(Received for publication, August 27, 1996, and in revised form, January 16, 1997)
From the Laboratory of Host Defenses, NIAID, National Institutes of Health, Bethesda, Maryland 20893
The NADPH oxidase of phagocytes generates microbicidal oxidants in response to a variety of stimuli. Its activation and assembly involve multiple SH3 domain interactions among several oxidase components. Here we present evidence that the cytosolic oxidase-associated protein, p40phox, mediates down-regulation of NADPH oxidase through interactions with its SH3 domain. Recombinant p40phox was produced in several eukaryotic expression systems (insect, mammalian, and yeast) to explore its role in oxidase function in relation to domains involved in interactions with other factors, p47phox and p67phox. p40phox inhibited oxidase activity in vitro when added to neutrophil membranes and recombinant p47phox, p67phox, and p21rac. Co-transfection of p40phox into K562 cells resulted in significant decreases (~40%) in whole cell oxidase activity. Furthermore, the isolated SH3 domain of p40phox was even more effective in inhibiting whole cell oxidase activity, consistent with experiments showing that this domain binds to the same proline-rich target in p47phox (residues 358-390) that interacts with p67phox. In contrast, deletion of the carboxyl-terminal domain of p40phox that binds to p67phox did not relieve its oxidase inhibitory effects. Thus, p40phox appears to down-regulate oxidase function by competing with an SH3 domain interaction between other essential oxidase components.
Production of superoxide anion by phagocytic blood cells involves assembly of an activated NADPH oxidase complex of membrane-bound and cytosolic components (for review, see Ref. 1). Five distinct NADPH oxidase components have been identified; deficiencies or defects in any one of four essential proteins result in impaired oxidase activity associated with chronic granulomatous disease (CGD)1 (1). Two are subunits of cytochrome b558 (p22phox and gp91phox) that donate electrons to molecular oxygen to yield superoxide, and two are cytosolic proteins, p47phox and p67phox, that associate with the membrane-bound flavocytochrome during oxidase activation. The fifth essential component is a Ras-related GTPase, p21rac, that renders the enzyme sensitive to guanine nucleotides (2-4). Recently, yet another factor, p40phox, was identified in a high molecular weight cytosolic complex with p47phox and p67phox (5-7). The importance of the p40phox-p67phox association is evident in resting neutrophils from p67phox-deficient CGD patients, where the absence of p67 phox was correlated with reduced levels of p40 phox (5, 7-9).
The deduced structure of p40 phox reveals homology with p67phox and p47phox (5), which both contain two Src homology 3 (SH3) domains that also occur in various intracellular signaling proteins in eukaryotes (10, 11). The SH3 domain of p40phox has its greatest homology to the second SH3 domain of p67phox (44% identity), although p40phox and p47phox also exhibit homology in regions N-terminal to their SH3 domains. Recently, the role of SH3 domains in NADPH oxidase function has been the subject of intense investigation (12-23). Several SH3 interactions were shown to participate in translocation and assembly of the active membrane-bound enzyme in whole cells (12, 13, 16-18). The first p47phox SH3 domain engages a membrane-bound proline-rich target in p22phox (17), while the second SH3 domain of p67 phox binds to a carboxyl-terminal p47phox target and is critical for p67phox membrane translocation (12-14, 16-19). Recent work has shown that the SH3 domain of p40phox also binds the same carboxyl-terminal target in p47phox that binds p67phox (21, 22). Other SH3 domain interactions have been demonstrated, including intramolecular contacts within p47phox (12, 15-17); several of these associations may be regulated during oxidase activation.
Despite work in several laboratories on the interactions between oxidase components, the function of p40phox has remained unclear (5-9, 21-25). Antibodies against p40phox were shown to inhibit NADPH oxidase in vitro (25), although the effects of p40phox on oxidase activity have not been examined directly, and CGD patients with lesions in the p40phox gene have not been described. Furthermore, p40 phox is not needed to reconstitute high levels of oxidase activity in cell-free systems containing purified p21rac, p47phox, p67phox, and relipidated flavocytochrome b558 (26, 27). Like several of the core oxidase components, p40phox expression is myeloid-specific (5). The fact that p40phox is bound in a complex with p67phox and p47phox that translocates to the membrane upon activation suggests that p40phox could have a role in modulating the respiratory burst. Here we report about interactions of p40phox with other oxidase components and explore the effects of these interactions on oxidase activity, both in cell-free and whole cell reconstituted systems. Based on these observations, we provide evidence for yet another SH3 domain-mediated interaction, which in this case appears to inhibit oxidase function.
