Phosphorylation of threonine 154 in p40^phox is an important physiological signal for activation of the neutrophil NADPH oxidase

The neutrophil nicotinamide adenine dinucleotide phosphate [NADPH] oxidase plays an important role in the mechanisms that neutrophils use to kill pathogens.^1,2  In the resting state of the neutrophil, the components of the NADPH oxidase are segregated between membrane-bound cytochrome b₅₅₈, comprising gp91^phox and p22^phox, and 4 cytosolic components, p47^phox, p67^phox, p40^phox, and guanosine diphosphate (GDP)-bound Rac1/2.^3-6  Upon neutrophil stimulation by chemoattractants, cytokines, or pathogens, the cytosolic components translocate to the membrane and interact with the cytochrome, with each other, and with phospholipids, resulting in formation of an active NADPH oxidase complex.^7-10  This active complex catalyses the transfer of electrons across the membrane, from NADPH to O₂, to generate the superoxide anion, which can then be converted to other reactive oxygen species (ROS) that together form one of the key weapons that are used to kill pathogens. The importance of the NADPH oxidase is highlighted by chronic granulomatous disease (CGD), which is characterized by genetic deficiency in one of the oxidase components, leading to recurrent bacterial and fungal infections.¹¹ 

The key steps underlying oxidase activation include the guanosine-5′-triphosphate loading of Rac, allowing its translocation to the membrane and its interaction with the tetratricopeptide repeat domain of p67^phox; the interaction of p67^phox with gp91^phox; and the phosphorylation of p47^phox.^8,12,13 Phosphorylation of p47^phox results in the relief of an auto-inhibitory conformation and exposure of the tandem SH3 domain to allow its interaction with p22^phox. The C-terminus of p47^phox also binds to the C-terminal SH3 domain of p67^phox, and thus p47^phox phosphorylation is thought to play a major role in orchestrating assembly and subsequent activation of the oxidase. At the membrane, the phox homology (PX) domain of p47^phox interacts with phosphatidylinositol 3,4-bisphosphate (PtdIns(3,4)P₂) and phosphatidic acid,^14,15  which may also contribute to efficient complex assembly and activation,¹⁶  though the physiological importance of these interactions have yet to be defined.

The p40^phox subunit is thought to be an obligate binding partner for p67^phox, but its role in oxidase activation has been controversial.^17,18  However, the recent development of mouse models that either lack p40^phox (p40^phox−/−)¹⁹  or carry a point mutation in the p40^phox PX domain (p40^phoxR58A/−),²⁰  together with the recent discovery of a CGD patient expressing a PX domain mutant of p40^phox (p40^{phoxR105Q/−}),²¹  has now firmly established the importance of this subunit in oxidase activation by physiological stimuli. Further, this work has shown that PtdIns3P binding to the PX domain of p40^phox plays an important role in oxidase activation to several agonists, in particular during the assembly of the oxidase on phagosomes,^20,21  though the molecular mechanism through which this interaction regulates oxidase activity is still uncertain. Moreover, oxidase defects in p40^phoxR58A/−neutrophils are not as severe as those in p40^phox−/− neutrophils, in particular the extracellular ROS responses to the formylated bacterial peptide (fMLP) are unaffected in the former but severely reduced in the latter.¹⁹  Unfortunately, loss of p40^phox also reduces expression of its binding partner, the p67^phox subunit (by approx 55%), and, therefore, since p67^phox is an essential component of the oxidase, it is difficult to assess the extent to which the p40^phox−/− phenotype is caused by non-PX domain functions of p40^phox.¹⁹ 

p40^phox also contains a conserved SH3 domain and 2 conserved phosphorylation sites, threonine (T) 154 and serine (S) 315.²²  Good evidence has been provided that phosphorylation of p40^phox on both of these sites occurs during stimulation of neutrophils or neutrophil-like cells by phorbol 12-myristate 13-acetate (PMA) or fMLP.^22-24  Further, both of these sites can be phosphorylated in vitro by protein kinase C (PKC), an enzyme known to be important in the phosphorylation of p47^phox and in the activation of the oxidase.²²  In a cell-free system, using recombinant, phosphorylated p47^phox as the activator of the NADPH oxidase, p40^phox phosphorylated on T154 was shown to inhibit oxidase activity,²⁵  but the in vivo functions of either T154 or S315 phosphorylation remain to be defined.

