Changing the Conformation State of Cytochrome b 558 Initiates NADPH Oxidase Activation REGULATION*

Phagocyte NADPH oxidase generates O 2 . for defense mechanisms and cellular signaling. Myeloid-related proteins MRP8 and MRP14 of the S100 family are EF-hand calcium-binding proteins. MRP8 and MRP14 were co-isolated from neutrophils on an anti-p47 phox matrix with oxidase cytosolic factors and identified by mass spectrometry. MRP8 and MRP14 are absent from Epstein-Barr virus-immortalized B lymphocytes, and, coinciden-tally, these cells display weak oxidase activity compared with neutrophils. MRP8/MRP14 that was purified from neutrophils enhanced oxidase turnover of B cells in vitro , suggesting that MRP8/MRP14 is involved in the activation process. This was confirmed ex vivo by co-transfection of Epstein-Barr virus-transformed B lymphocytes with genes encoding MRP8 and MRP14. In a semi-recombinant cell-free assay, recombinant MRP8/ MRP14

The phagocyte respiratory burst depends on the NADPH reduction of molecular oxygen and generation of superoxide anion O 2 . upon activation of the cell with soluble (chemotactic factors) or particulate (bacteria or fungi) stimuli (1). The O 2 . -generating NADPH oxidase (EC 1.6.99.6) is a heterogeneous complex compartmentalized in resting cells, between plasma membrane, cytosol, and the cytoskeleton, which assembles at the plasma membrane level once activated. Cytochrome b 558 is the membrane redox core of the oxidase complex, which, after activation, transfers electrons from NADPH to molecular oxygen. It is constituted of gp91 phox (the ␤-catalytic subunit) and p22 phox (␣-subunit), which is necessary in phagocytes to stabilize the 1/1 (␣/␤) heterodimer (2). Three soluble factors, p67 phox , p47 phox , and p40 phox , participate in the activation process, as does the monomeric G protein Rac1/2. Another monomeric G protein, Rap1A, has been identified in the membrane and proposed as a possible regulation factor of the oxidase complex (3,4). In phagocytes, O 2 . and derivatives are synthesized in large amounts and used as bactericidal tools, but uncontrolled production is also injurious to tissues and triggers inflammatory reactions (5,6). Phagocyte NADPH oxidase activity is transitory; it is submitted to a strict up-and down-regulation process. Up to now, oxidase activation has been investigated, but the molecular support of down-regulation has not. In fact, the role of a membrane-associated GTPase-activating protein has been described and related to the deactivation of NADPH oxidase (7,8). Recent observations would support an allosteric regulation of NADPH oxidase activity where cytochrome b 558 is the catalytic subunit. In the oxidase assembly, p67 phox is the limiting factor (9 -11). The binding of p67 phox to cytochrome b 558 mediates the transition from an inactive to an active conformation of cytochrome b 558 (11). In the process, p47 phox and Rac1/2 are introduced as positive effectors, whereas the p40 phox function remains undetermined. In chronic granulomatous disease (CGD) 1 , there is no oxidase activity (12).
All the components of the phagocyte NADPH oxidase complex are expressed in Epstein-Barr virus-immortalized B lymphocytes. However, in these cells, the activity of stimulated oxidase is 100 times inferior to that of human neutrophils (13). Both neutrophils and EBV-B cells are qualitatively similar in terms of the nature of their oxidase constituents, but they differ in the concentration and stoichiometry of the cytosolic factors compared with cytochrome b 558 (14). We have recently demonstrated that the weak oxidase activity measured in intact EBV-B lymphocytes was not only a result of the low expression of cytochrome b 558 , but also of a possible defect or constraint in the assembly process because of an unfavorable membrane environment (15). In fact, once cytochrome b 558 is extracted from the membrane of both cell types, the turnover of reconstituted NADPH oxidase is similar. It is also possible that such a low activity may result from down-regulation because regulatory factors normally present in phagocytes are absent in the B cells.
