The adaptor protein p40^phox as a positive regulator of the superoxide-producing phagocyte oxidase

Abstract

Activation of the superoxide-producing phagocyte NADPH oxidase, crucial in host defense, requires the cytosolic proteins p67^phox and p47^phox. They translocate to the membrane upon cell stimulation and activate flavocytochrome b₅₅₈, the membrane-integrated catalytic core of this enzyme system. The activators p67^phox and p47^phox form a ternary complex together with p40^phox, an adaptor protein with unknown function, comprising the PX/PB2, SH3 and PC motif- containing domains: p40^phox associates with p67^phox via binding of the p40^phox PC motif to the p67^phox PB1 domain, while p47^phox directly interacts with p67^phox but not with p40^phox. Here we show that p40^phox enhances membrane translocation of p67^phox and p47^phox in stimulated cells, which leads to facilitated production of superoxide. The enhancement cannot be elicited by a mutant p40^phox carrying the D289A substitution in PC or a p67^phox with the K355A substitution in PB1, each being defective in binding to its respective partner. Thus p40^phox participates in activation of the phagocyte oxidase by regulating membrane recruitment of p67^phox and p47^phox via the PB1–PC interaction with p67^phox.

Introduction

Professional phagocytes, including neutrophils and macrophages, play a crucial role in host defense against microbial infection. During phagocytosis of invading microorganisms or upon cell stimulation with soluble agents, phagocytes produce superoxide (O₂^–), a precursor of microbicidal oxidants, by catalysis of an activated NADPH oxidase (Sumimoto et al., 1997; Babior, 1999; Clark, 1999; Nauseef, 1999). The soluble stimulants include phorbol myristate acetate (PMA) and chemoattractants, the latter of which bind to their own receptors to activate the oxidase via the trimeric GTPase G_i (Bokoch, 1995). The significance of the oxidase in host defense is exemplified by recurrent and life-threatening infections that occur in patients with chronic granulomatous disease (CGD), whose phagocytes are defective in the superoxide-producing activity (Roos et al., 1996).

The redox core of the phagocyte NADPH oxidase is a membrane-spanning flavocytochrome b₅₅₈, comprising the two subunits gp91^phox and p22^phox. The cytochrome is catalytically inactive in resting cells but acquires the capacity to transfer electrons from NADPH to molecular oxygen upon cell stimulation, leading to superoxide production (Roos et al., 1996; Babior, 1999). The activation of cytochrome b₅₅₈ requires stimulus-induced membrane translocation of cytosolic proteins, namely the small GTPases Rac and two SH3 domain-harboring oxidase factors p47^phox and p67^phox (Sumimoto et al., 1997; Babior, 1999; Clark, 1999; Nauseef, 1999). In CGD neutrophils lacking cytochrome b₅₅₈, neither p47^phox nor p67^phox can be recruited to the membrane upon cell stimulation (Heyworth et al., 1991). In p47^phox-deficient phagocytes, membrane targeting of p67^phox does not occur, whereas p47^phox is independently targeted in p67^phox-defective cells (Heyworth et al., 1991; Dusi et al., 1996). Thus p47^phox appears to directly interact with cytochrome b₅₅₈ and recruits p67^phox to the membrane. Indeed the SH3 domains of p47^phox bind to a proline-rich region (PRR) of p22^phox, the small subunit of cytochrome b₅₅₈ (Leto et al., 1994; Sumimoto et al., 1994). Although the SH3 domains are normally masked via an intramolecular interaction, the conformation of p47^phox becomes changed upon cell stimulation to render the domains in a state accessible to p22^phox, which leads to the activation of the NADPH oxidase (Sumimoto et al., 1994; Ago et al., 1999). The P156Q substitution in the p22^phox PRR, a mutation that has occurred in a case of CGD (Dinauer et al., 1991), results in not only impaired interaction between p22^phox and p47^phox in vitro (Sumimoto et al., 1994), but also defective translocation of p47^phox to the membrane in vivo (Leusen et al., 1994). It is also known that p67^phox constitutively associates with p47^phox via a tail-to-tail interaction: the C-terminal SH3 domain of p67^phox directly binds to a PRR in the C-terminal region of p47^phox (Leusen et al., 1994; Ito et al., 1996). On the other hand, membrane translocation of Rac is independent of the presence of cytochrome b₅₅₈, p47^phox and p67^phox: even in CGD neutrophils lacking one of these proteins, the GTPase is normally recruited to the membrane (Heyworth et al., 1994).

Another SH3 domain-containing oxidase factor, p40^phox, has been identified as a protein that constitutively associates with p67^phox in the cytosol of resting phagocytes (Someya et al., 1993; Wientjes et al., 1993; Tsunawaki et al., 1994). It comprises three portions: a PX (Phox) domain [also known as PB2 (Phox and Bem 2) domain] at the N-terminus, an SH3 domain in the middle and the C-terminal region containing a PC (Phox and Cdc) motif (Ponting, 1996; Sumimoto et al., 1997; Mizuki et al., 1998; Nakamura et al., 1998; Hiroaki et al., 2001) (Figure 1). The PC motif is present in a variety of signaling proteins such as the yeast polarity protein Cdc24p, the protein kinase MEK5 and Zip, a protein that interacts with the ζ isoform of protein kinase C, and serves as a target of PB1 (Phox and Bem 1) domain, thereby mediating modular protein–protein interactions (Ito et al., 2001; Terasawa et al., 2001). The protein p67^phox contains a PB1 domain between the two SH3 domains, which recognizes the PC motif of p40^phox and plays an essential role in interaction between p67^phox and p40^phox (Ito et al., 2001). Thus, in a ternary oxidase complex of resting phagocytes, p67^phox is thought to tether p47^phox to p40^phox: the C-terminal SH3 domain binds to the PRR of p47^phox and the PB1 domain directly interacts with the PC motif of p40^phox (Figure 1).

