Ala: Alanine; Arg: Arginine; Asp: Aspartic acid; CGD: Chronic Granulomatous Disease; CH50: Complement Hemolysis 50%; CYBA: Cytochrome b alpha; CYBB: Cytochrome b beta; DHR: Dihydrorhodamine-1,2,3; DFP: Diisopropyl fluorophosphate; FAD: Flavine Adenine Dinucleotide; fMLP: formyl-methionyl-leucyl-phenylalanine; GEF: Guanine nucleotide Exchange Factor; GTP: Guanosine 5’-triphosphate; Lys: Lysine; mAb: Monoclonal antibody; MR: Magnetic Resonance; NADPH: Nicotinamide Adenine Dinucleotide Phosphate (reduced); NBT: Nitro-Blue Tetrazolium; NCF: Neutrophil Cytosolic Factor; Nox: NADPH oxidase; O₂^-: Superoxide; pAb: Polyclonal Antibody; PAF: Platelet-Activating Factor; PAGE: Polyacrylamide Gel Electrophoresis; PB1: Phox and Bem-1; PBS: Phosphate-Buffered Salt; phox: Phagocyte oxidase; PI3: Phospho-inositol-3; PMA: Phorbol Myristate Acetate; ROS: Reactive Oxygen Species; S.D.: Standard Deviation; SDS: Sodium dodecylsulphate; STZ: Serum-Treated Zymosan; TPR: Tetratricopeptide; Val: Valine; X-CGD: X chromosome-linked CGD

Phagocytic leukocytes protect us against bacteria, yeasts and fungi by ingesting these micro-organisms, followed by intracellular killing, or by attachment and extracellular killing. In this process, release of stored bactericidal proteins as well as generation of reactive oxygen species (ROS) by the phagocytes is essential [1]. Superoxide (O₂^-), as a precursor of other ROS, is produced by the leukocyte NADPH oxidase. This enzyme consists of two membrane-bound components, glycoprotein (gp)91^phox (phox from phagocyte oxidase), also called Nox2, and p22^phox, together forming flavocytochrome b₅₅₈, and three cytosolic proteins called p40^phox, p47^phox and p67^phox. The gp91^phox protein is the catalytic subunit; it contains an NADPH binding site, one FAD and two heme prosthetic groups. The p22^phox protein stabilizes gp91^phox in membranes and also provides a docking site for the cytosolic p47^phox subunit. The three cytosolic components form a tight complex that changes its conformation and translocates to the gp91^phox/p22^phox complex in the plasma membrane upon cell activation, e.g. after contact with micro-organisms [2]. Superoxide production also requires membrane translocation and activation of the small Rho GTPase Rac (preferentially Rac1 in macrophages and Rac2 in neutrophils), which subsequently binds to the tetratricopeptide regions (TPR) in p67^phox (Supplementary Figure S1) and to the plasma membrane [2,3]. Following assembly and activation of the cytosolic subunits on flavocytochrome b₅₅₈, the NADPH binding site of gp91^phox becomes available for NADPH in the cytosol. NADPH donates two electrons to gp91^phox, which are then transported within the protein to FAD, thereafter to the hemes, and finally to molecular oxygen at the other side of the membrane. Electron transfer requires Rac-activated p67^phox binding to gp91^phox as well as interactions between Rac and gp91^phox [4]. In this way, superoxide is generated within the phagosome or on the cell surface, in close proximity to the ingested or attached micro-organisms.

Genetic failure of superoxide generation leads to a rare syndrome of recurrent, life-threatening infections called Chronic Granulomatous Disease (CGD) [5]. The most common form of CGD (about 70% of all cases) is due to mutations in CYBB, the gene that encodes gp91^phox [6]. Since CYBB is located on the X chromosome, this form of CGD is found almost exclusively in males. The autosomal forms of CGD are less common, with mutations in NCF1 (p47^phox) in about 20% of cases and mutations in CYBA (p22^phox) or NCF2 (p67^phox) each in about 5% of cases [7]. A single patient with mutations in NCF4 (p40^phox) has also been described [8]. Usually, these mutations lead to complete absence of the protein involved, but a few cases are known with diminished expression of gp91^phox (resulting in diminished NADPH oxidase activity) or normal expression of completely inactive gp91^phox [9]. We describe here three unusual CGD patients from two families with an identical mutation in NCF2, leading to diminished to near normal p67^phox expression and substantial residual NADPH oxidase activity in their neutrophils. Expression of the mutated p67^phox protein in a cellular test system proved this mutation to be the cause of the disease.