Myeloid-related proteins, MRP8 and MRP14, are two proteins of the S100 family (S100A8 and S100A9, respectively) expressed by myeloid cells and some secretory epithelium (16). In myeloid cells, they represent ϳ45% of the cytosolic proteins in neutrophils (17) and 1% in monocytes. MRP8 and MRP14 are not expressed by resident macrophages, although they may be present in macrophages of inflammatory lesions (18). MRPs are known to be markers of inflammation. Their expression by infiltrating neutrophils was assumed to reflect activation stages (19). In inflammation, high levels of MRP8 and MRP14 were found in the extracellular medium of stimulated neutrophils (20) and these proteins could be markers of inflammatory diseases such as arthritis and ulcerative colitis (17,21). Whether the complex is released by dying cells or is a result of active secretion has not been clarified. The two proteins are deposited onto endothelium venules through a specific MRP14heparan sulfate proteoglycan interaction (22). It has been postulated that MRP8 and MRP14 participate in neutrophil migration, phagocytosis, and activation (21). However, their specific function remains unclear. MRP8 and MRP14 are believed to be present as a 1/1 non-covalent heterodimer, the process of dimer formation being calcium-dependent. Four MRP14 isoforms have been identified, although this is not the case for MRP8 (23). Several studies have reported that both proteins also associate with membrane and cytoskeleton in a calcium-dependent manner (24).
Recent studies have suggested that MRP8 and MRP14 play a role in potentiating activation of O 2 . -generating oxidase in bovine neutrophils (25). Doussière et al. (26) also reported that MRP8/MRP14 interacts preferentially with the cytosolic factor p67 phox , which translocates to plasma membrane upon stimulation. Moreover, various phosphorylated states of MRPs may discriminate subcellular compartmentation (27). Using a proteomics approach, we identified MRP8 and MRP14 associated with oxidase cytosolic factors in a complex isolated from neutrophil cytosol. In contrast, there was neither MRP8 nor MRP14 in the complex isolated from EBV-B cell cytosol. MRP8/MRP14 purified from human neutrophils and recombinant MRP8/MRP14 were shown to complement the cytosol of EBV-B cells and to restore a full oxidase activity, as measured in human neutrophils. These observations were confirmed ex vivo after MRP-encoding gene transfection. We demonstrated further that this effect was mediated through a specific interaction with cytochrome b 558 and a change of conformation that initiated oxidase activation.
Lymphoid Cell Lines and Neutrophils-Citrate-sterile venous blood was drawn from healthy patients after obtaining their informed consent. Neutrophils from buffy coats and B lymphocytes were isolated according to previously used methods (31). Lymphocytes were immortalized with the B95-8 strain of Epstein-Barr virus. The EBV-B lymphocyte cell lines were kept in culture using RPMI 1640 supplemented with 10% (v/v) fetal calf serum, 2 mM L-glutamine at 37°C in 5% CO 2 atmosphere. The cytosolic fractions from the 3 mM diisopropyl fluorophosphate-treated and purified neutrophils and EBV-B cells were prepared as described (32,33). EBV-B lymphocytes from a p47 phox -deficient CGD patient were kindly provided by M.-A. Pocidalo (INSERM U479, Hôpital Bichat, Paris, France).
Isolation of the p47 phox , p67 phox , p40 phox Cytosolic Activation Complex-Anti-p47 phox immunoglobulins were immobilized onto a Carbolink TM coupling matrix. Cytosol from either EBV-B lymphocytes or neutrophils was used to affinity-purify the p47 phox -, p67 phox -, and p40 phox -activating factors as a complex (33,34). The cytosol (50 -100 mg, i.e. 2-5 ϫ 10 9 cell eq) was preincubated first with an uncoupled Carbolink TM matrix and then with a Carbolink TM matrix coupled with nonspecific immunoglobulins. Unbound proteins were loaded on the anti-p47 phox matrix previously equilibrated in PBS, with overnight recycling at 4°C. After washing in PBS, bound proteins were eluted either with 0.1 M glycine, pH 3, or with 1 M NaCl and then 0.1 M glycine, pH 3, or with 2 mM competitor peptide (used to generate the anti-p47 phox antibodies). Glycine eluates were immediately buffered with 1 M Tris-HCl, pH 9.5, dialyzed against PBS containing a mixture of protease inhibitors (10 M N3 ␣ -p-tosyl-L-lysine chloromethyl ketone, 1.8 M leupeptin, 1.5 M pepstatin) and stored at Ϫ80°C until further use. To verify the specificity of the recovered proteins, a control experiment was conducted with cytosol from p47 phox -deficient EBV-B lymphocytes.
