A domain of p47phox that interacts with human neutrophil flavocytochrome b558.

The NADPH-dependent oxidase of human neutrophils is a multicomponent system including cytosolic and membrane proteins. Activation requires translocation of cytosolic proteins p47phox, p67phox, and Rac2 to the plasma membrane and association with the membrane flavocytochrome b to assemble a functioning oxidase. We report the location of a region in p47phox that mediates its interaction with flavocytochrome b. From a random peptide phage display library, we used biopanning with purified flavocytochrome b to select phage peptides that mimicked potential p47phoxbinding residues. Using this approach, we identified a region of p47phox from residue 323 to 342 as a site of interaction with flavocytochrome b. Synthetic peptides 315SRKRLSQDAYRRNS328, 323AYRRNSVRFL332, and 334QRRRQARPGPQSPG347 inhibited superoxide (O∸2) production in the broken cell system with IC50 of 18, 57, and 15 μM, respectively. 323AYRRNSVRFL332 and its derivative peptides inhibited phosphorylation of p47phox. However, the functional importance of this peptide was independent of its effects on phosphorylation, since 323AYRRNAVRFL332 inhibited O∸2 production, but had no effect on phosphorylation. None of the peptides blocked O∸2 production when added after enzyme activation, suggesting that they inhibited the assembly, rather than the activity, of the oxidase. Furthermore these peptides inhibited membrane association of p47phox in the broken cell translocation assay and O∸2 production by electropermeabilized neutrophils, thereby supporting the interpretation that this region of p47phox interacts with flavocytochrome b.

The NADPH-dependent oxidase of human neutrophils is a multicomponent system including cytosolic and membrane proteins. Activation requires translocation of cytosolic proteins p47 phox , p67 phox , and Rac2 to the plasma membrane and association with the membrane flavocytochrome b to assemble a functioning oxidase. We report the location of a region in p47 phox that mediates its interaction with flavocytochrome b. From a random peptide phage display library, we used biopanning with purified flavocytochrome b to select phage peptides that mimicked potential p47 phox binding residues. Using this approach, we identified a region of p47 phox from residue 323 to 342 as a site of interaction with flavocytochrome b. Synthetic peptides 315 SRKRLSQD-AYRRNS 328 , 323 AYRRNSVRFL 332 , and 334 QRRRQARPG-PQSPG 347 inhibited superoxide (O 2 . ) production in the broken cell system with IC 50 of 18, 57, and 15 M, respectively. 323 AYRRNSVRFL 332 and its derivative peptides inhibited phosphorylation of p47 phox . However, the functional importance of this peptide was independent of its effects on phosphorylation, since 323 AYRRNA-VRFL 332 inhibited O 2 . production, but had no effect on phosphorylation. None of the peptides blocked O 2 . production when added after enzyme activation, suggesting that they inhibited the assembly, rather than the activity, of the oxidase. Furthermore these peptides inhibited membrane association of p47 phox in the broken cell translocation assay and O 2 . production by electropermeabilized neutrophils, thereby supporting the interpretation that this region of p47 phox interacts with flavocytochrome b.
Human polymorphonuclear leukocytes (PMNs) 1 play an important role in host defense against invading microorganisms. PMNs possess an NADPH-dependent oxidase which is capable of generating superoxide anion (O 2 . ) and other microbicidal oxygen-derived species (e.g. H 2 O 2 , HOCl) when activated by various particulate and soluble stimuli (1,2). The NADPH oxidase is a multicomponent enzyme system which is unassembled in resting PMNs but assembles on the plasma membrane in activated PMNs (3,4). The critical importance of the PMN NADPH oxidase in normal host defense is most dramatically illustrated by the frequent and severe infections seen in patients with chronic granulomatous disease (5,6). The PMNs from such patients lack a functionally competent oxidase and, when stimulated, fail to generate O 2 . .