Cell purification

Blood samples were obtained from healthy controls, patients and their family members by their physician, following the procedures and appropriate consent protocols approved by the Human Subjects Committee of the hospitals involved. Total leukocytes were obtained by lysis of the erythrocytes in the pellet fraction with a non-fixing lysis solution of 155 mM NH₄Cl, 10 mM NaHCO₃ and 0.1 mM EDTA. Neutrophils were purified by centrifugation of the leukocyte fraction over a layer of Percoll with a specific gravity of 1.077 g/ml. The cells in the pellet (neutrophils) were suspended in Hepes medium [132 mM NaCl, 6 mM KCl, 1 mM MgSO₄, 1.2 mM KH₂PO₄, 20 mM Hepes, 5.5 mM glucose and 0.5% (wt/vol) human albumin (pH 7.4)], and the cells in the ring fraction (mononuclear leukocytes) were used for RNA purification.

NADPH oxidase tests

Oxygen consumption by neutrophils activated with serum-treated zymosan (STZ) or phorbol myristate acetate (PMA) was measured with an oxygen electrode [10]. The dihydrorhodamine-1,2,3 (DHR) test was performed with total leukocytes as described in Köker et al. [11]. This test measures the oxidation of DHR by hydrogen peroxide to the fluorescent compound rhodamine-1,2,3 on a per-cell basis in a flow cytometer by gating of the neutrophils on the basis of forward and side scatter. The nitro-blue tetrazolium (NBT) slide test was performed as described by Meerhof and Roos [12]. This test was performed with purified neutrophils and measures in a semi-quantitative way the reduction of NBT by superoxide to dark blue precipitates in each cell. The formation of superoxide by purified neutrophils was also measured by reduction of ferricytochrome c to ferrocytochrome c, followed in a spectrophotometer at 550 nm [13]. Finally, the secretion of hydrogen peroxide by purified neutrophils was evaluated in a 96-well plate with Amplex Red (Molecular Probes, Life Technologies, Carlsbad, CA, USA) and horse-radish peroxidase [14]. The resulting resorufin was measured over a period of 30 min in a plate reader (Genios Plus, Tecan, Männedorf, Switzerland) at 590 nm (excitation at 535 nm). The steepest part of the slope was used for calculating the maximal rate of H₂O₂ production.

Expression of NADPH oxidase components

The expression of gp91^phox, p22^phox, p47^phox and p67^phox was analyzed in a flow cytometer with permeabilized and fixed blood cells as described [11]. Western blot analysis was performed after SDS-10% PAGE of DFP-treated neutrophil lysates in 2-mercapto-ethanol, transfer to nitrocellulose, blocking with 5% (w/v) milkpowder and incubation with mAb anti-p47^phox (Santa Cruz Biotechnology, Santa Cruz, CA, USA; mouse-anti-human p47^phox, clone 10, cat. no. SC-17845) and pAb anti-p67^phox (Merck Millipore, Billerica, MA, USA; rabbit-anti-human p67^phox, cat.no. 07-002). Conjugates were fluorescently labelled (LI-COR Biosciences, Lincoln, NE, USA), detected by scanning with the Odyssey Infrared Imagine System and quantified with Odyssey Application Software V3.0 (LI-COR).

Translocation of p67^phox to the membrane in intact neutrophils

Neutrophils (5×10⁶/ml, 1 ml) were stimulated with PMA (100 ng/ml) or STZ (1 mg/ml), washed once, and resuspended in 0.5 ml PBS with 0.1 mM diisopropyl fluorophosphate (DFP) for 10 min at 4°C. The cells were centrifuged and the pellet was resuspended in 100 μl of digitonin for 10 min at 4°C; for PMA-stimulated neutrophils a concentration of 150 μM digitonin in PBS was used, and for STZ stimulation a concentration of 300 μM digitonin in PBS. Thereafter, the cells were centrifuged (20 sec 20,000xg), the pellets were resuspended in Laemmli sample buffer and a Western blot was performed as described above. At these concentrations of digitonin, >90% of LDH was released from the cells and less than 5% of the protease content (data not shown), indicating proper separation of the cytosol and the rest of the cell.