The physiological function of the SH3 domain of p40^phox is also unclear. A structure has been resolved for the SH3 domain of p40^phox in complex with the polyproline (PP) motif of p47^phox.²⁶  Among the key residues in the SH3 domain of p40^phox that interact with the PP motif of p47^phox is tryptophan (W) 207, a residue that is evolutionarily highly conserved and that interacts with the backbone of arginine (R) 368 in p47^phox through hydrophobic interactions. p40^phox-W207 to R mutation, which is predicted to disrupt the interaction with its PP binding target in vivo, fails to reconstitute full NADPH oxidase activity in permeabilized “core” neutrophils in response to PMA, as well as in the COS^phox system in response to immunoglobulin-opsonized sheep red blood cells (IgG-SRBCs).^27,28  Further, using a cell-free system for oxidase activation, the SH3 domain of p40^phox was shown to interact with the PP motif of p22^phox and through this interaction was able to substitute for p47^phox.²⁹  The W207R mutation was also found to disrupt this interaction. However, the in vivo binding target(s) of the SH3 domain of p40^phox, as well as its true physiological function, still remain to be established.

A major obstacle to fully understanding p40^phox function and indeed, more generally, the signaling mechanisms that underpin NADPH oxidase activation is the limited half-life of neutrophils ex vivo and the difficulty in introducing defined proteins in a robust and controlled manner. To try to overcome this obstacle, we set up a system for retroviral transduction of bone marrow progenitors and reconstitution of the bone marrow in irradiated recipient mice to introduce wild-type (wt) or mutated versions of p40^phox in neutrophils on a p40^phox−/− genetic background. We then performed functional studies to elucidate the role of the SH3 domain and phosphorylation sites of p40^phox in NADPH oxidase activation in fully differentiated, mouse bone marrow–derived neutrophils.

fMLP, PMA, luminol, and horseradish peroxidase (HRP) were from Sigma-Aldrich. Murine tumor necrosis factor (TNF)α was from R&D Systems. All other cytokines were from Peprotech. Dulbecco's phosphate-buffered saline (PBS) with Ca²⁺ and Mg²⁺ was from Sigma-Aldrich. Tissue culture reagents were from Invitrogen. All buffer components were from Sigma-Aldrich and were endotoxin-free or low-endotoxin, as available.

For mouse bone marrow–reconstitution experiments, donor mice were p40^phox−/− mice¹⁹  or PKCδ^−/− mice.³⁰  Recipient mice were Rag2⁰IL2rg⁰ mice.³¹  The p40^{phoxR58A/R58A} mice were described previously.³²  All animals were housed in the small animal barrier unit at the Babraham Institute under specific pathogen-free conditions or in individually ventilated cages in the small animal unit. This work was approved by Home Office Project License PPL 80/1875 and PPL 80/2335. All procedures were carried out under Home Office Personal Individual License PIL 80/10 089.

Plat-E cells³³  were cultured in Dulbecco Modified Eagle Medium containing 4500 mg/L glucose, l-glutamine, and pyruvate and supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. Fetal liver (FL) cells were cultured in StemPro-34 medium supplemented with the provided nutrient supplement, 1% Glutamax, and 1% penicillin/streptomycin/amphotericin B (StemPro-34 complete medium). In addition, a cocktail of cytokines was used to support hematopoietic stem and progenitor cell proliferation: 100 ng/mL mSCF, 100 ng/mL thrombopoietin, 10 ng/mL human interleukin-6, 6 ng/mL murine interleukin-3, and 20 ng/mL Flt3-ligand (Martin Turner, Babraham Institute, personal communication).

The bicistronic retroviral transfer vector pMIG R1-internal ribosome entry site-green fluorescent protein (IRES-GFP) has been described previously.³⁴  A full-length murine p40^phox IMAGE cDNA clone was obtained from RZPD Deutsches Ressourcenzentrum für Genomforschung GmbH and sequence verified. To obtain pMIG R1-p40^phox-IRES-GFP, polymerase chain reaction (PCR) fragments of p40^phox containing engineered restriction sites BglII and EcoRI were digested and cloned into pMIG R1-IRES-GFP. To create pMIG R1-p40^phox ΔSH3-IRES-GFP, PCR fragments were generated from p40^phox 5′ and 3′ of the SH3 domain. PCR fragments were digested with BglII and SnaBI (5′ fragment) or SnaBI and EcoRI (3′ fragment) and cloned into pMIG R1-IRES-GFP. All point mutants of p40^phox that were used in this study were created by site-directed mutagenesis (Stratagene), according to the manufacturer's instructions. Each mutant was verified by sequencing of the p40^phox coding sequence and recloned into their original vectors.