Essential components of the NADPH oxidase include plasma membrane and cytosolic proteins. The key plasma membrane component is a heterodimeric flavocytochrome b which is composed of a 91-kDa glycoprotein (gp91 phox ) and a 22-kDa protein (p22 phox ) (7-10). Flavocytochrome b serves to transfer electrons from NADPH to molecular oxygen, resulting in the generation of O 2 . . In PMN membranes, a low molecular weight GTP-binding protein, Rap1A, is associated with flavocytochrome b and plays an important role in NADPH oxidase regulation in vivo (11)(12)(13). Cytosolic proteins p47 phox , p67 phox , and a second low molecular weight GTP-binding protein, Rac2, are absolutely required for NADPH oxidase activity (14 -18), and these three proteins associate with flavocytochrome b to form the functional NADPH oxidase (19 -21). Additionally, a cytosolic protein, p40 phox , has recently been identified, but its role in oxidase function is not completely defined (22). According to the current model of NADPH oxidase assembly, p47 phox and p67 phox translocate en bloc to associate with flavocytochrome b during PMN activation (23,24). Rac2 translocates simultaneously but independently of the other two cytosolic components to associate with the membrane-bound NADPH oxidase (25,26). Studies of oxidase assembly in PMNs of patients with various forms of chronic granulomatous disease suggest that p47 phox binds directly to flavocytochrome b (20), and at least six regions of flavocytochrome b have been identified as potential sites for interaction with p47 phox , including four sites on gp91 phox and two sites on p22 phox (27)(28)(29)(30)(31)(32)(33)(34). In contrast, the complementary sites of interaction presented on p47 phox have not been fully characterized (29,33,34). In previous studies, peptides mimicking p47 phox residues 323 AYR-RNSVRFL 332 inhibited phosphorylation of p47 phox , O 2 . production, and translocation of cytosolic components in the broken cell system (35), suggesting that this might be a possible site of interaction between p47 phox and flavocytochrome b.
In the present work, we used an approach combining the screening of a random peptide phage display library with the functional analysis of synthetic peptides to define residues in p47 phox which interact with flavocytochrome b. Our data indicate that the region encompassing amino acids 323-342 comprises a functionally important domain in the association of p47 phox with flavocytochrome b.

EXPERIMENTAL PROCEDURES
Materials-NADPH, ferricytochrome c (horse heart, type VI), superoxide dismutase, and FAD were obtained from Sigma. GTP␥S (5Јtriphosphate) was obtained from Boehringer Mannheim and arachidonic acid (sodium salt) from NuChek Prep (Elysian, MN). NHS-LCbiotin was obtained from Pierce, and streptavidin was from Fisher. Dextran, Ficoll, and Percoll were obtained from Pharmacia Biotech Inc., and Hypaque was obtained from Winthrop Laboratories (New York).
Preparation and Fractionation of PMNs-Human PMNs were isolated from heparinized venous blood using sequential dextran sedimentation, differential density sedimentation in Hypaque-Ficoll gradients, and hypotonic lysis of erythrocytes as described previously (35,36). Two different methods were employed for the isolation of PMN membranes and cytosol with no difference in the experimental results. Purified PMNs were treated with 2 mM diisopropyl fluorophosphate for 20 min at 4°C, washed, resuspended in relaxation buffer (KCl, 100 mM; NaCl, 3 mM; MgCl 2 , 3.5 mM; EGTA, 1.25 mM; Pipes, 10 mM, pH 7.3) and disrupted by N 2 cavitation (35). Nuclei and unbroken cells were pelleted (200 ϫ g, 6 min, 5°C) and the supernatant loaded on top of an isotonic discontinuous gradient of Percoll as described previously (37). Cytosol was collected from the top of the gradient and clarified before use. The plasma membrane fraction was collected, Percoll removed by centrifugation, and the membrane washed prior to use. Alternatively, purified PMNs (5 ϫ 10 8 /ml in 150 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 15 g/ml leupeptin, 10 g/ml chymostatin, 1 mM EGTA, and 10 mM Hepes, pH 7.0) were disrupted by N 2 cavitation as described previously. Membrane and cytosolic fractions were prepared from the cavitate by sequential centrifugation as described by Fujita et al. (36).
Preparation and Purification of Flavocytochrome b-Human neutrophil flavocytochrome b was purified from ϳ10 10 neutrophils following the methods of Parkos et al. (7,38). The purified flavocytochrome b was Ͼ95% pure, as determined by SDS-polyacrylamide gel electrophoresis and Coomassie Blue staining (7,38), and immunoblotting with anticytochrome b antibodies confirmed the presence of both gp91 phox and p22 phox (not shown).
Random Sequence Peptide Bacteriophage Display Library Analysis-Fifty g of human neutrophil flavocytochrome b was biotinylated to a final biotin/protein ratio of 20:1 (39). Three rounds of biopanning were performed as described previously (39) using 10 g/pan of biotinylated flavocytochrome b in 0.2% Triton X-100 and 75 l (round 1 only) of a nonapeptide phage display library (J404-3) (40). The bacteriophage eluates of the first two pans were amplified in K91 Escherichia coli cells on solid Luria-Bertani (LB) agar dishes containing 100 g/ml kanamycin to approximately 10 13 plaque-forming units. The phage were then extracted (41), and one-third of the eluate was used in the next round. The third round eluate was plated in K91 cells on LB plates, and plaques were then randomly picked and sequenced using a gene III-specific primer (40). Dried sequencing gels were analyzed on a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA).