Fig. 1. Protein–protein interactions in a complex of the phagocyte NADPH oxidase factors p67^phox, p47^phox and p40^phox. The NADPH oxidase proteins p67^phox, p47^phox and p40^phox form a complex in resting phagocytes. The three proteins harbor multiple modular domains: p67^phox comprises a domain containing four units of TPR (tetratricopeptide repeat), the first SH3 domain, PB1 domain and the second SH3 domain; p47^phox contains PX or PB2 domain, two SH3 domains and the C-terminal PRR; and p40^phox harbors PX/PB2 and SH3 domains, followed by PC motif. In the oxidase complex, p67^phox directly binds to p47^phox via a tail-to-tail interaction between the C-terminal SH3 of p67^phox and the PRR of p47^phox; and p67^phox also directly binds to p40^phox via a novel modular interaction between the PB1 domain of p67^phox and the PC motif-containing region of p40^phox. Thus p67^phox tethers p47^phox to p40^phox.

The role of p40^phox in the phagocyte NADPH oxidase activation has remained largely elusive, although it probably resides in a complex of the oxidase as described above. The oxidase can be activated in a cell-free system reconstituted with cytochrome b₅₅₈, cytosolic proteins (p47^phox, p67^phox and Rac in the GTP-bound form) and anionic amphiphiles such as arachidonate and sodium dodecyl sulfate (Bromberg and Pick, 1985; Hata et al., 1998). It is well known that the cell-free activation does not require p40^phox. The effect of the addition of p40^phox to the cell-free system, however, seems controversial: Cross (2000) has shown that p40^phox only slightly elevates the oxidase activity under cell-free conditions, while other investigators have reported that p40^phox represses the oxidase activation at high concentrations (Sathyamoorthy et al., 1997). Although the cell-free system has greatly contributed to understanding of the molecular mechanism for activation of the NADPH oxidase, results obtained using this system sometimes seem inconsistent with those using intact cells. For instance, p67^phox lacking the C-terminal SH3 domain is incapable of activating the NADPH oxidase when it is expressed in cells, while the mutant p67^phox almost fully activates the oxidase under cell-free conditions (de Mendez et al., 1994).

Here we describe the role of p40^phox in activation of the phagocyte NADPH oxidase at the cellular level. This protein enhances membrane translocation of the cytosolic activators p67^phox and p47^phox, and subsequent activation of the NADPH oxidase by 2- to 3-fold, when cells are stimulated with PMA. The enhancements are totally dependent on the binding of p40^phox to p67^phox, which is mediated via the PB1–PC interaction. Intriguingly, in cells treated with a G_i-activating peptide instead of PMA, p40^phox serves as a much more efficient activator of the oxidase, thereby underscoring its importance in a physiological context that cannot be reproduced under cell-free conditions.

Results

Expression of p40^phox leads to enhanced production of superoxide in PMA-stimulated cells

To investigate the role of p40^phox in activation of the phagocyte NADPH oxidase, we transfected the plasmid vector pREP4 encoding the full-length cDNA of p40^phox or vector alone to K562 cells that express functional cytochrome b₅₅₈, p47^phox and p67^phox (Ago et al., 1999; Koga et al., 1999), and cloned seven independent transformants, each expressing p40^phox at the same level (for details see Materials and methods). The transformants stably express p40^phox as estimated by flow cytometric analysis (Figure 2A). The p40^phox-expressing cells harbored the same amounts of functional gp91^phox as the p40^phox-deficient cells, containing the vector pREP4 that did not encode this protein (Figure 2B). In addition, analyses by immunoblot (Figure 2C) and flow cytometry (data not shown) revealed that no difference existed between these cells in expression of the proteins p22^phox, p47^phox, p67^phox and the small GTPase Rac.

Fig. 2. Stable expression of p40^phox in K562 cells transfected with pREP4 encoding the full-length of this protein. (A) Expression of p40^phox in the K562 cells. Parent K562 cells stably expressing gp91^phox, p67^phox and p47^phox were transfected with pREP4 encoding the full-length p40^phox (p40^phox) or vector alone (vector). The cells were stained with anti-p40^phox serum (filled histogram) or pre-immune serum (open histogram) and analyzed by flow cytometry. For details see Materials and methods. (B) Expression of functional gp91^phox. The cells were stained with the monoclonal antibody 7D5 to detect gp91^phox in functional cytochrome b₅₅₈ (filled histogram) or control IgG (open histogram) and analyzed by flow cytometry. (C) Expression of p47^phox, p67^phox, p40^phox and p22^phox. The lysates of the K562 cells were analyzed by immunoblot with an anti-p47^phox, an anti-p67^phox, an anti-p40^phox or an anti-p22^phox antibody, as described in Materials and methods.

