p40(phox) down-regulates NADPH oxidase activity through interactions with its SH3 domain.

The NADPH oxidase of phagocytes generates microbicidal oxidants in response to a variety of stimuli. Its activation and assembly involve multiple SH3 domain interactions among several oxidase components. Here we present evidence that the cytosolic oxidase-associated protein, p40(phox), mediates down-regulation of NADPH oxidase through interactions with its SH3 domain. Recombinant p40(phox) was produced in several eukaryotic expression systems (insect, mammalian, and yeast) to explore its role in oxidase function in relation to domains involved in interactions with other factors, p47(phox) and p67(phox). p40(phox) inhibited oxidase activity in vitro when added to neutrophil membranes and recombinant p47(phox), p67(phox), and p21rac. Co-transfection of p40(phox) into K562 cells resulted in significant decreases ( approximately 40%) in whole cell oxidase activity. Furthermore, the isolated SH3 domain of p40(phox) was even more effective in inhibiting whole cell oxidase activity, consistent with experiments showing that this domain binds to the same proline-rich target in p47(phox) (residues 358-390) that interacts with p67(phox). In contrast, deletion of the carboxyl-terminal domain of p40(phox) that binds to p67(phox) did not relieve its oxidase inhibitory effects. Thus, p40(phox) appears to down-regulate oxidase function by competing with an SH3 domain interaction between other essential oxidase components.

Despite work in several laboratories on the interactions between oxidase components, the function of p40 phox has remained unclear (5)(6)(7)(8)(9)(21)(22)(23)(24)(25). Antibodies against p40 phox were shown to inhibit NADPH oxidase in vitro (25), although the effects of p40 phox on oxidase activity have not been examined directly, and CGD patients with lesions in the p40 phox gene have not been described. Furthermore, p40 phox is not needed to reconstitute high levels of oxidase activity in cell-free systems containing purified p21rac, p47 phox , p67 phox , and relipidated flavocytochrome b 558 (26,27). Like several of the core oxidase components, p40 phox expression is myeloid-specific (5). The fact that p40 phox is bound in a complex with p67 phox and p47 phox that translocates to the membrane upon activation suggests that p40 phox could have a role in modulating the respiratory burst. Here we report about interactions of p40 phox with other oxidase components and explore the effects of these interactions on oxidase activity, both in cell-free and whole cell reconstituted systems. Based on these observations, we provide evidence for yet another SH3 domain-mediated interaction, which in this case appears to inhibit oxidase function.

EXPERIMENTAL PROCEDURES
Recombinant Protein Expression-The full coding sequence of p40 phox was cloned from a human leukocyte cDNA library (Clontech, Palo Alto, CA; catalog No. HL4014AB) using the polymerase chain reaction. Amplification primers were targeted to sequences beginning 21 bases upstream from the start codon and ending 30 bases after the stop codon of the p40 phox cDNA (5). The p40 phox polymerase chain reaction product was first cloned into the TA cloning vector (Invitrogen, San Diego, CA) for sequencing and then subcloned (EcoRI fragment) into pGEX3X for production of recombinant p40 phox -glutathione S-transferase (GST) fusion protein in Escherichia coli. The fusion protein was affinity purified * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Production of recombinant p40 phox baculovirus for infection of cultured Sf9 (Spodoptera frugiperda) cells was performed by methods described earlier (29), with the following modifications. The p40 cDNA was subcloned (EcoRI fragment) into pVL1392 (Invitrogen) and used for co-transfection of Sf9 cells along with linearized, deleted baculovirus DNA (BaculoGold DNA, Pharmingen, San Diego, CA) according to the manufacturer's protocols. High titer p40 phox baculovirus was used to infect large scale suspension cultures (100 ml) grown at 2.5 ϫ 10 6 cells/ml in Sf900-II SFM medium (Life Technologies, Inc.). The protein was harvested from the cleared soluble fraction of sonicated cells harvested 3 days after infection as described (29). These preparations were analyzed by SDS-PAGE and immunoblotting and were used in cell-free oxidase reconstitution studies without further purification. Affinitypurified full-length p40 phox used for immunization was obtained by binding to a recombinant GST fusion protein constructed with the carboxyl-terminal domain of p47 phox (residues 358 -390) (12).
