Kinase-dependent activation of the leukocyte NADPH oxidase in a cell-free system. Phosphorylation of membranes and p47(PHOX) during oxidase activation.

The leukocyte NADPH oxidase catalyzes the 1-electron reduction of oxygen to O2− at the expense of NADPH: 2 O2 + NADPH → 2 O2− + NADP+ + H+. The oxidase is dormant in resting cells but acquires activity when the cells are stimulated with a suitable agent. Activation in whole cells is accompanied by extensive phosphorylation of p47PHOX, an oxidase subunit located in the cytosol of resting cells that during oxidase activation migrates to the plasma membrane to complex with cytochrome b558, an oxidase-specific flavohemoprotein. Oxidase activation can be mimicked in a cell-free system using an anionic amphiphile as activating agent. We now report a cell-free system in which the oxidase can be activated in two stages using phosphorylated p47PHOX. The first stage, which effects a change in the membrane, requires ATP and GTP and is blocked by the protein kinase inhibitor GF-109203X, suggesting a protein kinase requirement. The second stage requires phosphorylated p47PHOX and GTP, but no ATP, and is unaffected by GF-109203X; assembly of the oxidase may take place during this stage. Activation is accomplished by p47PHOX phosphorylated by protein kinase C but not protein kinase A or mitogen-activated protein kinase. We believe that activation by phosphorylated p47PHOX is more physiological than activation by amphiphiles, because the mutant p47PHOX S379A, which is inactive in whole cells, is also inactive in this system but works in systems activated by amphiphiles.

The leukocyte NADPH oxidase is an enzyme found in neutrophils and certain other leukocytes that catalyzes the oneelectron reduction of oxygen to O 2 Ϫ at the expense of NADPH (1): 2 O 2 ϩ NADPH 3 2 O 2 Ϫ ϩ NADP ϩ ϩ H ϩ . The O 2 Ϫ produced by this enzyme is itself weakly microbicidal but serves as the precursor of a complex battery of highly reactive oxidants that act as powerful microbicidal agents. These oxidants are major components of the system used to defend the host against invading pathogens.
The NADPH oxidase consists of four polypeptides that have been identified through their absence from phagocytes that are unable to manufacture O 2 Ϫ (2-5) plus a fifth polypeptide, p40 PHOX (6,7), whose function is unclear. The oxidase is dormant in resting cells but acquires catalytic activity when the cells are exposed to any of a variety of stimuli. Activation involves the transfer to the plasma membrane of cytochrome b 558 , a flavohemoprotein located in the membranes of the secretory vesicles and specific granules in the resting neutrophil (8 -10). A cytosolic complex consisting of the oxidase components p47 PHOX , p67 PHOX , and p40 PHOX then associates with the cytochrome to assemble the active oxidase (6,(11)(12)(13)(14)(15).
The mechanism of activation of the oxidase is unclear. The phosphorylation of p47 PHOX is a well-recognized concomitant of oxidase activation in whole cells (16 -22), but to date a causeand-effect relationship between the phosphorylation of this oxidase component and the activation of the enzyme has not been definitively established. The oxidase can be activated in a cell-free system, but the activating agent usually employed is an anionic amphiphile such as arachidonic acid or SDS (23,24). At least two examples of phosphorylation-mediated oxidase activation in a cell-free system have been reported, however. In 1985, Cox et al. (25,26) reported that the phosphorylation of resting neutrophil membranes with protein kinase C led to a low level of oxidase activity. More recently, McPhail and associates (27) showed that ATP increased by a factor of Ϸ2.5 the rate of O 2 Ϫ production in a cell-free system that had been activated by 0.1 mM each of phosphatidic acid and diacylglycerol, suggesting the possibility that a phosphatidic acid-activated protein kinase participated in oxidase activation in their system. We describe here the activation of the leukocyte NADPH oxidase in a cell-free system by p47 PHOX that had been pre-phosphorylated by protein kinase C.

Preparation of Neutrophil Fractions
Neutrophil cytosol and membrane were prepared as described previously (8). Briefly, neutrophils were obtained from normal subjects by dextran sedimentation and Ficoll-Hypaque fractionation of freshly drawn citrate-coagulated blood. The neutrophils were suspended at a concentration of 10 8 cells/ml in a modified relaxation buffer (100 mM KCl, 3 mM NaCl, 3.5 mM MgCl 2 , 10 mM PIPES buffer, pH 7.3), and plasma membrane and cytosol were prepared by nitrogen cavitation and centrifugation through Percoll. Both cytosol and membrane were divided into aliquots and stored at Ϫ70°C until use. Cytosol deficient in p47 PHOX was obtained from neutrophils isolated from p47 PHOX -deficient chronic granulomatous disease patients as described above.