The full coding sequence of p40phox was cloned from a human leukocyte cDNA library (Clontech, Palo Alto, CA; catalog No. HL4014AB) using the polymerase chain reaction. Amplification primers were targeted to sequences beginning 21 bases upstream from the start codon and ending 30 bases after the stop codon of the p40phox cDNA (5). The p40phox polymerase chain reaction product was first cloned into the TA cloning vector (Invitrogen, San Diego, CA) for sequencing and then subcloned (EcoRI fragment) into pGEX3X for production of recombinant p40phox-glutathione S-transferase (GST) fusion protein in Escherichia coli. The fusion protein was affinity purified from induced bacteria on glutathione-Sepharose as described elsewhere (28).
Production of recombinant p40phox baculovirus for infection of cultured Sf9 (Spodoptera frugiperda) cells was performed by methods described earlier (29), with the following modifications. The p40 cDNA was subcloned (EcoRI fragment) into pVL1392 (Invitrogen) and used for co-transfection of Sf9 cells along with linearized, deleted baculovirus DNA (BaculoGold DNA, Pharmingen, San Diego, CA) according to the manufacturer's protocols. High titer p40phox baculovirus was used to infect large scale suspension cultures (100 ml) grown at 2.5 × 106 cells/ml in Sf900-II SFM medium (Life Technologies, Inc.). The protein was harvested from the cleared soluble fraction of sonicated cells harvested 3 days after infection as described (29). These preparations were analyzed by SDS-PAGE and immunoblotting and were used in cell-free oxidase reconstitution studies without further purification. Affinity-purified full-length p40phox used for immunization was obtained by binding to a recombinant GST fusion protein constructed with the carboxyl-terminal domain of p47phox (residues 358-390) (12).
Antiserum PreparationGST-p40phox fusion protein preparations from inclusion bodies were subjected to SDS-PAGE and visualized by staining with Coomassie blue. The full-length protein was excised and used to immunize rabbits (primary immunization, 150 µg in complete Freund's adjuvant, followed by two boosts in incomplete Freund's adjuvant at 2-week intervals). Subsequent boosts used affinity-purified baculovirus-derived p40phox eluted from SDS-PAGE gels. The antiserum obtained after 10 weeks of immunization cross-reacted specifically with p40phox in neutrophil cytosol and recombinant p40phox from Sf9 cells.
Mammalian Expression VectorsThe
p40phox cDNA was excised from the TA-p40
vector with KpnI (5) and XbaI (3
) and subcloned
into the episomal vector pCEP4 (Invitrogen) restricted with
KpnI and NheI. The carboxyl-terminal-deleted form
of p40phox (residues 1-260) was constructed by
restriction of full-length cDNA with KpnI (5
polylinker) and BglII (internal) sites and subcloned into
KpnI and BamHI-digested pCEP4. The segment
encoding the SH3 domain of p40phox (residues
172-229) was obtained by polymerase chain reaction amplification. pREP
episomal expression vectors carrying p47phox and
p67phox cDNAs were described earlier (13,
30).
Co-transfection of pCEP or pREP plasmids containing p67phox, p47phox, and p40phox cDNAs (20 µg each) into transduced K562 (gp91phox-expressing) cells was performed by electroporation as described previously (16, 30). In control experiments, pCEP4CAT vector (Invitrogen) was used in place of pCEP40. At 48 h post-transfection, 105 cells/ml were selected for 5 days in complete medium containing 250 µg/ml hygromycin B. Production of p40phox in hygromycin-resistant cells was confirmed by immunoblotting of cytosolic fractions with rabbit p40 antiserum (1:1000 dilution). All functional assays were performed within 2 weeks of hygromycin selection. Whole cell superoxide production in response to PMA stimulation (2 µg/ml) was determined as the superoxide dismutase-inhibitable chemiluminescence detected with an enhancer-containing luminol-based detection system as described (DIOGENESTM, National Diagnostics, Atlanta, GA) (16, 30), using a multiwell plate-reading luminometer (Luminoskan, Labsystems). Both experimental and superoxide dismutase-inhibited reactions were monitored at 1-min intervals for at least 15 min at 37 °C following PMA activation. The DIOGENESTM reagent is >1000 times more sensitive to superoxide than hydrogen peroxide due to the presence of a redox active compound that facilitates direct transfer of electrons from superoxide to luminol (Dr. G. Kitzler, National Diagnostics). As noted earlier (16, 30), chemiluminescence signals were completely abolished in the presence of superoxide dismutase and were catalase-insensitive (G. Kitzler), indicative of a superoxide-specific assay.