Retroviral vectors were produced by transient transfection of the packaging cell line Plat-E, using Lipofectamine 2000 (Invitrogen), with minor modifications. Briefly, transfer vector DNA and Lipofectamine were combined in Optimem medium in a ratio of 1:3 (wt/vol) for 20 minutes at room temperature (RT). Optimem/DNA/Lipofectamine were then added to cells suspended in Plat-E medium lacking antibiotics. Per 1.2 × 10⁷ cells, 16.8 μg of DNA and 50.4 μL Lipofectamine was used. Twenty-four hours after transfection, medium containing DNA-Lipofectamine 2000 complexes was removed and replaced by StemPro-34 complete medium. Forty-eight hours after transfection, the virus-containing supernatant was harvested and centrifuged at 800 × g for 5 minutes at 4°C to pellet cell debris. The supernatant was filtered through a 0.45-μm polyamide filter (Sartorius) and directly used for transduction of FL cells.

p40^phox−/− FL cells were harvested from E14.5 embryos, cultured for 24 hours, and transduced on non-tissue culture–treated plates coated with 12.5 ng/mL Retronectin (Takara Bio Inc.) at a concentration of 5 × 10⁵ cells/mL. Seventeen hours after transduction, cells were washed in Hank's Buffered Salt Solution, resuspended in PBS/10% FBS, and injected into 8-week-old lethally irradiated recipient mice (1-2 × 10⁶ FL cells/mouse). Recipients were irradiated the day before injection (2 × 5 Gy from a ¹³⁷Cs source, separated by 3 hours). wt or PKCδ^−/− bone marrow was isolated under sterile conditions and suspended in PBS/10% FBS. To increase the number of experimental animals, 5 × 10⁶ bone marrow cells were injected per lethally irradiated recipient mouse. To monitor the efficiency of reconstitution of the bone marrow compartment by transduced FL cells, samples of peripheral blood were stained with neutrophil indicator Gr1-phycoerythrin (BD Biosciences) and subjected to flow cytometric analysis of GFP marker gene expression.

Mouse neutrophils (bone marrow–derived neutrophils [BMNs]) were isolated from bone marrow as previously described.³²  Purity of neutrophils was determined by flow cytometric analysis of their distinct forward scatter/side scatter properties and was generally between 70% and 90%. Cells were resuspended at a concentration of 6.25 × 10⁶/mL in Dulbecco's PBS with Ca²⁺ and Mg²⁺, 1 g/L glucose, and 4 mM sodium bicarbonate (D-PBS⁺) and primed with 1000 U/mL mouse TNFα and 100 ng/mL granulocyte-macrophage colony-stimulating factor for 1 hour at 37°C with occasional gentle mixing.

Staphylococcus aureus (S aureus) bacteria from the Wood 46 strain were cultured and opsonized as described.³²  A quantity of 10 μL of SRBCs in Alsevers (TCS Biosciences) were washed (4 minutes, 4000 × g) and resuspended in 1 mL of D-PBS⁺ containing 0.1% fatty acid- and endotoxin-free bovine serum albumin (D-PBS⁺⁺). SRBCs were opsonized by using a subagglutinating concentration of rabbit anti-SRBC IgG while rotating end-over-end for 20 minutes at RT. SRBCs were washed twice and resuspended in 800 μL of D-PBS⁺⁺.

Where indicated, primed BMNs (typically isolated from a single reconstituted mouse) were pre-incubated with vehicle control (0.1% dimethyl sulfoxide), selective PKC inhibitors bisindolylmaleimide I (BIM-1), Gö 6976, or Gö 6983 (1μM; Calbiochem), or PI3K inhibitor wortmannin (100 nM or 300 nM) for 10 minutes before stimulation. Rate kinetics of ROS production were measured using a luminol-based assay in polystyrene 96-well plates (Berthold Technologies Ltd) as described previously.³²  Assays were conducted in the absence or presence of exogenously added HRP (18.75 U/mL) to measure intracellular or total ROS, respectively. fMLP was used at 10μM, PMA at 100 nM, and particles were used at a BMN/particle ratio of 1:20. Light emission was recorded by a Berthold Microlumat Plus luminometer (Berthold Technologies). Data output is relative light units per second or total relative light units integrated over the indicated periods of time. To compare the average responses of wt and mutated p40^phox across different agonists, a logarithmic transformation of the raw data were performed before analysis to stabilize the errors, followed by a repeated measures analysis of variance (ANOVA), taking into account the variability between the responses to different agonists.