Preparation of Recombinant GST-p47 PHOX Fusion Proteins
Recombinant fusion proteins composed of an upstream glutathione S-transferase (GST) linked to a downstream p47 PHOX , either the wild type protein or the inactive mutant p47 PHOX S379A, were isolated from Escherichia coli that had been transformed with pGEX-1T plasmids containing cDNA inserts encoding the downstream proteins, as previously reported (11). The fusion proteins were purified by affinity chromatography on glutathione-agarose as described elsewhere. Before use, excess glutathione was removed from the solution of purified recombinant protein by dialysis against relaxation buffer. The concentration of proteins was determined with the Bio-Rad assay kit using bovine serum albumin as a standard.

In Vitro Phosphorylation and Measurement of 32 P Incorporation
Labeling of p47 PHOX with PKM, PKA, MAP kinase, or combinations of kinases was performed by incubating a reaction mixture containing 1 g of recombinant p47 PHOX , 1 mM ATP, 5 Ci of [␥-32 P]ATP (Amersham Corp.), 10 mM MgCl 2 , 0.1 g of the indicated kinase(s), and relaxation buffer, pH 7.3, in a total volume of 30 l for 30 min at 37°C. To label with PKC, 1 g of recombinant p47 PHOX was incubated for 30 min at 37°with 0.1 g of PKC in 10 mM magnesium acetate, 1 mM ATP, 5 Ci [␥-32 P]ATP, 0.5 mM CaCl 2 , 50 g/ml phosphatidylserine, and 5 g/ml diolein in a total volume of 30 l. After terminating the phosphorylation reactions by the addition of 10 l of 4 ϫ SDS-sample buffer, the samples were subjected to SDS-PAGE using an 8% running gel according to the method of Laemmli (28). 32 P-Labeled proteins on the dried gels were detected by autoradiography, and 32 P was quantified by excising the labeled bands from the dried gel and measuring their radioactivity using Cerenkov counting. To determine background, a piece of nitrocellulose of similar size was excised from a 32 P-free portion of the gel and counted.

Preparation of Phosphorylated GST-p47 PHOX
Phosphorylation of recombinant GST-p47 PHOX was typically carried out as described above, except that radioactive ATP was omitted and 10 -50 g of fusion protein and a reaction volume of 100 l were employed. Incubations were carried out in Eppendorf tubes for 30 min at 37°C. Each incubation was terminated by the addition of 1.0 ml of ice-cold MTPBS (150 mM NaCl, 16 mM Na 2 HPO 4 , 4 mM NaH 2 PO 4 , pH 7.3) and 100 l of packed MTPBS-washed GSH-agarose beads. The tubes were then rotated end-over-end for 1 h at 4°C and then spun for a few seconds at maximum speed in an Eppendorf centrifuge to sediment the GSH-agarose beads. After washing the beads with four 1-ml portions of ice-cold MTPBS, the bound phosphorylated p47 PHOX fusion protein was eluted by incubating for 30 min at 4°C with 200 l of 50 mM Tris⅐HCl, pH 8.0, 5 mM GSH, 0.2 M NaCl. Before use, the eluted fusion protein was dialyzed against relaxation buffer. To compare the phosphorylation of wild type GST-p47 PHOX and GST-p47 PHOX S379A, 50 g of each of the two proteins were phosphorylated as described above except in the presence of 10 Ci of [␥-32 P]ATP, then analyzed by SDS-PAGE, transferred to nitrocellulose, and detected and quantified by autoradiography. The labeled bands were then excised from the blot and analyzed by two-dimensional peptide mapping as described (29,30). The autoradiogram indicated that the levels of phosphorylation of the wild type and mutant proteins were similar, and the two-dimensional peptide maps of the wild type and mutant proteins were virtually identical (Fig. 1).

