylated by "conventional" PKC, mitogen-activated protein kinase (MAP kinase), and protein kinase A (PKA) (16). To map the p47 phox amino acid residues that were phosphorylated by PKC , we analyzed the phosphorylated protein by CNBr cleavage and tryptic peptide mapping. Cleavage of p47 phox by CNBr followed by Tricine-SDS-PAGE showed that the phosphorylated sites were located in a single peptide of ϳ14 kDa (Fig. 3, autoradiography) corresponding to the C-terminal portion of the protein because it was recognized by an Ab directed against the 10 carboxyl aa 380 -390 of p47 phox (Fig. 3, Western blot).
Analysis of the two-dimensional phosphopeptide mapping of p47 phox phosphorylated by PKC (Fig. 4A) showed the presence of only four phosphopeptides, compared with the much larger number of phosphorylated peptides found in the phosphopeptide map of the p47 phox phosphorylated by "conventional" PKCs ( Fig.  4B). This result suggests that PKC specifically phosphorylates a subgroup of serine residues on p47 phox .
In previous studies on "conventional" PKCs, five phosphorylated serine residues were shown to be critical for p47 phox activation (16). To investigate whether PKC phosphorylation targets   include one or more of these critical residues, we analyzed the phosphomaps of p47 phox mutants in which the serine residues mentioned above were replaced by alanine residues S303/304A-, S315A-, S320A-, and S328A-p47 phox . One of the major phosphopeptides resulting from wild-type p47 phox phosphorylation by PKC was absent in the phosphopeptide map of mutant S303/ 304A-p47 phox (Fig. 4C), indicating that the peptide containing serine residues 303/304 is a target of PKC . Interestingly, the labeling of the other phosphopeptides of S303/304A-p47 phox was reduced, suggesting that phosphorylation by PKC may be sequential and that phosphorylation of serine residues 303/304 occurs first. Accordingly, the phosphorylation map of the S315A mutant showed the disappearance of a weakly labeled peptide but not other changes in the phosphorylation pattern (Fig. 4D). The phosphorylation map of both the S320A (Fig. 4E) and S328A (not shown) mutants were identical with that of wild-type p47 phox , indicating that corresponding serine residues do not constitute phosphorylation targets for PKC .
In stimulated PMN, PKC and p47 phox translocate to plasma membrane fractions, and PKC phosphorylates p47 phox Stimulation of superoxide production by fMLP in PMN is associated with the phosphorylation and translocation of p47 phox to the plasma membrane (reviewed in Ref. 6). We reported previously that fMLP induced a translocation of PKC to the plasma membrane of human PMN (32). Here, we studied whether this PKC translocation paralleled that of p47 phox , which could be consistent with a coordinated spatial redistribution of the two proteins. In PMN stimulated by fMLP, PKC translocated to the plasma membrane in a rapid and sustained manner (Fig. 5). The membrane translocation of p47 phox induced by fMLP was similar to that of PKC . Densitometric analysis of the immunoblots showed a good correlation in the membrane translocation of the two proteins (r ϭ 0.88, p Ͻ 0,01). In addition, a myristoylated membrane-permeable peptide antagonist (10 M) corresponding to the pseudo-substrate region of PKC inhibited membrane translocation of p47 phox by ϳ45% (Fig. 6, upper panel). This peptide, which was shown to inhibit protein kinase activity in vitro (46,47), was used to demonstrate that PKC is involved in the regulation of integrin-dependent adhesion and chemotaxis of intact human neutrophils (31). A control myristoylated peptide directed against PKC , used at the same concentration (10 M), was ineffective on fMLP-induced translocation of phosphorylated p47 phox . This peptide was chosen as a control because the PKC isoform seems not to be expressed in human PMN, as determined by immunoblotting experiments (23,25). The PKC peptide antagonist did not alter the translocation of p47 phox induced by PMA, indicating that this peptide did not interfere with PMA-sensitive PKCs (i.e., "conventional" and "novel" PKCs). The PKC antagonist (10 M) also inhibited the fMLP-induced phosphorylation of p47 phox in 32 P-loaded PMN, but not the p47 phox phosphorylation induced by PMA (Fig. 6, lower panel), whereas the control peptide was without effect. These results are consistent with the hypothesis that fMLP-activated PKC phosphorylates and activates p47 phox in intact PMN.