S aureus phagocytosis assays were performed as described previously.³²  Fixed neutrophils were stained for phox components p40^phox (Millipore) or p47^phox (Millipore; 1 hour, RT). After washing in PBS/0.5% bovine serum albumin, cells were stained with AlexaFluor405- or AlexaFluor568-conjugated goat anti-rabbit antibodies. Coverslips were mounted on glass microslides with Aqua-Polymount antifading solution (Poly-Science Inc) and fluorescence visualized using a Zeiss LSM 510 META point-scanning confocal microscope with an integrated camera, using a Plan/Apochromat 63×/1.4 oil objective. Cytosolic phox levels and phagosomal accumulation were quantified using LSM 510 software. Given the variability in the data obtained, heterogeneous random error was assumed and a logarithmic transformation of the data was carried out before analysis to stabilize any errors. Outliers were then removed, and a 2-way ANOVA was performed to determine whether the differences between the means of the datasets were statistically significant, taking into account the experimental variability.

Where indicated, primed BMNs were pretreated with vehicle control or inhibitors and stimulated with agonists as described in section Measurement of ROS production. A quantity of 5 × 10⁵ BMNs were sonicated into 1× sodium dodecyl sulfate (SDS) loading buffer, subjected to sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), transferred, and immunoblotted for phospho-p40^phox (T154; Cell Signaling Technology), p40^phox, p47^phox, or p67^phox (Millipore). Signal was detected by enhanced chemiluminescence and quantified using Aida Image Analyser 2.2.

We previously showed that p40^phox−/− mice exhibit a CGD-like phenotype, with profound defects in neutrophil ROS responses to both soluble and particulate stimuli.¹⁹  Neutrophils from these mice also show a substantial reduction in expression of p67^phox. We aimed to use p40^phox−/− tissue as a vehicle for the expression and functional analysis of p40^phox mutants in fully differentiated neutrophils. To this end, we used retroviral transduction of p40^phox−/− FL cells followed by repopulation of the bone marrow compartment of lethally irradiated recipient mice. We established this method by first expressing wt, untagged p40^phox in p40^phox−/− FL cells using the retroviral vector pMIG R1-IRES-GFP³⁴  (the retroviral constructs and p40^phox mutants used in this study are listed in Figure 1A-B). These transduced FL cells were then used to repopulate irradiated mice and to isolate fully differentiated neutrophils from their bone marrow compartment. Flow cytometric analysis of peripheral blood from these chimeric mice revealed that the average percentage of Gr1-positive cells expressing GFP varied between 73.1% and 82.5%, indicating successful reconstitution of the bone marrow compartment by transduced FL cells (Figure 1C). We also analyzed expression of phox components in the BMNs isolated from these mice and found that p40^phox introduced by pMIG R1-IRES-GFP expressed at levels similar to endogenous p40^phox in wt neutrophils (Figure 1D). Importantly, p67^phox expression was also rescued to levels that were equivalent to endogenous p67^phox in wt neutrophils (Figure 1D).

We characterized ROS responses in p40^phox−/− neutrophils reconstituted with p40^phox-IRES-GFP. We measured the rate of ROS formation in TNFα/granulocyte-macrophage colony stimulating factor primed BMNs stimulated with the chemotactic peptide fMLP or serum-opsonized S aureus. We also measured ROS formation stimulated by the phorbol ester PMA, which is envisaged to bypass receptor activation. Expression of untagged p40^phox fully rescued the ROS responses to fMLP, S aureus, and PMA (Figure 2A-C). Expression of a GFP-tagged p40^phox only partially rescued these responses, indicating an N-terminal GFP-tag partially inhibits p40^phox function (data not shown). We have also noted that neutrophils from p40^phox+/− heterozygotes displayed normal ROS responses to a variety of different agonists, despite reductions in p40^phox expression, suggesting these responses are insensitive to approximately 50% reductions in the levels of p40^phox (supplemental Figure 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article). Taken together, these results indicate that the introduction of untagged p40^phox mutants, through retroviral transduction of progenitor cells and subsequent repopulation of the bone marrow compartment, may be a useful tool for assessing their impact on neutrophil oxidase activation.