When the K562 cells expressing p40^phox were stimulated with PMA, they produced a large amount of superoxide, which was estimated by superoxide dismutase (SOD)-inhibitable cytochrome c reduction (Figure 3A). We also tested four other clones (data not shown) and found that the cells exhibited essentially the same superoxide-producing activity of 4.2 ± 0.14 nmol/min per 10⁶ cells (n = 5). On the other hand, the activity of the cells without p40^phox was 1.8 ± 0.12 nmol/min per 10⁶ cells (n = 5 independent clones) (Figure 3A; data not shown). Using another method, SOD-inhibitable chemiluminescence, we also tested the effect of p40^phox in activation of the phagocyte NADPH oxidase (Figure 3B). In response to PMA, seven independent clones of the p40^phox-expressing K562 cells all produced 2- to 3-fold more superoxide than the cells lacking p40^phox (Figure 3B). Thus expression of p40^phox leads to enhanced production of superoxide in PMA-stimulated cells, raising the possibility that p40^phox facilitates the assembly of the phagocyte oxidase factors.

Fig. 3. p40^phox-enhanced superoxide production and membrane translocation of p47^phox and p67^phox upon cell stimulation with PMA. (A) Superoxide production by the K562 cells with stable expression of p40^phox or without p40^phox was measured as reduction of cytochrome c. The cells (5.0 × 10⁵ cells/ml) were stimulated with PMA (200 ng/ml) (arrow), and the superoxide-producing activity was measured by determining the rate of ferricytochrome c reduction at 550 to 540 nm. The reduction was stopped by addition of superoxide dismutase (SOD) (50 µg/ml) (arrowhead). (B) Superoxide production by the K562 cells with p40^phox or without p40^phox was measured as change of chemiluminescence. The K562 cells (5.0 × 10³ cells) were stimulated with PMA (200 ng/ml), and the chemiluminescence change was continuously monitored with DIOGENES. SOD (50 µg/ml) was added where indicated (arrowhead). For details see Materials and methods. (C) Membrane translocation of p47^phox, p67^phox, p40^phox and Rac upon PMA stimulation. The K562 cells with or without p40^phox (1.0 × 10⁷ cells) were incubated with PMA (200 ng/ml) for the indicated time. After the incubation, the cells were sonicated and ultracentrifuged as described in Materials and methods. The membrane fractions were analyzed by immunoblot with the antibody against p47^phox, p67^phox, p40^phox, Rac or p22^phox. The data are representative of results from five independent experiments.

p40^phox facilitates membrane translocation of p47^phox and p67^phox but not of Rac upon cell stimulation

It is well established that stimulus-elicited translocation of p47^phox and p67^phox from the cytosol to the membrane is required for activation of the phagocyte NADPH oxidase (Roos et al., 1996; Clark, 1999; Nauseef, 1999). Since p40^phox is constitutively associated with p67^phox and enhances the oxidase activation as shown above (Figure 3A and B), it seems likely that p40^phox affects the membrane recruitment of p47^phox and p67^phox. To test this possibility, we fractionated the membrane of the K562 cells with or without p40^phox and estimated the amounts of p47^phox and p67^phox. When the p40^phox-lacking cells were stimulated with PMA, both p47^phox and p67^phox translocated to the membrane in a time-dependent manner (Figure 3C). In the cells expressing p40^phox, this protein as well as p47^phox and p67^phox was targeted upon cell stimulation to the membrane (Figure 3C). The amounts of p47^phox and p67^phox at the membrane were increased by 2- to 3-fold, compared with those of the p40^phox-deficient cells (Figure 3C). On the other hand, p40^phox did not affect PMA-elicited membrane targeting of Rac (Figure 3C). Thus the enhancement of the oxidase activation by p40^phox is likely to be due to the p40^phox-induced facilitation of membrane translocation of p47^phox and p67^phox.

p40^phox carrying the D289A substitution is incapable of binding to p67^phox and fails to enhance superoxide production and membrane translocation of p47^phox and p67^phox

We next tested whether the interaction of p40^phox with p67^phox participates in the p40^phox-enhanced translocation of p47^phox and p67^phox. As we have previously shown, this interaction is mediated via a direct binding of the p40^phox PC motif to the p67^phox PB1 domain (Nakamura et al., 1998; Ito et al., 2001). Consistent with previous observations, the maltose-binding protein (MBP)-tagged full-length protein of p40^phox, MBP–p40^phox (WT), directly interacted with the p67^phox PB1 domain as a glutathione S-transferase (GST) fusion (GST–p67-PB1) as analyzed by a pull-down assay (Figure 4A). To confirm the role for the PC motif of p40^phox, we purified a mutant p40^phox carrying the amino acid substitution of Ala for Asp289, an invariant residue among PC motifs (Nakamura et al., 1998); the corresponding residue in the PC motif of the yeast polarity protein Cdc24p plays an essential role in its binding to the PB1 domain of Bem1p (Ito et al., 2001). As shown in Figure 4A, the D289A substitution resulted in a complete loss of the interaction of p40^phox with p67^phox.