Yeast Two-hybrid Expression StudiesThe yeast Gal4
two-hybrid expression system (31-33) from Clontech (San Diego, CA) was
used to study interactions of NADPH oxidase components as fusion
proteins produced in yeast. In this system, two proteins of interest
are expressed as hybrids of the Gal4 DNA binding domain (pGBT9) or the
Gal4 transcriptional activation domain (pGADGH or pGAD424). An
interaction between the two hybrid proteins results in transactivation
of the HIS3 and lacZ reporter genes. Full-length
coding sequences of p40phox and
p47phox were subcloned into the EcoRI
site of pGADGH in frame with the Gal4 transcriptional activation
domain. Full-length p67phox cDNA was derived
from pREP10/67 (13). Deleted forms of p67phox
cDNA were made as follows: p67NT (residues 1-246) was excised from
full-length cDNA using EcoRI and BglII sites
and ligated into EcoRI and BamHI (or
BglII) sites in pGBT9 and pGAD424. The region between the
two SH3 domains of p67phox (CT), residues
293-465, was polymerase chain reaction-amplified and subcloned into
pGBT9. Sequence encoding the p47phox
proline-rich tail region (residues 351-390) was amplified and subcloned into pGBT9 and pGAD424. p40CT (residues 1-260) and p40CT
(residues 260-340) were obtained by restricting
p40phox cDNA in the pGADGH with
BglII (internal site) and SpeI (5
MCS) or
SalI (3
MCS) and subcloning
SpeI-BglII or BglII-SalI
fragments into appropriately cut pGADGH. Expression vectors were
confirmed by DNA sequencing and in some cases by Western blotting. All
manipulations in yeast strains HF7c or CG1945 were according to
manufacturers' directions. Interactions between fusion proteins were
monitored by histidine prototropy and
-galactosidase activity
detected with
5-bromo-4-chloro-3-indole-
-D-galactopyranoside within
lysed colonies that were adsorbed onto nitrocellulose filters (33). All
constructs were co-expressed along with appropriate control (nonrecombinant or empty) vectors to test for false positive
signals.
Cell-free superoxide production was determined from the
superoxide dismutase-inhibitable reduction of cytochrome c
as described earlier (13, 29). Reactions (100 µl) contained varying
amounts of Sf9 cell cytosols derived from uninfected or
p40phox baculovirus-infected cultures (0-20
µg) along with 5 × 105 cell equivalents of
deoxycholate-solubilized neutrophil membranes, 0.8 µg each of
recombinant p47phox, and
p67phox (29) and 1 µg of the mutant form of
p21rac1 (Q61L) preloaded with GTPS. This mutant form of
p21rac1 was expressed in pGex-2T (4) and was more active than
wild-type p21rac1 as noted earlier with p21rac2 (34). Negative
control reactions contained 5 µg of superoxide dismutase, in which
case cytochrome c reduction did not exceed 5% of the
activity observed in the absence of superoxide dismutase. Maximum rates
of superoxide generation were calculated from a linear least-squares
fit of 5 consecutive 1-min data points based on reactions performed in
duplicate.
Attempts to isolate intact recombinant p40phox from E. coli were not successful due to the proteolytic susceptibility of GST-p40phox fusion protein. The affinity-purified fusion protein contained a predominant breakdown product ~46 kDa in size, suggesting cleavage of a ~19-kDa fragment from the N terminus of p40phox. However, a nondegraded form of the GST-p40phox fusion protein was purified from insoluble inclusion bodies of induced cells by preparative SDS-PAGE and was used for primary immunization of rabbits against p40phox.
The baculovirus expression system was used for production of native,
full-length recombinant p40phox, since earlier
work showed that this system was efficient in production of other
cytosolic oxidase proteins (29). Sonicated lysates from
p40phox baculovirus-infected Sf9 cells (72 h
post-infection) contained a prominent 40-kDa species (Fig.
1A), which demonstrated
immuno-cross-reactivity with a 40-kDa band in neutrophil cytosol (Fig.
1B). The Sf9 lysates were considerably enriched in
p40phox in comparison with neutrophil cytosol on
a specific weight basis. These Sf9 preparations were used without
further purification as a source of p40phox in
subsequent cell-free oxidase reconstitution studies.
