Mutagenesis of p22 phox Histidine 94

The NADPH oxidase is a multicomponent enzyme that transfers electrons from NADPH to O2 to generate superoxide (O 2 ⨪ ), the precursor of microbicidal oxygen species that play an important role in host defense. Flavocytochromeb 558, a heterodimeric oxidoreductase comprised of gp91 phox and p22 phox subunits, contains two nonidentical, bis-histidine-ligated heme groups imbedded within the membrane. Four histidine residues that appear to serve as noncovalent axial heme ligands reside within the hydrophobic N terminus of gp91 phox , but the role of p22 phox in heme binding is unclear. We compared biochemical and functional features of wild type flavocytochrome b 558 with those in cells co-expressing gp91 phox with p22 phox harboring amino acid substitutions at histidine 94, the only invariant histidine residue within the p22 phox subunit. Substitution with leucine, tyrosine, or methionine did not affect heterodimer formation or flavocytochrome b 558 function. The heme spectrum in purified preparations of flavocytochromeb 558 containing the p22 phox derivative was unaffected. In contrast, substitution of histidine 94 with arginine appeared to disrupt the intrinsic stability of p22 phox and, secondarily, the stability of mature gp91 phox and abrogated O 2 ⨪ production. These findings demonstrate that His94 p22 phox is not required for heme binding or function of flavocytochrome b 558 in the NADPH oxidase.

The NADPH oxidase is a multicomponent enzyme that transfers electrons from NADPH to O 2 to generate superoxide (O 2 . ), the precursor of microbicidal oxygen species such as hydrogen peroxide and hypochlorous acid that play an important role in host defense. The oxidase is comprised of four cytosolic protein subunits, p67 phox , p47 phox , p40 phox , and Rac2, and the membrane-associated flavocytochrome b 558 , a heterodimer of gp91 phox and p22 phox , two integral membrane proteins. Accumulated studies have established a model for the assembly and activation of the oxidase by which the cytosolic oxidase protein components, segregated from the membrane-associated oxidase proteins in resting phagocytes, translocate to flavocytochrome b 558 following an inflammatory stimulus, thereby forming the catalytically active enzyme complex. The physiological significance of the phagocyte NADPH oxidase is illustrated in patients with the inherited disorder chronic granulomatous disease (CGD), 1 who suffer from recurrent bacterial and fungal infections caused by impaired phagocyte killing. CGD results from a genetic defect in any one of the four essential NADPH oxidase components, p47 phox , p67 phox , gp91 phox , or p22 phox . The redox center of the NADPH oxidase is flavocytochrome b 558 (1)(2)(3)(4). In phagocytic cells, the gp91 phox and p22 phox subunits co-purify unless denatured, indicating a close association with one another (5). Heterodimer formation occurs post-translationally (6). Inhibition of heme biosynthesis results in a marked decrease in flavocytochrome expression, (6,7), suggesting that heme incorporation influences heterodimer formation. Mutations in CGD patients resulting in loss of expression of either flavocytochrome b 558 subunit lead to a marked decrease in expression of the other (8), indicating that heterodimer formation is important for the normal stability of gp91 phox and p22 phox in phagocytic cells. Thus, it is difficult to distinguish among mutations that directly affect heme incorporation, those that impair heterodimer formation by other mechanisms, and those that render the protein intrinsically unstable.