We investigated the function of the 2 documented phosphorylation sites of p40^phox, threonine 154, and serine 315²²  by creating bone marrow chimaeras expressing p40^phox-T154A-IRES-GFP or p40^phox-S315A-IRES-GFP in their hematopoietic lineage. BMNs were isolated from these mice, and we compared both their p40^phox expression levels and ROS responses to equivalently generated BMNs expressing either wt p40^phox-IRES-GFP or IRES-GFP alone (Figure 3). Both p40^phox-T154A and p40^phox-S315A were expressed to very similar levels as p40^phox, and, importantly, both mutants also rescued expression of endogenous p67^phox (Figure 3A). Substitution of threonine 154 to alanine led to significant decreases in p40^phox-dependent ROS responses to fMLP (40%), IgG-SRBCs (67%), and S aureus (63%) but no clear defect in the response to PMA (Figure 3B-E). In contrast, substitution of serine 315 to alanine resulted only in a small decrease in the ROS response to PMA (32%) but had no effect on the ROS responses to each of the physiological agonists tested (Figure 3B-E).

We investigated the role of phosphorylation on threonine 154 further by creating mouse chimaeras expressing p40^phox-T154E-IRES-GFP. The substitution of threonine by an acidic glutamate residue is often found to mimic the charge created by phosphorylation; however, we found p40^phox-T154E behaved almost identically to p40^phox-T154A in all of the BMN ROS assays we conducted, suggesting this substitution was unable to mimic phosphorylation in this context (supplemental Figure 2).

We also investigated the function of the SH3 domain of p40^phox in mouse neutrophils by creating bone marrow chimaeras expressing a mutant of p40^phox lacking this domain (p40^phox-ΔSH3-IRES-GFP) or carrying a mutation predicted to disrupt its binding to PP targets (p40^phox-W207Y-IRES-GFP). We had initially attempted to express p40^phox-W207R, because previous work had established that the tryptophan to arginine substitution blocks p40^phox SH3 activity,^27,28  but this mutant was poorly expressed in trial experiments (data not shown). We therefore made an alternative tryptophan 207 to tyrosine mutation (W207Y), introducing an acidic hydroxyl group predicted to disrupt hydrophobic interactions between the SH3 domain of p40^phox and a PP target (supplemental Figure 3) and which also preserved expression of the protein (see below).

Both p40^phox-W207Y and p40^phox-ΔSH3 were expressed in p40^phox−/− BMNs to similar levels as wt p40^phox (Figure 4A). In addition, both mutants of p40^phox rescued expression of endogenous p67^phox (Figure 4A). The p40^phox-dependent ROS response to PMA was reduced by approximately 30%-40% in BMNs expressing p40^phox-W207Y or p40^phox-ΔSH3 (Figure 4E). However, ROS responses to fMLP, S aureus, and IgG-SRBCs were unaffected in cells expressing p40^phox-W207Y and significantly elevated in cells expressing p40^phox-ΔSH3 (Figure 4B-D).

Given that mutation of the p40^phox SH3 domain did not reduce ROS responses to physiological agonists and that deletion of the domain indeed enhanced ROS responses to some stimuli, we postulated that the SH3 domain might play an auto-inhibitory role in oxidase activation, which may be alleviated by phosphorylation on threonine 154. To test this, we created the T154A mutation in the context of a deleted SH3 domain. p40^phox-ΔSH3/T154A was expressed normally in p40^phox−/− BMNs and this mutant fully rescued p67^phox expression (supplemental Figure 4). The defects in ROS responses in BMNs expressing p40^phox-ΔSH3/T154A were generally slightly greater than the defects seen in BMNs expressing p40^phox-T154A (supplemental Figure 4 and Figure 3), indicating SH3 domain deletion does not relieve the inhibition caused by a block in phosphorylation of threonine 154.

We used a phospho-specific antibody raised against phospho-threonine 154 to measure the phosphorylation of this site in BMNs stimulated with various agonists. Stimulation with PMA, fMLP, S aureus, or IgG-SRBCs resulted in substantial increases in T154 phosphorylation that were consistent with the time courses of oxidase activation to these agonists (Figure 5). T154 is located within a polybasic motif (RRLRPRTRKIK) that is a consensus site for phosphorylation by PKC enzymes (consensus-S/T-X-R/K-). Further, previous work has shown that a PKC can phosphorylate this site in vitro^23,25  and that broad-range inhibitors of PKC block PMA-induced phosphorylation of p40^phox in neutrophils.^22,23  We therefore sought to establish whether PKCs were responsible for phosphorylation of T154 in our assays using previously well characterized inhibitors of these enzymes.