Baculovirus-derived p40phox also exhibited
binding to immobilized GST fused with the proline-rich, C-terminal
(33-residue) segment of p47phox, while not
binding to unfused GST (Fig. 1C), consistent with interactions detected in the yeast two-hybrid system (21, 22) (see
below). This affinity system was scaled up to purify
p40phox used in later immunizations.
Recombinant p40phox was also expressed in the
cell line K562. These cells lack the neutrophil oxidase components
p47phox, p67phox, and
gp91phox and proved to be an efficient
transfectable model for reconstitution of NADPH oxidase following
transduction with gp91phox retrovirus and
co-transfection of p47phox and
p67phox cDNAs in Epstein-Barr virus-based
episomal expression vectors (pREP4 or pREP10) (16, 30). To confirm
expression of p40phox in K562 cells, protein
lysates from hygromycin-resistant K562 cells transfected with
pCEPp40phox or control vectors were compared by
p40phox immunoblotting. Endogenous
p40phox was not detected in K562 cells
transfected with pCEPCAT (control) (Fig. 2, lane
1), while production of recombinant p40phox
in K562 cells transfected with pCEP40phox was
evident by detection of a 40-kDa immunoreactive species at levels
similar to neutrophil cytosol (Fig. 2, lanes 2 and
5). Co-expression of recombinant
p40phox was confirmed in hygromycin-selected
K562 cells co-transfected with three episomal vectors,
pCEP4/p40phox,
pREP4/p47phox, and
pREP10/p67phox (Fig. 2, lane 3). The
diminished levels of p40phox seen in this case
were consistent with previous experience, where lower expression was
observed with co-transfection of several constructs bearing the same
selection marker (30). In the absence of
pCEP40phox, immunoblotting of K562 cells did not
reveal any endogenous p40phox stabilized by the
other recombinant oxidase components (Fig. 2, lane 4) as was
observed in mature CGD neutrophils. These results, showing the absence
of endogenous p40phox in K562 cells and
efficient production of p40phox with this
vector, established the utility of these transfectable cells for
exploring the role of p40phox in whole cell
NADPH oxidase function.
Whole Cell Oxidase Activity
Retroviral gp91-transduced K562
cells were co-transfected by electroporating 20 µg each of
pREP4/p47phox,
pREP10/p67phox, and
pCEP4/p40phox (or pCEP4CAT (control)) plasmids.
Immunoblotting of protein lysates from hygromycin-resistant cells
confirmed expression of all three recombinant oxidase proteins in these
experiments (Figs. 2 and 3). Oxidase activity in
response to PMA activation was monitored by chemiluminescence. Four
independent co-transfection experiments were performed and the results
of one representative experiment are shown in Fig. 3. A significant
inhibition of oxidase activity was observed in the presence of
p40phox, typically resulting in 35-40%
reduction in chemiluminescence signals in comparison to control
transfections with pCEP4CAT (Fig. 3). The extent of oxidase inhibition
attributed to p40phox expression was significant
when considering that this vector lacks a unique selection marker.
Since the levels of p47phox and
p67phox detected were comparable in cells
expressing p40phox or CAT (Fig. 3,
inset), variable expression of the other oxidase components
did not account for the reduced oxidase activity in the presence of
p40phox. While the yield of active oxidase in
the presence of p40phox based on the magnitude
of the chemiluminescence peak was consistently diminished, the overall
kinetic profiles of PMA-elicited respiratory bursts appeared to be
similar to control reactions, showing maximum activity 5-10 min after
stimulation.
Cell-free NADPH Oxidase Reconstitution Studies
The effects of
p40phox on NADPH oxidase activity reconstituted
in vitro were examined following preincubation of
recombinant baculovirus-derived p40phox lysates
with one or the other pure cytosolic factors
(p67phox, p47phox).
Measurements of superoxide production in a system containing recombinant p47phox,
p67phox, and p21rac1 (Q61L) showed a
significant dose-dependent inhibition of oxidase activity
associated with increasing amounts of p40phox
Sf9 cell lysates, in comparison with the same quantities of lysates from uninfected Sf9 cells (Fig. 4). These inhibitory
effects were evident irrespective of the order of addition of the other
cytosolic factors shown to bind p40phox.
Predictably, low levels of oxidase inhibition (<25%) were also seen
with excessive amounts of control uninfected lysates, although these
effects were likely due to effects of high protein concentrations on
free arachidonate levels.