MRP8 and MRP14 Purification-MRP8 and MRP14 were purified from the cytosol of unstimulated neutrophils as described (35). Briefly, neutrophil cytosol was submitted to 70% (w/v) (NH 4 ) 2 SO 4 precipitation. The 10,000 ϫ g centrifugation supernatant was dialyzed against 50 mM Tris-HCl, pH 8.5, containing 1 mM DTT, 1 mM EDTA, and 1 mM EGTA, and fractionated through FPLC onto a monoQ anion exchange column. The bound MRP8 and MRP14 were eluted with 0.13 M NaCl, as shown on the elution chromatogram (Fig. 1A, left panel), and then dialyzed against PBS. The major elution peak was analyzed by SDS-PAGE (Fig.  1A, right panel, lane 1). This peak presented two peptide bands with an apparent molecular mass of 16 and 11 kDa, identified by Western blotting as MRP14 and MRP8, respectively (Fig. 1A, right panel, lane 2). This fraction was called MRP8/MRP14. The absence of contaminating p47 phox , p67 phox , and Rac in the purified fraction was checked by Western blot (Fig. 1B, lanes 1) versus a positive control performed on neutrophil cytosol (Fig. 1B, lanes 2). The purified proteins were stored at Ϫ80°C until further use.
Oxidase Activity Reconstitution in a Cell-free Assay with Purified Cytochrome b 558 -Cytochrome b 558 was purified from the plasma mem-branes of 10 10 PMA-stimulated neutrophils and from 10 10 EBV-B lymphocytes and relipidated with L-␣-phosphatidylcholine II-S as reported (11). It was then quantified by reduced-minus-oxidized difference spectra using an absorption coefficient of 106 mM Ϫ1 ⅐cm Ϫ1 (15). Oxidase activity was reconstituted by incubating purified cytochrome b 558 (0.215 pmol) with cytosol isolated from neutrophils or from EBV-B lymphocytes (300 g) in the presence of 10 M FAD, 40 M GTP␥S, 5 mM MgCl 2 and an optimal amount of arachidonic acid (80 -100 nmol) in a final volume of 100 l of PBS. In some experiments the effect of MRP8/ MRP14 purified from neutrophil cytosol was investigated. MRP8/ MRP14 (0.75-1.3 g) was preincubated in the presence of 500 nM CaCl 2 at room temperature for 20 min. Then, reconstitution was performed by adding first calcium-loaded MRPs, then purified cytochrome b 558 , and finally cytosol. In the semi-recombinant system, the activity was reconstituted with purified cytochrome b 558 from neutrophils (0.215 pmol ϭ 2.15 nM) and recombinant proteins: rRac1 (100 nM), rp67 phox (0 -300 nM), rp47 phox (185 nM), and a mixture (1/1) of rMRP8 and rMRP14 (300 nM) (rMRP8/rMRP14) preincubated for 20 min with 500 nM calcium as previously done. The oxidase activity was estimated by measuring the reduction of ferricytochrome c in the absence or the presence of superoxide dismutase at 550 nm (͚ 550 nm ϭ 21.1 mM Ϫ1 ⅐cm Ϫ1 ) and expressed as turnover, mol of O 2 . ⅐s Ϫ1 ⅐mol of heme b Ϫ1 (15).
High Resolution Two-dimensional Gel Electrophoresis-The neutro-phil cytosolic activating factors p47 phox , p67 phox , and p40 phox , isolated as a complex on the Carbolink TM affinity matrix (75 g), were submitted to high resolution two-dimensional gel electrophoresis as reported (36). Proteins (75 g) were precipitated with 7% (v/v) perchloric acid, washed with cold acetone, and solubilized in 2% (v/v) IPG buffer 3-10 containing 8 M urea, 4% (p/v) chaps, 18 mM DTT, and traces of bromophenol blue. The strips were rehydrated with the 100,000 ϫ g ultracentrifugation supernatant (1 h, 20°C). Proteins were first separated according to their isoelectric point along linear pH gradient strips (pH 3-10). Electrophoresis was performed at 50 A/strip at 20°C. Then the strips were equilibrated, first in a solution containing 0.375 M Tris, 6 M urea, 2% (w/v) SDS, 20% (v/v) glycerol, and 130 mM DTT at pH 8, and then in a similar buffer in which DTT was replaced by 135 mM iodoacetamide. After 20 min of incubation at room temperature, the proteins were separated according to their molecular mass using standard SDS-PAGE, and silver-stained as previously described (37), or processed for mass spectrometry (MS) analysis. Trypsin Digestion-Protein bands of interest were excised from a Coomassie Blue-stained 11% SDS-PAGE (38) and washed with 25 mM NH 4 HCO 3 , pH 8.0, for 10 min and then with 50% (v/v) acetonitrile in 25 mM NH 4 HCO 3 , pH 8.0, for 10 min at room temperature. This step was repeated three times, and finally the sample was washed in ultrapure water and completely vacuum dried. "In-gel" tryptic digestion was performed for 4 h at 37°C in 10-20 l of 25 mM NH 4 HCO 3 , pH 8.0, with 0.3-0.5 g of trypsin per sample, depending on the gel volume and protein amount.