The gp91 phox subunit of flavocytochrome b 558 , a membrane glycoprotein that is a member of the Nox flavocytochrome family (9), contains binding sites in its C-terminal portion for the electron donor NADPH and the redox intermediate FAD (2,4,10). The two nonidentical, bis-histidine-ligated heme groups of flavocytochrome b 558 are imbedded within the membrane, and four histidine residues that appear to serve as noncovalent axial heme ligands reside within the hydrophobic N terminus of gp91 phox (11)(12)(13)(14). The role of the p22 phox subunit in O 2 .
production is less well characterized. A proline-rich motif in the C terminus of p22 phox functions as an important site of interaction with an SH3 domain in the cytosolic oxidase protein p47 phox (15)(16)(17). Although some have proposed that p22 phox also plays a role in heme binding, this hypothesis is controversial. The p22 phox subunit contains a single invariant histidine residue (18) that has been proposed to serve as an axial ligand for shared binding of one of the two heme groups between gp91 phox and p22 phox (12). This histidine residue, located at codon 94, is conserved in human, bovine, porcine, murine, and rattus (19) and lies within a region with sequence similarity to cytochrome c oxidase and myoglobin (20). Quinn et al. (12) reported that gp91 phox and p22 phox each bind heme based on heme staining detected at ϳ22 kDa after electrophoresis of flavocytochrome b 558 on lithium dodecylsulfate polyacrylamide gels. Indirect evidence of a role for p22 phox His 94 in heme binding came from the identification of a CGD patient with a p22 phox H94R point mutation that resulted in undetectable levels of flavocytochrome b 558 (21). However, we previously reported evidence indicating that both of the two heme groups of flavocytochrome b 558 are entirely located with gp91 phox based on analysis of membranes prepared from a Cos7 cell line expressing transgenic gp91 phox in the absence of p22 phox (22). More recently, four histidines that appear to be involved in axial heme ligation were identified within gp91 phox by sitedirected mutagenesis (14), and histidines at these positions are conserved in all other NOX family members (9). To investigate the basis whereby arginine substitution of p22 phox His 94 results in absence of flavocytochrome b 558 expression and to definitively determine whether or not His 94 is involved in heme binding and flavocytochrome b 558 function, we exploited nonphagocytic cell lines that are void of endogenous p22 phox expression to express site-directed mutants of p22 phox at position 94. Herein we report that His 94 of p22 phox is not essential for NADPH oxidase function, heme binding, or flavocytochrome b 558 assembly in that substitution with leucine has no effect on these parameters. In contrast to replacement of His 94 with leucine, methionine, or tyrosine, arginine substitution appears to disrupt the intrinsic stability of p22 phox ; therefore, the decreased expression of flavocytochrome b 558 in H94R-CGD appears to be a secondary consequence of disturbed p22 phox expression.

EXPERIMENTAL PROCEDURES
Materials-Fetal bovine serum was purchased from Hyclone Laboratories (South Logan, UT). Dulbecco's modified Eagle's medium/Ham's F-12 medium, F12K medium, and RPMI 1640 tissue culture medium were purchased from Invitrogen. Monoclonal antibodies 449 and 48 specific for p22 phox and gp91 phox , respectively, were kindly provided by A. Verhoeven and D. Roos (Central Laboratory of the Netherlands Blood Transfusion Service, University of Amsterdam, Amsterdam, the Netherlands). The polyclonal anti-p47 phox and polyclonal anti-p67 phox antibodies were kindly provided by D. Lambeth (Emory University, Atlanta, GA). Monoclonal anti-Rac-1 antibody was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). All of the other reagents were purchased from Sigma unless otherwise specified.
Cell Lines-The monkey Cos7 cell line expressing transgenic human gp91 phox , Cos 91, was made previously by stable transfection of the full-length gp91 phox cDNA into parental Cos7 wild type cells and was maintained as described (22). The NIH3T3 cell line was maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum and was stably transduced with an MFG retroviral vector containing the gp91 phox cDNA (vector kindly provided by Adrian Thrasher, University College, London, UK). The CHO K1 line was maintained in F12K medium supplemented with 10% fetal calf serum and sodium bicarbonate (Invitrogen) and was stably transfected with a full-length gp91 phox cDNA cloned into the NotI site of pEF pac (23) using Lipo-fectAMINE 2000 (Invitrogen) according to the manufacturer's instructions. The RAW cell line, a murine macrophage cell line (obtained from Donald Durden, Indiana University School of Medicine, Indianapolis, IN), was maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. A human promyelocytic cell line (PLB-91), a derivative of the PLB-985 cell line with a gene-targeted disruption of the endogenous CYBB and constitutive expression of a wild type gp91 phox transgene, was maintained as described (23).