Mapping of Interacting Domains within p40phox, p47phox, and p67phox using the Yeast 2-hybrid System
The yeast two-hybrid system was used to delineate
interacting domains within cytosolic oxidase proteins that were shown
to exist as a complex in resting neutrophils (5-7). This system has
advantages over other in vitro binding assays, since the
interactions between proteins occurs in vivo and no protein
purification or renaturation from SDS-PAGE gels is required (31-33).
While the system permits qualitative assessment of weak protein-protein interactions (Kd < ~106
M), quantitative binding data are not obtained by this
approach. The p40phox cDNA sequences
corresponding to various domains tested were subcloned into activation
domain vectors pGADGH or pGAD424. Coding sequences for
p67phox and p47phox
domains were engineered in frame with the Gal4 binding domain in pGBT9.
As summarized in Table I, a positive interaction was observed between full-length p40phox and
p67phox, which was abolished by deletion of 80 residues from the carboxyl-terminal end of
p40phox (p40
CT). Consistent with this,
residues 260-340 (CT) of p40phox were
sufficient to observe an interaction with full-length
p67phox. The corresponding domain of
p67phox that interacts with
p40phox was identified within a 172-residue
segment (residues 293-465) bounded by the two SH3 domains, consistent
with similar two-hybrid studies reported recently (22). The interaction
between p40CT and residues 293-465 of p67phox
was also confirmed when these cDNAs were expressed in the
reciprocal vectors (data not shown). There was no evidence for
interaction between the isolated SH3 domain of
p40phox and full-length
p67phox or p67NT (residues 1-246), although a
weak positive signal detected in a prolonged
-galactosidase assay
was observed with co-expression of full-length
p40phox and the amino-terminal domain (246 residues) of p67phox (NT) or a smaller fragment
(residues 155-246, data not shown). Others have characterized an
interaction between 67 NT and p40phox by surface
plasmon resonance, although binding of residues 293-465 of
p67phox with p40phox was
not examined in this case (24).
The interaction between p40phox and p47phox was mapped to the SH3 domain (residues 172-229) of p40phox and the proline-rich tail region (residues 351-390) of p47phox, consistent with yeast two-hybrid experiments reported elsewhere (21-23). An SH3 domain interaction between GRB-2 and human SOS-1 was also demonstrated by yeast two-hybrid expression (33). The p40phox SH3 domain interaction with the C-terminal domain of p47phox was confirmed by expression in reciprocal vectors (data not shown).
Delineation of NADPH Oxidase Inhibitory Domains of p40phoxTo explore the significance of interactions
between p40phox and other essential components
on whole cell NADPH oxidase function, truncated forms of
p40phox were co-expressed in the episomal vector
pCEP4 along with full-length p47phox and
p67phox; this work used the deleted form of
p40phox, p40CT, lacking the
p67phox binding domain (C-terminal 80 residues)
and pCEP40SH3, encoding the isolated SH3 domain that interacts with p47
phox. As shown in Fig. 5, there
was no significant difference between oxidase activity observed in
cells expressing full-length p40phox or
p40phox
CT, indicating that the oxidase
inhibitory effects of p40phox do not involve the
p67phox interaction. In contrast, expression of
the isolated SH3 domain of p40phox, which we and
others (21-23) have shown interacts with the proline-rich C terminus
of p47phox, resulted in inhibition of oxidase
activity that was even greater than observed with full-length
p40phox (60% versus 35% inhibition,
respectively).
Previous studies using purified proteins establish five core NADPH oxidase components as necessary and sufficient for reconstitution of high levels of oxidase activity in vitro (26, 27). Recently, p40phox was identified in resting neutrophil cytosol as yet another protein bound within a high molecular complex (~250 kDa) with two essential cytosolic components, p67phox and p47phox. The three proteins appear to co-translocate to the membrane as a complex that interacts with cytochrome b558 during oxidase activation (5, 9). We have now shown that p40phox can significantly inhibit oxidase activity, whether added to the other cytosolic components in a cell-free assay system or when co-transfected into K-562 cells which lack the endogenous protein.