Peptide Mass Fingerprinting by Matrix-assisted Laser Desorption/ Ionization Time-of-flight (MALDI-TOF) Mass Spectrometry-Mass
spectra of the tryptic digest were acquired on a Biflex (Bruker Daltonik, Bremen, Germany) MALDI-TOF mass spectrometer equipped with gridless delayed extraction. The tryptic digests (0.5 l) were deposited onto the target disk on a thin dry layer of matrix (mixture 4:3 (v/v) of a saturated solution of ␣-cyano-4-hydroxy-trans-cinnamic acid in acetone, and a solution of nitrocellulose (10 mg of nitrocellulose in 1 ml of 50% (v/v) isopropanol and 50% (v/v) acetone). Samples had been washed with 5 l of 0.1% (v/v) trifluoroacetic acid before drying, as previously described by Garin (36). The MALDI spectra were analyzed by comparing a list of mass-to-charge ratios (peptide mass fingerprint) acquired for each digested protein to available data bases at prospector.ucsf.edu/ ucsfhtml4.0u/msfit.htm. Proteins were identified by MS/MS analysis when no consistent hit was found.
Peptide Sequencing by Tandem Mass Spectrometry (MS/MS)-The tryptic digest was extracted with 5% formic acid and then with pure acetonitrile. The digest solution and extracts were vacuum dried, dissolved in 10 l of 10% formic acid, and desalted with ZipTip TM (Millipore). After elution with 5-10 l of 50% acetonitrile combined with 0.1% formic acid, the peptide solution was introduced into a glass capillary (MDS Protana) for nanoelectrospray ionization. Tandem mass spectrometry experiments were carried out on a Q-TOF hybrid mass spectrometer (Micromass) to obtain amino acid sequence information. MS/MS sequence information was used for data base searching using the MS-Pattern programs (prospector.ucsf.edu/ucsfhtml4.0u/mspattern.htm) or Peptide Search (www.narrador.embl-heidelberg.de/).
Atomic Force Microscopy (AFM) Experimentation-For AFM experiments, the medium contained cytochrome b 558 (0.2 pmol) purified from neutrophils or from EBV-B lymphocytes and incorporated into L-␣phosphatidylcholine liposomes, 40 M GTP␥S, 5 mM MgCl 2 , and 10 M FAD (11). Aliquot fractions (10 l) were collected from the mixture before and after incubation with 100 nmol of arachidonic acid at room temperature, deposited on a mica surface, and allowed to adhere for 1 min. After three washings with 200 l of distilled water and overnight desiccation, samples were observed using AFM. In some experiments, the medium also contained purified MRP8/MRP14 (0.75 g), preloaded or not with Ca 2ϩ (as described for oxidase activity reconstitution) or cytosol (from neutrophils or EBV-B lymphocytes; 300 g). The AFM measurements were carried out in air at room temperature using a ThermoMicroscopes Explorer AFM in the low amplitude true noncontact mode, as previously described (11). Liposome height was determined on x images (where x represents the number of liposome images acquired for the same preparation, x Ͼ 3) of three different experiments using the ThermoMicroscopes SPMLab software after calculation of the distance between two selected points.