The PKC family of enzymes comprises 10 different isoforms identified to date, which are classified according to their primary structure and mode of regulation (Table 1).³⁵  The PKC inhibitor Gö 6976 inhibits conventional (c) PKCs over other PKCs and is generally accepted to be a selective cPKC inhibitor (Table 1). BIM-1 inhibits cPKCs and novel (n) PKCs with similar potency, and Gö 6983 inhibits cPKCs, nPKCs, and atypical (a) PKCs with similar potency (Table 1). We used all 3 PKC inhibitors at a concentration of 1μM, which falls within the range of concentrations characterized in the literature.³⁶  Phosphorylations on T154 induced by PMA and S aureus were substantially and selectively inhibited by BIM-1 and Gö 6983 (Figure 5B). fMLP-induced phosphorylation of T154 was only modestly inhibited by each of the inhibitors tested and IgG-SRBC–induced phosphorylation was more potently inhibited by each of these inhibitors (Figure 5B). The most parsimonious explanation of these results is that the T154 phosphorylation induced by PMA or S aureus depends heavily on nPKC activation but that more redundancy is likely to exist between PKC isoforms, including cPKC and possibly other protein kinases, in the responses to fMLP or IgG-SRBCs. Further, the effects of these PKC inhibitors on ROS responses to these agonists were equivalent to, or greater than, would be predicted on the basis of their effects on T154 phosphorylation (supplemental Figure 5), consistent with previous data indicating PKC enzymes are also important in the activation of the oxidase through phosphorylation of p47^phox, particularly with respect to fMLP.^37-39 

To date, the only nPKC that is known to be expressed in BMNs is PKCδ.⁴⁰  We therefore sought to confirm a role for this PKC isoform in the phosphorylation of T154 in p40^phox stimulated by PMA and S aureus using BMNs isolated from mice lacking this protein (PKCδ^−/−). Phosphorylation of T154 was almost completely abolished in response to PMA and S aureus in PKCδ^−/− neutrophils (Figure 5C), whereas loss of PKCδ had a much smaller effect on this response in fMLP-stimulated cells (Figure 5C), consistent with the inhibitor studies described in the previous paragraph. In addition, the ROS response to S aureus was reduced to approximately 50% of wt in PKCδ^−/− BMNs, which is equivalent to the reduction in the ROS response caused by the T154A mutation (supplemental Figure 5). The ROS response to PMA, however, was almost completely abolished by loss of PKCδ (supplemental Figure 5), suggesting this enzyme is also required for other critical inputs into the oxidase downstream of PMA, such as the phosphorylation of other oxidase components.

A previous study has suggested that T154 phosphorylation induced by IgG targets is partially inhibited by the general PI3K inhibitor wortmannin.²⁴  We found that 100 nM and 300 nM wortmannin substantially inhibited T154 phosphorylation in response to S aureus (approximately 60%-75%), partially inhibited T154 phosphorylation in response to fMLP and IgG-SRBCs (approximately 20%-30%), and had no effect on T154 phosphorylation in response to PMA (Figure 5B). These results suggest PI3Ks may be differentially involved upstream of PKC-mediated T154 phosphorylation.

T154 phosphorylation induced by S aureus was normal in BMNs isolated from p40^{phoxR58A/R58A} mice and p40^phox−/− BMNs reconstituted with p40^phox-S315A or p40^phox-ΔSH3 (Figure 6). p40^phox−/− BMNs reconstituted with p40^phox-T154A, -ΔSH3/T154A, or -T154E displayed no detectable phosphorylation, which served to confirm the specificity of the antibody used (Figure 6). The lack of effect of the R58A mutation on T154 phosphorylation suggests the wortmannin sensitivity identified above is not mediated via a requirement for PtdIns3P-binding to p40^phox.

We had noted previously that translocation of p47^phox to S aureus–containing phagosomes is highly dependent on the presence of p40^phox (supplemental Figure 6), and so we used this system to investigate whether the T154A mutation might exert its effects on ROS production through impaired assembly of the oxidase complex. After ingestion of S aureus, wt p40^phox and p40^phox-T154A localized to phagosomal membranes to the same relative extents (Figure 7A-C). However, there was a marked decrease in the translocation of endogenous p47^phox in neutrophils expressing p40^phox-T154A (Figure 7B-C). This decrease was very similar in scale to the decrease in ROS effected by the T154A mutation, suggesting they may be causally linked.

