Characterization of a Mutation in the Phox Homology Domain of the NADPH Oxidase Component p40phox Identifies A Mechanism for Negative Regulation of Superoxide Production*

The phagocyte oxidase (Phox) protein p40phox contains a Phox homology (PX) domain which, when expressed alone, interacts with phosphatidylinositol 3-phosphate (PtdIns (3)P). The functions of the PX domain in p40phox localization, association with the cytoskeleton, and superoxide production were examined in transgenic COS-7 cells expressing gp91phox, p22phox, p67phox, and p47phox (COSphox cells). Full-length p40phox exhibited a cytoplasmic localization pattern in resting cells. Upon stimulation with phorbol 12-myristate 13-acetate or fMet-Leu-Phe, p40phox translocated to plasma membrane in a p67phox- and p47phox-dependent manner. Heterologous expression of p40phox markedly enhanced superoxide production in phorbol 12-myristate 13-acetate - and fMet-Leu-Phe-stimulated COSphox cells. Unexpectedly, mutation of Arg-57 in the PX domain to Gln, which abrogated PtdIns (3)P binding, produced a dominant inhibitory effect on agonist-induced superoxide production and membrane translocation of p47phox and p67phox. The mutant p40phox (p40R57Q) displayed increased association with actin and moesin and was found enriched in the Triton X-100-insoluble fraction along with p67phox and p47phox. The enhanced cytoskeleton association of p67phox and p47phox and the dominant inhibitory effect produced by the p40R57Q were alleviated when a second mutation at Asp-289, which eliminated p40phox interaction with p67phox, was introduced. Likewise, cytochalasin B treatment abolished the dominant inhibitory effect of p40R57Q on superoxide production. These findings suggest a dual regulatory mechanism through the PX domain of p40phox; its interaction with the actin cytoskeleton may stabilize NADPH oxidase in resting cells, and its binding of PtdIns (3)P potentiates superoxide production upon agonist stimulation. Both functions require the association of p40phox with p67phox.

Phox homology (PX) 3 domains are evolutionarily conserved protein modules of 120 -140 amino acids that bind phosphoinositides. Initially named for their presence in the two cytosolic factors of NADPH oxidase, p47 phox and p40 phox (1), PX domains have been identified in more than 150 eukaryotic proteins including the sorting nexins (SNX1-15), vacuolar sorting and morphogenesis proteins (Vam7p, Vps5p, and Vps17p), yeast bud-emergence proteins (Bem1p and Bem3p), and phospholipase D2 (2,3). The PX domains from these proteins interact with a variety of phosphoinositides.
Published studies have shown that the PX domain in p40 phox binds phosphatidylinositol 3-phosphate (PtdIns (3)P), and the PX domain in p47 phox preferentially interacts with phosphatidylinositol 3,4-phosphates (4 -6). A proposed function of the PX domain is membrane targeting of proteins containing this structural module. In studies using a green fluorescence protein (GFP)-fused PX domain of p40 phox , membrane localization was observed in a phosphatidylinositol 3-kinasedependent manner (4,5). Membrane binding of the PX domains involves electrostatic interaction as well as membrane penetration by hydrophobic residues in the PX domain-containing proteins (7). Structural analysis of the PX domain of p40 phox reveals a positively charged binding pocket for the negatively charged PtdIns (3)P. Binding of p40 phox to the phosphoinositide requires three conserved arginine residues (Arg-57, Arg-58, and Arg-105) that stabilize a critical lipid binding loop within the PX domain (8). Mutation of any one of the three arginines can cause a significant reduction in binding of PtdIns (3)P (4,5).
Studies have been conducted for the function of p40 phox in NADPH oxidase activation since its initial discovery as a p67 phox -associated protein (9 -11). These studies have resulted in different and sometimes conflicting observations. Evidence supporting a positive regulatory role of p40 phox came from studies using both cell-free reconstitution and whole-cell assays. The possible mechanisms for p40 phox -mediated potentiation of NADPH oxidase include increasing the affinity of p47 phox for flavocytochrome b 558 (12), binding to membraneassociated PtdIns (3)P through its PX domain (5) and cooperation with p67 phox for membrane translocation of the cytosolic complex (13). Other investigators, using essentially the same cells and cell-free reconstitution assays, found p40 phox to be a negative regulator for NADPH oxidase. The negative regulatory mechanisms include SH3 domain-mediated interference of p40 phox association with other cytosolic factors (14) and inhibition of p67 phox membrane translocation (15). More recent studies have examined the roles of p40 phox in NADPH oxidase activation using transfected cells and mouse models. Suh et al. (16) reported that p40 phox is required for Fc␥R receptor-mediated superoxide generation after phagocytosis, a function that was lost when critical residues for PtdIns (3)P binding were mutated. Ellson et al. (17) found that neutrophils from p40 phox knock-out mice displayed defective oxidant production in response to several types of stimuli. Moreover, replacement of the mouse p40 phox gene with one that contains a Arg-58 to Ala mutation caused embryonic lethality in homozygous offspring, with the heterozygous mice displaying compromised ability to kill Staphylococcus aureus (18). These findings demonstrate a physiological function of p40 phox in regulating NADPH oxidase activation that involves its PX domain.