Site-directed Mutagenesis and Expression Plasmids-A series of p22 phox point mutations at His 94 were introduced into a full-length wild type p22 phox cDNA cloned into the EcoRI site of Bluescript KS (ϩ) (Stratagene, La Jolla, CA) using the Sculptor in vitro mutagenesis system (Amersham Biosciences) according to the manufacturer's instructions. Mutant p22 phox cDNAs were verified by dideoxynucleotide sequencing (performed by the Biochemistry Biotechnology Facility, In-diana University, Indianapolis, IN) and were subcloned into the pEF-PGKneo mammalian expression vector. The mutant p22 phox constructs were resequenced to confirm the mutations after subcloning and were transfected into Cos 91 and 3T3 91 cells using LipofectAMINE Plus (Invitrogen) and into CHO 91 cells using LipofectAMINE 2000 (Invitrogen) according to the manufacturer's instructions. All of the preparations and manipulations of plasmids were performed using standard protocols (24). Following transfection, the clones were selected by limiting dilution in 1.6 mg/ml G418 (Invitrogen) and were screened for p22 phox expression by immunoblot. To minimize any clone to clone variation in recombinant p22 phox expression or NADPH oxidase activity, three or four independent clones determined to express transgenic p22 phox were pooled and used for subsequent analysis. In some experiments, wild type p22 phox or H94R p22 phox cDNA constructs were harvested 18 h post-transfection and analyzed for transient expression of p22 phox .
For transient expression in selected cell lines, p47 phox cloned into the NotI site of the pRK5 expression vector and p67 phox cloned into the pcDNA3.1 expression vector (kindly obtained from David Lambeth, Emory University, Atlanta, GA) were transfected into CHO derivatives using LipofectAMINE 2000 (Invitrogen). The cells were harvested for further analysis 18 h post-transfection.
Protein Expression and Immunoblot Analysis-Proteins were extracted as described previously (14) and analyzed by SDS-PAGE. Following transfer to nitrocellulose membranes, the blots were probed with monoclonal antibodies 49 and 449 for expression of recombinant gp91 phox and p22 phox , respectively (25). In some experiments the blots were also analyzed with a polyclonal anti-p22 phox antibody, which recognizes both murine and human p22 phox (18).
Northern Blot Analysis-Northern blot analysis was performed as described previously. Briefly, total RNA was extracted from cells using RNAzol B (Tel-Test, Inc., Friendswood, TX) according to the manufacturer's instructions. The RNA was electrophoresed on a formaldehyde agarose gel and transferred to nylon membrane by capillary action overnight at room temperature. The membrane was probed with radio-

FIG. 2. Analysis of endogenous p22 phox expression in 3T3 and CHO cells.
A, RNA extracts were resolved by Northern blot and probed with radiolabeled full-length human p22 phox cDNA. Extracts from RAW cells, a murine macrophage cell line, are included as a positive control for p22 phox expression. The results are representative of two independent experiments. B, RT-PCR was performed on total RNA extracts with primers specific for murine p22 phox cDNA or with primers specific for a sequence of p22 phox common to both murine and human p22 phox . RT-PCR was repeated with primers for HPRT as a control for RNA quality and equal loading. The results are representative of two independent experiments. labeled full-length human p22 phox cDNA and exposed to film for various amounts of time before developing in an X-Omat processor. The membranes were stripped and reprobed with cDNA for ␤-actin to control for equal loading.
RT-PCR-RT-PCR was performed on total RNA extracted as described above using a kit obtained from Promega (Madison, WI) according to the manufacturer's instructions with primers specific for murine p22 phox cDNA or with primers specific for a sequence of p22 phox common to both murine and human. RT-PCR was repeated with primers for HPRT as a control for RNA quality and equal loading.