Fig. 4. Effect of the D289A substitution in the PC motif of p40^phox on its interaction with p67^phox both in vivo and in vitro. (A) Direct interaction of p40^phox with p67^phox. GST-tagged PB1 domain of p67^phox (GST–p67-PB1) or GST alone was incubated with MBP–p40^phox (WT) or MBP–p40^phox (D289A) and proteins were pulled down with glutathione–Sepharose-4B. The precipitated proteins were subjected to SDS–PAGE, followed by staining with CBB, as described, see Materials and methods. Positions for marker proteins are indicated in kDa. (B) Expression of p40^phox (D289A) in K562 cells. K562 cells stably expressing gp91^phox, p67^phox and p47^phox were transfected with pREP4 encoding p40^phox with the D289A substitution, p40^phox (D289A). The K562 cells were stained with anti-p40^phox serum (filled histogram) or pre-immune serum (open histogram) and analyzed by flow cytometry. (C) Interaction of p40^phox with p67^phox in the K562 cells. The cell lysates of the K562 cells were analyzed by immunoprecipitation with an anti-p67^phox or control IgG (cont.) (left panel), or an anti-p47^phox antibody or control IgG (cont.) (right panel) followed by immunoblot with the indicated antibody, as described under Materials and methods.

To examine the effect of the D289A substitution in the PC motif at the cellular level, we prepared seven clones of the K562 cells stably expressing p40^phox (D289A). The amount of p40^phox as well as those of other oxidase proteins including gp91^phox, p22 ^phox, p47^phox and p67^phox in the p40^phox (D289A)-expressing cells was identical to that in the cells expressing the wild-type p40^phox (WT) (Figure 4B and C; data not shown). An immunoprecipitation assay demonstrated that p67^phox binds to p40^phox (WT) but not to p40^phox (D289A) in vivo (Figure 4C), indicating that the binding of p40^phox to p67^phox is mediated solely via the PC motif even at the cellular level. Although p47^phox was involved in a complex with p40^phox and p67^phox (Figure 4C), it should be noted that the interaction between p47^phox and p67^phox occurs in a manner independent of the presence of p40^phox (Figure 4C).

When cells were stimulated with PMA, p40^phox (D289A) was incapable of enhancing the activation of the NADPH oxidase: the amount of superoxide produced by the cells expressing p40^phox (D289A) was similar to that by the cells without p40^phox (Figure 5A). The amounts of p67^phox and p47^phox in the membrane of the p40^phox (D289A)-expressing cells were half-to-one-third of those of the p40^phox (WT)-expressing ones. In addition, p40^phox (D289A) did not translocate to the membrane in response to PMA (Figure 5B), suggesting that its localization is totally dependent on the interaction with p67^phox. The present observations thus indicate that membrane translocation of p47^phox and p67^phox is facilitated by the binding of p40^phox to p67^phox, which leads to enhanced activation of the NADPH oxidase.

Fig. 5. Effect of the D289A substitution in the PC motif of p40^phox on superoxide production and membrane translocation of p47^phox and p67^phox upon PMA stimulation. (A) Superoxide production by the K562 cells with p40^phox (WT) or p40^phox (D289A) was measured as change of chemiluminescence. The K562 cells (5.0 × 10³ cells) were stimulated with PMA (200 ng/ml) and the chemiluminescence change was measured. Each histogram indicates the average from five independent experiments, with bars representing SD. (B) Membrane translocation of p47^phox, p67^phox and p40^phox upon PMA stimulation. The K562 cells stably expressing p40^phox (WT) or p40^phox (D289A) (1.0 × 10⁷ cells) were stimulated with PMA (200 ng/ml) for the indicated time and the amounts of p47^phox, p67^phox and p40^phox in the membrane fractions were analyzed by immunoblot as described under Materials and methods.

p67^phox carrying the K355A substitution is incapable of interacting with p40^phox, and fails to enhance superoxide production and membrane translocation of p47^phox and p67^phox

To further study the importance of the interaction between p40^phox and p67^phox, we used a mutant p67^phox with the substitution of Ala for Lys355 (Figure 6A), a conserved residue among PB1 domains (Ito et al., 2001; Terasawa et al., 2001). In an in vitro pull-down assay, GST–p67-PB1 (K355A) was incapable of interacting with MBP–p40^phox, while the wild-type PB1 domain fully bound to p40^phox (data not shown). Similarly, p67^phox (K355A) could not associate with p40^phox in vivo (Figure 6B). On the other hand, both p67^phox (WT) and p67^phox (K355A) were co-immunoprecipitated with p47^phox (Figure 6B), indicating that p40^phox binds to p47^phox indirectly but via p67^phox (see Figure 1).

Fig. 6. Effect of the K355A substitution in the PB1 domain of p67^phox on its interaction with p40^phox. (A) Expression of p67^phox in K562 cells. K562 cells stably expressing gp91^phox, p47^phox and p40^phox were transfected with pREP4 encoding p67^phox (WT) or p67^phox (K355A). The K562 cells were stained with anti-p67^phox serum (filled histogram) or pre-immune serum (open histogram), and analyzed by flow cytometry. (B) In vivo interaction of p67^phox with p40^phox in the K562 cells. The cell lysates of the K562 cells were analyzed by immunoprecipitation with the anti-p67^phox or control IgG (cont.) (left panel), or the anti-p47^phox antibody or control IgG (cont.) (right panel) followed by immunoblot (Blot) with the indicated antibody.