Transfection of EBV-B Lymphocytes-Electroporation was used for transfection of pRc-CMV-MRP8, pRc-CMV-MRP14, and pRc-CMV-GFP in EBV-B lymphocytes. Briefly, EBV-B lymphocytes were harvested, washed three times, and suspended at 10 7 cells/300 l in RPMI 1640 medium supplemented with 2 mM L-glutamine at room temperature. Cells were then mixed with 50 g of plasmid DNA and electroporated FIG. 1. Co-purification of MRP8 and MRP14 from neutrophil cytosol by FPLC monoQ anion chromatography. A, chromatogram illustrating elution of purified MRP8 and MRP14 during the FPLC monoQ anion chromatography (left panel). Proteins of neutrophil cytosol (162.5 mg) were precipitated by 70% (w/v) of (NH 4 ) 2 SO 4 ; the 10,000 ϫ g centrifugation supernatant was used after dialysis for fractionation onto FPLC monoQ anion exchange chromatography. Elution was carried out with 0.13 M NaCl. The purified fraction (arrow on the graph) was submitted to SDS-PAGE (right panel, lane 1) and Western blotting (right panel, lane 2), using polyclonal antibodies that bind with both MRP8 and MRP14. The immune complexes were detected with a secondary antibody coupled to alkaline phosphatase, then stained with nitro blue tetrazolium. B, control of the absence of contaminant cytosolic Phox proteins in the purified fraction containing MRP8 and MRP14. The fraction containing MRP8 and MRP14 (MRP8/MRP14) purified as described (A) (lanes 1, 10 g/lane) was submitted to SDS-PAGE followed by Western blot. The absence of p47 phox , p67 phox , and Rac and the presence of MRP8 and MRP14 in this fraction were verified by using specific polyclonal antibodies. Neutrophil cytosol was used as a positive control (lane 2, 5 g/lane). The immune complexes were detected by ECL.
once in a 4-mm gap electroporation cuvette at 220 V for 10 ms (BTX ECM 830 TM , Electro Square Porator). Then they were immediately diluted in 6 ml of complete medium containing 10% (v/v) fetal calf serum. FACS analysis of transfected cells (FACScalibur, Becton Dickinson) showed a 14% transfection efficiency 48 h after transfection. Selection of transfected cells began at 48 h after transfection by cultivating cells in the presence of Geneticin (G418) (500 g/ml). Superoxide production of transfected cells was measured by chemiluminescence.
Measurement of NADPH Oxidase Activity in EBV-B Lymphocytes Using Chemiluminescence-Lymphocytes (5 ϫ 10 5 to 2 ϫ 10 6 ) suspended in 50 l of PBS were added to 200 l of PBS containing 0.9 mM CaCl 2 , 0.5 mM MgCl 2 , 20 mM glucose, 20 M Luminol, and 10 units/ml horseradish peroxidase. Superoxide production was measured by chemiluminescence after adding 10 l of a 2 g/ml PMA solution (39). Photon emission was recorded at 37°C for 1 h with a Luminoscan (Labsystem, Pontoise, France).
SDS-PAGE and Western Blotting-The proteins were fractionated by 10 or 11% SDS-PAGE (38) and electrotransferred to nitrocellulose as previously described (40). Immunodetection was performed using polyclonal antibodies raised against MRP8 and MRP14, and specific polyclonal antibodies directed against p47 phox , p67 phox , or p40 phox . The two subunits of cytochrome b 558 were detected with mouse monoclonal antibodies directed against gp91 phox (monoclonal antibody 54.1) and p22 phox (monoclonal antibody 44.1). The immune complexes were detected with goat anti-rabbit secondary antibody coupled to alkaline phosphatase or peroxidase. The bound phosphatase activity was measured by staining with nitro blue tetrazolium. The bound peroxidase activity was detected using ECL reagents.
Biochemical Assays-Proteins were measured with the Bradford assay using bovine serum albumin as a protein standard (41).
Statistical Analysis-Data were given as mean Ϯ S.D. Statistical analysis was performed using the unpaired t test. The results are reported when significantly different (p Ͻ 0.05) from the controls.

MRP8 and MRP14 Co-purify with the Cytosolic Activating
Factors p67 phox , p47 phox , and p40 phox -The NADPH oxidasesoluble components, present in the cytosol of neutrophils or EBV-B lymphocytes, were isolated on an anti-p47 phox affinity matrix ( Fig. 2A). SDS-PAGE analysis of the eluate protein content showed that several polypeptides other than Phox proteins were present (Fig. 2B). Similar protein profiles were observed with other elution conditions (data not shown). The specificity of bound proteins was confirmed by isolating the complex from p47 phox -deficient B lymphocytes. In this case, no significant associated protein could be detected in the eluted fraction on Coomassie Blue-stained gel (Fig. 2B, lane 3). The three factors (p67 phox , p47 phox , and p40 phox ) were eluted from the matrix with glycine, pH 3, in neutrophil and B lymphocyte complexes (Fig. 2C). The immunoreactive band migrating between p67 phox and p47 phox corresponded to the immunoglobulin heavy chain stripped from the affinity matrix. Moreover, two low molecular weight polypeptide bands, with an apparent molecular mass of 11 and 16 kDa, were only found in the complex isolated from neutrophil cytosol (Fig. 2B, lane 1) and were completely absent from EBV-B cell cytosol (Fig. 2B, lane  2). These proteins were then identified by immunoblot as MRP8 and MRP14 (Fig. 2C). These belong to the S100 family and are also named S100A8 and S100A9, respectively. The identification of these two proteins was confirmed by MALDI-TOF mass spectrometry. The 11-kDa protein (Fig. 2B, lane 1) was clearly recognized by mass fingerprint as MRP8, with 48% of sequence recovery (Table I). The 16-kDa protein (Fig. 2B,  lane 1) was identified by sequencing a peptide with tandem mass spectrometry. The peptide sequence obtained was LGH-PDTLNQGEFK and corresponded to residues 26 -38 of MRP14.