Studies with inhibitory peptides and antisense RNA indicate that PKC participates in the fMLP-dependent activation of NADPH oxidase
The data obtained above suggest that PKC regulates NADPH oxidase activation. Incubation of PMN in the presence of the myristoylated peptide antagonist of PKC for 15 min induced a concentration-dependent inhibition of the fMLP-induced respiratory burst, as assessed by a chemiluminescence assay (Fig. 7). This effect was not due to a cytotoxic effect of the peptide because peptide-treated PMN excluded trypan blue (data not shown). Unlike the PKC antagonist, the control PKC peptide antagonist did not inhibit fMLP-and PMA-induced PMN superoxide production, but rather slightly potentiated their effect. The concentration of the PKC antagonist used here (5-15 M) did not appear to inhibit other PKC isoforms. This was first suggested by the observation that the myristoylated peptide did not alter PMA-induced superoxide production, in agreement with previous data (31). In addition, measurement of in vitro phosphorylation of endogeneous p47 phox upon stimulation of conventional PKC by diglycerides and calcium in a cytosolic fraction of PMN was not altered by low peptide concentrations (5-15 M) (Fig. 8). By contrast, this low peptide concentration did inhibit p47 phox phosphorylation induced by a PKC activator (PA). Higher peptide concentrations (20 -50 M) inhibited the activation of conventional PKC (Fig. 8). Taken together, our results indicate that the down-regulation of PMN respiratory burst by the myristoylated PKC -derived peptide results from inhibition of PKC function.
To further examine the functional contribution of PKC , we down-regulated its expression with antisense phosphorothioate oligonucleotides. For this purpose, HL-60 cells differentiated into neutrophils were used rather than fresh blood PMN because this latter cell type is inappropriate for the antisense strategy, given their low transcriptional activity and short survival. After differentiation of HL-60 cells into neutrophil-like cells with DMSO, cells were incubated with 5 and 10 M PKC sense (control) and antisense oligonucleotides. Immunoblot experiments (Fig. 9, lower  panel) showed that this treatment efficiently inhibited the expression of PKC relative to control cells (cells incubated with the sense oligonucleotide). The antisense oligonucleotide did not alter the expression of other PKC isoforms (␣, ␤, and ␦) or of p47 phox (Fig. 9, lower panel), indicating the specificity of the oligonucleotide treatment toward PKC . Both concentrations of PKC antisense oligonucleotide significantly inhibited the fMLP-promoted production of superoxide anion relative to controls (Fig. 9, upper  panel). In contrast, the PMN respiratory burst induced by PMA was not altered. This fact is consistent with the observation that the expression level of "conventional" and "novel" PKCs was not altered by the oligonucleotide and shows that the NADPH oxidase complex remained fully functional in these experiments. These results, obtained with the two different approaches, strongly indicate that PKC is a major effector of fMLP receptors in the signaling pathway leading to NADPH oxidase activation.

Discussion
Classical chemoattractants such as fMLP, the complement-derived C5a, or platelet-activating factor induce various rapid PMN responses such as adhesion, chemotaxis, exocytosis, and production   of superoxide, which all contribute to the bactericidal function of PMN (1). Signaling pathways triggered by chemoattractant receptors are mediated through activation of a pertussis-toxin-sensitive G protein and involve the activation of tyrosine kinases and the generation of second messengers by various effectors. Among the multiple effectors involved in PMN functions, PKC play an important role in the activation of superoxide-generating NADPH oxidase (48). p47 phox , one of the major proteins forming the oxidase complex, is a substrate of PKC. Phosphorylation of p47 phox is essential for enzyme activation and assembly in PMN (6). In the present study, we show that p47 phox is a good substrate for PKC . PKC selectively targeted only a few serine residues on p47 phox , mainly serines 303/304 and 315. In addition, using two different strategies based on the inhibition of PKC activity or enzyme expression, we provide the first evidence that PKC is indeed involved in the regulation of fMLP-induced PMN respiratory burst.