Cell-free Activation of the Respiratory Burst Oxidase
The activity of protein kinase-activated NADPH oxidase was measured by chemiluminescence (31). Two types of assays were used as follows: a one-stage assay in which recombinant GST-p47 PHOX was added at the start of the incubation, and a two-stage assay in which the oxidase was partly activated in an initial incubation carried out in the absence of added GST-p47 PHOX and then fully activated and assayed for activity in a second incubation initiated by adding phosphorylated GST-p47 PHOX , NADPH, and the detection system to the initial incubation mixture. The assays were conducted by the following procedures.
One-stage Chemiluminescence Assay-The complete reaction mixture contained 2.5 ϫ 10 7 cell eq cytosol, 1.5 ϫ 10 7 cell eq membrane, 50 M GTP␥S, and the unphosphorylated GST-p47 PHOX mixture (5 g of recombinant unphosphorylated GST-p47 PHOX , 1 mM ATP, 5 units of protein kinase C, 10 mM MgCl 2 , 0.5 mM CaCl 2 , 25 g of phosphatidylserine, and 2.5 g of diacylglycerol, adding the lipids as mixed liposomes prepared by dissolving 1.0 mg of phosphatidylserine and 0.1 mg of diacylglycerol in chloroform, removing the chloroform under a stream of nitrogen, and then sonicating the dried lipids for 2 min on ice in 0.8 ml of 20 mM Tris buffer, pH 7.4) or GST-p47 PHOX P 6 (3 g GST-p47 PHOX that had been phosphorylated with protein kinase C) in 0.38 ml (final volume) relaxation buffer. A chemiluminescence detection mixture containing 18 g of horseradish peroxidase, 10 M luminol, and 0.16 mM NADPH (final concentrations) was then added to the reaction mixtures, either immediately (for GST-p47 PHOX P 6 -containing reactions) or after incubating for 5 min at 37°C (for reactions containing the unphosphorylated GST-p47 PHOX mixture). Oxidase activity was then determined by measuring chemiluminescence at room temperature in a Monolight 2010 luminometer (Analytical Luminescence Laboratories, San Diego) at successive 10-s intervals; the final volume of the assay was 0.5 ml.
One-stage Cytochrome c Assay-An identical protocol was used for the one-stage cytochrome c assay, except that the assay mixture contained 1.5 ϫ 10 7 cell eq solubilized membrane (prepared by mixing 100 l (1.25 ϫ 10 8 cell eq) of membrane suspension with 50 l of glycerol, 50 l of relaxation buffer, 25 l of octylglucoside (10%, w/v), and 25 l of sodium deoxycholate (10%, w/v) and incubating the mixture on ice for 15 min); the detection mixture contained 0.1 mM cytochrome c and 0.16 mM NADPH (final concentrations); the final volume was 0.75 ml; and cytochrome c reduction was followed at 550 nm for 5 min at room temperature in a Uvikon 941 dual-beam recording spectrophotometer (Kontron Instruments, Milan), reading against a reference containing the same components plus 45 g of superoxide dismutase.
Two-stage Chemiluminescence Assay-A reaction mixture containing 2.5 ϫ 10 7 cell eq cytosol and/or 1.5 ϫ 10 7 cell eq membrane as indicated plus 50 M GTP␥S in a total volume of 0.35 ml (the initial incubation) was incubated for 20 min at 37°C. The second incubation was then started by adding the chemiluminescence detection mixture (final total volume 0.5 ml), with or without 3 g of GST-p47 PHOX P 6 , and oxidase activity was determined by chemiluminescence as described above. In both the one-stage and two-stage chemiluminescence assays, light emission was followed until shortly past the point where it reached a maximum; this maximum, expressed in relative light units/s (RLU/s), is the luminescence value reported in the tables and figure legends. Deviations from these general procedures are indicated in the legends to the figures and tables.
The two chemiluminescence assays are diagrammed in Scheme I.

Depletion of ATP and Measurement of ATP Concentration
To deplete it of ATP, cytosol (1 ϫ 10 8 cell eq) was supplemented with 0.1 M glucose, 6 mM MgCl 2 , and 0.1 mg of hexokinase (final concentrations) and incubated for 30 min at room temperature. ATP concentrations were measured by an ATP bioluminescent assay kit (FL-AA, Sigma), calibrating by comparison with the chemiluminescence signal from a standard curve established for known concentrations of ATP.

Separation of Membrane from Cytosol after Initial Incubation in the Two-stage Assay
Membranes from the initial incubation were reisolated by layering the incubation mixture over a discontinuous sucrose gradient composed of 1 ml of 15% (w/v) sucrose layered over 0.5 ml of 50% (w/v) sucrose, both in relaxation buffer, and centrifuging at 105,000 ϫ g for 30 min at 4°C. After centrifugation, the contents of the centrifugation tube were carefully withdrawn from the bottom, discarding the first 250 l and saving the next 300 l as the preincubated membrane.

Electrophoresis and Immunoblotting
Protein samples containing GST-p47 PHOX were subjected to SDS-PAGE on 8% polyacrylamide gels using the Laemmli buffer system (28). The separated proteins were electrophoretically transferred onto a nitrocellulose sheet (32) and probed with a 1:5000 dilution of partially purified rabbit polyclonal antibody raised against C-terminal decapeptide from p47 PHOX and finally detected with a 1:2000 dilution of alkaline phosphatase-labeled goat anti-rabbit Ig antibody (Sigma) using 5-bromo-4-chloro-3-indolyl phosphate/nitroblue tetrazolium as substrate (Bio-Rad). To establish whether purification over GSH-agarose was able to separate the PKC in the phosphorylation mixture from the newly phosphorylated GST-p47 PHOX P 6 , blots prepared from the passthrough and eluate fractions obtained from the GSH-agarose beads were probed with a pan-PKC antibody (5 g/ml), visualizing the bands as described for the anti-p47 PHOX antibody.

Inhibitor Activity
The activities of inhibitors in the cell-free system were investigated by evaluating the incorporation of 32 P from [␥-32 P]ATP into proteins by gel electrophoresis and autoradiography. Reaction mixtures contained 1.5 ϫ 10 6 cell eq cytosol, 5 Ci of [␥-32 P]ATP, and where indicated 2 ϫ 10 6 cell eq membrane, 50 M GTP␥S, 0.25 M calyculin A (an inhibitor of protein phosphatases 1 and 2A (33)), and 5 M GFX (an inhibitor of protein kinases (34)) in a final volume of 30 l. Incubations were conducted for 20 min at 37°C. The reactions were then terminated with 2 ϫ sample buffer, and the reaction mixtures were subjected to SDS-PAGE on a 10% polyacrylamide gel. The gels were dried and the labeled proteins detected by autoradiography. The autoradiogram (Fig. 2) showed that in this system both inhibitors exerted their expected effects.