Because there are several isoforms of MRP8 and MRP14, the neutrophil-isolated complex was submitted to two-dimensional gel electrophoresis to determine the specific isoforms co-purifying with Phox proteins (Fig. 3A). Interestingly, two spots with a different isoelectric point were identified as MRP8 by immunoblot (Fig. 3B, left panel), whereas only one spot corresponded to MRP14 (Fig. 3B, right panel).
FIG. 2. Affinity purification of p47 phox , p67 phox , and p40 phox cytosolic activating factors as a complex, from the cytosol of either neutrophils or EBV-B lymphocytes. A, elution profile. The isolation of cytosolic factors of oxidase as a complex was carried out on a specific anti-p47 phox affinity matrix (Carbolink TM ), as described under "Experimental Procedures" and as previously reported (33,34). Briefly, the cytosol recovered from neutrophils, corresponding to 50 -100 mg of proteins, was incubated successively with the uncoupled matrix, then with the non-immune matrix, and finally with the specific anti-p47 phox matrix. It was eluted after an overnight incubation at 4°C by: 0.1 M, pH 3 glycine (Ⅺ); 1 M NaCl followed by 0.1 M, pH 3 glycine (OE); or 2 mM competitor peptide (ADLILNRSSESTKRKLASAV), which was used to generate the anti-C-terminal peptide antibodies against p47 phox (q). Glycine eluates were immediately buffered with 1 M Tris-HCl, pH 9.5. For each elution condition, the eluted proteins were fractionated by SDS-PAGE. The p67 phox , p47 phox , and p40 phox were then identified by Western blotting. On the x axis, the fraction size is 700 l for glycine and NaCl/glycine elutions, and 2 ml for peptide elution. B, SDS-PAGE (Coomassie Blue-stained gel) of the purified cytosolic complex (10 -15 g) isolated from neutrophil cytosol (lane 1), from B lymphocyte cytosol (lane 2) on the anti-p47 phox matrix and eluted by 0.1 M, pH3 glycine. The specificity of the binding was verified by loading the cytosol isolated from p47 phox -deficient EBV-B lymphocytes on the anti-p47 phox matrix; the glycine-eluted complex was analyzed by SDS-PAGE (lane 3). C, Western blot of the purified cytosolic complex (10 -15 g) isolated either from neutrophil cytosol (lanes 1) or B lymphocyte cytosol (lanes 2) on the anti-p47 phox matrix and eluted by 0.1 M, pH 3 glycine. Immunodetection of p67 phox , p47 phox , p40 phox , and MRP8/MRP14 was carried out using specific polyclonal antibodies. The immune complexes were stained with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate . Hc corresponds to the heavy chain of IgG.