Despite recent progress in p40 phox research, the function of p40 phox in resting cells remains undefined. p40 phox was originally discovered as a p67 phox -associated protein (9 -11). In unprimed neutrophils, p40 phox forms a complex with p67 phox , whereas p47 phox was not a part of the complex (19). Neutrophils from chronic granulomatous disease (CGD) patients who lack p67 phox contain very little p40 phox (11,15), suggesting that interaction between the two cytosolic factors helps to stabilize their structures. Moreover, p40 phox , like the other cytosolic factors, associates with the actin cytoskeleton in resting neutrophils and with membrane skeleton in activated neutrophils (20). One of the proteins that helps to mediate protein association with the actin cytoskeleton is moesin, which interacts with the PX domain of p40 phox (21). Based on these findings, we speculate that the PX domain in p40 phox may have dual regulatory functions through its interaction with the actin cytoskeleton and with PtdIns (3)P. In the current study, we employed a COS-7-based whole-cell reconstitution system (22) to examine the effects of a full-length p40 phox and a PX domain mutant on NADPH oxidase activity. We observed that expression of the wild type p40 phox could enhance superoxide generation in response to both PMA and fMet-Leu-Phe (fMLF), a finding consistent with recent publications suggesting that p40 phox enhances NADPH oxidase activation (16 -18). Surprisingly, an Arg to Gln mutation at position 57 (R57Q), which abolishes p40 phox interaction with PtdIns (3)P through its PX domain (4), switched p40 phox to a different mode of action. It not only abrogated the potentiation effect but also produced a dominant inhibitory effect on superoxide generation. We found an increased association of p40R57Q with actin and moesin compared with the wild type p40 phox . In cells expressing p40R57Q, more cytosolic factors were targeted to the Triton X-100-insoluble fraction than in cells expressing the wild type p40 phox . The dominant inhibitory effect of p40R57Q was eliminated when the cells were treated with cytochalasin B, which prevents actin polymerization, or when the association of p40 phox with p67 phox was eliminated. These intriguing findings suggest that p40 phox can positively and negatively regulate NADPH oxidase through its PX domain interaction with PtdIns (3)P and with the actin cytoskeleton.
Plasmid Constructs-Preparation and characterization of the expression constructs of formyl peptide receptor (FPR), protein kinase C␦ and p40 phox were described in a previous publication (23). The full-length cDNA encoding the human p40 phox was subcloned in-frame with GFP in pEGFP-N1 vector (Clontech, Palo Alto, CA) to produce a p40 phox protein fused to the N terminus of GFP. Point mutations of p40 phox were generated with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). The oligonucleotide primers used for the PCR-based mutagenesis were 5Ј-GTACCTCATCTACCAAC-GCTACCGCCAGTTC-3Ј and its reverse and complementary sequence for the R57Q mutant and 5Ј-CTGAATTACCGGG-CCGCTGAGGGGGATC-3Ј and its reverse and complement primer for the D289A mutant. Both primer pairs were used for construction of the double mutant p40R57Q/D289A. All DNA constructs were verified by automated sequencing.
Cell Culture and Transient Transfection-The transgenic COS phox and COS 91/22 cells were generated as described previously (22). COS 91/22 expresses gp91 phox and p22 phox . Subsequent transfection resulted in COS phox , which expresses p67 phox and p47 phox in addition to gp91 phox and p22 phox (22). The stable cell lines were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum and antibiotics for proper selection (22). Cells plated in 90-mm (diameter) tissue culture dishes (0.5-1 ϫ 10 6 cells per dish) were transfected using Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's instructions. A total of 6.5 g of DNA was used in each transfection. Transient transfection efficiency of 45-50% was routinely obtained based the expression of a co-transfected GFP construct using flow cytometry.
The human myelomonoblastic cell line PLB-985 was maintained in RPMI 1640 medium (Invitrogen) supplemented with 10% heat-inactivated fetal bovine serum, 100 IU/ml penicillin, 50 g/ml streptomycin, and 2 mM L-glutamine. The cells were grown in suspension at a density between 2 ϫ 10 5 /ml and 1 ϫ 10 6 /ml. Cell line Nucleofector Kit V (Amaxa Biosystems, Cologne, Germany) was used for transient transfection of 3 ϫ 10 6 PLB-985 cells with 5 g of DNA using Program C-023. Transfection efficiency was ϳ30% as determined by flow cytometry based on the fluorescence of a co-expressed green fluorescent protein.
Measurement of NADPH Oxidase Activity-Superoxide produced by COS phox and PBL-985 cells was determined using an isoluminol-enhanced chemiluminescence assay, as previously described (23,24). Oxidant production was inhibited by superoxide dismutase (250 units) as reported previously (23). The assay buffer contained horseradish peroxidase (see below) to offset the possible effect of myeloperoxidase. Briefly, COS phox cells were harvested with enzyme-free cell dissociation buffer (Invitrogen). Both COS phox and PLB-985 cells were collected by centrifugation and resuspended in RPMI 1640 containing 0.5% bovine serum albumin at 1-3 ϫ 10 6 cells/ml. Cells were incubated in the dark with 100 M isoluminol and 40 units/ml horseradish peroxidase at room temperature for 10 min, and 200-l aliquots were transferred into 6-mm diameter wells of a 96-well, flat-bottom, white tissue culture plate (E&K Scientific, Campbell, CA). Chemiluminescence (CL) was measured at 37°C in a Wallac 1420 Multilabel Counter (PerkinElmer Life Sciences). The CL CPS were continually recorded at 1-min intervals for 5-15 min before and 20 -40 min after stimulation with PMA (200 ng/ml) or fMLF (1 M). The relative amount of superoxide produced was calculated based on the integrated CL during the first 20 min after agonist stimulation.
Western Blotting-Protein samples were loaded on a 12% SDS-polyacrylamide gel. After electrophoresis, proteins were transferred to nitrocellulose membranes (Schleicher & Schuell). The blots were blocked with 5% nonfat dry milk in TBS/T buffer (20 mM Tris-HCl, pH 7.6, 137 mM NaCl, 0.1% Tween 20) for 2 h at RT. The blots were washed with TBS/T and incubated with primary antibodies (0.2-1 g/ml) overnight at 4°C. Anti-rabbit (Bio-Rad) or anti-mouse (Calbiochem) peroxidase-conjugated secondary antibodies were added to the membranes at a dilution ratio of 1:3000, and incubation was continued to for 1 h at RT. The protein bands on the membrane were visualized by chemiluminescence (Pierce).