RESULTS
Activation of the Leukocyte NADPH Oxidase by Phosphorylation-In a cell-free system, the leukocyte NADPH oxidase has customarily been activated by certain anionic amphiphiles, including arachidonic acid and SDS. In an earlier study in which we supplemented the cell-free system with extra p47 PHOX (added as the GST fusion protein) (35), we found that the enzyme can also be activated by PKC. We have now employed the cytochrome c assay to obtain better quantitation of O 2 Ϫ production by the PKC-activated system (36). The results obtained with this assay (Table I) were qualitatively similar to those obtained with the chemiluminescence assay. In particular, maximum rates of O 2 Ϫ production were achieved only in the presence of added PKC and lipids. The sustained maximum rate of O 2 Ϫ production of 3.8 Ϯ 0.4 nmol/min/10 7 cell eq membrane in the PKC-activated system was Ϸ15-20% of the typical rate seen in the detergent-activated system.
Activation of the cell-free oxidase by PKC may represent a more physiological process than activation by anionic amphiphiles, because in intact cells, as in the PKC-activated cell-free system, oxidase activation is associated with a PKC-dependent event (in the case of intact cells, the phosphorylation of the oxidase subunit p47 PHOX ). Further evidence that the PKCactivated system may mimic the physiological route of oxidase activation was provided by experiments with a p47 PHOX mutant in which serine 379 was replaced by alanine (GST-p47 PHOX S379A). This mutant was shown to support O 2 Ϫ production by a cell-free system activated with anionic amphiphiles, but it did not restore oxidase activity to intact p47 PHOX -deficient B lymphoblasts (31). We found that GST-p47 PHOX S379A supported only very low levels of oxidase activity in the PKC-activated cell-free system (Fig. 3). In this respect, activation of the oxidase by PKC resembled activation of the oxidase in intact cells, supporting the physiological nature of the PKC-dependent activation process in the cell-free system.
An understanding of the mechanism of oxidase activation by PKC in this system is confounded to a certain extent by the lipid and calcium requirements of the kinase. To eliminate these from consideration, we examined the activation of the cell-free oxidase by PKM, an active fragment of PKC that no longer requires lipids (phosphatidylserine and diacylglycerol) or calcium. We found that the oxidase was activated by PKM (not shown), indicating that the exogenous lipids and calcium were required only to support PKC function and not to activate the oxidase per se.
In the foregoing experiments, the protein kinase was added directly to the assay mixtures. It was therefore hard to know whether O 2 Ϫ production in those experiments was related to the phosphorylation of p47 PHOX or to the phosphorylation of other proteins in the assay mixture. To address this question, we conducted experiments in which we replaced the combination SCHEME I of unphosphorylated GST-p47 PHOX plus PKC by GST-p47 PHOX that had been phosphorylated in isolation and then purified away from the phosphorylating kinase (i.e. PKC). Preliminary experiments showed that under the conditions used in our experiments, GST-p47 PHOX was completely phosphorylated by a 20-min incubation with PKC and could be cleanly separated from PKC by purification over glutathione-agarose (not shown). When the phosphorylated GST-p47 PHOX (henceforth designated GST-p47 PHOX P 6 ) was added to the assay mixture, the oxidase was activated without the need of additional PKC (Fig. 4). Furthermore, the extent of activation increased with increasing amounts of GST-p47 PHOX P 6 . These results strongly suggest that at least one of the events required for oxidase activation in this system is the phosphorylation of p47 PHOX .
Earlier studies showed that besides PKC, both mitogen-activated protein kinase (MAP kinase) and the catalytic subunit of protein kinase A (PKA) were able to phosphorylate p47 PHOX (37). These findings raised questions as to whether phosphorylation of p47 PHOX by either of these two kinases affected the subsequent phosphorylation of p47 PHOX by PKC or the ability of phosphorylated p47 PHOX to activate the leukocyte NADPH oxidase. The stoichiometry of phosphorylation of GST-p47 PHOX by the three kinases is shown in Fig. 5. As we found previously, 2 treatment of GST-p47 PHOX by PKC led to the incorporation of 6 phosphates/mol of protein. Treatment with PKM resulted in the same stoichiometry. Only 2 phosphates/mol of protein, however, were incorporated by GST-p47 PHOX treated with PKA, whereas MAP kinase treatment resulted in the incorporation of slightly less than 1.5 phosphates/mol of protein. Pretreatment of GST-p47 PHOX with PKA had no effect on the final amount of phosphate incorporated into the protein after subsequent treatment with PKM, consistent with earlier results showing that the serine residues phosphorylated by PKA were also phosphorylated by PKM (37). Pretreatment with MAP