These studies confirm and extend the findings of others by defining precise binding sites between the cytosolic oxidase factors and exploring the role of p40phox in whole cell oxidase function. Work in several laboratories suggested that the primary association in this cytosolic oxidase complex occurs between p67phox and p40phox. In vitro binding studies suggested that the entire pool of cytosolic p40phox was bound in a tight complex with p67phox in resting neutrophils (6, 7). Deficiencies in either p47phox or p67phox seen in autosomal recessive forms of CGD resulted in diminished translocation of p40phox to the membrane (9), although only the patients deficient in p67phox showed dramatically reduced levels of total cellular p40phox, suggesting that p67phox stabilizes p40phox in neutrophils.
Direct interactions have also been demonstrated between p47phox and p40phox, although these interactions were not readily detected by methods involving immobilized proteins, such as surface plasmon resonance or blotting techniques (8, 24). We detected this interaction in solution with recombinant fusion proteins and in the yeast two-hybrid system, as shown by others (21-23). Early work identified p40phox by two-hybrid interactions using p47phox as bait to screen a cDNA library from Epstein-Barr virus-transformed B-cells (23). The importance of the p47phox interaction is evident in our current findings which correlate delineation of binding sites in p40phox with whole cell oxidase function. The p40phox interaction with p47phox was critical for inhibition of NADPH oxidase activity, while deletion of the p67phox binding site in p40phox did not affect its ability to inhibit oxidase activity in whole cells. These observations were not reconciled with recent findings showing in vitro inhibition of the oxidase by antibodies directed against p40phox (C-terminal, p67phox-binding domain), although these investigators did not study the proteins directly (25).
The NADPH oxidase inhibition we observed by p40phox is thought to involve the interaction between its SH3 domain and a target within p47phox previously shown to participate in assembly of p67phox in the active oxidase complex (12-14, 16, 18). The isolated SH3 domain of p40phox inhibited whole cell oxidase activity to an extent that exceeded the inhibition by full-length p40phox. The demonstration of an interaction between the SH3 domain of p40phox and the proline-rich carboxyl-terminal domain of p47phox (residues 358-390) was supported by earlier work in which p47phox was shown to bind to a central 115-residue segment of p40phox, half of which included the SH3 domain (22, 23). More recent studies showed the SH3 domain of p40phox binds to the proline-rich tail region of p47phox and thereby inhibits binding of the C-terminal p67phox SH3 domain to the same region p47phox, when over-expressed in yeast (21). These authors suggested that competitive binding between p40phox and p67phox could modulate oxidase activity. Our observations of inhibitory effects of p40phox or its isolated SH3 domain on whole cell oxidase function lend further support to this competitive binding model.
Our results also provide additional evidence for the importance of the tail-tail interaction between p47phox and p67phox in oxidase activation. The role of this C-terminal domain interaction had been disputed because neither tail domain was necessary for cell-free reconstitution of NADPH oxidase activity (8, 13). However, work in whole transfected cells showed that deletions or mutations in the carboxyl-terminal domains of either protein dramatically affected whole cell oxidase activity and p67phox membrane translocation (13, 16, 18). Since the tail-tail interaction is not essential for cell-free oxidase activity and cell-free oxidase inhibition was not influenced by the order of addition of the three cytosolic proteins, it is unclear whether p40phox inhibits the cell-free system by the same mechanism suggested from the whole cell studies.
In summary, we have examined the role of interactions between
p40phox and two other essential oxidase
components, both in vitro and in transfected cells where
expression of NADPH oxidase components could be genetically
manipulated. Based on these findings, we suggest a model outlined in
Fig. 6 in which clear distinctions can be made in terms
of function between two interactions of p40phox.
In one case, the interaction of p40phox with
p67phox appears to play no obvious role in the
oxidase activation process, although its importance is evident in
neutrophils where p40phox stability is affected
by the absence of p67phox seen in CGD (5, 7-9).
In the second, the interaction between p47phox
and p40phox is not readily detected in resting
neutrophil cytosol but was shown to play a predominant role during
activation through its inhibition of productive interactions between
two other essential oxidase components. These findings again illustrate
the central role played by multiple SH3 domain interactions in
modulation of NADPH oxidase activity. In this case, we demonstrated
down-regulation by an accessory SH3 domain-containing protein
(p40phox), which by virtue of its homology to
one essential oxidase component (p67phox)
competes for a common target site on another. This SH3 target site in
p47phox is flanked by sites that were recently
shown to be phosphorylated during oxidase activation (35). Future work
will examine whether phosphorylation alters binding affinities of these
competing SH3 domains and whether these changes are critical in the
activation process.
We thank Dr. C. K. Kwong for providing the Rac61L vector and Dr. H. L. Malech for providing the gp91phox retrovirus.