Immunoprecipitation-Twenty-four hours after transfection the cells were lysed in a buffer containing 20 mM Tris-HCl, pH 7.4, 1 mM dithiothreitol, 100 mM NaCl, 1 mM EDTA, 5 mM MgCl 2 , 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and 1ϫ protease inhibitor mixture set I (Calbiochem). For immunoprecipitation with moesin and actin, a buffer containing 1% sodium deoxycholate, 10 mM Tris, pH 7.4, 0.1% SDS, 150 mM NaCl, 1% Nonidet P-40, 2 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 1ϫ protease inhibitor mixture set I was used instead. The cell lysates were cleared of debris by centrifugation at 14,000 ϫ g for 10 min at 4°C. Protein content in the cell lysate was measured using a DC Protein Assay (Bio-Rad) and standardized before immunoprecipitation with the anti-FLAG monoclonal antibody (5 g/ml) at 4°C overnight. Protein A/G PLUS-agarose was added to the samples for 1.5 h at 4°C. The beads were washed twice in washing buffer (20 mM Tris-HCl, pH 7.4, 1 mM dithiothreitol, 100 mM NaCl, 1 mM EDTA, 5 mM MgCl 2 , 0.1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride) and then once in PBS. The beads were resuspended in 50 l of 2 ϫ SDS-PAGE loading buffer and boiled for 5 min. The samples were analyzed by Western blotting.
Cell Fractionation-Cell fractionation was performed as described (25), with a modification in buffer composition. Briefly, 24 h after transfection, the cells were lysed in a buffer containing 20 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1 mM EDTA, 5 mM MgCl 2 , and 1% Triton X-100 at 4°C for 20 min. Cell lysates were then centrifuged at 14,000 ϫ g for 15 min to separate Triton X-100-soluble and -insoluble fractions. The insoluble fraction was dissolved in 500 l of 1 ϫ SDS-PAGE loading buffer and boiled for 5 min. The samples were analyzed by Western blotting.
The lysate was then subjected to 3 cycles of freeze/thaw in liquid nitrogen and at 37°C. Samples were then centrifuged, and the pellets were washed twice in the hypotonic buffer and resuspended in the same buffer containing 1% Triton X-100. The samples were mixed for 30 min at 4°C to dissociate membranebound proteins and then spun down at 14,000 rpm for 10 min at 4°C. The supernatant were collected as the Triton-soluble membrane fraction. The proteins in the sample were detected by Western blotting.
Immunofluorescence Microscopy-Confocal microscopy was performed using indirect immunofluorescence. Six hours after transfection, cells were seeded on glass coverslips pre-coated with 50 g/ml poly-L-lysine (Sigma) and grown for 18 h in Dulbecco's modified Eagle's medium supplemented with 10% heatinactivated fetal bovine serum. Cells were stimulated with or without PMA (200 ng/ml) for 5 min, washed 3 times in PBS, and fixed with 3% paraformaldehyde in PBS for 15 min at RT. Cells were washed 3 more times in PBS at RT and permeabilized with 0.2% Triton X-100 in PBS for 15 min at RT. Coverslips were blocked with 5% bovine serum albumin in PBS for 1 h at RT. Cells were washed 3 times in PBS and incubated with the primary antibodies in PBS containing 5% bovine serum albumin overnight at 4°C. Anti-p47 phox and anti-p67 phox were used at 2 g/ml each. After washing 5 times with PBST (0.2% Tween 20 in PBS) at RT, cells were incubated with rhodamine red-Xconjugated goat anti-rabbit IgG (secondary antibody; Jackson ImmunoResearch Laboratories, West Grove, PA) at 1.5 g/ml for 1 h at RT. After additional washes with PBST and H 2 O, coverslips were mounted on glass slides using the ProLong Gold antifade reagent with 4Ј,6-diamidino-2-phenylindole (Molecular Probes). Fluorescence images were captured with a Zeiss LSM 510 confocal microscope equipped with heliumneon, argon, and krypton laser sources.
Statistic Analysis-Data were analyzed by paired Student's t test using the PRISM software (Version 4.0, GraphPad, San Diego, CA).

Localization of the Full-length p40 phox in Transfected Cells-
The PX domain of p40 phox is thought to preferentially bind PtdIns (3)P for its membrane targeting (4,6,26). Previous studies have shown that an isolated PX domain fused to a GFP was localized primarily in early endosome (4,5), an intracellular organelle enriched with PtdIns (3)P (27). In this study we examined the full-length p40 phox for its intracellular localization and redistribution before and after agonist stimulation. A fulllength p40 phox fused to GFP (p40 phox -GFP) was transfected into COS phox cells, a stable cell line of COS-7 that expresses gp91 phox , p22 phox , p47 phox , and p67 phox but lacks p40 phox (22). Imaging analysis of the transfected COS phox cells (Fig. 1, A-F) revealed cytoplasmic localization of the GFP fluorescence in resting state (Fig. 1, A and C). Slightly more intense fluorescence was observed in the perinuclear region and in membrane ruffles. In comparison, an antibody against early endosome antigen 1 (EEA-1) stained punctate structures in unstimulated cells (Fig. 1B). There were very few punctate structures with both green (p40 phox -GFP) and red (anti-early endosome antigen 1) fluorescence (Fig. 1C). Upon stimulation with PMA ( Fig.  1, D-F) or fMLF (data not shown), there was a marked increase in plasma membrane-associated green fluorescence (Fig. 1, D and F). PMA stimulation did not increase or decrease doublestained fluorescence in the periphery of the cells (Fig. 1F), indicating the absence of fusion between early endosome and the plasma membrane. To determine whether agonist-induced membrane translocation of p40 phox requires p67 phox and p47 phox , p40 phox -GFP was expressed in COS 91/22 cells, a stable cell line of COS-7 expressing gp91 phox and p22 phox but not p67 phox and p47 phox (22). As shown in Fig. 1, G-L, the GFP fluorescence remained cytoplasmic in the resting state (G and I) as well as following PMA stimulation (J and L). This result suggests that p40 phox membrane translocation requires the presence of p67 phox and p47 phox , which is consistent with the notion that p40 phox translocates to plasma membrane in a complex with p67 phox and p47 phox . A recent study conducted by Ueyama et al. (28) showed that in the RAW267.4 macrophage cell line, which contains very low level of endogenous p67 phox , PMA and fMLF was unable to induce membrane translocation of p40 phox . Our result is in agreement with their observation.