We were able to successfully express p40^phox in p40^phox−/− mouse neutrophils through retroviral transduction of FL cells and repopulation of the hematopoietic compartment in lethally irradiated mice. Remarkably, heterologous expression of p40^phox driven by the murine stem cell virus promoter in pMIG R1-IRES-GFP reconstituted normal levels of p40^phox expression and normal p40^phox-dependent ROS responses in BMNs isolated from these chimeric mice. This provided us with an opportunity to investigate the function of conserved domains in p40^phox by reconstituting p40^phox−/− BMNs with p40^phox mutants. It is possible that the expression levels of p40^phox and its direct binding partner p67^phox are mutually self-regulating,^5,41-43, and thus it remains to be established how applicable this approach might be for probing the function of other oxidase subunits for which the appropriate knockout mice are available, such as gp91^phox or p47^phox. In any event, this represents a powerful methodology for expressing desired proteins in fully differentiated mouse neutrophils.

Our data suggest that the SH3 domain of p40^phox does not play an essential role in neutrophil oxidase activation. The introduction of a W207Y mutation predicted to disrupt SH3 domain binding to PP motifs caused a decrease in the ROS response to PMA but had no effect on ROS responses to physiological stimuli. This is consistent with a recent study that reported addition of p40^phox-W207R to permeabilized neutrophils could only partially restore a p40^phox-dependent ROS response to PMA.²⁷  It contrasts, however, with observations that the W207R mutation also reduces the ability of p40^phox to fully support IgG-SRBC–stimulated production of ROS in the model COS^phox cell system.²⁸  We also report that deletion of the entire SH3 domain causes elevated ROS production in response to fMLP, S aureus, and IgG-SRBCs, suggesting that any role for the SH3 domain in oxidase activation is likely to be that of an “inhibitory gate.” A previous study conducted by Sathyamoorthy and coworkers¹⁸  also reported an inhibitory role for the SH3 domain of p40^phox in whole-cell oxidase activity in transfecto, although in their assay system, full-length p40^phox was also found to inhibit oxidase activity. It is currently difficult to reconcile all of this data, suggesting we still do not have a clear understanding of the physiological role of this domain.

Our studies pinpoint PKC-mediated phosphorylation of p40^phox on T154 as a key regulatory step in the activation of the NADPH oxidase. T154A mutation resulted in substantial defects in ROS responses to fMLP, S aureus, and IgG-SRBCs but did not clearly affect the ROS response to PMA. Mutation of the other documented phosphorylation site on p40^phox, S315, only resulted in a modest defect in the ROS response to PMA. Using a combination of pharmacological and genetic inhibition of PKC isoforms, we were able to establish that phosphorylation of T154 in response to PMA or S aureus is entirely dependent on PKCδ but that probably a combination of both PKCδs and cPKCs are responsible for T154 phosphorylation in response to fMLP or IgG-SRBCs. Further, we identified wortmannin-sensitive pathways that lie upstream of PKC-mediated phosphorylation of T154, particularly in response to S aureus. cPKCs and PKCδs have previously been implicated in phosphorylation of p47^phox and subsequent activation of the NADPH oxidase, and in some cases PKCδ has been placed downstream of PI3K activation,^37,44  and thus our data extend the known influence of these enzymes in pathways regulating oxidase activation.

Our discovery of an important, positive role for T154 phosphorylation in the activation of the oxidase is in apparent disagreement with a study by Lopes and coworkers,²⁵  which demonstrates that in a cell-free system using membranes and recombinant proteins, p40^phox phosphorylated on T154 but not unphosphorylated p40^phox inhibits activation of the NADPH oxidase. This discrepancy is likely to reflect a difference in the experimental settings used. Lopes and coworkers use p47^phox phosphorylated in vitro by PKC as the activator of the NADPH oxidase. It is possible that in their in vitro recombinant assay system, the regulatory mechanism that depends on T154 phosphorylation is absent, lies upstream of p47^phox phosphorylation, or is bypassed by addition of excess amounts of phosphorylated p47^phox. In this regard, we also note that T154 phosphorylation is not essential for PKCδ-mediated ROS formation in PMA-stimulated BMNs.