For immunofluorescence, neutrophils were incubated with PMA (100 ng/ml) or left untreated for 10 minutes at 37°C in suspension. Next, the cells were allowed to adhere for 10 minutes on fibronectin (10 ng/ml)-coated glass covers, followed by a 10-minute incubation with STZ (1 mg/ml) or PMA (100 ng/ml), or left untreated. Thereafter, the cells were fixed with 3.7% (w/v) formaldehyde for 10 minutes and permeabilized with 0.5% (w/v) Triton X-100 for 10 minutes. To visualize p67^phox protein, the cells were incubated with the corresponding rabbit-anti-human antibody (Merck) for 30 minutes at room temperature, followed by a 30-minute incubation with a secondary goat-anti-rabbit-Ig ALEXA-568 antibody (Invitrogen). Coverslips were mounted with Vectashield (Vector Laboratories Inc., Peterborough, UK) on microscope slides and imaged with a confocal microscope through a 63x oil-objective (LSM510 META; Carl Zeiss MicroImaging, Inc.).

Expression and functional testing of recombinant p67^phox in K562 cells

K562 cells, immature myeloid cell line cells that constitutively express p22^phox, were first stably transfected with gp91^phox cDNA in pEF-PGKpac [15] and then with p47^phox cDNA in pEF-PGKhygro [16]. Cells were selected as individual clones in 2 μg/ml puromycin and 250 μg/ml hygromycin for 3 weeks. A clone with high recombinant gp91^phox and p47^phox expression (K562-91-47) was used for further studies, and immunoblots were made as described [8]. K562-91-47 cells were transiently transfected with Amaxa by means of Nucleofector Kit V and protocol T-16kit V (Walkersville, MD, USA). Superoxide production by these cells was determined with isoluminol chemiluminescence in the presence of HRP as described [17].

Mutation analysis

Genomic DNA was isolated from total leukocytes by standard procedures and analyzed for mutations in NCF2 exons and exon-intron boundaries by PCR amplification of each exon with its intronic boundaries, followed by bi-directional sequencing. The PCR conditions were as follows: 50 cycles of 5 s at 95°C, 30 s at 60°C and 15 s at 72°C, and for exon 16 50 cycles of 5 s at 95°C, 30 s at 52°C and 15 s at 72°C. The PCR products were sequenced with the Big dye terminator sequencing kit v1.1 (Applied Biosystems, Foster City, CA, USA).

Total mRNA was purified from the mononuclear leukocyte fraction and converted into cDNA by means of Superscript III first-strand synthesis system for RT-PCR (Invitrogen, Carlsbad, CA, USA).

Family A, a presumably non-consanguineous Turkish immigrant family with three daughters and one son, originally from the province of Tokat, Turkey, and now living in London, UK

Patient A1 (eldest female sibling), born 1970: This lady presented at age 17 with a 4-year history of recurrent cutaneous abscesses. These were controlled with antibiotics alone, and she managed minor flares of these on her own without the need to seek medical advice. She also has a chronic inflammatory, discoid lupus-like rash on her face. At age 30, she had an episode of peripheral ulcerative keratitis with adjacent conjunctival granulomata. The keratitis itself was non-granulomatous on biopsy. There was a recurrent episode of keratitis at age 35 (during the second trimester of pregnancy). Both episodes responded well to topical steroids and chloramphenicol drops. Since the diagnosis of CGD was made, the patient has been on trimethoprim-sulfamethoxazole and itraconazole prophylaxis. Neutrophil testing (ferricytochrome c reduction assay) revealed about 10% of normal NADPH oxidase activity with PMA.

Patient A2 (male sibling), born 1987: The male sibling was diagnosed with CGD at birth by NBT slide testing. He completed all childhood vaccinations without complications, but suffered from recurrent oral ulceration, leg ulcers, folliculitis and skin abscesses throughout childhood, controlled with repeated courses of topical and systemic antibiotics. Levels of all immunoglobulin sub-classes were normal. The patient has been taking trimethoprim-sulfamethoxazole and itraconazole prophylaxis during the last twelve years. At age 19, he had a short episode of what was thought to be inflammatory bowel disease, with diarrhoea and rectal bleeding, although an MR-imaging of the abdomen at age 21 demonstrated no small or large bowel inflammation. He is currently in symptomatic remission with no history of other bacterial infections. Other members of this family did not present with medical problems.