Measurement of NADPH Oxidase Activity-Superoxide dismutaseinhibitable NADPH oxidase activity was measured in transgenic cell lines following stimulation with arachidonic acid as described previously (26). The assays were performed at 37°C using a Thermomax microplate reader (Molecular Devices, Inc., Sunnyvale, CA) to measure the reduction of cytochrome c resulting from the production of O 2 .. An extinction co-efficient of 21.1 mM Ϫ1 cm Ϫ1 for the reduced form of cytochrome c was used to calculate the amount of O 2 . generated by the transgenic cells. The data were analyzed using SOFTMAX version 2.02, and the V max was calculated over a 6-min interval. Statistical analyses were performed using GraphPad Instat, version 2.0 (GraphPad Software Inc., San Diego, CA). Alternatively, qualitative assessment of NADPH oxidase activity in single cells was performed using the nitro blue tetrazolium test, as described previously (23). For some transgenic cell lines, a cell-free assay was used to measure electron transfer and O 2 . production. The transfer of electrons from NADPH to FAD was determined by the iodonitrotetrazolium (INT) violet diaphorase assay as described (27). For the cell-free analysis of NADPH oxidase activity, the superoxide dismutase-inhibitable cytochrome c reduction was measured in triplicate in 96-well microtiter plates as described previously (28). Neutrophil cytosol was employed as a source of cytosolic oxidase proteins in combination with membranes prepared from transgenic Cos7 cell lines as described previously (29) or partially purified and relipidated flavocytochrome b 558 from CHO cell lines, as described previously (22,30). Protein concentration was determined with BCA reagents (Pierce) according to the manufacturer's instructions.
Analysis of Heme Spectrum-The membranes were harvested, and flavocytochrome b 558 was partially purified and relipidated as described previously (22,30). Dithionite-reduced minus oxidized difference spectroscopy was performed using a PerkinElmer Lambda 18 spectrophotometer. Flavocytochrome b 558 heme content was estimated using an extinction co-efficient of 21.6 mM Ϫ1 cm Ϫ1 at 559 nm, as described previously (31).

Arginine, but Not Other Amino Acid Substitutions of p22 phox His 94 , Affects the Stable Expression of Flavocytochrome b 558 in
Cos7 Cells-To investigate the role of p22 phox His 94 in flavocytochrome b 558 function, we first utilized a p22 phox -deficient Cos7 cell line that was identified in a previous study (22) to express engineered mutant derivatives of p22 phox in a null background. We used site-directed mutagenesis to introduce point substitutions into p22 phox cDNA at codon 94, the only invariant histidine residue within p22 phox . His 94 was mutated FIG. 3. Flavocytochrome b 558 expression and heme spectra in transgenic 3T3 or CHO cells expressing recombinant p22 phox mutants. A, Triton X-100 extracts, derived from wild type 3T3 or CHO cells and 3T3 91 or CHO 91 cells stably expressing recombinant wild type or mutant p22 phox , as indicated, were resolved by SDS-PAGE and immunoblotted using mAb 449 (upper panels), which detects p22 phox , or mAb 48 (lower panels), which detects gp91 phox . Triton X-100 extracts derived from X-CGD PLB-985 cells stably expressing recombinant wild type gp91 phox (PLB-91) were included as positive controls for gp91 phox expression. The results are representative of at least three independent experiments. B, Triton X-100 extracts derived from wild type CHO cells, CHO 91 cells, and derivative cell lines transiently expressing recombinant wild type or mutant p22 phox , as indicated, were resolved by SDS-PAGE and immunoblotted using mAb 449 which detects p22 phox . The results are representative of two independent experiments. C, solubilized samples of partially purified flavocytochrome b 558 prepared from membranes of CHO cell lines, as indicated, were analyzed by difference spectroscopy as described under "Experimental Procedures." The data are representative of duplicate analyses from two independent membrane preparations.
to tyrosine or methionine residues, which are alternative heme ligands in some cytochromes (32); leucine, which is nonpolar and therefore likely to be well tolerated within the hydrophobic domain in which His 94 is located; or arginine, because it mimics a mutation previously identified in a p22 phox -deficient CGD patient.