kinase, however, increased by Ϸ1 mol/mol of protein the final amount of phosphate incorporated into GST-p47 PHOX by subsequent treatment with PKM. As to the abilities of GST-p47 PHOX phosphorylated by various kinases to support oxidase activity in this system, the results are those expected if the protein had only been exposed to PKC (or PKM) (Fig. 6). GST-p47 PHOX phosphorylated with either MAP kinase or PKA was no more active than unphosphorylated GST-p47 PHOX , whereas the activity of GST-p47 PHOX phosphorylated by PKM was the same as that of PKC-treated GST-p47 PHOX regardless of whether or not the protein had been previously treated with MAP kinase or PKA. In summary, phosphorylation of GST-p47 PHOX by MAP kinase or PKA had little effect on oxidase activation in this system. Furthermore, other kinases in the incubation mixture did not appear to catalyze the incorporation of additional phosphate into GST-p47 PHOX once it had been fully phosphorylated by PKM (i.e. converted to GST-p47 PHOX P 6 ), because no radioactivity was found in GST-p47 PHOX P 6 reisolated on glutathione-agarose and counted after incubation for 30 min in an oxidase assay mixture containing 5 Ci of [␥-32 P]ATP (data not shown).
Although further phosphorylation of GST-p47 PHOX P 6 did not take place during oxidase activation, there was an additional ATP-dependent step in the activation sequence. This was shown in experiments using cytosol that had been depleted of  Ϫ production by the protein kinase C-activated cell-free system as measured by the cytochrome c assay The incubations were conducted as described under "Experimental Procedures," measuring O 2 Ϫ production by the one-stage cytochrome c assay. The results are presented as the mean Ϯ range of two experiments.  ATP by hexokinase plus glucose. Treatment of cytosol with the hexokinase-glucose combination reduced the ATP concentration in the treated cytosol to 67.3 Ϯ 12.5 nM (mean Ϯ S.D., n ϭ 3), a value 20 -40-fold below the K m for PKC (38). This ATPdepleted cytosol was as active as untreated cytosol when used in the standard SDS-activated cell-free system (data not shown). In the GST-p47 PHOX P 6 -activated system, however, the depleted cytosol was only 13.0 Ϯ 1.2% (mean Ϯ range, n ϭ 2) as active as untreated control. These findings suggest that at least two kinase-dependent reactions participate in protein kinasedependent oxidase activation, one involving the phosphorylation of p47 PHOX and the other the phosphorylation of a substrate (or substrates) yet to be identified.
ATP-dependent Activation of the Leukocyte NADPH Oxidase Is a Multistep Process-Measurements of activity as a function of time showed that the activation of the oxidase by phosphorylation occurs with a very long lag, peak activity not being achieved until half an hour has passed (Fig. 7). This is in contrast to oxidase activation in whole cells and in the amphiphile-activated cell-free system, in which the lag is less than 5 min in length. The lag seen in the protein kinase-dependent activation system was not shortened by preincubating GTP␥S (39 -41) with membranes alone, cytosol alone, or the combination of cytosol plus GST-p47 PHOX P 6 and PKM. The lag fell to less than 8 min, however, when cytosol and membranes together were preincubated with GTP␥S before starting O 2 Ϫ production with GST-p47 PHOX P 6 and NADPH. The fall in the lag was even sharper when cytosol plus membranes were preincu-bated with GTP␥S in the presence of calyculin A, a protein phosphatase inhibitor, O 2 Ϫ production in this reaction being started as before by the addition of GST-p47 PHOX P 6 and NADPH to the preincubation mixture. Okadaic acid, another protein phosphatase inhibitor (33,42), had the same effect in this system as calyculin A (not shown). In the calyculin A-containing assay, the maximum rate of O 2 Ϫ production was 25-30% that seen in a similar assay mixture activated by SDS, the latter amounting to 379,000 Ϯ 41,800 S.D. relative luminescence units/s (n ϭ 3). These results indicate that the protein kinase-dependent activation of the cell-free system can be divided into three distinct sets of events as follows: 1) events that occur during the initial incubation (i.e. the preincubation), 2) the phosphorylation of p47 PHOX , and finally, 3) events that occur during the second incubation (i.e. after the addition of GST-p47 PHOX P 6 and NADPH).
To determine which of the two fractions, membrane or cytosol, is altered by the initial incubation, experiments were carried out in which membranes and cytosol were incubated together in the presence of GTP␥S, then separated, combined with their unincubated complementary fractions plus GST-p47 PHOX P 6 and NADPH, and immediately assayed for O 2 Ϫ production. Activity was only seen in assay mixtures containing preincubated membranes (Table II). An assay mixture containing preincubated cytosol plus unincubated membranes showed no O 2 Ϫ production. The initial incubation therefore causes a modification of some kind affecting the membranes.