Peptide Synthesis-Peptides SRKRLSQDAYRRNS, QRRRQARPG-PQSPG, RQARPGPQ, KLSYRPRDSNE, and AVEGGMKPVKLLVGC were synthesized by the Montana State University Peptide Synthesis Facility. Peptide sequences and purity (Ͼ95%) were confirmed by mass spectrometry. The p47 phox peptides PPRRSSIRNA, HQRSRKRLSQD, AYRRNSVRFL, and gp91 phox peptide RGVHFIF were synthesized at the Core Peptide Facility at the University of Iowa and displayed Ͼ95% purity via high performance liquid chromatography analysis. The p47 phox (323-332)-related peptides (Tables I and II) were obtained from Macromolecular Resources (Colorado State University, Fort Collins, CO), and sequence and purity were confirmed by mass spectrometry.
Broken Cell NADPH Oxidase Reconstitution and Translocation Assay-NADPH oxidase activity was measured using two different versions of the broken cell system, and results with the two systems were similar. In some studies, the broken cell system used arachidonic acid as the activating agent, and activity was determined as described by Nauseef et al. (35). In other experiments, activity was measured in an SDS-activated system as described by DeLeo et al. (32). To analyze peptide inhibition, peptides dissolved in broken cell assay buffer were added to the reaction mixture at the indicated concentrations before, and after the 2-min incubation with agonist and maximum rate of O 2 . production was determined as described (32). Two broken cell translocation assays were used with similar results. Studies of AYRRNSVRFL and related peptides were performed using 91 M arachidonic acid as the activating agent, and translocation was determined as described by Park and Babior (42). In other experiments, the broken cell translocation assay of Verhoeven et al. (43) was used.
Analysis of Phosphorylated Proteins-Conditions used to assess the effect of various peptides on the phosphorylation of p47 phox mirrored closely those for quantification of O 2 . generation and have been described in detail previously (35). Relative quantitation of p47 phox phosphorylation was obtained by densitometric analysis of resultant autoradiographs.
Electrophoresis and Immunoblotting-Proteins were separated by SDS-polyacrylamide gel electrophoresis and electroblotted as described previously (21,35). Immunoblots were probed with primary antibody, rabbit anti-p47 phox (21,35,44), and followed by either iodinated protein A or alkaline phosphatase-conjugated goat anti-rabbit IgG antibody (Bio-Rad). Relative amounts of translocated proteins were quantitated by densitometric analysis of the immunoreactive p47 phox detected by autoradiography or colorimetric assay.
Preparation and Electropermeabilization of Human Neutrophils-Histopaque-purified neutrophils were washed once with ice-cold permeabilization buffer (140 mM KCl, 10 mM Hepes, 10 mM D-glucose, 1 mM MgCl 2 , 0.193 mM CaCl 2 , and 1 mM EGTA, pH 7.2) and then resuspended in the same buffer supplemented with 100 M GTP␥S, 1 mM ATP, and 2 mM NADPH, to a final concentration of 10 7 cells/ml and stored on ice. Aliquots of 800 l were placed in 0.4-cm electroporation cuvettes and permeabilized with a Bio-Rad gene pulser using two consecutive pulses (with brief stirring between pulses) of 3.75 kV/cm with a 25-microfarad capacitor ( ϭ 0.3-0.5 ms) at 4°C. The cells were then incubated on ice for 5 and 30 min with or without 1 mM of the appropriate peptide. Subsequently, the cells were diluted into supplemented permeabilization buffer containing 100 M cytochrome c Ϯ 1 mM peptide (final concentration of 5 ϫ 10 6 cells/ml) and then stimulated with 1 g/ml phorbol 12-myristate 13-acetate (reference cuvette contained 25 g/ml superoxide dismutase). The maximal rate of superoxide dismutaseinhibitable reduction of cytochrome c was determined on a Cary dualbeam spectrophotometer (Varian, Melbourne, Australia) at 550 nm and 25°C for 10 min. Electropermeabilized neutrophils retained 92.3 Ϯ 1.9 and 49.8 Ϯ 6.3% of their ability to produce O 2 . when incubated for 5 and 30 min after pulsing, respectively, compared with control, nonpermeabilized cells and were found to be 98% permeable to trypan blue 5 min after permeabilization.