We next investigated the effect of the K355A substitution in the p67^phox PB1 domain on superoxide production and membrane translocation of p47^phox and p67^phox. As shown in Figure 7A, the amounts of p47^phox and p67^phox in the membranes of the p67^phox (K355A)-expressing cells were half-to-one-third of those of the p67^phox (WT)-expressing cells. In the cells expressing p67^phox (K355A), p40^phox failed to translocate to the membranes. As shown in Figure 7B, the amount of superoxide produced by the cells expressing p67^phox (K355A) was half-to-one-third of that by those expressing p67^phox (WT). In response to PMA, the cells with p67^phox (K355A) produced the same low amount of superoxide irrespective of the presence of p40^phox (Figure 7B). These observations further supported the idea that the p40^phox–p67^phox interaction enhances the membrane translocation of p47^phox and p67^phox, thereby facilitating the assembly of the active oxidase complex.

Fig. 7. Effect of the K355A substitution in the PB1 domain on membrane translocation of p47^phox and p67^phox and superoxide production upon PMA stimulation. (A) Membrane translocation of p47^phox, p67^phox and p40^phox upon PMA stimulation. The K562 cells stably expressing p67^phox (WT) or p67^phox (K355A) (1.0 × 10⁷ cells) were stimulated with PMA (200 ng/ml) for the indicated time and the amounts of p47^phox, p67^phox and p40^phox in the membrane fractions were analyzed by immunoblot. (B) Superoxide production by the K562 cells with p67^phox (WT) or p67^phox (K355A) was measured as change of chemiluminescence. The K562 cells (5.0 × 10³ cells) were stimulated with PMA (200 ng/ml) and the chemiluminescence change was measured. The chemiluminescence change by K562 cells without the expression of p40^phox and with p67^phox (WT) or p67^phox (K355A) was also measured. Each histogram indicates the average from five independent experiments, with bars representing SD.

p40^phox greatly enhances superoxide production in cells stimulated with the muscarinic receptor peptide m4I3C(14), an agent that can activate the trimeric GTPase G_i

It is well documented that, in neutrophils, chemoattractants such as N-formyl peptides induce superoxide production via the trimeric GTPase G_i (Bokoch, 1995). The chemotactic receptor, however, is not expressed in K562 cells. To further clarify the role of p40^phox in the oxidase activation, we used the muscarinic receptor peptide m4I3C(14) as a stimulant.

In response to m4I3C(14), K562 cells expressing p40^phox produced superoxide (Figure 8A). The superoxide production, albeit 10-fold less than that by PMA-stimulated cells, is likely to be catalyzed by the phagocyte NADPH oxidase, since the peptide did not act on parental K562 cells that lack oxidase proteins (data not shown) or cells expressing all oxidase factors except p67^phox (see later). Furthermore, we did detect, although qualitatively, the peptide-induced membrane translocation of the cytosolic oxidase proteins (data not shown). The treatment of the K562 cells with pertussis toxin (PTX) resulted in a loss of the m4I3C(14)-induced superoxide production, while the toxin did not affect that induced by PMA (Figure 8B). The peptide thus activates the oxidase via G_i.

Fig. 8. p40^phox-enhanced superoxide production upon cell stimulation with the muscarinic receptor peptide m4I3C(14). (A) Superoxide production by the K562 cells with stable expression of p40^phox or without p40^phox. The K562 cells (5.0 × 10³ cells) were stimulated with the muscarinic receptor peptide m4I3C(14) (m) (200 µM) and the chemiluminescence change was continuously monitored with DIOGENES, and SOD (50 µg/ml) was added where indicated (arrowhead). For details see Materials and methods. (B) Effect of pertussis toxin (PTX) on superoxide production upon stimulation with m4I3C(14) or PMA. The K562 cells (5.0 × 10³ cells) were pretreated with PTX (18 µg/ml) and stimulated with m4I3C(14) (200 µM) or PMA (200 ng/ml) and superoxide production was measured as change of chemiluminescence. Each histogram indicates the average from five independent experiments, with bars representing SD. (C) Effect of the D289A substitution in the PC motif of p40^phox on superoxide production upon stimulation with m4I3C(14). The K562 cells with p40^phox (WT) or p40^phox (D289A) or without p40^phox (5.0 × 10³ cells) were stimulated with 200 µM m4I3C(14) and superoxide production was measured as change of chemiluminescence. Each histogram represents the average from five independent experiments, with bars representing SD. (D) Effect of the K355A substitution in the PB1 domain of p67^phox on superoxide production upon stimulation with m4I3C(14). The K562 cells with p67^phox (WT) or p67^phox (K355A) or without p67^phox (5.0 × 10³ cells) were stimulated with 200 µM m4I3C(14), and superoxide production was measured as change of chemiluminescence. Each histogram indicates the average from five independent experiments, with bars representing SD.