Family B, Turkish family living in the province of Adana, Turkey, with no obvious relation to family A

The patient in this family is the only sibling, a girl born in 1990, whose parents are first cousins, with no history of early death in the family. She was referred to hospital at 8 years of age with diffuse pustular and eczematous lesions of the scalp skin, which were treated with systemic antibiotics, but without complete cure. CGD was diagnosed by impaired NBT test (all cells weakly positive). T cell function and lymphocyte subsets as well as CH50 were also normal. Her serum immunoglobulin levels were IgG 2380 mg/l, IgA 273 mg/l, IgM 73 mg/l, and IgE 71 mg/l, which is within normal limits. When she was 8 months of age, strabismus in the left eye was noticed after convulsions. During ophthalmological inspection, chorioretinitis and decreased vision in the left eye was diagnosed. She suffered repeatedly from chorioretinitis attacks in both eyes and has severe bilateral uveitis. At present, that has resulted in an almost 75% loss of vision in her left eye. She has been doing well for 10 years on prophylactic trimethoprim-sulfamethoxazole and itraconazole, at half dose during the last five years. Eczematous lesions of the scalp skin disappeared with Fe⁺⁺ supplements. DHR analysis showed 5-10% of normal NADPH oxidase activity after stimulation of her neutrophils with PMA [11].

All three patients had considerable NADPH oxidase activity, measured as oxygen consumption and as hydrogen peroxide release, in neutrophils activated by various stimuli (Figure 1A). This activity was about 50% of control values with unopsonized zymosan, 60-70% with serum-treated zymosan (STZ), 15-25% with phorbol-myristate acetate (PMA) and 20% with formyl-methionyl-leucyl-phenylalanine (fMLP) in platelet-activating factor (PAF)-primed neutrophils. The difference in residual NADPH oxidase activity in the patients’ cells activated with STZ as compared with PMA was highly significant, both in the oxygen consumption assay (p=0.009) and in the H₂O₂ release assay (p<0.0001). For comparison, neutrophils from two “classical” CGD patients (one with a one-nucleotide insertion in CYBA and another with a p.Arg102X nonsense mutation in NCF2) were also tested, in the same assay run. These neutrophils showed only 6% residual oxidase activity with unopsonized zymosan and 3% or less of control values with the other stimuli (Supplementary Table 1). The parents and sisters of patients A1 and A2 showed normal hydrogen peroxide release from their neutrophils (not shown). The mother of patient B showed normal hydrogen peroxide generation by her neutrophils in the DHR test (not shown). Western blots of neutrophil lysates from all three patients showed substantial expression of p47^phox and p67^phox (Figure 1B), as well as of gp91^phox and p22^phox (not shown).

Figure 1: Characteristics of patients’ neutrophils. (A) NADPH-oxidase activity. Neutrophils were incubated with serum-treated zymosan (STZ, 1 mg/ml, 70 particles per neutrophil) or PMA (100 ng/ml), and the oxygen consumption was measured with an oxygen electrode (10). The maximal rate of oxygen consumption of patient cells is displayed in nmoles/min/10⁶ cells. Alternatively, the release of hydrogen peroxide from the cells was measured with the Amplex Red assay (14) after stimulation with zymosan (1 mg/ml), STZ (1 mg/ml), PMA (100 ng/ml or PAF (1 μM) followed by fMLP (1 μM). The maximal rate of H₂O₂ release is displayed as nmoles/min/10⁶ cells. Open bars, control neutrophils; closed bars, patient neutrophils (patients A1, A2, and B). Mean ± SEM of 3 (oxygen consumption) or 4 (H₂O₂ release) independent experiments (in the H₂O₂ release assay, one patient was tested twice). Significance of differences was calculated with the paired, two-tailed ttest. (B) Western blot of p47^phox and p67^phox. Neutrophils from a control donor (lane 1), patient B (lane 2), patient A1 (lane 3), patient A2 (lane 4) and a CGD patient with a p.Trp137Arg mutation in p67^phox (lane 5) were lysed and subjected to SDS-PAGE as described under Methods. The proteins were blotted onto nitrocellulose, treated with antibodies to p47^phox and to p67^phox and visualized by fluorescence. The lower (green) band indicates the presence of p47^phox, the upper (red) band the presence of p67^phox.