Antiserum Preparation-GST-p40 phox fusion protein preparations from inclusion bodies were subjected to SDS-PAGE and visualized by staining with Coomassie blue. The full-length protein was excised and used to immunize rabbits (primary immunization, 150 g in complete Freund's adjuvant, followed by two boosts in incomplete Freund's adjuvant at 2-week intervals). Subsequent boosts used affinity-purified baculovirus-derived p40 phox eluted from SDS-PAGE gels. The antiserum obtained after 10 weeks of immunization cross-reacted specifically with p40 phox in neutrophil cytosol and recombinant p40 phox from Sf9 cells.
Functional Studies in Transfected Cells-Co-transfection of pCEP or pREP plasmids containing p67 phox , p47 phox , and p40 phox cDNAs (20 g each) into transduced K562 (gp91 phox -expressing) cells was performed by electroporation as described previously (16,30). In control experiments, pCEP4CAT vector (Invitrogen) was used in place of pCEP40. At 48 h post-transfection, 10 5 cells/ml were selected for 5 days in complete medium containing 250 g/ml hygromycin B. Production of p40 phox in hygromycin-resistant cells was confirmed by immunoblotting of cytosolic fractions with rabbit p40 antiserum (1:1000 dilution). All functional assays were performed within 2 weeks of hygromycin selection. Whole cell superoxide production in response to PMA stimulation (2 g/ml) was determined as the superoxide dismutase-inhibitable chemiluminescence detected with an enhancer-containing luminol-based detection system as described (DIOGENES TM , National Diagnostics, Atlanta, GA) (16,30), using a multiwell plate-reading luminometer (Luminoskan, Labsystems). Both experimental and superoxide dismutase-inhibited reactions were monitored at 1-min intervals for at least 15 min at 37°C following PMA activation. The DIOGENES TM reagent is Ͼ1000 times more sensitive to superoxide than hydrogen peroxide due to the presence of a redox active compound that facilitates direct transfer of electrons from superoxide to luminol (Dr. G. Kitzler, National Diagnostics). As noted earlier (16,30), chemiluminescence signals were completely abolished in the presence of superoxide dismutase and were catalase-insensitive (G. Kitzler), indicative of a superoxide-specific assay.
Yeast Two-hybrid Expression Studies-The yeast Gal4 two-hybrid expression system (31)(32)(33) from Clontech (San Diego, CA) was used to study interactions of NADPH oxidase components as fusion proteins produced in yeast. In this system, two proteins of interest are expressed as hybrids of the Gal4 DNA binding domain (pGBT9) or the Gal4 transcriptional activation domain (pGADGH or pGAD424). An interaction between the two hybrid proteins results in transactivation of the HIS3 and lacZ reporter genes. Full-length coding sequences of p40 phox and p47 phox were subcloned into the EcoRI site of pGADGH in frame with the Gal4 transcriptional activation domain. Full-length p67 phox cDNA was derived from pREP10/67 (13). Deleted forms of p67 phox cDNA were made as follows: p67NT (residues 1-246) was excised from fulllength cDNA using EcoRI and BglII sites and ligated into EcoRI and BamHI (or BglII) sites in pGBT9 and pGAD424. The region between the two SH3 domains of p67 phox (CT), residues 293-465, was polymerase chain reaction-amplified and subcloned into pGBT9. Sequence encoding the p47 phox proline-rich tail region (residues 351-390) was amplified and subcloned into pGBT9 and pGAD424. p40⌬CT (residues 1-260) and p40CT (residues 260 -340) were obtained by restricting p40 phox cDNA in the pGADGH with BglII (internal site) and SpeI (5Ј MCS) or SalI (3Ј MCS) and subcloning SpeI-BglII or BglII-SalI fragments into appropriately cut pGADGH. Expression vectors were confirmed by DNA sequencing and in some cases by Western blotting. All manipulations in yeast strains HF7c or CG1945 were according to manufacturers' directions. Interactions between fusion proteins were monitored by histidine prototropy and ␤-galactosidase activity detected with 5-bromo-4-chloro-3-indole-␤-D-galactopyranoside within lysed colonies that were adsorbed onto nitrocellulose filters (33). All constructs were co-expressed along with appropriate control (nonrecombinant or empty) vectors to test for false positive signals.