MRP8/MRP14 Increases NADPH Oxidase Activity, in Vitro and ex Vivo-MRP8 and MRP14 were purified from the cytosol of human neutrophils in one step on a monoQ anion exchange matrix (FPLC) (Fig. 1), as described under "Experimental Procedures." The fraction eluted at 0.13 M NaCl was called purified MRP8/MRP14. The MRP effect on reconstituted oxidase activity was investigated because they were present in the complex purified from cytosol of neutrophils, but absent from that of B cells. The purified MRP8/MRP14 preparation was first reactivated in the presence of calcium; reconstitution was performed by adding MRP8/MRP14, optimal amount of cytochrome b 558 purified from neutrophils, and the cytosol of EBV-B lympho-cytes (33). Results are shown in Fig. 4A. Optimal oxidase activity was measured in the presence of cytochrome b 558 and neutrophil cytosol. It corresponded to a turnover of the reconstituted enzyme ranging close to 189 mol of O 2 . ⅐s Ϫ1 ⅐mol of heme b Ϫ1 (Fig. 4A, lane 5). The turnover decreased to 30 -40% of the control when B-cell cytosol was used instead of neutrophil cytosol (Fig. 4A, lane 2 (Fig. 2B, lane 1) was excised from the Coomassie Blue gel, digested by trypsin, and analyzed by MALDI-TOF mass spectrometry, as described under "Experimental Procedures." Experimental mass fingerprint was compared with data bases of predicted peptide profiles: 6 of 8 peptides matched with the mass fingerprint of MRP8, with 48% sequence recovery.  3. Two-dimensional gel electrophoresis analysis of the cytosolic activating complex isolated from PMNs. Two-dimensional gel electrophoresis was carried out, as described under "Experimental Procedures." The cytosolic activating factor complex (75 g), isolated from PMNs and eluted with 0.1 M glycine pH 3 ( Fig. 2A), was used for high-resolution two-dimensional gel electrophoresis. A, proteins were separated first according to their isoelectric point on immobilized linear pH gradient, 3-10, and then by standard SDS-PAGE, after which the gels were silver-stained. B, Western blot of the twodimensional gel. MRP8 and MRP14 were immunodetected using specific polyclonal antibodies directed against MRP8 (left panel) or MRP14 (right panel). The immune complexes were stained with nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate. . production was determined after adding an optimal amount of arachidonic acid by the superoxide dismutase-sensitive cytochrome c reduction at 550 nm. In some experiments, oxidase activity was reconstituted in the presence of cytosol (300 g) isolated from PMNs (lane 5) or EBV-B lymphocytes (lanes 2 and 4), and/or in the presence of an optimal amount of MRP8/MRP14 (0.78 g/assay) (lanes 4 and 3). In the latter case, the MRP8/MRP14 preparation (0.78 g in 40 l of PBS) was preincubated with 500 nM CaCl 2 at room temperature for 20 min. Oxidase activity was expressed as turnover (mol of O 2 . ⅐s Ϫ1 ⅐mol of heme b Ϫ1 ). Inset, determination of the amount of MRP8/MRP14 to be added to the medium, so as to reconstitute an optimal NADPH oxidase activity in vitro. NADPH oxidase activity was present (Fig. 4A, lane 4 versus 5). Purified MRP8/MRP14 was preincubated with increasing concentrations of calcium (0 -1 mM), before being incubated with cytochrome b 558 and cytosol (data not shown), to verify the calcium dependence of this effect. The results demonstrated that, in the absence of calcium, MRP8/MRP14 had no impact on the oxidase activity. The optimal amount of calcium-reactivated MRP8/MRP14 was determined with each B-cell cytosol preparation used for the cell-free assay (Fig. 4A, inset). Furthermore, the results demonstrated that Ca 2ϩ -loaded MRP8/MRP14 was able, on its own, to considerably increase O 2 . production of relipidated and purified cytochrome b 558 in the absence of cytosol (Fig. 4A, lane 3  versus 1), suggesting a direct MRPs/cytochrome b 558 interaction. The absence of p47 phox , p67 phox , and Rac from the fraction containing the purified cytochrome b 558 (Fig. 4B, left panel) was assessed by Western blot and compared with positive control (neutrophil cytosol) (Fig. 4B, right panel). Because of the sensitivity of the immunoblotting assay, if potential contaminant cytosolic Phox proteins were present, their concentration would be more than 10 8 times inferior to the one necessary for cytochrome b 558 activation in a cell-free assay. The specificity of the interaction between MRP8/MRP14 and cytochrome b 558 was investigated by adding purified neutrophil cytochrome b 558 to the reconstitution medium containing Ca 2ϩloaded MRP8/MRP14, either after (Fig. 5A, black circles) or before (Fig. 5B, black circles) adding EBV-B lymphocyte cytosol. In the absence of MRP8/MRP14, oxidase activity ranged between 50 and 