We started DNA analysis by sequencing the exons and intron-exon boundaries of CYBB, as well as the first 600 nucleotides of its promoter region, because mutations are known in this gene to cause diminished expression of gp91^phox and diminished NADPH oxidase activity [9]. However, no mutations were found in CYBB in these patients. We then investigated whether a common dinucleotide deletion in NCF1 was present, because deficiency of p47^phox is known to leave some residual NADPH oxidase activity [18]. However, a gene scan [19] failed to detect this GT deletion in NCF1. We then sequenced the relevant parts of CYBA (from gDNA) and NCF1 (from cDNA) but found no mutations or indications for mRNA missplicing. Finally, in NCF2 we did find a homozygous c.605C>T mutation in all three patients, predicting p.Ala202Val in p67^phox (Figure 2A). The parents and sisters of patients A1 and A2 were heterozygotes for the c.605C>T mutation, as were the parents of patient B. In more than 100 healthy controls we did not observe this mutation. To investigate whether the mutant p67^phox mRNA was as stable as the wild-type p67^phox mRNA, we amplified the relevant part of p67^phox cDNA in the mother of patient B and found both cDNA species to be present in similar amounts (Figure 2B). Moreover, the wild-type and the mutated cDNA amplicon had a similar size, which rules out activation of a cryptic splice site by the mutation.

Figure 2: DNA and RNA sequencing of NCF2 (end exon 6). Figure 2A shows the sequence obtained from genomic DNA of family A, with patients A1 and A2 being homozygous for the c.605C>T mutation and all other family members heterozygous for this mutation. Figure 2B shows the sequence obtained with cDNA of patient B and her mother. The c.605C peak and the c.605T peak in the mother have a similar height, indicating that the mutated c.605T mRNA is as stable as the wild-type c.605C mRNA.

The question remained whether this mutation in p67^phox was really the cause of the diminished NADPH oxidase activity in the neutrophils of the patients. To investigate this, we expressed the mutant p67^phox Ala202Val and the wild-type p67^phox in K562 cells stably transfected with p47^phox and gp91^phox and expressing endogenous Rac and p22^phox. As shown in Figure 3A, both mutant and wild-type p67^phox proteins were expressed in similar amounts in these cells. For comparison, we also expressed p67^phoxVal204Ala in the K562 cells, because this mutation, which – like Ala202Val – resides in the p67^phox "activation domain" that is critical for activation of electron transport in gp91^phox, has been shown in an in vitro system to lack all oxidase-activating potency [20,21]. Moreover, an Ala202Asn mutation in p67^phox markedly reduces NADPH oxidase activity in a gp91^phox-dependent whole cell system stimulated with PMA [21]. The p67^phoxVal204Ala protein we used also contained a C-terminal myc tag, which does not have any effect on the superoxide production supported by p67^phox wt-tagged protein [16]. Figure 3B shows that both the p67^phox Ala202Val and the p67^phox Val204Ala variant were far less effective than the wild-type p67^phox in inducing NADPH oxidase activity in PMA-activated K562 cells. In three separate experiments, the Ala202Val variant induced 2.7 ± 1% (S.D.) of wild-type p67^phox-induced oxidase activity, whereas the Val204Ala variant induced 1.0 ± 0.8% of wild-type p67^phox-induced oxidase activity.

Figure 3: Expression and function of p67^phox in K562 cells. K562 cells containing all NADPH oxidase components except p67^phox were transfected with cDNA encoding p67^phoxAla202Val (p67A202), p67^phoxVal204Ala with a myc-tag (p67V204myc) or p67^phoxAla202 (p67wt). A representative Western blot in Figure 3A shows that all proteins were expressed in similar amounts in the cells. Figure 3B shows that upon activation with PMA, the mutated proteins did not support the oxidase activity in K562 cells, in contrast to the wild-type protein.

Finally, we studied the translocation of p67^phox to the cell membrane after NADPH oxidase activation of neutrophils with two different assays, as described under Methods. The results are shown in Figure 4 and indicate that with PMA as the stimulus, the translocation of the p67^phox protein from the cytosol to the membrane was clearly diminished, whereas with STZ, the translocation was close to normal. In control experiments with classical X-CGD neutrophils (without expression of gp91phox) the translocation of p67^phox was completely absent with either PMA or STZ (Supplementary Figure S2).