An Arg to Gln Mutation of p40 phox (R57Q) Produces a Dominant Negative Effect in Superoxide Production-We recently reported that NADPH oxidase activation through FPR could be reconstituted in COS phox cells through expression of FPR along with selected signaling molecules (23). Heterologous expression of p40 phox significantly enhanced fMLF-induced superoxide (Fig. 2B, solid line) as compared with vector control (dotted line). The potentiation effect of p40 phox was maximal at an input plasmid DNA concentration of 1.5 g. A slight decline (ϳ10%) from the maximal superoxide production was observed at the input DNA concentration of 2.5 g (data not shown). The change in superoxide production followed a similar time course in the presence or absence of p40 phox , indicating that p40 phox  increased oxidant production without altering its kinetics. COS phox cells expressing p40 phox also produced more superoxide when stimulated with PMA ( Fig. 2C, solid line), suggesting that p40 phox regulates an NADPH oxidase activation pathway shared by fMLF and PMA. Release of superoxide was not detectable in the presence of superoxide dismutase (SOD) (Fig.  2, B and C, triangles).
Mutation of selected amino acids (Arg-57, Arg-58, and Arg-105) in the PX domain of p40 phox abolishes its interaction with PtdIns (3)P (4,6,8,26). We prepared an R57Q mutation ( Fig.  2A) and examined its effect in the context of a full-length p40 phox . When transfected into COS phox cells, p40R57Q was expressed at a level similar to that of the wild type p40 phox ( Fig.  2A). However, cells co-transfected with p40R57Q failed to respond to fMLF with superoxide production (Fig. 2B, dashed  line). Likewise, expression of p40R57Q markedly reduced the PMA-stimulated oxidant production (Fig. 2C, dashed line). Given that all COS phox cells in the sample could respond to PMA and only 45-50% of the cells were transfected with the p40R57Q construct, the actual inhibition by p40R57Q in the transfected cells could be greater than shown in Fig. 2C. Mutation of Arg-58 to Gln produced a similar inhibitory effect when expressed in COS phox cells (data not shown). The R57Q mutant was further characterized in subsequent experiments.
The difficulty associated with transfecting suspension cells and the presence of endogenous p40 phox in many hematopoietic cell lines prevented us from conducting an extensive investigation of the PX domain in leukocytes. To determine whether the effect produced by p40R57Q is an isolated phenomenon only seen in COS phox cells or is applicable to leukocytes, we used nucleofection to deliver the plasmid encoding p40 phox or p40R57Q into the PLB-985 myelomonoblastic leukemia cell line (29,30). As shown in Fig. 3A, a low level of endogenous p40 phox was detected in undifferentiated PLB-985 cells. An increase in p40 phox expression was evident after nucleofection, with up to 30% of the cells receiving the plasmid of interest based on flow cytometry analysis (data not shown). As expected, exogenous expression of p40 phox in the PLB-985 cells caused an increase in superoxide production (Fig. 3, B and C, solid lines), whereas expression of p40R57Q led to a decrease in superoxide production (dashed lines). These changes were observed in both fMLF-and PMA-stimulated cells (Fig. 3, B and  C). Because of the low transfection efficiency, actual inhibition in the transfected cells could be greater that what was observed.
We next examined whether p40R57Q acted as a dominant negative mutant. COS phox cells were transfected with a fixed FIGURE 4. A dominant inhibitory effect of p40R57Q in superoxide production. COS phox cells were transiently transfected with expression constructs encoding FPR, G␣ i2 , protein kinase C␦, and either an empty vector (Ϫ) or a vector for p40 phox or p40R57Q. The amount of p40 phox and p40R57Q vectors used in each transfection is indicated (in g). When necessary, empty vector was added to bring the total amount of DNA in transfection equal for all samples. A transfection efficiency of ϳ50% was obtained in these experiments. Twenty-four hours after transfection, cells were harvested for determination of the protein expression level using Western blotting (A), as described in Fig. 1, and for stimulation with 1 M of fMLF (B) or 200 ng/ml of PMA (C). Superoxide produced in a 20-min period was recorded and integrated as described under "Experimental Procedures." The integrated (Intl.) CL was shown as % change relative to vectortransfected cells (set as 100%). Data shown in B and C are the means Ϯ S.E. from three independent experiments. p Ͻ 0.05 (*) and p Ͻ 0.01 (**), compared with control. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. amount of plasmid DNA coding for the wild type p40 phox and variable amounts of plasmid DNA coding for the mutant. The resulting changes in p40 phox expression levels were determined with an anti-p40 phox antibody (Fig. 4A). When the transfected cells were stimulated with fMLF (Fig. 4B), the R57Q mutant of p40 phox could overcome the potentiation effect of the wild type p40 phox . At a p40R57Q:p40 phox DNA input ratio of 4:1, the mutant brought superoxide production below the level obtained without p40 phox . In cells stimulated with PMA, similar changes were observed (Fig. 4C), but the magnitude of potentiation and inhibition was smaller because only half of the cell population was affected by the wild type and mutant constructs due to a ϳ50% transfection efficiency. These results demonstrate a dominant inhibitory effect of p40R57Q on NADPH oxidase activation.