Intriguingly, in response to m4I3C(14), K562 cells expressing p40^phox produced superoxide ∼20-fold as much compared with the cells lacking this adaptor protein (Figure 8C). In addition, the peptide served as a weak agonist when cells expressed p40^phox (D289A) but not the wild-type p40^phox (Figure 8C). Consistent with this, only a small amount of superoxide was generated in cells expressing p67^phox (K355A) instead of the wild-type p67^phox, while no superoxide production was detected in cells lacking p67^phox (Figure 8D). Thus, in cells stimulated with the G_i activator m4I3C(14), p40^phox greatly enhances activation of the phagocyte NADPH oxidase, which is dependent on its interaction with p67^phox.

Discussion

Although p40^phox is known as a protein that constitutively associates with the phagocyte oxidase activator p67^phox, the role of p40^phox in the oxidase activation has remained elusive. In the present study, we show that p40^phox facilitates the activation of the phagocyte NADPH oxidase at the cellular level (Figures 3 and 8). The facilitation is thought to be caused by enhancing the stimulus-induced recruitment of p67^phox and p47^phox to the membrane, an essential step for the oxidase activation, without affecting the membrane translocation of the small GTPase Rac (Figure 3). We also demonstrate that both enhanced recruitment of the oxidase activators and facilitated activation of the oxidase are totally dependent on the binding of p40^phox to p67^phox, which is mediated via a novel modular interaction between the p40^phox PC motif and the p67^phox PB1 domain (Figures 5, 7 and 8).

One of the reasons why the role of p40^phox has been obscure for a long time is due to the fact that p40^phox is dispensable for NADPH oxidase activation under both cell-free and whole-cell conditions (de Mendez et al., 1994; Sathyamoorthy et al., 1997). In addition, it has been reported that the presence of p40^phox marginally affects the oxidase activity in a cell-free activation system: this protein only slightly enhances the activation (Cross, 2000). Although studies under cell-free conditions have shed light on the understanding of the molecular mechanism for the oxidase activation, results obtained in such studies sometimes seem discrepant with those observed in a whole-cell system. For instance, p67^phox binds to p47^phox via the C-terminal SH3 domain of p67^phox, a domain which is required for the oxidase activation in a whole-cell system (de Mendez et al., 1994) but not for the cell-free activation (de Mendez et al., 1994; Leusen et al., 1995; Hata et al., 1998). Here we show, using cells stably expressing p40^phox, that p40^phox enhances activation of the NADPH oxidase. When cells are stimulated with PMA, a potent activator of protein kinase C, expression of p40^phox leads to a 2- to 3-fold enhancement (Figure 3). On the other hand, the oxidase activation is much more drastically (∼20-fold) facilitated by p40^phox in response to the muscarinic receptor peptide m4I3C(14), that acts as an activator of G_i (Figure 8). The peptide probably functions as a more physiological stimulant, since PTX blocks the peptide-induced superoxide production but not the PMA-elicited one (Figure 8). It seems probable that, in in vivo activation of the oxidase, p40^phox plays a role that is much more important than expected from the results of experiments with PMA. Alternatively, the significance of p40^phox could be dependent on the types of stimulants or signaling pathways for the oxidase activation. In either case, p40^phox is deeply involved in activation of the phagocyte NADPH oxidase.

Although no case has been thus far reported of CGD with the defect of p40^phox, it is conceivable that its defect causes less severe forms of compromised host defense and hence has escaped the clinical screening for CGD. Alternatively, since the phagocyte oxidase is known to participate not only in host defense but also inflammation, an aberrance in inflammatory response, rather than vulnerability to infection, may be manifested in the cases with defect of p40^phox. It is thus intriguing to search single nucleotide polymorphisms or mutations in the p40^phox gene NCF4 showing significant association with those displaying these symptoms. However, one should note that extra-phagocytic roles for p40^phox and hence its involvement in other diseases are plausible, because mouse Ncf4 gene is also expressed in T cells and neurons, with no expression of p67^phox (Mizuki et al., 1998). In any case, pathophysiological roles for p40^phox remain largely elusive.

In contrast with the present results, Sathyamoorthy et al. (1997) have reported that transient expression of p40^phox leads to a repressed activation of the NADPH oxidase when cells are stimulated with PMA. The reason for this discrepancy is presently unknown. They have also shown that expression of the SH3 domain of p40^phox by itself represses the oxidase activation more efficiently than the full-length p40^phox does (Sathyamoorthy et al., 1997). Since the SH3 domain of p40^phox is capable of binding to p47^phox but to a much lesser extent than the C-terminal SH3 domain of p67^phox (Ito et al., 1996), overexpression of the p40^phox SH3 domain may compete with the SH3 domain of p67^phox to prevent p67^phox from interacting with p47^phox, thereby inhibiting the oxidase activation.