Cell-free NADPH Oxidase (Cytochrome c Reduction) Assay-Cell-free superoxide production was determined from the superoxide dismutaseinhibitable reduction of cytochrome c as described earlier (13,29). Reactions (100 l) contained varying amounts of Sf9 cell cytosols derived from uninfected or p40 phox baculovirus-infected cultures (0 -20 g) along with 5 ϫ 10 5 cell equivalents of deoxycholate-solubilized neutrophil membranes, 0.8 g each of recombinant p47 phox , and p67 phox (29) and 1 g of the mutant form of p21 rac1 (Q61L) preloaded with GTP␥S. This mutant form of p21 rac1 was expressed in pGex-2T (4) and was more active than wild-type p21rac1 as noted earlier with p21 rac2 (34). Negative control reactions contained 5 g of superoxide dismutase, in which case cytochrome c reduction did not exceed 5% of the activity observed in the absence of superoxide dismutase. Maximum rates of superoxide generation were calculated from a linear least-squares fit of 5 consecutive 1-min data points based on reactions performed in duplicate.

RESULTS
Expression of Recombinant p40 phox -Attempts to isolate intact recombinant p40 phox from E. coli were not successful due to the proteolytic susceptibility of GST-p40 phox fusion protein.
The affinity-purified fusion protein contained a predominant breakdown product ϳ46 kDa in size, suggesting cleavage of a ϳ19-kDa fragment from the N terminus of p40 phox . However, a nondegraded form of the GST-p40 phox fusion protein was purified from insoluble inclusion bodies of induced cells by preparative SDS-PAGE and was used for primary immunization of rabbits against p40 phox . The baculovirus expression system was used for production of native, full-length recombinant p40 phox , since earlier work showed that this system was efficient in production of other cytosolic oxidase proteins (29). Sonicated lysates from p40 phox baculovirus-infected Sf9 cells (72 h post-infection) contained a prominent 40-kDa species (Fig. 1A), which demonstrated immuno-cross-reactivity with a 40-kDa band in neutrophil cytosol (Fig. 1B). The Sf9 lysates were considerably enriched in p40 phox in comparison with neutrophil cytosol on a specific weight basis. These Sf9 preparations were used without further purification as a source of p40 phox in subsequent cell-free oxidase reconstitution studies. Baculovirus-derived p40 phox also exhibited binding to immobilized GST fused with the proline-rich, C-terminal (33-residue) segment of p47 phox , while not binding to unfused GST (Fig. 1C), consistent with interactions detected in the yeast two-hybrid system (21, 22) (see below). This affinity system was scaled up to purify p40 phox used in later immunizations.
Recombinant p40 phox was also expressed in the cell line K562. These cells lack the neutrophil oxidase components p47 phox , p67 phox , and gp91 phox and proved to be an efficient transfectable model for reconstitution of NADPH oxidase following transduction with gp91 phox retrovirus and co-transfection of p47 phox and p67 phox cDNAs in Epstein-Barr virus-based episomal expression vectors (pREP4 or pREP10) (16,30). To confirm expression of p40 phox in K562 cells, protein lysates from hygromycin-resistant K562 cells transfected with pCEPp40 phox or control vectors were compared by p40 phox immunoblotting. Endogenous p40 phox was not detected in K562 p40 phox Down-regulates NADPH Oxidase through Its SH3 Domain cells transfected with pCEPCAT (control) (Fig. 2, lane 1), while production of recombinant p40 phox in K562 cells transfected with pCEP40 phox was evident by detection of a 40-kDa immunoreactive species at levels similar to neutrophil cytosol (Fig. 2,  lanes 2 and 5). Co-expression of recombinant p40 phox was confirmed in hygromycin-selected K562 cells co-transfected with three episomal vectors, pCEP4/p40 phox , pREP4/p47 phox , and pREP10/p67 phox (Fig. 2, lane 3). The diminished levels of p40 phox seen in this case were consistent with previous experience, where lower expression was observed with co-transfection of several constructs bearing the same selection marker (30). In the absence of pCEP40 phox , immunoblotting of K562 cells did not reveal any endogenous p40 phox stabilized by the other recombinant oxidase components (Fig. 2, lane 4) as was observed in mature CGD neutrophils. These results, showing the absence of endogenous p40 phox in K562 cells and efficient production of p40 phox with this vector, established the utility of these transfectable cells for exploring the role of p40 phox in whole cell NADPH oxidase function.