Expression of p40R57Q Reduces Membrane Translocation of the Cytosolic Factors-A hallmark of NADPH oxidase activation is membrane translocation of the cytosolic factors and their interaction with flavocytochrome b 558 . We sought to determine whether the R57Q mutant of p40 phox could affect this important process of NADPH oxidase activation. At resting state (Fig. 5A), there was minimal membrane association of p40 phox and p47 phox in vector-or p40 phox -transfected COS phox cells, which was consistent with the accepted notion that p40 phox is present as a cytosolic protein in unstimulated phagocytes. In COS phox cells expressing p40 phox , enhanced membrane translocation of both p40 phox (green fluorescence) and p47 phox (red fluorescence) was observed upon PMA stimulation (Fig. 5B). In cells transfected to express p40R57Q, PMAinduced membrane translocation of the mutant p40 phox was greatly impaired. Reduced membrane translocation of p47 phox was also observed in cells expressing p40R57Q. We next examined whether membrane translocation of p67 phox was affected by the R57Q mutation. As shown in Fig. 5C (unstimulated cells) and Fig.5D (PMA-stimulated cells), the PMA-induced p67 phox translocation (red fluorescence) was markedly impaired in cells expressing p40R57Q. This result is expected since p40 phox is closely associated with p67 phox and translocates to plasma membrane in a complex consisting of p67 phox and p47 phox .
To determine whether inhibition of p47 phox and p67 phox membrane translocation is a mechanism for the p40R57Q mutant to negatively regulate NADPH oxidase activation, COS phox cells transfected with wild type p40 phox or p40R57Q were stimulated with either PMA or vehicle control and subjected to cellular fractionation. The membrane fraction was collected and analyzed by SDS-PAGE and Western blotting with antibodies against p40 phox , p47 phox , p67 phox , and p22 phox (a membrane marker for loading control). As shown in Fig. 6A, PMA potently increased the level of membrane-associated p40 phox . In unstimulated cells, a small amount of p47 phox and p67 phox was found in the membrane fraction. PMA stimulation caused an increase in membrane-associated p47 phox and p67 phox without p40 phox . In cells expressing the wild type p40 phox , PMA stimulation resulted in a more potent increase in p47 phox and p67 phox membrane association. Densitometry analysis of the blots were shown in Fig. 6B (p40 phox ), Fig. 6C (p47 phox ), and Fig. 6D (p67 phox ). Because the PMA-induced changes only occurred in about half of the cell population that were transfected with the p40 phox expression plasmids and untransfected cells lacking p40 phox could also respond to PMA, the actual effect of p40 phox could be greater. Indeed, after normalization against the transfection efficiency, the enhancement effect of p40 phox on p47 phox membrane translocation became more apparent (Fig. 6C, solid bar in the fourth group). In contrast, expression of p40R57Q reduced membrane association of p47 phox . When the proportion of untransfected cells, in which p47 phox membrane translocation was not affected by p40R57Q, was taken into consideration, the inhibitory effect of p40R57Q was more prominent (Fig. 6C, solid bar in the last group). A similar effect on p67 phox membrane translocation was observed. Whereas p40 phox potentiated p67 phox membrane translocation, p40R57Q reduced this response to PMA (Fig.  6D). Therefore, results from the biochemical characterization corroborate with data from imaging analysis and together support the conclusion that the R57Q mutation of p40 phox has a negative effect on p47 phox and p67 phox membrane translocation.
Suppression of Superoxide Production by p40R57Q Requires Its Interaction with p67 phox -Published studies demonstrate that p40 phox is tightly associated with p67 phox , and p40 phox can be co-purified with p67 phox in resting neutrophils. In CGD neutrophils lacking p67 phox , there is a concomitant reduction in cellular content of p40 phox (9,10,31,32). The tight association between p40 phox and p67 phox involves their PB1 domains, and mutation of selected residues in these domains, such as Asp-289 in p40 phox , can abolish this interaction (13). To determine whether the inhibitory effect of p40R57Q involves its interaction with p67 phox , we prepared an R57Q/D289A double mutant of p40 phox . The double mutant could be readily expressed in COS phox cells, as determined in Western blotting using cell lysate prepared from the transfected cells (Fig. 7A). However, the immunoprecipitated, FLAG-tagged double mutant failed to associate with p67 phox or p47 phox in co-immunoprecipitation and Western blotting assays (Fig. 7B). Interestingly, the mutation at Asp-289 abolished the dominant inhibitory effect of p40R57Q in fMLF-and PMA-induced superoxide production (Fig. 7, C and D, respectively). This result indicates that an interaction between p40 phox and p67 phox is required for the inhibitory effect of p40R57Q.
Expression of p40R57Q Increases the Association of the Cytosolic Factors with Actin Cytoskeleton-In resting cells, the p40 phox /p67 phox complex was primarily associated with the cytoskeleton, whereas p47 phox was found in both the soluble fraction and cytoskeleton fraction (20,33). Using Western blotting for detection of proteins in the Triton X-100-insoluble fraction, which is enriched with cytoskeletal proteins such as FIGURE 6. Western blot analysis of the effects of p40 phox and p40R57Q on membrane translocation of p47 phox and p67 phox . A, COS phox cells were transiently transfected with either empty vector or expression constructs of p40 phox or p40R57Q. Twenty-four hours after transfection, cells were collected, and the membrane fractions were prepared as described under "Experimental Procedures." The relative levels of p40 phox , p47 phox , and p67 phox in the membrane fractions were determined using antibodies against p40 phox , p47 phox , and p67phox, respectively. An anti-p22 phox was used to detect equal loading of membrane proteins. Three independent experiments were conducted, and a representative set of blots is shown. The membrane-associated p40 phox (B), p47 phox (C), and p67 phox (D) was quantified against p22 phox based on relative intensity of the Western blot bands in A using Quantity One software (Bio-Rad, Version 4.3.1). In C and D, both raw data (unprocessed, open bars) and normalized data (processed against a 50% transfection efficiency, solid bars) are presented. In PMA-stimulated samples, only ϳ50% of the cells were transfected and affected by the p40 phox or p40R57Q constructs. OCTOBER 12, 2007 • VOLUME 282 • NUMBER 41 filamentous actin, we examined the potential effect of p40R57Q on cellular distribution of the cytosolic factors. In unstimulated COS phox cells, expression of wild type p40 phox slightly increased the contents of p47 phox and p67 phox in the Triton X-100-insoluble fraction (Fig. 8). However, in cells expressing p40R57Q, significantly more p67 phox and p47 phox were recovered in the Triton X-100-insoluble fraction along with the mutant p40 phox . This result indicates that p40R57Q can promote cytoskeleton association of the cytosolic factors, thereby altering their cellular distribution profile.

