Activation of the Leukocyte NADPH Oxidase by Phorbol Ester Requires the Phosphorylation of p47 PHOX on Serine 303 or 304*

The leukocyte NADPH oxidase is an enzyme in phagocytes and B lymphocytes that when activated catalyzes the production of O 2 . from oxygen and NADPH. During oxidase activation, serine residues in the C-terminal quarter of the oxidase component p47 PHOX become extensively phosphorylated, the protein acquiring as many as 9 phosphate residues. In a study of 11 p47 PHOX mutants, each containing an alanine instead of a serine at a single potential phosphorylation site, we found that all but S379A corrected the defect in O 2 . production in Epstein-Barr virus (EBV)-transformed p47 PHOX -defi-cient B cells (Faust, L. P., El Benna, J., Babior, B. M., and Chanock, S. J. (1995) J. Clin. Invest. 96, 1499–1505). In particular, O 2 . production was restored to these cells by the mutants S303A and S304A. Therefore, apart from serine 379, whose state of phosphorylation in the activated oxidase is unclear, no single potential phosphorylation site appeared to be essential for oxidase activation. We now report that the double mutant p47 PHOX S303A/S304A

the mutants S303A and S304A. Therefore, apart from serine 379, whose state of phosphorylation in the activated oxidase is unclear, no single potential phosphorylation site appeared to be essential for oxidase activation. We now report that the double mutant p47 PHOX S303A/S304A was almost completely inactive when expressed in EBV-transformed p47 PHOX -deficient B cells, even though it was expressed in normal amounts in the transfected cells and was able to translocate to the plasma membrane when the cells were stimulated. In contrast, the double mutant p47 PHOX S303E/S304E was able to support high levels of O 2 . production by EBV-transformed p47 PHOX -deficient B cells. The surprising discovery that the double mutant S303K/S304K was also able to support considerable O 2 . production suggests either that the effect of phosphorylation is related to the increase in hydrophilicity around serines 303 and 304 or that activation involves the formation of a metal bridge between the phosphorylated serines and another region of the protein.
The leukocyte NADPH oxidase is a membrane-associated enzyme in phagocytes and B lymphocytes that catalyzes the production of O 2 . from oxygen using NADPH as electron donor Dormant in resting cells, it acquires catalytic activity when the cells are exposed to appropriate stimuli. Activation involves the transfer of cytosolic subunits designated p47 PHOX and p67 PHOX to the plasma membrane, where they associate with a flavocytochrome known as cytochrome b 558 to assemble the active oxidase (2). During oxidase activation in whole cells, p47 PHOX becomes phosphorylated on numerous serine residues that lie between Ser 303 and Ser 379 in the C-terminal quarter of this 390-residue molecule (3)(4)(5)(6)(7). Studies conducted to date have identified the phosphorylated serines (7)(8)(9), examined their susceptibility to phosphorylation by various protein kinases that occur in neutrophils (9,10), and shown that in all likelihood no single phosphorylated serine is indispensable for oxidase activity (8). The present report is concerned with the phosphorylation of p47 PHOX Ser 303 and Ser 304 in relation to the activation of the leukocyte NADPH oxidase.

MATERIALS AND METHODS
The mutants p47 PHOX S303A/S304A, S303D/S304D, S303E/S304E, and S315A/S320A were constructed by single strand mutagenesis of the p47 PHOX cDNA as described previously (8). Mutagenesis was performed on cDNA cloned into pBluescript II KSϩ. The mutated cDNAs were then subcloned into the mammalian expression vector EBOpLPP. Mutations that altered the sequence of the protein were accompanied by silent mutations of nearby restriction sites, introduced to aid in screening. The mutations were confirmed by sequencing in the departmental facility.
The S303K/S304K, S303A/S304E, and S303E/S304A mutations were constructed by a PCR 1 strategy. In each case, the 5Ј end of the "forward" primer spanned a unique NarI site 16 bases upstream of the 303 codon, the appropriately altered 303 and 304 codons, and 18 additional bases of native downstream sequence. A 19-base "reverse" primer that spanned an NaeI site about 180 base pairs downstream from the 304 codon was used for all three PCR mutagenesis reactions. PCR was performed with Pfu polymerase (Stratagene) using the wild-type p47 PHOX cDNA as template. The PCR products were isolated from agarose gels using QIAEXII (Qiagen) and digested with NarI and NaeI. The resulting fragments were ligated into the NarI and NaeI sites of the p47 PHOX S303A/S304A plasmid, a procedure that introduced the desired mutation into the product and destroyed a BssHII site that was introduced during the construction of the p47 PHOX S303A/S304A cDNA. All PCR mutations were confirmed and the constructs documented to be error-free by sequencing across the full NarI-NaeI span. Table I lists the primers and plasmids used in preparing the p47 PHOX mutants.