RESULTS
Biopanning with Purified Flavocytochrome b-To identify the region or regions in p47 phox that associate with flavocytochrome b, a random sequence nonapeptide bacteriophage display library was screened with purified flavocytochrome b. The predicted amino acid sequences from 94 of these affinity selected bacteriophage were analyzed and three dominant consensus motifs were identified (Fig. 1). When compared with the amino acid sequence of p47 phox , these motifs mapped to residues 323-342 in p47 phox , a region which contains multiple sites for serine/threonine kinase phosphorylation and includes a previously identified area of potential interaction with flavocytochrome b at residues 323-332 (35,45). The strongest homology among phage peptides was evident in those representative of the p47 phox region 338 QARPG 342 , and eight phage isolates contained three to four residues identical to this region. Additionally, seven phage isolates contained peptides that were homologous to p47 phox residues 332 LQQRRRQ 338 (six of these contained three-residue matches), which immediately precede the 338 QARPG 342 region. In combination, 21 phage isolates were representative of the p47 phox region 332 LQQRRRQA-RPG 342 (Fig. 1). Conservative substitutions and one residue shifts give many of these phage peptides even more similarity to the indicated region. This is especially evident for residues 335-337, where arginine appears to be substituted by a positively charged histidine or lysine.
A third consensus sequence mapped to p47 phox 324 YRRNS 328 and was represented by 21 phage peptides, and conservative substitutions or one-residue shifts give many phage representative of this area an even greater similarity. Nauseef et al. (35) previously found that p47 phox residues 323 AYRRNSVRFL 332 represented a functionally important domain involved in NADPH oxidase assembly. Certain residues within this region appear to be critically important for structural constraints, as the exclusion of these residues renders the peptide ineffective at inhibiting O 2 . production in the broken cell assay (35). Thus, the mapping of this site using a random sequence library provides direct evidence confirming the biological relevance of this site.
Peptide Inhibition in the Broken Cell Assay-To determine if the p47 phox regions selected from the phage display library approach were relevant for the association of p47 phox with flavocytochrome b and, thus, NADPH oxidase activity, peptides representative of these regions were synthesized and analyzed in a broken cell NADPH oxidase assay. In a previous study, p47 phox peptides 299 PPRRSSIRNA 308 and 312 HQRSRSRKRL-SQD 322 did not affect O 2 . production significantly in the broken cell assay at concentrations up to 500 M (35). These peptides encompass potential phosphorylation sites, are also positively charged, and precede the 323-342 region of p47 phox mapped by the phage display library. In contrast, the peptides mimicking portions of p47 phox residues 315-347, 315 SRKRLSQDAYR-RNS 332 , 323 AYRRNSVRFL 332 , and 334 QRRRQARPGPQSPG 347 , all inhibited O 2 . production in the broken cell assay in a dosedependent manner with IC 50 values of 18, 57, and 15 M, respectively (see Fig. 2 and Table I). Surprisingly, a shorter peptide derived from 334 QRRRQARPGPQSPG 347 , 337 RQAR-PGPQ 344 , was much less inhibitory (IC 50 ϳ750 M) (see Fig. 2). Apparently, the loss of key flanking residues and/or the potential phosphorylation site at Ser 345 , or most likely, loss of conformational constraint rendered 337 RQARPGPQ 344 ineffective at inhibiting O 2 . production in the broken cell system. Charge alone does not seem to be the major determinant of inhibition in the broken cell assay. A control, unrelated peptide (KLSR-PRDSNE) and two p47 phox peptides ( 299 PPRRSSIRNA 308 and 312 HQRSRSRKRLSQD 322 ) all contain a similar percentage of positively charged residues (27.3, 30.0, and 46.2%, respectively) as do 334 QRRRQARPGPQSPG 347 and 315 SRKRLSQ-DAYRRNS 332 (28.6 and 35.7%, respectively), but had no inhibitory activity in the broken cell system at concentrations up to 500 M.