Whole Cell Oxidase Activity-Retroviral gp91-transduced K562 cells were co-transfected by electroporating 20 g each of pREP4/p47 phox , pREP10/p67 phox , and pCEP4/p40 phox (or pCEP4CAT (control)) plasmids. Immunoblotting of protein lysates from hygromycin-resistant cells confirmed expression of all three recombinant oxidase proteins in these experiments (Figs. 2 and 3). Oxidase activity in response to PMA activation was monitored by chemiluminescence. Four independent cotransfection experiments were performed and the results of one representative experiment are shown in Fig. 3. A significant inhibition of oxidase activity was observed in the presence of p40 phox , typically resulting in 35-40% reduction in chemiluminescence signals in comparison to control transfections with pCEP4CAT (Fig. 3). The extent of oxidase inhibition attributed to p40 phox expression was significant when considering that this vector lacks a unique selection marker. Since the levels of p47 phox and p67 phox detected were comparable in cells expressing p40 phox or CAT (Fig. 3, inset), variable expression of the other oxidase components did not account for the reduced oxi-dase activity in the presence of p40 phox . While the yield of active oxidase in the presence of p40 phox based on the magnitude of the chemiluminescence peak was consistently diminished, the overall kinetic profiles of PMA-elicited respiratory bursts appeared to be similar to control reactions, showing maximum activity 5-10 min after stimulation.
Cell-free NADPH Oxidase Reconstitution Studies-The effects of p40 phox on NADPH oxidase activity reconstituted in vitro were examined following preincubation of recombinant baculovirus-derived p40 phox lysates with one or the other pure cytosolic factors (p67 phox , p47 phox ). Measurements of superoxide production in a system containing recombinant p47 phox , p67 phox , and p21 rac1 (Q61L) showed a significant dose-dependent inhibition of oxidase activity associated with increasing amounts of p40 phox Sf9 cell lysates, in comparison with the same quantities of lysates from uninfected Sf9 cells (Fig. 4). These inhibitory effects were evident irrespective of the order of addition of the other cytosolic factors shown to bind p40 phox . Predictably, low levels of oxidase inhibition (Ͻ25%) were also seen with excessive amounts of control uninfected lysates, although these effects were likely due to effects of high protein concentrations on free arachidonate levels.
Mapping of Interacting Domains within p40 phox , p47 phox , and p67 phox using the Yeast 2-hybrid System-The yeast two-hybrid system was used to delineate interacting domains within cyto- Protein lysates (5 g) described in A were transferred to nitrocellulose and probed with anti-p40 phox serum detected by enhanced chemiluminescence. PMN, human neutrophil cytosol. C, binding of p40 phox to the proline-rich COOH terminus of p47 phox (residues 358 -390) fused to GST (GST-p47PR). p40 phox -containing Sf9 cytosol (p40/BV Sf9) was incubated for 20 min with ϳ5 g of fused or unfused GST proteins immobilized on glutathione-agarose followed by separation of bound protein complexes (Bd) from unbound (Unb) proteins in total cytosol as described (28). Top, fast green (total protein) staining of nitrocellulose blot after SDS-PAGE. Bottom, immunodetection of p40 phox on the same blot. Arrowheads indicate migration of p40 phox ; sizes of molecular weight standards are indicated on the right in kilodaltons.