FIG. 5. Stoichiometry of phosphorylation of GST-p47 PHOX by various kinases and kinase combinations.
Phosphorylation of GST-p47 PHOX was carried out, and stoichiometry of phosphorylation was determined as described in the text. Values shown represent the mean Ϯ range of two to four separate determinations.

FIG. 6. Activation of the leukocyte NADPH oxidase by recombinant GST-p47 PHOX phosphorylated by various kinases and combinations of kinases.
The incubations were carried out as described under "Experimental Procedures," using the one-stage chemiluminescence assay. The reaction mixtures contained recombinant GST-p47 PHOX phosphorylated with various kinases and kinase combinations as indicated. Results are shown as the percent Ϯ 1 S.D. of control activity (n ϭ 3), where the control activity was that seen with GST-p47 PHOX P 6 (i.e. GST-p47 PHOX phosphorylated with protein kinase C). The absolute value for the control activity, representing the maximum rate of light emission observed during the course of a 40-min incubation, was 45800 Ϯ 4100 RLU/s (mean Ϯ 1 S.D., n ϭ 3). otide binding protein (GNBP) is required for the assembly of the oxidase, an event that probably does not occur in this experimental system until GST-p47 PHOX P 6 is added (13,43). To identify the stage(s) where the protein kinase-dependent activation process requires a GNBP and to determine whether activation depends on transfer of the ␥-phosphate of GTP, experiments were conducted in which incubation protocols were varied and analogs of GTP were employed. The results are shown in Table III. Oxidase activity was nil when the two incubations were conducted in the absence of GTP␥S or in the presence of GDP␤S, confirming that a GNBP is required for oxidase activity. A GNBP appeared to be required in both the initial and second incubations, because the addition of GTP␥S after the initial incubation was complete (i.e. at the same time as the addition of GST-p47 PHOX P 6 ), although leading to O 2 Ϫ production, resulted in a much lower activity than when GTP␥S was present for both the initial and second incubations. The activity generated when GTP␥S was added at the same time as GST-p47 PHOX P 6 probably reflected the extent to which the events of the initial incubation took place during the second incubation. A role for a GNBP in the second incubation was further supported by an experiment in which activated membranes reisolated from an initial GTP␥S-containing incubation were used in a second incubation with fresh cytosol containing either GTP␥S or GDP␤S. O 2 Ϫ production was brisk in the sec-ond incubation that contained GTP␥S, but nil in the second incubation that contained GDP␤S. Finally, the ability of Gp-pNHp to support oxidase activation indicates that activation does not require the cleavage of the bond between the ␤ and ␥ phosphates of GTP. Protein Kinase Inhibitors and the Phosphorylation of p47 PHOX -GFX is a powerful inhibitor of protein kinase C and other protein kinases (34). As expected from the ATP requirement shown in previous experiments, GFX was able to inhibit the activation of the oxidase in this system. O 2 Ϫ production was also inhibited by H-7, another protein kinase inhibitor (44) (not shown). Table IV shows that GFX inhibited oxidase activation considerably if present during the entire activation process but had no effect if added after the initial incubation, suggesting that ATP was required during the initial incubation but not during the second incubation. The possibility that GFX was acting exclusively against the phosphorylation of p47 PHOX was ruled out by the use of GST-p47 PHOX P 6 , the fully phosphorylated fusion protein. This result confirms a requirement for ATP beyond that necessary for the phosphorylation of p47 PHOX and strongly suggests that the action of one or more protein kinases FIG. 7. O 2 ؊ production by GST-p47 PHOX P 6 -activated leukocyte NADPH oxidase as a function of time.
Incubations were carried out as described in the text, using the one-stage or the two-stage chemiluminescence assay as indicated in the figure. Where shown, calyculin A was added to the first stage of the two-stage incubation at a concentration of 0.25 M. Results are representative of three separate experiments.