The adaptor protein p40^phox is present in a complex containing not only p67^phox but also p47^phox. In this complex, p40^phox directly binds to p67^phox via the PB1–PC interaction, while p40^phox interacts with p47^phox in an indirect manner: p67^phox tethers p40^phox to p47^phox (Figure 1). If the association between p40^phox and p67^phox is absent, p40^phox and p47^phox cannot reside in the same complex (Figures 4 and 6). Thus membrane translocation of p47^phox is not enhanced by p40^phox under conditions where the association between p40^phox and p67^phox is specifically disrupted (Figures 5 and 7). On the other hand, the membrane translocation of p40^phox requires p67^phox. When a mutant p67^phox (K355A), that cannot interact with p40^phox, is expressed instead of the wild-type p67^phox, p40^phox is incapable of translocating to the membrane upon cell stimulation (Figure 7). This is consistent with a previous observation that, in neutrophils of a p67^phox-deficient CGD patient, p40^phox is not targeted to the membrane in response to cell stimuli (Dusi et al., 1996). Even in these cells, stimulus-induced translocation of p47^phox can still be observed (Dusi et al., 1996). Taken together, p47^phox plays an essential role in membrane recruitment of the p67^phox–p47^phox–p40^phox complex, an event which is strongly enhanced by p40^phox.

This study clearly demonstrates that p40^phox facilitates membrane targeting of the essential oxidase factors p67^phox and p47^phox, which requires the interaction of p40^phox with p67^phox. The molecular mechanism whereby p40^phox functions, however, is not fully understood at present. In the N-terminal region, p40^phox harbors the PX/PB2 domain (Figure 1), a module that interacts with phosphoinositides (Ago et al., 2001; Kanai et al., 2001; Ellson et al., 2001b). The PX domain of p40^phox specifically binds to phosphatidylinositol-3-phosphate [PtdIns(3)P] in vitro, and the domain expressed as a fusion to green fluorescent protein localizes to early endosomes where this phosphoinositide is enriched (Ago et al., 2001; Kanai et al., 2001; Ellson et al., 2001b). It is possible that the phospholipid-binding activity of p40^phox is involved in its membrane localization. In this context, it is intriguing to note a current report showing that PtdIns(3)P is enriched in phagosomes (Ellson et al., 2001a), where the NADPH oxidase is activated. It has been also reported that p40^phox can interact with coronin, an actin-binding protein (Grogan et al., 1997). Interaction of p40^phox with cytoskeletal elements may participate in a process of membrane translocation of this protein. In addition, stimulus-induced phosphorylation of p40^phox (Bouin et al., 1998) possibly plays a role, an event which has escaped analysis in a cell-free system. These possibilities should be tested in future studies.

The PB1 domain and PC motif prevail in a variety of proteins and mediate protein–protein interactions: the PB1 domain of Bem1p, scd2, atypical protein kinase Cζ or p67^phox recognizes and binds to the PC motif of Cdc24p, scd1, ZIP or p40^phox, respectively (Ito et al., 2001). In the budding yeast Saccharomyces cerevisiae, the interaction between Bem1p and Cdc24p is essential for polarity establishment of cells (Ito et al., 2001; Butty et al., 2002). The present study provides another example of functional interactions mediated via these modules, i.e. the binding of p40^phox to p67^phox plays a significant role in activation of the phagocyte NADPH oxidase, an enzyme involved in host defense. Further studies are required to clarify the importance of other PB1–PC interactions.

Materials and methods

Plasmid constructions

cDNA fragments encoding p40^phox, p47^phox and p67^phox were prepared as described previously (Hata et al., 1998; Nakamura et al., 1998; Ago et al., 1999) and cloned into the vectors pREP (Invitrogen), pGEX-4T1 (Amersham Pharmacia Biotech) and pMALc2 (New England Biolabs). Mutations leading to the indicated amino acid substitutions were introduced by PCR-mediated site-directed mutagenesis. All the constructs were sequenced to confirm their identities.

Expression of p40^phox in K562 cells containing cytochrome b₅₅₈, p47^phox and p67^phox

The doubly transduced K562 cells that stably express both cytochrome b₅₅₈ and p67^phox were prepared using the monoclonal antibody 7D5 to detect gp91^phox in functional cytochrome b₅₅₈ (Yamauchi et al., 2001) and anti-p67^phox polyclonal antibodies raised to a C-terminal peptide of this protein (Imajoh-Ohmi et al., 1992), as previously described (Ago et al., 1999). The cells were electroporated in the presence of the full-length cDNA of p47^phox in pREP9. After incubation with G418 (1.0 mg/ml), the cells were cloned by limiting dilution and checked by flow cytometry with an anti-p47^phox monoclonal antibody (Transduction Laboratories). The stable transformants were further transfected with pREP4 encoding the wild-type p40^phox or a mutant one carrying the D289A substitution. The transfected cells were cloned by limiting dilution in the presence of hygromycin B (100 µg/ml) and expression of p40^phox was confirmed by flow cytometry with anti-p40^phox polyclonal antibodies raised to an N-terminal peptide of this protein (Tsunawaki et al., 1994).

Expression of p67^phox in K562 cells containing cytochrome b₅₅₈, p47^phox and p40^phox

The doubly transduced K562 cells that stably express both functional cytochrome b₅₅₈ and p47^phox were prepared as previously described (Koga et al., 1999). The cells were electroporated in the presence of pREP9 (Invitrogen) encoding the full-length cDNA of p40^phox or vector alone. After incubation with G418 (1.0 mg/ml), the K562 cells were cloned by limiting dilution. The stable transformants were further transfected with pREP10 encoding the wild-type p67^phox or a mutant one carrying the K355A substitution. The transfected cells were cloned by limiting dilution in the presence of hygromycin B (100 µg/ml), and expression of p67^phox was confirmed by flow cytometry with the anti-p67^phox antibodies.