Regulation of NADPH Oxidase by the PX Domain of p40 phox
Given that mutation at Asp-289 abolished the dominant inhibitory effect of p40R57Q (Fig. 7), we next examined whether the double mutant was able to retain the cytosolic factors in the Triton X-100-insoluble fraction. As shown in Fig. 8, the D289A mutation caused a marked decrease in the amount of p67 phox and p47 phox in the Triton X-100-insoluble fraction, whereas it had a smaller effect on the retention of p40 phox in the Triton X-100-insoluble fraction. These results are consistent with data from the functional assay shown in Fig. 7 and together demonstrate a correlation between increased actin cytoskeleton association of the cytosolic factors and reduced NADPH oxidase activity in cells expressing the p40R57Q mutant. The above results also indicate that the interaction between p40 phox and p67 phox is necessary for inhibition of superoxide production as well as increased association of p67 phox and p47 phox with the actin cytoskeleton in the presence of p40R57Q.
The effect of the R57Q mutation on p40 phox association with the actin cytoskeleton was further examined using co-immunoprecipitation of a FLAG-tagged p40 phox or a similarly tagged R57Q mutant and Western blotting detection of actin. As shown in Fig. 9A, significantly more actin was found associated with p40R57Q than with p40 phox in COS phox cells (left panel). The same experiment was repeated in COS 91/22 cells, which lack p67 phox and p47 phox (Fig. 9B). In the absence of these cytosolic factors, there was a decrease in the overall association of both p40 phox and its R57Q mutant with actin. Still, relatively more actin association was detected with the p40R57Q mutant than with p40 phox . This result corroborates with the findings in the Triton X-100-insoluble fraction. Likewise, we found in the transfected COS phox cells an increased association of p40R57Q with moesin, an actin-binding protein that serves to bridge the interaction between actin and many proteins (Fig. 9C). When expressed in COS 91/22 cells, the R57Q mutant still associated more strongly with moesin than did the wild type p40 phox , although there was a decrease in the overall level of association (Fig. 9D). Taken together these results suggest a shift to increased actin cytoskeleton association when a PtdIns (3)P binding site in p40 phox was mutated. It is also evident that both p67 phox and p47 phox contribute to cytoskeleton association, as previously reported (20,33).
Cytochalasin B Eliminates the Dominant Inhibitory Effect of p40R57Q on Superoxide Production-Cytochalasin B is a fungal metabolite that binds actin filaments and prevents actin polymerization. It is widely used in phagocyte studies due to its potentiation effect on superoxide production and degranulation. How cytochalasin B enhances oxidant production remains incompletely understood. We treated the transfected COS phox cells with cytochalasin B (5 g/ml, 10 min) or with vehicle control before PMA stimulation. As shown in Fig. 10, cytochalasin B increased the relative level of superoxide production by ϳ3-4-fold. More interestingly, the dominant inhibitory effect produced by p40R57Q was completely abolished in the presence of cytochalasin B. This result further support the notion that increased association of p40R57Q with the actin cytoskeleton is responsible for its dominant inhibitory effect.

DISCUSSION
In this study we attempted to characterize the PX domain in the full-length p40 phox for p40 phox intracellular localization, its interaction with the actin cytoskeleton, and its role in superoxide production. Mutations were introduced to abolish PtdIns (3)P binding through the PX domain and to eliminate interac- FIGURE 7. Requirement of p67 phox association for the inhibitory effect of p40R57Q. A, COS phox cells were transfected with expression vectors coding for FPR, G␣ i2 , protein kinase C␦, the empty vector, and the wild type or mutant p40 phox constructs as indicated. Twenty-four hours after transfection, cell lysate was prepared, and equal expression of p40 phox and the indicated mutants was determined using Western blotting with an anti-p40 phox antibody. B, COS phox cells were transfected with the above constructs of p40 phox , which were tagged with FLAG. Immunoprecipitation (IP) was carried out using an anti-FLAG monoclonal antibody, and Western blotting (IB) was conducted with the respective antibodies against p40 phox , p47 phox , and p67 phox as described above. The p40R57Q/D289A double mutant was unable to interact with p67 phox or p47 phox . The transfected cells were stimulated with either 1 M fMLF (C) or 200 ng/ml of PMA (D). Superoxide produced in a 20-min time span was recorded as described under "Experimental Procedures." The integrated chemiluminescence (Int. CL) was shown as % change relative to vector-transfected cells (set as 100%). Data shown in C and D are the mean Ϯ S.E. from three independent experiments. **, p Ͻ 0.01, compared with control. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
tion with p67 phox through the PB1 motif in the C terminus of p40 phox . We took advantage of the COS phox cell line, which exhibits many properties similar to those of neutrophils in terms of NADPH oxidase activation, including responsiveness to PMA with superoxide dismutase-inhibitable superoxide production, membrane translocation of the cytosolic factors, and requirement for the Rac small GTPase (22,23,34). The absence of endogenous p40 phox allowed us to examine the effects of the p40 phox mutations on NADPH oxidase activity with relative ease. We observed an unexpected and remarkable change caused by a point mutation in the PX domain of p40 phox . The R57Q mutation not only abolished the potentiation effect of the wild type p40 phox on NADPH oxidase activity but also strongly suppressed superoxide production in fMLF-and PMA-stimulated cells. Our experimental data suggest that a mechanism for the observed inhibition is increased association of the cytosolic factors with the actin cytoskeleton, which is mediated by the R57Q mutant of p40 phox . This study also identifies the association between p40 phox and p67 phox as being important for the dominant inhibitory effect of the R57Q mutation. These results, when considered together with recent observations made by other investigators, suggest a dual regulatory function of p40 phox in NADPH oxidase activation.