In an attempt to further elucidate the mechanism by which T154 phosphorylation regulates the oxidase, we sought to establish if this event is linked to the function of other domains in p40^phox. R58A mutation in the PX domain of p40^phox stops p40^phox binding to PtdIns3P and causes significant defects in ROS responses to phagocytosis of both S aureus and IgG-opsonized targets.^20,32  p40^phox-R58A was phosphorylated normally on T154 in response to S aureus, indicating PtdIns3P-binding to p40^phox is not a prerequisite for phosphorylation of this site. Further, p40^phox-ΔSH3 or p40^phox-S315A were also phosphorylated normally in response to S aureus, suggesting T154 phosphorylation is also independent of the SH3 domain or phosphorylation on S315. Further, deletion of the SH3 domain did not overcome the inhibition of ROS responses caused by the T154A mutation, suggesting this phosphorylation event does not mediate its effects through derepression of an SH3 domain interaction. It remained possible that T154 phosphorylation is required for binding of p40^phox to PtdIns3P. Indeed, phosphorylation of p40^phox on T154 could be a trigger that alleviates the closed conformation of p40^phox in which PX-PB1 domain interaction masks the PX domain from binding to its target.⁴⁵  Although we have not directly tested this hypothesis, the normal translocation of p40^phox-T154A to S aureus phagosomes, as opposed to the approximate 40% decrease in translocation of p40^phox-R58A,⁴⁶  argue against phosphorylation of p40^phox being upstream of its binding to PtdIns3P. In agreement with these observations, p40^phox-T154A translocates to early endosomes after stimulation of the macrophage-like cell line RAW 264.7 with arachidonic acid, indicating that a p40^phox unable to be phosphorylated on T154 can still bind to PtdIns3P-rich structures.¹⁶ 

Our observation that T154 phosphorylation plays a partial but important role in oxidase activation by stimuli that direct oxidase assembly on the plasma membrane (fMLP), internalized phagosomes (S aureus), or a combination of the 2 (IgG-SRBCs) suggests this event lies downstream of the differential signaling webs engaged by these stimuli (emanating from activation of G protein coupled receptors, toll-like receptors, integrins, or Fcγ receptors). This in turn suggests T154 phosphorylation may directly participate in the efficient assembly of the soluble oxidase subunits around cytochrome b₅₅₈. In support of this concept, we found that T154A mutation in p40^phox significantly reduced the accumulation of p47^phox on S aureus phagosomes to a similar degree as it affected ROS formation. Given that p47^phox is accepted to be a critical regulator of oxidase assembly and activity,^8,10  it is possible that T154 phosphorylation on p40^phox induces a more effective interaction between p47^phox and the other subunits of the oxidase complex and, through this, regulates oxidase activity. Unfortunately, T154 resides in a flexible linker between the PX and SH3 domains of p40^phox that is invisible in the existing crystallographic structure of the full-length protein,⁴⁵  and thus it is difficult to speculate meaningfully as to how T154 phosphorylation might achieve this. Lopes and coworkers,²⁵  as well as Bouin and coworkers,²²  found that phosphorylation of p40^phox on T154 conferred protection against cleavage by thrombin, suggesting that unphosphorylated and phosphorylated p40^phox indeed differ in their conformation. However, we found that phosphorylation on T154 could not be mimicked by substitution with a glutamate residue, suggesting phosphorylation may provide additional contacts with other protein domains. Clearly, more work will need to be done to understand the mechanism by which T154 phosphorylation on p40^phox regulates oxidase activity, but our data identify for the first time that this is an important physiological event.

We thank Chris Ellson, Gordon John Ferguson, and Suhasini Kulkarni for helpful discussion and critical reading of the manuscript. We would also like to thank Tom Henley and Elena Vigorito for helpful advice and technical assistance, John Coadwell and Simon Andrews for bioinformatic support, and Anne Segonds-Pichon for statistical analysis. We also would like to thank Geoff Morgan and Arthur Davies for support with flow cytometry and staff in the small animal unit and small animal barrier unit at the Babraham Institute for animal husbandry.

This work was supported by the BBSRC and Wellcome Trust (WT085889MA). T.C. is a UCB BBSRC-CASE award student.

Contribution: T.A.M.C. designed and performed experiments, analyzed and interpreted data, and wrote the manuscript; K.E.A. designed and performed experiments and analyzed and interpreted data; Y.H., Q.X., and O.R. provided valuable reagents; and L.R.S. and P.T.H. designed experiments, analyzed and interpreted data, and wrote the manuscript.

Conflict-of-interest disclosure: The authors declare no completing financial interests.

Correspondence: Phillip Hawkins, Inositide Laboratory, Babraham Institute, Babraham Research Campus, Cambridge CB22 3AT, United Kingdom; e-mail: phillip.hawkins@bbsrc.ac.uk.