We examined the expression of recombinant p22 phox after stable transfection of wild type or mutant p22 phox cDNA constructs into the Cos7 cell line in the absence of gp91 phox . We had previously observed that the individual flavocytochrome subunits can be stably expressed in the absence of the other in this heterologous cell line (22), in contrast to phagocytes. Protein levels of the p22 phox mutants H94L and H94Y were similar to wild type p22 phox as determined by Western blot analysis (Fig. 1A), whereas levels of the p22 phox mutant H94M were moderately decreased as compared with wild type p22 phox (Fig.  1A). However, substitution of the His 94 with arginine resulted in a markedly decreased level of p22 phox expression (Fig. 1A), suggesting that this point mutation disrupted the overall stability of the p22 phox polypeptide. Consistent with this, Northern blot analysis showed that expression of vector-derived H94R p22 phox mRNA was similar to other His 94 p22 phox mutants (data not shown).
We also analyzed expression of wild type or mutant p22 phox cDNA constructs in Cos 91 cells, in which recombinant gp91 phox is constitutively expressed from a stable transgene in a Cos7 cell line deficient in endogenous p22 phox (22). Co-expression of gp91 phox and wild type p22 phox has been shown to increase the expression of gp91 phox in PLB-985 and Cos7 cells and to in-crease maturation of gp91 phox from the 65-kDa precursor to the fully glycosylated ϳ 91-kDa form (33). As in the p22 phox -transfected Cos7 cells, Western blot analysis showed that protein levels of the p22 phox mutants H94Y (Fig. 1B), H94L, or H94M (data not shown) were similar to wild type p22 phox when expressed in Cos 91 cells, but the H94R substitution dramatically decreased the level of p22 phox protein expression (Fig. 1B). Furthermore, in contrast to Cos 91 cells expressing H94Y p22 phox (Fig. 1B), H94L, or H94M (data not shown), co-expression of mutant p22 phox H94R failed to increase maturation of transgenic gp91 phox from its 65-kDa precursor to its fully glycosylated 91-kDa form, as shown by Western blot (Fig. 1B). This carbohydrate maturation step occurs in the endoplasmic reticulum and Golgi compartments of granulocytes after association of p22 phox with the 65-kDa high mannose precursor of gp91 phox (33). Taken together, these data suggest that a histidine at position 94 does not appear to be required for p22 phox to interact with gp91 phox . However, an arginine substitution at position 94 results in intrinsic p22 phox instability and, secondarily, an instability of gp91 phox .