Flow cytometric analysis

Cells were washed with phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na₂HPO₄ and 1.5 mM KH₂PO₄ pH 7.4), and fixed with 2.0% paraformaldehyde for 1 h at 25°C. For detection of cytosolic proteins, the fixed cells were permeabilized with PBS including 0.1% Triton X-100 and 2.0% bovine serum albumin (BSA). The cells were incubated with the indicated antibody, washed twice with 2.0% BSA in PBS, and indirectly labeled with fluorescein isothiocyanate-conjugated goat-anti mouse or rabbit IgG (BioSource International). Cells were washed twice with PBS and analyzed by FACS scan.

Activation of the phagocyte NADPH oxidase

Superoxide production was determined as SOD-inhibitable chemiluminescence, as previously described (Ago et al., 1999; Koga et al., 1999). Cells in HEPES-buffered saline (120 mM NaCl, 5 mM KCl, 5 mM glucose, 1 mM MgCl₂, 0.5 mM CaCl₂ and 17 mM HEPES pH 7.4) were stimulated at 37°C with PMA (200 ng/ml), and the reaction was terminated by the addition of SOD (50 µg/ml). The chemiluminescence was counted with an enhancer-containing luminol-based detection system (DIOGENES; National Diagnostics) using a luminometer (Auto Lumat LB953; EG&G Berthold).

Alternatively, the superoxide-producing activity was measured by determining the rate of SOD-inhibitable cytochrome c reduction at 550–540 nm using a dual-wavelength spectrophotometer (Hitachi 557) (Sumimoto et al., 1994).

Translocation of p47^phox, p67^phox and p40^phox to the membrane in stimulated K562 cells

Membrane translocation of cytosolic oxidase factors were determined by the method of Leusen et al. (1994) with minor modifications. Briefly, cells were stimulated for the indicated time at 37°C with PMA (200 ng/ml). The incubation was terminated by the addition of ice cold PBS. The cells were resuspended in 3 ml of ice cold buffer A (75 mM NaCl, 170 mM sucrose, 1 mM MgCl₂, 0.5 mM EGTA, 10 µM ATP, 2 mM NaN₃, 5 µM GTPγS, 100 µg/ml of p-amidinophenyl methanesulfonyl fluoride hydrochloride and 20 mM HEPES pH 7.0) and lyzed by sonication. The sonicate was layered on a discontinuous sucrose gradient consisting of 4 ml of 40% (w/v) sucrose and 3 ml of 15% (w/v) sucrose with 1 mM MgCl₂, 40 mM NaCl, 0.5 mM EGTA and 5 µM 5′-3-O- (thio)-triphosphate (GTPγS). After ultracentrifugation for 45 min at 100 000 g, membrane-associated proteins (1.0 µg) were analyzed by immunoblot with the indicated antibody, which were developed using ECL-plus (Amersham Pharmacia Biotech).

An in vitro binding assay using purified recombinant proteins

The PB1 domain of p67^phox (amino acids 335–427) and the protein with the K355A substitution were expressed as GST fusion proteins in Escherichia coli and purified by glutathione–Sepharose-4B (Amersham Pharmacia Biotech) (Ago et al., 1999; Koga et al., 1999). MBP-tagged full-length p40^phox and the protein with the D289A substitution were also expressed and purified by amylose resin (New England BioLab). For in vitro pull-down binding assays, a pair of a GST fusion (250 µg) and an MBP-tagged protein (500 µg) were incubated for 30 min in 1 ml PBS containing 0.5% Triton X-100. Proteins were precipitated with glutathione–Sepharose-4B and eluted with 5 mM glutathione. The eluates were subjected to SDS–PAGE and stained with Coomassie Brilliant Blue (CBB).

An in vivo binding assay

Cells were lyzed with 1 ml of a lysis buffer (1% Triton X-100, 100 mM NaCl, 5 mM MgCl₂, 1 mM EDTA and 40 mM HEPES pH 7.4). The lysate was precipitated with the anti-p47^phox or anti-p67^phox antibody in the presence of protein G–Sepharose (Amersham Pharmacia Biotech). After washing with the lysis buffer, precipitated proteins were analyzed by immunoblotting.

Preparation of the muscarinic receptor peptide m4I3C(14)

m4I3C(14) is a 14-residue peptide fragment corresponding to the C-terminal portion of the third intracellular loop and the N-terminal portion of the sixth transmembrane helix of human m4 muscarinic acetylcholine receptor (Bonner et al., 1987) (residues 393–406: RERKVTRTIFAILL). This peptide, like mastoparan (Ross and Higashijima, 1994), activates trimeric G_i reconstituted in phospholipid vesicles in a similar manner as the parental m4 muscarinic acetylcholine receptor (K.Wakamatsu, in preparation) yet does not show any cell-damaging activity as observed for mastoparan (Nakajima et al., 2000). The peptide was synthesized by a standard fluoren-9-ylmethoxycarbonyl-based solid-phase method. The purity and the identity of the synthesized peptide were confirmed by analytical HPLC and time-of-flight mass spectroscopy.