EBV-transformed p47 PHOX -deficient B lymphocytes were co-transfected with SV40 plus wild-type or mutant p47 PHOX expression vectors as indicated, and expanded under hygromycin selection as described previously (8), except that the p47 PHOX -deficient cells were maintained at 0.5-1.0 ϫ 10 6 /ml before transfection and 10 6 cells/ml after transfec- tion. Transfected cells were assayed only when fewer than 10% of the cells took up trypan blue. The cell line used for these experiments contained an uncharacterized mutation in the p47 PHOX gene that prevented the expression of p47 PHOX .
Leukocyte NADPH oxidase activity was measured by chemiluminescence. Assays using whole cells were carried out as described elsewhere (11), except that 4 ϫ 10 6 cells and 10 IU of horseradish peroxidase were used in a final volume of 0.35 ml. The cell suspensions were placed in a 96-well microplate, warmed to 37°C, then activated at the same temperature with phorbol myristate acetate (1 g/ml). Chemiluminescence was then measured at 1-min intervals using a Luminoskan luminometer (Labsystems Research, Finland) at 37°C. For measurement of leukocyte NADPH oxidase activity in a cell-free system, reaction mixtures contained 1.6 ϫ 10 6 cell equivalents of neutrophil membranes, 9 ϫ 10 6 cell equivalents of B lymphocyte cytosol, 1 mM luminol, 5 IU of horseradish peroxidase, 90 M SDS, 160 M NADPH, and Hanks' balanced salt solution containing 0.5 mM CaCl 2 and 1 mM MgCl 2 , with or without 50 IU of superoxide dismutase, in final volume of 1 ml. Cytosol was prepared by sonicating a suspension of lymphoblasts in Dulbecco's phosphate-buffered saline for three 10-s intervals at 4°C, then removing particles by centrifugation for 15 min at the same temperature in an Eppendorf Microfuge. Initially, the assay mixture contained all the components except SDS and NADPH. The oxidase was then activated by adding SDS and incubating for 1 min at room temperature. O 2 .
production was then initiated with NADPH, and chemiluminescence was measured at successive 10-s interval using a Luminoskan luminometer at room temperature. The expression of p47 PHOX in the transfected cells and the translocation of p47 PHOX from cytosol to membranes was determined by immunoblotting (8). Fractions were subjected to SDS-polyacrylamide gel electrophoresis on 10% polyacrylamide gels using the Laemmli buffer system. The separated proteins were electrophoretically transferred onto a nitrocellulose membrane, which was blocked with Blotto and probed with a 1:5000 dilution of an antibody against the C-terminal decapeptide of WT p47 PHOX . In measurements of p47 PHOX expression in whole cells, p47 PHOX on the immunoblots was visualized with alkaline phosphatase-labeled goat anti-rabbit immunoglobulin antibodies followed by visualization with the 5-bromo-4-chloro-3-indolyl phosphate/ nitro blue tetrazolium detection system (12). Expression was measured by densitometry with a Zeinieh laser scanner, determining relative quantities of p47 PHOX from peak heights. In the translocation experiments, peroxidase-labeled goat anti-rabbit immunoglobulin antibodies and the ECL chemiluminescence detection system (Renaissance; Du-Pont, Boston, MA) were used for the visualization of p47 PHOX .