FIG. 2. Production of recombinant p40 phox in transfected K562
cells. Cells were co-transfected with various episomal vectors and selected as described under "Experimental Procedures." Cytosolic protein lysates (40 g) were subjected to SDS-PAGE, electroblotted, and probed with anti-p40 phox serum using enhanced chemiluminescence methods. Lanes 1 and 2, lysates from single transfections with pCEP4CAT (control) or pCEP4p40, respectively; lane 3, lysate from K562 cells cotransfected with p40 phox , p47 phox , and p67 phox cDNAs; lane 4, lysate from K562 cells co-transfected with p47 phox and p67 phox cDNAs; lane 5, human neutrophil cytosol. Sizes of molecular weight standards are indicated on the right in kilodaltons. p40 phox Down-regulates NADPH Oxidase through Its SH3 Domain solic oxidase proteins that were shown to exist as a complex in resting neutrophils (5)(6)(7). This system has advantages over other in vitro binding assays, since the interactions between proteins occurs in vivo and no protein purification or renaturation from SDS-PAGE gels is required (31)(32)(33). While the system permits qualitative assessment of weak protein-protein interactions (K d Ͻ ϳ10 Ϫ6 M), quantitative binding data are not obtained by this approach. The p40 phox cDNA sequences corresponding to various domains tested were subcloned into activation domain vectors pGADGH or pGAD424. Coding sequences for p67 phox and p47 phox domains were engineered in frame with the Gal4 binding domain in pGBT9. As summarized in Table I, a positive interaction was observed between fulllength p40 phox and p67 phox , which was abolished by deletion of 80 residues from the carboxyl-terminal end of p40 phox (p40⌬CT). Consistent with this, residues 260 -340 (CT) of p40 phox were sufficient to observe an interaction with fulllength p67 phox . The corresponding domain of p67 phox that interacts with p40 phox was identified within a 172-residue segment (residues 293-465) bounded by the two SH3 domains, consistent with similar two-hybrid studies reported recently (22). The interaction between p40CT and residues 293-465 of p67 phox was also confirmed when these cDNAs were expressed in the reciprocal vectors (data not shown). There was no evidence for interaction between the isolated SH3 domain of p40 phox and full-length p67 phox or p67NT (residues 1-246), although a weak positive signal detected in a prolonged ␤-galactosidase assay was observed with co-expression of fulllength p40 phox and the amino-terminal domain (246 residues) of p67 phox (NT) or a smaller fragment (residues 155-246, data not shown). Others have characterized an interaction between 67 NT and p40 phox by surface plasmon resonance, although binding of residues 293-465 of p67 phox with p40 phox was not examined in this case (24).
The interaction between p40 phox and p47 phox was mapped to the SH3 domain (residues 172-229) of p40 phox and the prolinerich tail region (residues 351-390) of p47 phox , consistent with yeast two-hybrid experiments reported elsewhere (21)(22)(23)). An SH3 domain interaction between GRB-2 and human SOS-1 was also demonstrated by yeast two-hybrid expression (33). The p40 phox SH3 domain interaction with the C-terminal domain of p47 phox was confirmed by expression in reciprocal vectors (data not shown).
Delineation of NADPH Oxidase Inhibitory Domains of p40 phox -To explore the significance of interactions between p40 phox and other essential components on whole cell NADPH oxidase function, truncated forms of p40 phox were co-expressed in the episomal vector pCEP4 along with full-length p47 phox and p67 phox ; this work used the deleted form of p40 phox , p40⌬CT, lacking the p67 phox binding domain (C-terminal 80 residues) and pCEP40SH3, encoding the isolated SH3 domain that interacts with p47 phox . As shown in Fig. 5, there was no significant difference between oxidase activity observed in cells expressing full-length p40 phox or p40 phox ⌬CT, indicating that the oxidase inhibitory effects of p40 phox do not involve the p67 phox interaction. In contrast, expression of the isolated SH3 domain of p40 phox , which we and others (21)(22)(23) have shown interacts with the proline-rich C terminus of p47 phox , resulted in inhibition of oxidase activity that was even greater than observed with full-length p40 phox (60% versus 35% inhibition, respectively).