Identification of Other p22 phox -deficient Cell Lines-To confirm that substitution of p22 phox His 94 does not necessarily affect the stable expression of flavocytochrome b 558 , we identified other nonphagocytic cell lines that were devoid of p22 phox expression. Northern blot analyses on NIH3T3 mouse fibroblasts and CHO cells established the absence of p22 phox mRNA ( Fig. 2A). Confirmation of the absence of p22 phox mRNA in these two cell lines was obtained by RT-PCR using murinespecific p22 phox primers (Fig. 2B) and p22 phox -specific primers common for both human and murine p22 phox (data not shown) because hamster-specific p22 phox primers are unavailable. Equal loading and integrity of RNA samples was confirmed by RT-PCR using HPRT primers (Fig. 2B) Transgenic wild type p22 phox or H94L p22 phox constructs were stably co-expressed with transgenic wild type gp91 phox in 3T3 cells and CHO cells. As in the Cos7 cell derivatives, expression levels of p22 phox H94L were similar to wild type p22 phox when expressed in 3T3 91 cells or CHO 91 cells (Fig.  3A) as determined by Western blot analysis. Likewise, expression of either H94L p22 phox or wild type p22 phox promoted maturation of transgenic gp91 phox from its 65-kDa precursor to its fully glycosylated 91-kDa form in both 3T3 91 and CHO 91 cells (Fig. 3A). There was a modest reduction in the amount of mature gp91 phox in CHO 91.22 H94L cells as compared with CHO 91.22 cells (Fig. 3A), suggesting that His 94 may have a subtle influence on assembly of flavocytochrome b 558 heterodimer in this cell background. However, the p22 phox H94L mutation does not appear to affect flavocytochrome b 558 expression in 3T3 cells (Fig. 3A) or Cos7 cells (data not shown). We also expressed H94R p22 phox in the CHO cell derivatives, both in the presence and the absence of the gp91 phox subunit. As in the Cos7 cells, H94R p22 phox was expressed at a markedly decreased level in both CHO cells FIG. 4. Cell-free analysis of NADPH oxidase activity in transgenic Cos7 cells expressing recombinant p22 phox mutants. NADPH oxidase activity was measured in a cell-free oxidase assay as described under "Experimental Procedures" using membranes isolated from transgenic Cos cells, as indicated, and neutrophil cytosol. The data are representative of three independent experiments. and CHO 91 cells (Fig. 3B), supporting the hypothesis that arginine substitution leads to intrinsic p22 phox instability. Effect of Amino Acid Substitution of p22 phox His 94 on Flavocytochrome b 558 Heme Spectrum-To determine whether the heme spectrum of recombinant flavocytochrome b 558 was altered by the H94L substitution in p22 phox , we analyzed the oxidation reduction difference spectra for partially purified flavocytochrome b 558 derived from transgenic CHO cell lines. The membranes from CHO cells expressing wild type p22 phox and wild type gp91 phox exhibited a characteristic flavocytochrome b 558 spectrum with peaks at 426 and 558 nm (Fig. 3C). The heme spectrum of the partially purified gp91 phox /p22 phox H94L heterodimer was qualitatively similar to that of the wild type flavocytochrome b 558 , although the total heme content was modestly decreased (Fig. 3C and Table I), consistent with the decreased level of gp91 phox protein expression seen on the immunoblot (Fig. 3A).
Expression of Mutant Transgenic p22 phox Does Not Affect Superoxide Production in Heterologous Cells-A cell-free NADPH oxidase assay was performed to measure superoxidegenerating activity in membranes prepared from transgenic flavocytochrome b 558 -expressing Cos7 cells when combined with neutrophil cytosol. Although, as previously shown, membranes from Cos 91 cells did not support O 2 . production (22) (Fig. 4), co-expression of H94L p22 phox or H94Y p22 phox with gp91 phox supported NADPH oxidase activity at a level similar to that of membranes derived from cells expressing wild p22 phox and gp91 phox (Fig. 4). These results indicate that a histidine at position 94 in p22 phox is not required for NADPH oxidase activity in this system. Not surprisingly, membranes from cells expressing H94R p22 phox did not support O 2 . production, consistent with the effect of this mutation on p22 phox polypeptide stability (Fig. 1) and, secondarily, gp91 phox expression (Fig. 1B). We also examined diaphorase and NADPH oxidase activity using neutrophil cytosol and purified flavocytochrome b 558 isolated from CHO cells expressing either wild type gp91 phox /p22 phox or gp91 phox /p22 phox H94L. The rate of transfer of electrons from NADPH to FAD, as determined by the INT diaphorase assay, and the formation of O 2 . by purified flavocytochrome b 558 from both CHO 91.22 and CHO 91.22 H94L membranes was similar to that seen previously with purified neutrophil flavocytochrome b 558 . Diaphorase activity, relative to heme content, was unaffected by the H94L p22 phox mutation (Table I), as was O 2 . production ( Table I). The rate per milligram of protein was decreased, consistent with the lower level of gp91 phox protein expression in CHO91.22 H94L cells (Fig. 3A). Thus, His 94 p22 phox does not appear to be directly involved in electron transport by flavocytochrome b 558 in the NADPH oxidase.