TABLE II
Expediting oxidase activation by preincubation of membranes with cytosol plus GTP␥S The two-stage chemiluminescence assay was conducted as described in the text, except that the initial incubations contained 0.25 M calyculin A, and membranes and cytosol were separated after the initial incubation as described under "Experimental Procedures," with the exception noted in the table. Fresh or preincubated materials were used in the second incubations as indicated in the table. Results are expressed as % of oxidase activity in a control experiment in which the membranes and cytosol were not separated after the initial incubation. The absolute value for the control experiment was 112,200 Ϯ 6,900 RLU/s. The results are expressed as the mean Ϯ range for two experiments. activation of the NADPH oxidase The two-stage chemiluminescence assay was conducted as described in the text, adding various guanine nucleotides at the times indicated in the table. In every case, the initial incubation contained 0.25 M calyculin A. Where indicated, membrane was reisolated from the initial incubation as described in the text and then mixed with fresh cytosol for the second incubation. Results are expressed as % of oxidase activity in a control experiment in which GTP␥S was added to the initial incubation. The absolute value for the control experiment was 117,700 Ϯ 400 RLU/s. The results are expressed as the mean Ϯ range for two experiments. is required during the initial incubation but that if phosphorylated p47 PHOX is supplied, further phosphorylation during the second incubation is unnecessary.
The kinase requirement during the initial incubation reflected more than just the preliminary phosphorylation of the p47 PHOX already present in the cytosol. This is indicated by the finding that cytosol deficient in p47 PHOX was able to function in the initial incubation, as indicated by the short lag and the high level of oxidase activity seen during the second incubation in experiments using the deficient cytosol (Fig. 8).
The use of GFX made it possible to ask whether unphosphorylated p47 PHOX could function in the second stage of the oxidase activation reaction. For this experiment, the initial incubation was carried out as described under "Experimental Procedures," but the second incubation received either no protein, unphosphorylated GST-p47 PHOX , or GST-p47 PHOX P 6 as indicated in Table V. GFX was added to the second incubations as indicated, to inhibit protein phosphorylation during the second stage of the reaction. The results showed that as compared with O 2 Ϫ production in a reaction mixture containing no added p47 PHOX , O 2 Ϫ production was more than doubled by the addition of unphosphorylated GST-p47 PHOX to the second incubation. This effect of unphosphorylated GST-p47 PHOX , however, was abolished by the simultaneous presence of GFX in the second incubation, suggesting that before it could stimulate O 2 Ϫ production, GST-p47 PHOX had to be phosphorylated. Confirming this conclusion were the results obtained with GST-p47 PHOX P 6 ( Table V). These results showed that the fully phosphorylated protein led to an even greater augmentation of O 2 Ϫ production than was seen with the unphosphorylated protein. Furthermore, in contrast to the results with unphosphorylated p47 PHOX , the increase seen with p47 PHOX P 6 was unaffected by the protein kinase inhibitor. These results indicate that 1) the form of p47 PHOX that was active in the protein kinase-dependent oxidase activating system was the phosphorylated protein, and 2) the phosphorylation of p47 PHOX was the only second stage phosphorylation required for the activity of that system.
Activity of Endogenous p47 PHOX -In most of the work described above, recombinant p47 PHOX P 6 was used in the assays. Neutrophil cytosol, however, contains an amount of p47 PHOX comparable to the amount of recombinant protein added to the incubation mixtures in the foregoing experiments, so to further evaluate the physiological significance of the protein kinasedependent oxidase activation system, it was necessary to determine whether this endogenous p47 PHOX could serve as a element of this system. For this purpose, cell-free oxidase activation by endogenous p47 PHOX phosphorylated with PKC was compared with cell-free oxidase activation by recombinant p47 PHOX P 6 . The detergent-activated cell-free system was also examined. The results (Fig. 9) showed that the peak rate of O 2 Ϫ production by the unsupplemented kinase-activated system was 60% greater than the rate seen in the GST-p47 PHOXsupplemented system, confirming that endogenous p47 PHOX could participate in the kinase-dependent oxidase activation system. O 2 Ϫ production by the detergent-activated system, however, was 5-fold greater than O 2 Ϫ production by the PKC-activated system (not shown), suggesting that additional facets of the kinase-dependent oxidase activation system remain to be investigated.

DISCUSSION
It has been known for many years that p47 PHOX , one of the cytosolic subunits of the leukocyte NADPH oxidase, becomes heavily phosphorylated on serine when the oxidase is activated (16 -22). More recent studies showed that the phosphates are confined to the C-terminal quarter of the molecule, identified many of the phosphorylation sites (37,45), and suggested a number of kinases, including PKC (22, 46 -53), the p21 rac/cdc42activated kinase (PAK kinase) (54), and other kinases yet to be characterized (55-58), as being potentially responsible for the  8. Activation of NADPH oxidase using p47 PHOX -deficient cytosol in the initial incubation. Incubations were carried out using a two-stage chemiluminescence assay as described in the text, with the following modifications. In the first stage, normal cytosol was replaced by p47 PHOX -deficient cytosol, and 0.25 M calyculin A was present. In the second stage, wild type GST-p47 PHOX P 6 or the inactive mutant GST-p47 PHOX P 6 S379A was added or GST-p47 PHOX P 6 was omitted, as indicated in the figure. Results are representative of two separate experiments.