DISCUSSION
Previous studies using purified proteins establish five core NADPH oxidase components as necessary and sufficient for reconstitution of high levels of oxidase activity in vitro (26,27). Recently, p40 phox was identified in resting neutrophil cytosol as yet another protein bound within a high molecular complex (ϳ250 kDa) with two essential cytosolic components, p67 phox and p47 phox . The three proteins appear to co-translocate to the membrane as a complex that interacts with cytochrome b 558 during oxidase activation (5,9). We have now shown that p40 phox can significantly inhibit oxidase activity, whether added to the other cytosolic components in a cell-free assay system or when co-transfected into K-562 cells which lack the FIG. 4. Effects of recombinant p40 phox on cell-free NADPH oxidase activity. In vitro reconstitution of NADPH oxidase activity was determined from the superoxide dismutase-sensitive reduction of cytochrome c as described under "Experimental Procedures." Recombinant p40 phox cytosol was preincubated 20 min at room temperature with recombinant p47 phox (Ⅺ) or p67 phox (छ) prior to addition of solubilized neutrophil membranes, arachidonic acid, substrates, and the other essential components (p21rac1 (Q61L), p67 phox , or p47 phox ). Control reactions (E) contained equivalent amounts of uninfected Sf9 cell cytosol in place of p40 phox cytosol. Activity is expressed relative to that observed in the absence of Sf9 cytosol (100%), which represented 5.7-5.9 nmol of superoxide/min in this experiment. The inhibitory effects of p40 phox shown here are consistent with results from four independent experiments.

TABLE I
Interaction of p40 phox with p67 phox and p47 phox The yeast two-hybrid system was used to identify interacting domains between p40 phox , p67 phox , and p47 phox . ϩ, positive interaction between domains tested; Ϫ, no interaction observed; ϩ/Ϫ, weak ␤-galactosidase activity with long incubation. p40 phox Down-regulates NADPH Oxidase through Its SH3 Domain endogenous protein.
These studies confirm and extend the findings of others by defining precise binding sites between the cytosolic oxidase factors and exploring the role of p40 phox in whole cell oxidase function. Work in several laboratories suggested that the primary association in this cytosolic oxidase complex occurs between p67 phox and p40 phox . In vitro binding studies suggested that the entire pool of cytosolic p40 phox was bound in a tight complex with p67 phox in resting neutrophils (6,7). Deficiencies in either p47 phox or p67 phox seen in autosomal recessive forms of CGD resulted in diminished translocation of p40 phox to the membrane (9), although only the patients deficient in p67 phox showed dramatically reduced levels of total cellular p40 phox , suggesting that p67 phox stabilizes p40 phox in neutrophils.
Direct interactions have also been demonstrated between p47 phox and p40 phox , although these interactions were not readily detected by methods involving immobilized proteins, such as surface plasmon resonance or blotting techniques (8,24). We detected this interaction in solution with recombinant fusion proteins and in the yeast two-hybrid system, as shown by others (21)(22)(23). Early work identified p40 phox by two-hybrid interactions using p47 phox as bait to screen a cDNA library from Epstein-Barr virus-transformed B-cells (23). The importance of the p47 phox interaction is evident in our current findings which correlate delineation of binding sites in p40 phox with whole cell oxidase function. The p40 phox interaction with p47 phox was critical for inhibition of NADPH oxidase activity, while deletion of the p67 phox binding site in p40 phox did not affect its ability to inhibit oxidase activity in whole cells. These observations were not reconciled with recent findings showing in vitro inhibition of the oxidase by antibodies directed against p40 phox (C-terminal, p67 phox -binding domain), although these investigators did not study the proteins directly (25).