O 2
. Production-In prior studies, we have shown that the activation of the leukocyte NADPH oxidase in neutrophils and EBV-transformed B lymphocytes is accompanied by the phosphorylation of serines 303 and 304 in the cytosolic oxidase component p47 PHOX (7). Consistent with the possibility that the phosphorylation of these serines is involved in oxidase activation, we reported that the conversion of either lymphocytes and by cells expressing wild-type or mutant forms of p47 PHOX are illustrated in Fig. 1; peak chemiluminescence values for the same cells are shown in Fig. 2. Inspection of the results obtained with the mutant in which serines 303 and 304 were converted to alanines showed that O 2 . production was nearly abolished by these mutations, supporting the idea that the phosphorylation of at least one of these serines is required for oxidase activation. A second double Ser 3 Ala mutant, p47 PHOX S310A/S315A, was fully active, indicating that the mere replacement with alanines of 2 serines in the region of the phosphorylation targets was not enough to abolish the activity of p47 PHOX . Further support for this idea was obtained when these two serines were replaced, not with alanines, but with negatively  charged residues. One of the effects of serine phosphorylation is the introduction of negative charge in a region of a polypeptide that was originally uncharged. Negative charge can also be introduced by replacing a serine with an aspartate or glutamate residue, and in fact studies on other proteins activated by serine phosphorylation have shown that the replacement of a phosphorylated serine with an aspartate residue can have the same functional consequences as the phosphorylation of that serine (13). To ascertain whether the introduction of negative charge in the vicinity of p47 PHOX Ser 303 -Ser 304 had the same effect on oxidase activity as the phosphorylation of serines 303 and 304, we measured O 2 . production in EBV-transformed B lymphocytes expressing the p47 PHOX double mutants S303D/ S304D and S303E/S304E. The results showed that, in contrast to the double p47 PHOX mutant S303A/S304A, the glutamate mutant was able to support O 2 . production. In cells that expressed p47 PHOX S303E/S304E, oxidase activation ( Fig. 1) and maximum rates of O 2 . production ( Fig. 2) were both similar to their counterparts in cells transfected with WT p47 PHOX . On the other hand, the results obtained with the cells transfected with p47 PHOX S303D/S304D were equivocal, not an unexpected outcome given that in steric terms, p47 PHOX S303S/S304S-OPO 3 ϭ is much more similar to p47 PHOX S303E/S304E than to p47 PHOX S303D/S304D. Replacement of one of the two serines with glutamate and the other with alanine led to p47 PHOX mutants that supported O 2 . production to the extent of about 25% of wild-type levels, a finding consistent with earlier work showing that mutants in which one of the two serines was replaced by alanine and the other was left alone were able to support O 2 . production, but at reduced levels (8).
Expression and Translocation of the p47 PHOX Mutants-One possible explanation for the inability of cells transfected with the S303A/S304A mutant to generate O 2 . is that the EBVtransformed p47 PHOX -deficient B cells were unable to express p47 PHOX S303A/S304A, the mutant that failed to support O 2 .
production. The findings presented in Fig. 3 and Table II . production in a cell-free system in which activation was accomplished by an anionic detergent (in this case, SDS), not phosphorylation. Fig. 4 shows that there was little difference in O 2 . production among cell-free systems containing the various forms of p47 PHOX . In particular, the cell-free system that contained p47 PHOX S303A/S304A produced O 2 . at the same rate as the system that contained WT p47 PHOX . Therefore the conformation of p47 PHOX S303A/S304A was sufficiently similar to the native conformation that the mutant protein functioned normally in the cell-free oxidase activating system, even though it was inactive in whole cells. Activation of the leukocyte NADPH oxidase is accompanied by the translocation of the oxidase components p47 PHOX and p67 PHOX from the cytosol to the plasma membrane, where they associate with the flavocytochrome b 558 (2,14). This association is thought to result at least in part from the appearance of a membrane-binding site on p47 PHOX when the oxidase is activated (12,15). We found that when p47 PHOX -deficient cells expressing the double mutants p47 PHOX S303A/S304A, p47 PHOX S303D/S304D, or p47 PHOX S303E/S304E were activated with phorbol ester, the mutant p47 PHOX polypeptides were transferred to the plasma membrane as efficiently as the WT protein (Fig. 5). The finding that p47 PHOX S303A/S304A translocates normally suggests that a negative electrostatic potential in the vicinity of residues 303-304 is not needed for the binding of p47 PHOX to the flavocytochrome, although it is required for the enzyme to express its catalytic activity.
p47 PHOX S303K/S304K, an Active Mutant-Besides adding local negative charge to p47 PHOX , the phosphorylation of serines 303 and 304 causes a large increase in the polarity of the polypeptide in the vicinity of those residues. To determine whether the effect of phosphorylation on oxidase activity was due to the negative electrostatic potential or the increase in  polarity, we examined the activity of the mutant p47 PHOX S303K/S304K, in which residues 303 and 304 carry positive, not negative, charges. The translocation of this mutant was normal (Fig. 6), and its activity was normal as well, somewhat to our surprise ( Figs. 1 and 2). Therefore the function of the Ser 303 and Ser 304 phosphates in the activation of the oxidase must be related to more than simply the acquisition of negative charge by the serine 303-304 region of p47 PHOX .