Intracellular Localization of Full-length p40 phox Differs from That of the Isolated PX Domain-We have shown that a fulllength p40 phox construct, when expressed in transfected cells, displayed a cytoplasmic localization pattern similar to that of p67 phox and p47 phox (Fig. 5). Upon agonist stimulation, a portion of the cytoplasmic p40 phox moved to plasma membrane ( Figs. 1 and 5). This pattern of distribution is drastically different from that of the isolated PX domain, which displayed a punctate, early endosome localization profile (4,5). PtdIns(3)P, to which the PX domain of p40 phox binds, is enriched in early endosome membrane (27). The observed localization in early endosome of transfected cells suggests that the PX domain, when expressed alone, retains an open conformation for access of PtdIns (3)P in the absence of agonist stimulation. The absence of early endosome localization with the full-length p40 phox suggests that its PX domain is not accessible to PtdIns (3)P. Therefore, any function of p40 phox in resting cells should be independent of PtdIns (3)P binding. A recent study by Ueyama et al. (28), published during the course of this work, provides a mechanism for the lack of PtdIns (3)P binding with the full-length p40 phox . Using deletion mutagenesis, the authors found that the C-terminal PB1 domain could fold back to mask the N-terminal PX domain in p40 phox (28). Subsequently, a structural analysis of the full-length p40 phox confirmed that the intramolecular interaction between the PB1 domain and the PX domain prevents access to membrane-associated PtdIns (3)P by the PX domain (35). Taken together, both imaging data and structural analysis confirm that in unstimulated cells, the PX domain in the full-length p40 phox does not interact with PtdIns (3)P in the membrane, suggesting that its functions in resting state and activated state are regulated differently.
Ueyama et al. (28) observed differential targeting of p67 phox by p40 phox (to early endosome) and by p47 phox (to plasma membrane). During Fc␥R-mediated phagocytosis, transient vesicular accumulation of GFP-p40 phox and its fusion with phagosome were observed. Therefore, the early endosome-targeted p67 phox , which constitutes a small fraction of the membrane translocated protein, may be available for NADPH oxidase activation. Interestingly, they have shown that arachidonic acid, but not PMA or fMLF, was able to alter the structure of fulllength p40 phox , allowing access of the PX domain to membrane-associated PtdIns (3)P (28). A similar finding was made in this study, in which we observed that PMA was unable to induce membrane translocation of p40 phox in the absence of p47 phox and p67 phox (Fig. 1). The mechanism by which arachidonic acid regulates structural changes of p40 phox is still undefined. Phosphorylation of p40 phox at Thr-154 and Ser-315 has been reported and was thought to be a regulatory mechanism for induced structural changes of p40 phox (36,37). Lopes et al. FIGURE 8. Effects of the R57Q mutation and R57Q/D289A double mutation on the retention of the cytosolic factors in Triton X-100 insoluble fraction. COS phox cells were transfected with expression vectors coding for FPR, G␣ i2 , protein kinase C␦, the empty vector, and the wild type or mutant p40 phox constructs. A, cells were lysed in a buffer containing 1% Triton X-100 at 4°C. The insoluble fraction was collected, resolved on SDS-PAGE, and blotted with antibodies against the respective cytosolic factors. An antibody against ␤-actin was used for equal loading control. The expression level of the wild type and mutant p40 phox constructs was determined by Western blotting (IB, top panel). Three independent experiments were conducted, and a representative set of blots is shown. Quantitative analysis of the relative levels of p40 phox (B), p47 phox (C) and p67 phox (D) in the Triton X-100 insoluble fraction was shown. (38) previously reported that phosphorylation of p40 phox at Thr-154 could cause inhibition of NADPH oxidase activity in cell-free assays. However, no functional changes were observed when Thr-154 and Ser-315 were mutated to Ala in the study conducted by Ueyama et al. (28) in transfected cells. Further study will be necessary to determine how a closed conformation of p40 phox is transformed during NADPH oxidase activation.
Increased Association with the Actin Cytoskeleton Is a Possible Mechanism for Inhibition of NADPH Oxidase through p40R57Q-The potent inhibition of NADPH oxidase brought upon by the R57Q mutation was unexpected. Because p40 phox is not essential for PMA-and fMLF-induced superoxide production in reconstituted COS phox cells (22,23), it would not be surprising if the mutation simply eliminated the potentiation effect through a change in the PtdIns (3)P binding site. Our attention was directed to the actin cytoskeleton as previous studies have shown that p40 phox , like p67 phox and p47 phox , interacts with the actin cytoskeleton although the biological consequence of this interaction was not entirely clear. Earlier studies have shown that functional cytosolic factors are found in the Triton X-100-insoluble fraction (33,39), suggesting a potential role of the actin cytoskeleton in the organization and redistribution of the cytosolic complex of phagocyte NADPH oxidase. Other studies have shown association of p40 phox along with p67 phox and p47 phox with the actin cytoskeleton and actinbinding proteins in unstimulated neutrophils, implying that such an association might stabilize the cytosolic factors in the resting state (20,21,40). Of interest is the finding that p40 phox interacts with moesin, an actin-binding protein and a member of the ezrin-radixin-moesin family, through its PX domain (21). The mutation at Arg-57 may alter the structure of the PX domain so that not only PtdIns (3)P binding is abolished but also the affinity for moesin is changed. Indeed, we observed increased association of p40R57Q with actin and moesin in co-immunoprecipitation assay as well as an enrichment of the mutant in the Triton X-100-insoluble fraction, which contains abundant polymerized actin cytoskeleton. These findings suggest the possibility that an aberrant p40 phox protein "entraps" the cytosolic complex through an enhanced interaction with the actin cytoskeleton, thereby preventing its membrane translocation and interaction with flavocytochrome b 558 .