Figure 4: Translocation of p67phox to the cell membrane in intact neutrophils. (A) Quantification of p67phox translocation in stimulated human neutrophils. Human neutrophils (5x10⁶/ml) were stimulated with PMA (100 ng/ml) for 10 min or with STZ (1 mg/ml) for 20 min. Separation of cytosol and the rest of the cells, followed by Western blot analysis of p67^phox was performed as described under Materials and Methods. The amount of p67^phox was quantified by means of fluorescently labeled conjugates, and detected by scanning with the Odyssey Infrared Imagine System and Odyssey Application Software V3.0. Black bars, control neutrophils; Red bars, patient neutrophils (patient A1, A2, and B). Mean ± SEM of 3 independent experiments for PMA. With STZ only the cells of patient B were tested. Significance of differences was calculated with the paired, two-tailed t-test. (B) Visualization of p67^phox translocation in stimulated human neutrophils. Neutrophils from a control donor and from patient B were incubated with PMA (100 ng/ml) or left untreated for 10 minutes at 37°C in suspension. The cells were then allowed to adhere on fibronectin-coated glass covers, followed by a 10-minute incubation with STZ (1 mg/ml) or left untreated. Thereafter, the cells were fixed with formaldehyde and permeabilized with Triton X-100. To visualize p67^phox protein, the cells were incubated with rabbit-anti-human-p67^phox, followed by incubation with a secondary goat-anti-rabbit-Ig ALEXA-568-labeled. Coverslips were mounted with Vectashield on microscope slides and imaged with a confocal microscope through a 63x oil-objective. Note that with PMA, p67^phox translocates to the plasma membrane of control neutrophils (arrows), but much less so to the plasma membrane of patient neutrophils. In contrast, translocation of p67^phox to the phagosomal membrane surrounding internalized STZ (arrowheads) is similar in control and patient neutrophils.

CGD patients with residual expression of NADPH oxidase components as well as residual NADPH oxidase activity are rare. Only four patients in three families have been described with low expression of p67^phox and/or low NADPH oxidase activity [22-24]. The first of these had a deletion of one amino acid (Lys58) on one allele of NCF2 and an undefined large deletion on the other allele [22]. The Lys58-deleted protein was expressed to a certain extent (tested on Western blot with a polyclonal antibody against p67^phox), but whether this was a normal expression (from one allele) or diminished expression could not be decided. The Lys58 deletion is in the fourth TPR domain and destroyed the interaction with Rac and the translocation of p67^phox to the membrane in PMA- or STZ-activated neutrophils [22]. The NADPH oxidase activity in the neutrophils of this patient was completely absent with all stimuli tested. However, in the so-called cell-free system with recombinant proteins and neutrophil membranes, SDS and GTPγS did induce the translocation of these cytosolic proteins, although the NADPH oxidase activity was still absent.

The second patient had a missense mutation in NCF2 that caused replacement of aspartic acid by valine at position 108 in p67^phox [23]. This Asp108Val replacement is between the third and fourth TPR region in p67^phox and left substantial residual NADPH oxidase activity (15-20% of normal) in the patient’s PMA-activated neutrophils (tested in the DHR assay). In a flow cytometric assay with permeabilized neutrophils, p67^phox was undetectable with a monoclonal antibody. The authors speculate that the mutation may have changed the conformation of p67^phox, rendering it undetectable with the monoclonal antibody used, but still able to interact to some extent with gp91^phox for inducing some NADPH oxidase activity.

In the last family, two brothers were recognized as CGD patients when they were already in their fifties [24]. They had a splice site mutation in NCF2 that gave rise to an in-frame deletion of exons 11 and 12 (amino acids 309-342, PB1 domain). The neutrophils from these patients showed 10-15% of normal NADPH oxidase activity after stimulation with PMA in the DHR assay and in the lucigenin-enhanced chemiluminescence assay. This result was reproduced in K562 cells that contained all NADPH oxidase components except p67^phox transduced with the Δexon11_12p67phox cDNA. The authors speculate that the p67^phox protein with the exon11_12 deletion was to some extent expressed and functional in the patients’ phagocytes.