DISCUSSION
In an earlier study of 11 p47 PHOX mutants containing an alanine instead of a serine at a single potential phosphorylation site, we found that all but p47 PHOX S379A were active when expressed in EBV-transformed p47 PHOX -deficient B cells (8). Therefore, apart from serine 379, no single phosphorylation site appeared to be essential for oxidase activation. As to Ser 379 , its state of phosphorylation in the activated oxidase is unclear: it became phosphorylated only to a very limited extent when neutrophils were activated, and p47 PHOX S379D was no more active than p47 PHOX S379A in the p47 PHOX -deficient cells. 2 Consequently, it was not possible to be absolutely certain from those results whether or not the phosphorylation of p47 PHOX , either at Ser 379 or at any other potential phosphorylation site, was required for the activation of the oxidase. Unlike Ser 379 , serines 303 and 304 are extensively phosphorylated during oxidase activation (7,9). In earlier work, we found that the oxidase could still be activated if either serines 303 or 304 were converted to alanine (8). The present results, however, show that if both of these serines are converted to alanine, the activation of the oxidase is almost completely abolished. These findings provide direct evidence that the phosphorylation of at least one of the serines of p47 PHOX , specifically, serines 303 or 304, is necessary for the activation of the leukocyte NADPH oxidase. Our results support similar conclusions drawn from earlier studies showing that p47 PHOX is phosphorylated when the oxidase is activated by a variety of stimuli (3,(5)(6)(7); that various protein kinase inhibitors can diminish or prevent O 2 .
production (16 -18) while protein phosphatase inhibitors augment O 2 . production (19 -23); and that recombinant p47 PHOX phosphorylated by protein kinase C can activate the oxidase in the cell-free system in the absence of added detergents (24 -26). The effect of the serine-to-alanine mutations is to substitute methyl groups for two of the hydroxymethyl groups known to be phosphorylated during oxidase activation in whole cells (7). The conversion of these serines to alanines destroyed the activity of p47 PHOX . In contrast, the conversion of the same serines to glutamate, a mutation that substituted negatively charged carboxymethyl groups for the two hydroxymethyl groups, gave rise to mutant protein with considerable activity. Since the phosphorylation of other serines in p47 PHOX was unaffected by the elimination of serines 303 and 304 (9), the simplest explanation for all these findings is that oxidase activation in whole cells requires a negative electrostatic potential in the vicinity of serines 303 and 304 of p47 PHOX , and that this potential is generated through the phosphorylation of one or both of those serines. The results with the S303K/S304K mutant, however, indicate that the acquisition of a negative electrostatic potential is not a sufficient explanation. They suggest instead that the purpose of the phosphorylation of Ser 303 and Ser 304 may be to change the polarity of the protein in the neighborhood of those serines. Alternatively, they may indicate that the phosphates are there to bind divalent metals such as Mg 2ϩ , an action that would confer on the serine 303-304 region a net positive charge similar to that in the S303K/S304K mutant. Experiments were carried out as described previously (8), except for changes in the immunoblotting procedure as described under "Materials and Methods." Results are representative of two experiments, each carried out with a separate transfection. The track labeled CTRL contained 1.5 ϫ 10 6 cell eq of cytosol. The remaining tracks each contained 1.25 ϫ 10 7 cell eq of membrane. Plus (ϩ) and minus (Ϫ) indicate cells activated with phorbol or resting cells, respectively.
FIG. 6. Translocation of p47 PHOX S303K/S304K during oxidase activation. Experiments were carried out as described previously (8), except for changes in the immunoblotting procedure as described under "Materials and Methods." Results are representative of two experiments, each carried out in duplicate with a separate transfection. The track labeled CTRL contained 1.5 ϫ 10 6 cell eq of cytosol. The remaining tracks each contained 1.25 ϫ 10 7 cell eq of membrane. Plus (ϩ) and minus (Ϫ) indicate cells activated with phorbol or resting cells, respectively.
How can these local effects promote the activity of the leukocyte NADPH oxidase? Because the mutants all translocate normally, it is clear that phosphorylation at Ser 303 -Ser 304 is not an obligatory precursor of translocation. In principle, then, phosphorylation at these positions can take place either before or after the translocation of p47 PHOX to the membrane. Earlier evidence showing that the final 1 or 2 phosphorylations take place after translocation (6) suggests that at least some p47 PHOX is phosphorylated after it is transferred from the cytosol to the membrane. It seems likely that the phosphorylation of Ser 303 and Ser 304 converts the membrane-bound subunit into a form that is able to render the oxidase catalytically active.