We also analyzed whole cell NADPH oxidase activity in CHO cells expressing recombinant flavocytochrome b 558 . We previously established that high level NADPH oxidase activity can be elicited by either arachidonic acid or phorbol myristate acetate in Cos7 cells expressing transgenic gp91 phox , p22 phox , p47 phox , and p67 phox (26). CHO derivatives that stably expressed transgenic wild type gp91 phox along with either transgenic wild type or H94L p22 phox were transiently transfected with vectors for expression of wild type p47 phox and wild type FIG. 5. Expression and function of recombinant oxidase proteins in CHO cells. A, Triton X-100 extracts derived from wild type CHO and CHO 91 cells stably expressing recombinant wild type or mutant p22 phox and transiently expressing recombinant wild type p47 phox and p67 phox , as indicated, were resolved by SDS-PAGE and immunoblotted using mAbs 449 and 48, which detect p22 phox and gp91 phox , respectively, and polyclonal antibodies that recognize p67 phox or p47 phox . Western blots were also immunoblotted with polyclonal anti-Rac 1 Ab as a loading control. Triton X-100 extracts derived from untransfected CHO cells and from Cos phox cells were included as negative and positive controls, respectively, for expression of oxidase proteins. The results are representative of four independent experiments. B, arachidonic acidstimulated O 2 . production was measured by a continuous cytochrome c reduction assay in intact CHO cells expressing recombinant wild type p47 phox and p67 phox in the presence or absence of recombinant wild type or p22 phox H94L-containing flavocytochrome b 558 , as indicated. The data are from one of four representative experiments.
p67 phox . After 18 h, the cells were harvested, and NADPH oxidase protein expression was verified by Western blotting (Fig. 5A). Arachidonic acid-elicited O 2 . production was measured by the cytochrome c reduction assay (Fig. 5B) (Fig.  5B). The nitro blue tetrazolium assay confirmed that H94L p22 phox supported whole cell NADPH oxidase activity in the presence of gp91 phox , p47 phox , and p67 phox (data not shown). Thus, the invariant histidine residue at position 94 of p22 phox is not required for O 2 . production in whole cells, consistent with the results from cell-free assays.

DISCUSSION
The phagocyte NADPH oxidase plays a critical role in the innate immune response by producing O 2 . , which is then con-verted to additional reactive oxygen species important for microbial killing. Flavocytochrome b 558 , a heterodimer comprised of two oxidase subunits, p22 phox and gp91 phox , serves as the redox center of the oxidase. The gp91 phox subunit contains regions that are homologous to the flavin-and NADPH-binding domains of ferredoxin NADPϩ reductase (34). It also contains two nonidentical heme groups with low midpoint redox potentials of Em7 ϭ Ϫ265 mV and Em7 ϭ Ϫ225 mV, respectively (11). Raman and electron paramagnetic resonance spectroscopy predicts that the hemes reside within the membrane in a hexacoordinated state with axial imidazole or imidazolate ligands supplied by histidine residues (35)(36)(37). Currently, there are conflicting views about the distribution and the exact location and coordination of the hemes within the flavocytochrome. Quinn et al. (12) suggested that gp91 phox and p22 phox each bind heme and proposed that one heme binds exclusively to gp91 phox , whereas the other heme is shared between gp91 phox and p22 phox . However, more recent evidence indicates that both hemes reside within gp91 phox (14,22). The p22 phox subunit of flavocytochrome b 558 has only a single invariant histidine at codon 94. Herein we provide evidence suggesting that although mutations in His 94 of p22 phox can affect stability of the p22 phox polypeptide, a histidine residue is not required for flavocytochrome b 558 heme binding or electron transfer.