Previous studies indicated that p47 phox peptide 323-332 is important in the assembly of a functioning NADPH oxidase (35). In order to define more precisely the residues in this region which are critical for oxidase assembly, we compared the effects of various peptides derived from p47 phox 323-332 on O 2 . generation in the broken cell system (Table I). Peptide 325-330, a hexapeptide including the putative phosphorylation site at Ser 328 , did not inhibit O 2 . production in the broken cell system. Similarly the addition of Tyr 324 or Phe 331 to RRNSVR did not make the peptide inhibitory. However, the addition of both aromatic residues to RRNSVR resulted in inhibitory activity, albeit still less than that of the parent peptide (Table I).     (27) and our previous studies (32). These data indicate that prior to full activation of the NADPH oxidase, peptides representative of regions on oxidase components that participate in protein-protein interactions block assembly. Inhibition of the Phosphorylation and Translocation of p47 phox -Previously, we reported that the inhibitory effects of p47 phox 323-332 on O 2 . generation are paralleled in assays of p47 phox phosphorylation and translocation in the broken cell system (35). Similar studies of this family of p47 phox 323-332related peptides were performed (Table I). In general, the effect of a given peptide on O 2 . generation was paralleled by its effect on translocation, supporting the idea that p47 phox 323-332 exerts its inhibitory effect on O 2 . production by blocking translocation of p47 phox and assembly of a functioning oxidase. In contrast, the effects of these peptides on phosphorylation of p47 phox did not parallel effects on O 2 . production. RRNSVR, YRRNSVR, and RRNSVRF reduced phosphorylation of p47 phox but did not block O 2 . production. The S328A mutation did not inhibit phosphorylation but did block O 2 . production and translocation. Taken together, these data suggest that in the broken cell system the interactive site(s) in p47 phox 323-332 which associate with flavocytochrome b do not depend on their being phosphorylated.
As discussed above, p47 phox peptides 315-332, 323-332, and 334 -347 all inhibited O 2 . production in the broken cell system when added prior to activation but were ineffective when added after assembly, suggesting that these peptides blocked assembly by interfering with the association of p47 phox with flavocytochrome b. To confirm that p47 phox peptides 315 SRKRLSQ-DAYRRNS 332 and 334 QRRRQARPGPQSPG 347 were blocking assembly of the NADPH oxidase rather than activity, these peptides were tested in the broken cell translocation assay at 100 M. Both peptides inhibited the translocation of p47 phox to the membrane compared to a control sample (Fig. 3). Peptide 334 QRRRQARPGPQSPG 347 inhibited translocation to ϳ4% of control (Fig. 3), whereas peptides 315 SRKRLSQDAYRRNS 332 and 323 AYRRNSVRFL 332 inhibited p47 phox translocation to the membrane to 35 and 16% ( Fig. 3 and Table I) of control, respectively. In contrast, peptides representative of p47 phox residues preceding and partially encompassing 315-342 ( 299 PPRRSSIRNA 308 and 312 HQRSRSRKRLSQD 322 ) did not inhibit O 2 . production or translocation of p47 phox to the membrane in broken cell assay systems (35).

Inhibition of Superoxide Production in Electropermeabilized
Neutrophils-The broken cell assay for O 2 . production does not mimic entirely, the NADPH oxidase assembly in the intact PMN (45). For that reason, the ability of active p47 phox peptides to inhibit in electropermeabilized neutrophils was evaluated. Peptides 315 SRKRLSQDAYRRNS 328 , 323 AYRRNSVRFL 332 , and 334 QRRRQARPGPQSPG 347 were incubated with the permeabilized neutrophils for 5 min and phorbol 12-myristate 13-acetate-stimulated O 2 . production by these cells was determined. We chose a 5-min treatment time in these experiments for two reasons: 1) the cells were determined to be fully permeable by this time (Ͼ98% permeable), and 2) at longer incubation times the permeabilized cells failed to retain a level of O 2 . generating capacity that reasonably reflected that in the intact cell (Ͼ92% of control activity was maintained after 5 min of incubation, whereas Ͻ50% activity was present after 30 min). As shown in Table II, the peptides that were inhibitory in the broken cell assay system also inhibited O 2 . production in permeabilized neutrophils (ϳ40 -50% inhibition compared with controls). Similarly, peptides that were inactive in the broken cell assay ( 337 RQARPGPQ 344 and 325 RRNSVR 330 ) also had no effect on permeabilized cells and served as negative controls (Table II). In addition, we also tested the gp91 phox carboxylterminal peptide RGVHFIF as a positive control, since it was shown previously by Kleinberg et al. (27) to inhibit oxidase activity in electropermeabilized cells. Our data (Table II) were consistent with their results. Two irrelevant peptides (AVEG-GMKPVKLLVGC and KLSYRPRDSNE) were used to deter-  mine nonspecific peptide effects in permeabilized cells, and their molecular weights (1500.3 and 1364.5, respectively) were similar to that of 323 AYRRNSVRFL 332 (1281.6), 315 SRKRLSQ-DAYRRNS 328 (1719.1), and 334 QRRRQARPGPQSPG 347 (1590.8). Thus, in support of our findings in the broken cell assays, our data indicate that the active p47 phox peptides identified represent sites which participate in the assembly of the active NADPH oxidase in vivo.