TABLE V
Oxidase activation by endogenous p47 PHOX ; requirement for phosphorylated p47 PHOX in the second stage of the protein kinasedependent oxidase activation system The two-stage chemiluminescence assay was conducted as described in the text, except that calyculin A (0.25 M) was added to the first stage and GST-p47 PHOX P 6 , unphosphorylated GST-p47 PHOX , or no protein was added to the second stage of the incubations as indicated. The protein kinase inhibitor GFX (5 M) was also added to the second stages of some of the incubations. The results are expressed as the mean Ϯ S.D. of n replicate incubations. The absolute value for the control experiment with no added GST-p47 PHOX was 25,300 Ϯ 4,900 S.D. RLU/s. GST phosphorylation of p47 PHOX . The question is still open, however, whether the phosphorylation of p47 PHOX actually plays a role in oxidase activation or is just an epiphenomenon; the only serine in the phosphorylated portion of the molecule that has been shown to be essential for oxidase activity is Ser-379, and it is not certain that Ser-379 is a functionally significant phosphorylation target (31). The experiments reported here have shown that the oxidase can be activated in a cell-free system by the addition of phosphorylated GST-p47 PHOX but not by the unphosphorylated molecule. They show further that p47 PHOX phosphorylated by PKC is competent to activate the oxidase but that p47 PHOX phosphorylated by PKA or MAP kinase is inactive. These findings strongly suggest that the phosphorylation of p47 PHOX that occurs during the activation of the leukocyte NADPH oxidase is of functional significance and that PKC is at least one of the enzymes capable of converting p47 PHOX into a functionally active molecule by phosphorylation. They further suggest that the kinase-dependent activation mechanism described here may reflect at least in part an oxidase activating system that operates in the whole cell, because of the extensive literature implicating phosphorylation in the activation of the oxidase and because a p47 PHOX mutant that is nonfunctional in the whole cell (i.e. p47 PHOX S379A (31)) is also nonfunctional in the kinase-dependent cell-free system, even though the phosphorylation of the p47 PHOX mutant appears to be normal both in vitro and in whole cells (37). The assay mixtures used in these experiments contained 2.5 ϫ 10 7 cell equivalents of cytosol. Assuming that the volume of a neutrophil is 500 fl, that its cytosol contains 20% protein (w/v), and that p47 PHOX represents 0.4% of this protein (43), it can be calculated that the cytosol contributed Ϸ10 g of unphosphorylated p47 PHOX to the reaction mixture. From this calculation and the above results, it appears that the oxidase activity elicited by GST-p47 PHOX P 6 is roughly comparable to that which would be observed in a system containing a similar amount of (unphosphorylated) p47 PHOX . Therefore on a weightfor-weight basis, GST-p47 PHOX P 6 seems to be as active as p47 PHOX that has been activated by endogenous posphorylation. This conclusion is supported by the results in Fig. 9, which shows that the activity achieved in an assay containing only endogenous p47 PHOX is comparable to that obtained in an assay that was supplemented with GST-p47 PHOX P 6 .
The foregoing results also show that the protein kinase-dependent activation of leukocyte NADPH oxidase in a cell-free system is a multi-step process that can be divided into three distinct stages: 1) the activation of the membrane, which appears to take place during the initial incubation; 2) the phosphorylation of p47 PHOX ; and finally 3) the assembly of the active oxidase on the activated membrane. The activation of the membrane seems to be the most complicated of these events. The participation of a GNBP in membrane activation is suggested by the finding that GTP is required in the initial incubation, and the involvement of one or more protein kinases is suggested by the ATP requirement in the initial incubation and by the ability of GFX and H-7, antagonists of protein kinase C, to inhibit membrane activation.
All these observations suggest the following as a possible route of activation for the oxidase: 1) processing of the membrane, a complex series of events involving a GNBP-dependent step and a protein kinase; 2) the activation of PKC, which phosphorylates p47 PHOX ; and finally 3) the assembly of the active NADPH oxidase on a membrane that somehow has been rendered capable of supporting oxidase activation by virtue of its newly acquired protein phosphate. With regard to membrane processing, the transfer to the membrane of Rac2, a GNBP, is already known to participate in oxidase activation (41,59), but many other possibilities can be considered, as for example the activation of the p21 rac/cdc42 -activated kinase (54), a single event that could explain the dependence of membrane processing on both a GNBP and a protein kinase; phosphorylation of a membrane-associated oxidase subunit; the participation of phosphatidic acid in ATP-dependent oxidase activation as described by McPhail and associates (27); and others. It is clear, however, that protein kinase-dependent oxidase activation requires an alteration in the membrane, at least in this system and perhaps in the intact cell as well.
In the activation of the cell-free system by amphiphiles, a GNBP (specifically, Rac1 or Rac2) is needed for the assembly of the oxidase (40,41). The GTP requirement in the second incubation implies that a GNBP is necessary for the events taking place in that incubation, most likely the assembly of the oxidase, although the possibility of other reactions taking place during this incubation cannot be excluded on the basis of the present results. With regard to ATP, the experiments with ATP-depleted cytosol and the protein kinase inhibitor GFX suggest that, apart from its role in the phosphorylation of p47 PHOX , this nucleotide is not required in the second incubation. A requirement for ATP in the second incubation cannot be conclusively ruled out, however, because it is conceivable that the very low level of ATP in the depleted cytosol is sufficient to support any (hypothetical) ATP-dependent reactions that may take place during the second incubation and that such hypothetical reactions may not be affected by GFX. FIG. 9. O 2 ؊ production by the leukocyte NADPH oxidase activated by p47 PHOX P 6 and by protein kinase C. For activation by protein kinase C, a reaction mixture containing 2.5 ϫ 10 7 cell eq cytosol, 1.5 ϫ 10 7 cell eq membrane, and 50 mM GTP␥S were treated with a mixture of 1 mM ATP, 0.5 units of PKC, 0.5 mM CaCl 2 , 25 g of phosphatidylserine, 2.5 g of diacylglycerol, and 0.25 M calyculin A in a total volume of 0.35 ml for 20 min at 37°C as indicated in the text (see "Two-stage Chemiluminescence Assay"). The control was preincubated for 20 min at 37°C in the absence of kinase, then supplied with chemiluminescence detection mixture, and immediately activated by adding 5 g of GST-p47 PHOX P 6 as indicated. Final volumes of the assay mixtures were 0.5 ml. Results are the mean Ϯ S.D. of three separate experiments.