We have previously observed that inhibition of heme biosynthesis with succinyl acetone results in markedly decreased expression of p22 phox and the mature 91-kDa form of gp91 phox , but not its 65-kDa high mannose precursor, suggesting that heme incorporation is required for heterodimer formation (6,7,33). It is therefore likely that heme insertion may impart a structural constraint to gp91 phox necessary for it to bind p22 phox . It has also been proposed that gp91 phox and p22 phox are linked directly through heme (12). We have also demonstrated that point mutation of any one of four histidine residues at positions 101, 115, 209, or 222 in gp91 phox disrupts heterodimer formation, maturation of gp91 phox , and NADPH oxidase activity. These results strongly suggest that these four histidines are the heme-binding ligands in flavocytochrome b 558 and that disruption of any individual heme ligand interferes with normal flavocytochrome b 558 maturation and oxidase function. In the current study, we provide evidence that point substitution of His 94 with leucine, methionine, or tyrosine does not affect expression of transgenic p22 phox and promotes maturation of gp91 phox in Cos7, 3T3, and CHO cells, which do not express endogenous p22 phox . Hence, mutation of p22 phox His 94 does not interfere with heterodimer formation. Furthermore, reduced minus oxidized spectra of partially purified flavocytochrome b 558 comprised of wild type gp91 phox and H94L p22 phox was qualitatively similar to that of the wild type flavocytochrome b 558 , thus providing direct evidence that p22 phox His 94 is not involved in heme binding.
NADPH oxidase-catalyzed O 2 . production involves the transfer of two electrons from NADPH to FAD, which are then transferred as single electrons to O 2 via two nonidentical bishistidine-ligated heme groups (11 558 . A, Kite-Doolittle plot of the p22 phox polypeptide calculated using a window of 19 amino acids. The p22 phox polypeptide contains three hydrophobic regions that are predicted to be transmembrane ␣ helices, indicated by boxes (for algorithms, see genome.cbs.dtu.dk). B, the hydropathy profiles of segments of p22 phox from residues 91-97 with amino acid substitutions for the histidine residue at position 94, as indicated. A three-point weighted moving average was used for determination of local hydropathic character. data in intact cells confirm that the histidine residue located at position His 94 of p22 phox is not required for O 2 . production.
Although our data show that p22 phox His 94 is not necessary for flavocytochrome b 558 expression, heme binding, or NADPH oxidase function, we observed that substitution of His 94 p22 phox with an arginine residue resulted in substantially decreased expression of recombinant p22 phox in either Cos7 or CHO cells. Expression of H94R p22 phox , which mimics a mutation previously identified in a CGD patient, was also associated with decreased expression of mature gp91 phox , and O 2 . production was virtually undetectable when transgenic Cos 91.22 H94R membranes were used in a cell-free assay of NADPH oxidase activity. Taken together, we infer that arginine substitution results in intrinsic p22 phox instability and, secondarily, instability of gp91 phox when expressed in human phagocytes, which leads to deficient NADPH oxidase activity and thus CGD. The p22 phox polypeptide contains three hydrophobic regions that are predicted to be transmembrane ␣ helices (Fig. 6A; for algorithms, see genome.cbs.dtu.dk). The histidine residue at position 94 is adjacent to a predicted transmembrane helix that spans residues 95-117. The hydropathy profile of the His 94containing segment of p22 phox from residues 91-97 (Fig. 6B) demonstrates that leucine and tyrosine have higher hydropathy indices than histidine; hence, substitution of His 94 with either leucine or tyrosine may have less potential to disrupt the stability and function of the p22 phox polypeptide. However, when an arginine residue, which has a lower hydropathy index than histidine, is substituted for His 94 adjacent to the predicted transmembrane ␣ helix, this region of p22 phox is now predicted to be hydrophilic. An arginine substitution may therefore produce more substantial conformational changes in p22 phox , which could result in misfolding and subsequent instability of p22 phox, leading to absent flavocytochrome b 558 expression and CGD.