DISCUSSION
The cytosolic NADPH oxidase protein p47 phox has been shown to associate with flavocytochrome b, and several sites have been identified in carboxyl-terminal domains of both gp91 phox and p22 phox that are important for this interaction (27)(28)(29)(30)(31). Additionally, our recent studies suggest that p47 phox also binds to a region close to the amino terminus of gp91 phox (32). Recent studies have examined which regions in p47 phox participate in the binding to flavocytochrome b. Two reports demonstrated that Src homology 3 domains in p47 phox may interact with p22 phox (29,33,34). In addition, Nauseef et al. (35) screened p47 phox peptides containing phosphorylation sites and found that one of these peptides (323-332) represents a functionally important domain in p47 phox . This peptide, 323 AY-RRNSVRFL 332 , inhibits O 2 . production, phosphorylation, and translocation of p47 phox in the broken cell system (35). Thus, the data suggested that this p47 phox domain might play a role in oxidase assembly, although no direct evidence supported this possibility.
In the present studies using random peptide phage display library analysis, we provide direct evidence that p47 phox residues 323-332, as well as the adjacent region extending to residue 342, are important in the binding of p47 phox to flavocytochrome b. Thus, the entire binding domain encompasses residues 323-342 of p47 phox , and synthetic peptides representing these adjacent sites and overlapping with the 323-332 region were potent inhibitors of NADPH oxidase activity.
production was shown to be dose-dependent and targeted assembly of the oxidase rather than activity, as demonstrated by the inhibition of translocation by these peptides and their ineffectiveness when added after assembly of the oxidase. . production in permeabilized neutrophils supports the data from the phage library analysis and broken cell system and confirms that p47 phox residues 323-342 are important for oxidase activation. Phosphorylation of p47 phox is required for activation of the NADPH oxidase in neutrophils (46 -48); however, this requirement for phosphorylation is not observed in the broken cell system (35,45). Previously, it has been suggested that phosphorylation of p47 phox may function, in part, to neutralize pos-itively charged regions of the protein, thus allowing it to interact with the membrane or target protein. In the broken cell system, the addition of anionic detergent (SDS) or arachidonic acid appears to bypass the need for p47 phox phosphorylation by imparting negative charge to the protein. Consistent with this hypothesis, the p47 phox region identified here as an important domain for association with flavocytochrome b and NADPH oxidase assembly (323-342), is within a larger region (314 -347) that contains 11 positively charged residues, one protein kinase C phosphorylation site (Ser 328 ), and is surrounded by several other potential sites for phosphorylation by protein kinase C and tyrosine kinase (49). Previously, Joseph et al. (50) reported that any polybasic peptide (Ն5 basic residues) could nonspecifically inhibit NADPH oxidase activity; however, they also found that an 11 residue peptide containing six lysines had no effect on oxidase activity. We have analyzed a number of polybasic peptides, including some peptides with the same number of basic residues as in our active peptides, that had no inhibitory effect on NADPH oxidase activity in a broken cell assay system (Ref. 35 and Table I) and in electropermeabilized neutrophils (Table II). Thus, the inhibition of NADPH oxidase activity by the active p47 phox peptides is specific to their sequences and/or the charge distribution represented by their sequences, and the apparent "nonspecific" inhibition of oxidase activity by polybasic peptides unrelated to the NADPH oxidase, as reported by Joseph et al. (50), is due to the blocking of specific binding interactions (in this case between p47 phox and gp91 phox ) that involve basic amino acid-enriched domains on one or both of the interacting proteins.
In conclusion, the data presented, based on the complementary approaches of random peptide phage display library analysis and peptide inhibition in both broken cell and permeabilized cell systems, demonstrate that p47 phox residues 323-342 interact with flavocytochrome b and are required for the association of these two NADPH oxidase components. The association is sequence-specific and appears to require certain conformational constraints. The elucidation of the sites of interaction between human neutrophil NADPH oxidase component proteins will lead to a further understanding of the regulation of this system.