In intact cells, p47 phox is phosphorylated on several serine residues located between positions 303 and 379. Previous studies have shown that p47 phox is phosphorylated in vitro by different types of protein kinases, including "conventional" PKC (14,16) such as PKC ␤ (17), MAP kinases (p42-ERK2 and p38) (16,49), PKA (16,50), p21-activated kinases (51), undefined protein kinases (17,52), and PKC (53). Phosphopeptide mapping of p47 phox revealed that "conventional" PKC, MAP kinase, and PKA phosphorylate different groups of serine residues, suggesting that these different kinases may have different regulatory effects on NADPH oxidase (16). The comparison of phosphopeptide maps described here further indicates that some specificity exists among members of the PKC family for the phosphorylation of p47 phox . Indeed, PKC selectively phosphorylated serine residues at position 303/304, and, to a lesser extent, at position 315, as well as two other residues that remain to be determined. PKC did not phosphorylate the other residues that were phosphorylated by "conven-tional" PKCs, indicating that PKC and conventional PKCs may have discrete specificities for p47 phox functions. In addition, the sites previously found to be phosphorylated by MAP kinase (S345/ 348A) and by PKA (S320A and one or both peptides containing S328A and S359/370A) (16) were not phosphorylated by PKC .
Site-directed mutagenesis studies using EBV-transformed lymphocytes expressing exogenous wild-type or p47 phox mutants in which serine residues were replaced by alanine residues indicated that each individual serine residue so far identified is phosphorylated independently (34). Our data regarding phosphorylation of the three p47 phox mutants (S315A, S320A, and S328A) by PKC are in agreement with this assumption. However, in the case of the p47 phox S303/304A mutant, the phosphorylation of some residues by PKC was strongly blunted. The decreased phosphorylation might result from an altered conformation of p47 phox caused by the mutation. However, this possibility can be ruled out because the p47 phox mutant S303/304A is still phosphorylated in PMA-activated EBV-lymphoblasts (34) and by "conventional" PKCs in vitro (data not shown). Alternatively, the strong decrease in the phosphorylation of the p47 phox mutant suggests that phosphorylation of S303/304A may facilitate the subsequent phosphorylation  of other residues. Finally, the double mutation S(303 ϩ 304)A inhibited superoxide production in B lymphoblasts (54), underlining the importance of the phosphorylation of these two serine residues in oxidase activation in intact cells. p47 phox is extensively phosphorylated during PMN stimulation (42) and plays the main role in transporting the cytosolic oxidase components to the plasma membrane (13). fMLP and PMA, two major activators of NADPH oxidase, induce the phosphorylation of several serine residues including S303/304A and S315A (42). In human PMN stimulated by fMLP, PKC was shown to be activated. To examine the contribution of PKC to NADPH oxidase activity, we used a synthetic peptide antagonist of PKC to downregulate the enzyme activity. The sequence of this peptide corresponds to the pseudosubstrate region of the N-terminal regulatory domain that maintains the enzyme in an inactive form in the absence of activator (55). This peptide, which was previously found to block PKC -dependent signal transduction in Xenopus oocytes and mouse fibroblasts (46 -56), inhibited ϳ50% of the fMLP-induced translocation of p47 phox and PMN respiratory burst. The specificity of this inhibition was suggested by the absence of alteration of PMA-induced respiratory burst. Because PMA is a direct activator of both "conventional" and "novel" PKCs (21,57), these data further suggest that the peptide antagonist does not interfere with oxidase activation mediated by these two classes of PKC. From these experiments, it appears that PKC may contribute substantially to the stimulation of PMN respiratory burst by chemoattractants. Consistent with this