The NADPH oxidase inhibition we observed by p40 phox is thought to involve the interaction between its SH3 domain and a target within p47 phox previously shown to participate in assembly of p67 phox in the active oxidase complex (12-14, 16, 18). The isolated SH3 domain of p40 phox inhibited whole cell oxidase activity to an extent that exceeded the inhibition by full-length p40 phox . The demonstration of an interaction between the SH3 domain of p40 phox and the proline-rich carboxyl-terminal do-main of p47 phox (residues 358 -390) was supported by earlier work in which p47 phox was shown to bind to a central 115residue segment of p40 phox , half of which included the SH3 domain (22,23). More recent studies showed the SH3 domain of p40 phox binds to the proline-rich tail region of p47 phox and thereby inhibits binding of the C-terminal p67 phox SH3 domain to the same region p47 phox , when over-expressed in yeast (21). These authors suggested that competitive binding between p40 phox and p67 phox could modulate oxidase activity. Our observations of inhibitory effects of p40 phox or its isolated SH3 domain on whole cell oxidase function lend further support to this competitive binding model.
Our results also provide additional evidence for the importance of the tail-tail interaction between p47 phox and p67 phox in oxidase activation. The role of this C-terminal domain interaction had been disputed because neither tail domain was necessary for cell-free reconstitution of NADPH oxidase activity (8,13). However, work in whole transfected cells showed that deletions or mutations in the carboxyl-terminal domains of either protein dramatically affected whole cell oxidase activity and p67 phox membrane translocation (13,16,18). Since the tail-tail interaction is not essential for cell-free oxidase activity and cell-free oxidase inhibition was not influenced by the order of addition of the three cytosolic proteins, it is unclear whether p40 phox inhibits the cell-free system by the same mechanism suggested from the whole cell studies.
In summary, we have examined the role of interactions between p40 phox and two other essential oxidase components, both in vitro and in transfected cells where expression of NADPH oxidase components could be genetically manipulated. Based on these findings, we suggest a model outlined in Fig. 6 in which clear distinctions can be made in terms of function between two interactions of p40 phox . In one case, the interaction of p40 phox with p67 phox appears to play no obvious role in the oxidase activation process, although its importance is evident in neutrophils where p40 phox stability is affected by the absence FIG. 5. Delineation of NADPH oxidase inhibitory domains of p40 phox . gp91 phox -transduced K562 cells co-transfected with p47 phox , p67 phox , and p40 phox full-length (p40FL), p40 phox residues 1-246 (p40⌬CT), or the p40 phox SH3 domain (p40SH3), were stimulated with 2 g/ml PMA and assayed for superoxide production by chemiluminescence. Superoxide production is expressed as the percent activity relative to control cells transfected with pCEP4CAT (pCAT). Each graph represents the mean of data from three independent transfection experiments, with bars representing the standard deviation of percent activity.
FIG. 6. Interactions of p40 phox with NADPH oxidase components shown in the context of other SH3 domain interactions. p40 phox interactions with both p47 phox and p67 phox have been identified and functionally characterized in whole cells. The interaction with p67 phox (dashed line) influences the stability of p40 phox in resting neutrophils (5-7, 9) but does not appear to affect oxidase activity. In contrast, the p40 phox SH3 domain interaction with p47 phox (gray line) inhibits oxidase activity by binding to the same site that binds an SH3 domain in p67 phox (12-14, 16, 18). This site also binds an SH3 domain within p47 phox (17), although the role of this interaction has not been clearly defined. See text for details. p40 phox Down-regulates NADPH Oxidase through Its SH3 Domain of p67 phox seen in CGD (5,(7)(8)(9). In the second, the interaction between p47 phox and p40 phox is not readily detected in resting neutrophil cytosol but was shown to play a predominant role during activation through its inhibition of productive interactions between two other essential oxidase components. These findings again illustrate the central role played by multiple SH3 domain interactions in modulation of NADPH oxidase activity. In this case, we demonstrated down-regulation by an accessory SH3 domain-containing protein (p40 phox ), which by virtue of its homology to one essential oxidase component (p67 phox ) competes for a common target site on another. This SH3 target site in p47 phox is flanked by sites that were recently shown to be phosphorylated during oxidase activation (35). Future work will examine whether phosphorylation alters binding affinities of these competing SH3 domains and whether these changes are critical in the activation process.