Arg-57 and Arg-58 are highly conserved among the PX domains identified so far (1,41). In p47 phox , the analogous residue of Arg-57 is Arg-42, which when mutated to a Gln results  D) were transfected with either empty vector or FLAG-tagged p40 phox or p40R57Q expression vectors. Twenty-four hours after transfection, p40 phox and p40R57Q proteins were immunoprecipitated from cell lysate with an anti-FLAG antibody. Immunoprecipitates (IP) were separated on SDS-PAGE, blotted to nitrocellulose membrane, and detected with antibodies against actin (A and B) and moesin (C and D). In all panels, an anti-p40 phox antibody was used to determine the level of p40 phox in the immunoprecipitates. Representative blots (IB) from three independent experiments are shown. The relative intensity of the Western blot bands was quantified with the Quantity One software (Bio-Rad, Version 4.3.1) and shown in the bar charts below each blot. in an autosomal recessive CGD (42). Studies of the isolated PX domain of p40 phox have shown that mutations of Arg-57 (4), Arg-58 (5,7,8), and Arg-105 (6) abolish the interaction of the PX domain with PtsIns (3)P. However, the absence of PtdIns (3)P binding cannot explain why replacement of the mouse p40 phox gene with a R58A mutant gene leads to embryonic lethality in homozygous offspring (18). Another mechanism, such as dominant inhibition of superoxide production, may be responsible for the deleterious effect in consideration of the important roles that reactive oxygen species play in organ development (43,44). There has been no reported clinical case of CGD that results from mutation in the p40 phox gene. The possibility that natural mutations at these sites lead to embryonic lethality due to defective superoxide production merits further investigation.
The Interaction between p40 phox and p67 phox Is Essential for Both Potentiation and Inhibitory Effects of p40 phox in Superoxide Production-We showed that the R57Q mutation abolished the potentiation effect of p40 phox on superoxide production as well as induced membrane translocation. This mutation, however, did not affect the interaction of p40 phox with p67 phox , which is mediated through a C-terminal PB1 interaction with the corresponding PB1 domain in p67 phox . Therefore, although the R57Q mutation can possibly destabilize the PX domain with respect to PtdIns (3)P binding (8), this point mutation does not seem to alter the global structure of p40 phox so as to weaken its interaction with p67 phox . This property of p40R57Q must be considered when evaluating its functional impact on NADPH oxidase activity. Indeed, p40 phox was originally identified as a cytosolic protein tightly associated with p67 phox (9 -11). As we have shown above (Fig. 7), when this association was disrupted by a second mutation at Asp-289, the resulting p40 phox double mutant could no long inhibit NADPH oxidase in superoxide production assays. Consistent with the functional change, dissociation of p67 phox from p40R57Q due to the D289A mutation caused a significant reduction in the amount of cytoskeletonassociated p67 phox and p47 phox found in the Triton X-100-insoluble fraction. Kuribayashi et al. (13) previously reported that the D289A mutation could abolish the potentiation effect of the wild type p40 phox , a finding confirmed in our study of FPRreconstituted COS phox cells (data not shown). Taken together, these experimental results support the notion that both the potentiation effect and the inhibitory effect of p40 phox require its association with p67 phox .
We have shown that, in the resting state, p40 phox as well as p40R57Q could be co-immunoprecipitated with p67 phox and p47 phox (Fig. 7). Upon agonist stimulation, p40 phox translocated to plasma membrane rather than early endosome (Fig. 1) and colocalized with p47 phox and p67 phox (Figs. 5 and 6). The absence of early endosome localization could indicate that the PX domain of p40 phox remains "closed" under the experimental conditions used in this study or the p47 phox -directed membrane translocation of the cytosolic complex is predominant. Interaction between p40 phox and p47 phox may be secondary to the p40 phox -p67 phox interaction, as reported previously (13,31,32) and confirmed in this study (Fig. 7). Our findings are consistent with one of the original observations that p40 phox remains associated with p67 phox and p47 phox in activated neu-trophils (9) and together suggest that p40 phox facilitates membrane targeting of p67 phox . The tight association between p40 phox and p67 phox apparently helps to stabilize their structures, as CGD patients with diminished p67 phox expression also display reduced p40 phox expression (9,11,15). Because of the association between these two cytosolic factors, it is possible that structural changes induced by the R57Q mutation could affect the structure and function of p67 phox and even p47 phox , thereby prohibiting their membrane translocation. In this regard it is notable that published reports showed that p40 phox dissociates from p67 phox in activated membranes (20). Another published study suggested that p40 phox stabilizes the resting state and should be dissociated from p67 phox for maximal oxidase activity (15).
In summary, the current study examines a PX domain mutation in the context of a full-length p40 phox and found that alteration of the PtdIns (3)P binding site may produce drastic changes in the way p40 phox regulates NADPH oxidase. Our observations suggest that, in addition to the potentiation effect through PtdIns (3)P binding, which has been confirmed both in cultured cells and in primary neutrophils (16 -18), the PX domain of p40 phox may negatively regulate NADPH oxidase through stabilization of the resting state. An increased association with the actin cytoskeleton combined with a possibly direct effect on p67 phox structure may contribute to the observed inhibitory effect. So far, most studies on NADPH oxidase have been focused on the activation mechanism. An understanding of how NADPH oxidase is negatively regulated may help to expand our knowledge with potential applications to the control of oxidant production.