assumption, the peptide antagonist inhibited the phosphorylation and translocation of p47 phox induced by fMLP but not by PMA. Further evidence for the involvement of PKC in NADPH oxidase regulation comes from the antisense-mediated inhibition of PKC expression in neutrophil-like cells from differentiated HL-60 cells. Again, this treatment, which caused a significant inhibition of the respiratory burst induced by fMLP, was ineffective on PMA-mediated superoxide production. The oligonucleotide treatment did not alter the expression of other PKC isoforms (␣, ␤, and ␦), further indicating its specificity for PKC . The sense sequence of PKC we used is also shared by a recently cloned human serine/threonine kinase termed kpm (58). However, the sense sequence within kpm is located more than 1500 bp away from the initiation codon, which makes the antisense oligonucleotides less effective than those targeting around the translational start codon, as in the case of PKC . In addition, kpm is expressed during mitosis and has been shown to regulate cell cycle; there is no evidence that kpm can be activated by PKC activators. Finally, an important contribution of PKC in respiratory burst is further illustrated by the strong inhibition of respiratory burst and p47 phox phosphorylation by the PKC peptide antagonist (pseudo-substrate region), the sequence of which is not present in kpm.
PKC can be activated in vitro by various lipidic second messengers that are generated by different signaling effectors. Among these messengers, phosphatidylinositol-3,4,5 trisphosphate is formed by phosphatidylinositol 3-kinase (PI3-kinase); PA is produced by phospholipase D (PLD) or DAG kinase; and ceramide is generated by sphingomyelinase. These signaling effectors are all stimulated in PMN by fMLP and may have a potential role in regulating the respiratory burst (48), although their respective contribution to the stimulation of PKC remains unknown. It was shown recently that in PMN from PI3-kinase-␥ knockout mice the PMN respiratory burst induced by fMLP was markedly inhibited (59,60). Interestingly, in this model the translocation of p47 phox as well as Rac to the plasma membrane was not decreased (59), indicating that signaling pathways other than PI3-kinase-␥ may be required for redistribution of p47 phox and Rac. The PLD pathway appears to be a possible candidate for this regulation. This hypothesis is supported by the observation that PLD is a major source of PA and diglycerides in stimulated PMN (45,61) and may thus contribute to the activation of PKC and other PKC isoforms. In addition, the PLD-derived PA production was associated with a priming of PMN respiratory burst (45,62,63), and this priming was accompanied by an enhanced phosphorylation of p47 phox (19,64). This study supports the hypothesis that several pathways might participate in p47 phox phosphorylation. In PMA-activated cells, "conventional" PKCs, and possibly "novel" PKCs, are activated and phosphorylate p47 phox . However, in more physiological conditions such as fMLP stimulation of neutrophils, other pathways, and so other PKC isoforms (such as PKC ), are activated and participate in p47 phox phosphorylation. The cooperativity of different protein kinases or multiple isoforms of PKC in phosphorylating p47 phox is explained by the fact that these kinases phosphorylate specific targets on p47 phox .
In conclusion, this study shows that p47 phox is a substrate for PKC in vitro. Phosphorylation of p47 phox occurs on the serine residue 303/304, on serine 315, and on two unidentified residues. The inhibition of PKC activity or enzyme expression in human PMN by the use of a specific antagonist or antisense oligonucleotide markedly inhibited fMLP-induced superoxide production, membrane translocation of PKC , and phosphorylation of p47 phox . These data provide evidence of the involvement of PKC in the signaling pathways leading to NADPH oxidase activation by fMLP receptors, but not by PMA, and suggest a role for PKC in regulating the bactericidal function of PMN.