The hypomorphic mutation in the three patients described in this article is in the so-called Activation Domain of p67^phox (Supplementary Figure S1). This stretch of twelve amino acids (199-210) is essential for the oxidase-inducing capacity of p67^phox[20]. In a cell-free oxidase system, it was found that mutations in this domain do not affect binding of p67^phox to p47^phox or to Rac but do inhibit the oxidase activity [20]. Alanine-202 is highly conserved in p67^phox from humans, mouse, chicken, frog, fish and lancelet, as well as in Noxa1 of humans, mouse and fish and in fungal NoxR [21]. Mutation of alanine-202 in p67^phox into leucine inhibits the cell-free oxidase activity induced by arachidonic acid by about 50%, and a Val204Ala mutant totally blocks this activity. This last mutant associates with the membrane (presumably with gp91^phox) as well as does the wild-type p67^phox [20]. Direct interaction of p67^phox with gp91^phox was shown by Dang et al. [25,26] by overlay techniques and GST pull-down assays, but the Activation Domain of p67^phox was not necessary for this reaction. Thus, the binding of p67^phox to gp91^phox is probably mediated by a site in p67^phox different from the Activation Domain, but the induction of oxidase activity in gp91^phox is strictly dependent on this domain. The site in gp91^phox interacting with the Activation Domain of p67^phox is not known.

The findings in our patients corroborate these notions and extend the findings to intact, primary phagocytes. We found partial inhibition of oxidase activity in the patients’ intact neutrophils with Ala202Val p67^phox, as was found with the recombinant Ala202Leu variant of p67^phox in the cell-free system [20]. Remarkably, much more oxidase activity was induced in the patients’ neutrophils with zymosan or STZ than with the soluble activators PMA or PAF/fMLP. This correlates with the normal translocation of p67^phox to the membrane after neutrophil activation with STZ and clearly diminished translocation after activation with PMA. It suggests that the Ala202Val mutation in p67^phox affects the translocation and – perhaps as a result – also the proper assembly or activation of the NADPH oxidase complex following translocation of the cytosolic components to flavocytochrome b₅₅₈ in the plasma membrane. This reduced translocation of Ala202Val-p67^phox might be due to reduced binding of p67^phox to gp91^phox, which would be in contrast to the conclusion drawn by Dang et al. from studies with purified neutrophil and recombinant proteins that the Activation Domain of p67^phox does not interact with gp91^phox [25,26]. Unfortunately, the 3D structure of the Activation Domain of p67^phox is unknown [27-30]. However, it is known that different stimuli induce different activation pathways in neutrophils, especially with respect to the synthesis of various lipid products needed for assembly of an active oxidase complex [31-34]. The type or amount of lipid mediators generated in STZ-activated neutrophils may have been sufficient for almost normal oxidase activation by Ala202Val-p67^phox, in contrast to the situation in PMA-activated neutrophils. Since lipid mediator generation in K562 cells may be different from neutrophils, this may also explain the low oxidase activation by Ala202Val-p67^phox in transfected K562 cells as compared to the patients’ neutrophils. Alternatively, since the signal transduction pathway induced by PMA (protein kinase C activation leading to p47^phox phosphorylation) is different from that induced by STZ (tyrosine phosphorylation of PI3 kinase leading to GEF and Rac activation), the translocation of p67^phox to gp91^phox and subsequent activation of gp91^phox may be differently affected by mutations in p67^phox. Thus, the p47^phox-dependent pathway induced by PMA may be more sensitive to mutations leading to conformational changes in p67^phox than the Rac-dependent pathway induced by STZ.

Our patients raise the question how much NADPH oxidase activity is required to be able to lead a normal life. The high residual oxidase activity in the patients’ neutrophils may have protected the patients to a certain extent from the full-blown CGD symptomatology. Indeed, their clinical problems were mild in comparison to those of oxidase-null CGD patients. On the other hand, it should be taken into account that we tested the neutrophil NADPH oxidase activity in in vitro assay systems, with strong stimuli. In vivo, more subtle stimuli may be encountered, with which the mutant p67^phox may be unable to properly activate the NADPH oxidase. Thus, high residual NADPH oxidase activity in vitro is no guarantee for protection against pathogenic infections in vivo, but it may help in ameliorating the symptoms [35] and increase the chance of survival [36].

DR and MYK are recipients of an EURO-CGD grant from the E-RARE program of the European Union.