70 mol of O 2 . ⅐s Ϫ1 ⅐mol of heme b Ϫ1 (Fig. 5, A and   B, open squares), ϳ40% of control value, as previously shown (Fig. 4A, lane 3). Adding cytochrome b 558 to the reconstitution medium containing Ca 2ϩ -loaded MRP8/MRP14 induced, by itself and in the absence of EBV-B lymphocyte cytosol, a slight but significant increase in oxidase activity (Fig. 5C, black circles). This activity remained stable after 70 min of incubation on ice. These data showed that, in the presence of MRP8/ MRP14, a different effect was observed depending on how cytochrome b 558 was added to the medium. A 2-fold increase in oxidase turnover was observed when the cytochrome b 558 fraction was added first to the MRP8/MRP14 fraction, before adding cytosol (Fig. 5B, black circles). On the other hand, the activity was similar to that reconstituted in the absence of MRP8/MRP14 when cytosol was incorporated before cytochrome b 558 (Fig. 5A, black circle versus open squares). These results suggest a specific and necessary interaction involving MRP8/MRP14 and relipidated cytochrome b 558 to restore a total oxidase activity in EBV-B cells. Moreover, these data show that Ca 2ϩ -loaded-MRP8/MRP14, on its own, is sufficient to obtain an active form of cytochrome b 558 . Effect of Recombinant MRP8 and MRP14 on NADPH Oxidase Activity Reconstituted in Vitro-A mixture (1/1) of recombinant MRP8 and MRP14 (rMRP8/rMRP14), prepared in E. coli (28,29) and preloaded with calcium, was used instead of MRP8/MRP14 purified from neutrophil cytosol, in the reconstitution experiment that was performed with an optimal amount of purified cytochrome b 558 and cytosol from EBV-B lymphocytes. Having rMRP8/rMRP14 in the system induced a 50% increase in reconstituted oxidase activity (88 Ϯ 7 versus 60 Ϯ 5 mol of O 2 . ⅐s Ϫ1 ⅐mol of heme b Ϫ1 ) (Fig. 6A, lane 6 versus 3).
However, adding either rMRP8 or rMRP14 alone had no effect (Fig. 6A, lanes 4 and 5). A significant rise in cytosol-independent activity was again measured in a medium containing exclusively cytochrome b 558 and rMRP8/rMRP14 (Fig. 6A, lane  2 versus 1). A similar MRP effect was observed ex vivo in EBV-B lymphocytes co-transfected with plasmids pRc-CMV-MRP8 and pRc-CMV-MRP14 encoding MRP8 and MRP14, respectively (Fig. 6B). Transfection efficiency was determined by FACS analysis (FACScalibur; Becton Dickinson) and showed that, 48 h after transfection, 14% of the cells expressed the transfected construct. After 19 days of selection for G418 resistance, superoxide production was measured by chemiluminescence upon PMA stimulation. NADPH oxidase activity was optimal 31 days after transfection (Fig. 6B). Expression of either MRP8 or MRP14 alone in EBV-B cells did not significantly increase oxidase activity versus control value. Co-expression of MRP8 and MRP14 is required for enhancing the NADPH oxidase activity of B cells. Conformation Change of Cytochrome b 558 Mediated by MRP8/MRP14 -The conformational change in cytochrome b 558 was investigated using AFM (Fig. 7) to estimate the potential interaction of Ca 2ϩ -loaded MRP8/MRP14 with cytochrome b 558 (Fig. 4, lane 3). AFM experiments were performed by incubating cytochrome b 558 liposomes with purified neutrophil MRP8/ MRP14 and addition of arachidonic acid for activation. The topography of native cytochrome b 558 liposomes was analyzed before and after "activation." Liposome height was determined as previously described (11), and reported in Fig. 7A. First, adding arachidonic acid to the medium containing cytochrome b 558 liposomes did not modify the liposome height, whatever the cellular origin of cytochrome b 558 ; liposome height reached ϳ6 nm. We found a similar distribution of cytochrome b 558 liposome height before and after the stimulus was added (Fig.  7B, left panel). The incubation of Ca 2ϩ -loaded MRP8/MRP14 with cytochrome b 558 purified from neutrophils or from EBV-B cells, followed by the addition of arachidonic acid, led to an increase in liposome height. Fig. 7B (right panel) clearly shows the difference in the distribution of liposome height values after activation in the system containing EBV-B cell cytochrome b 558 and Ca 2ϩ -loaded MRP8/MRP14. In the absence of calcium, there was no significant conformation change upon activation (data not shown). A statistically significant increase in liposome height was observed (8 versus 10 nm) (Fig. 7A) when MRP8/MRP14 was pre-incubated with arachidonic acid without calcium, and then incubated with cytochrome b 558 liposomes. These results confirm that MRP8/MRP14 interact directly with cytochrome b 558 purified from neutrophils and  1 and 2) or presence of EBV-B lymphocyte cytosol (300 g) (lanes [3][4][5][6]. Some experiments were carried out with recombinant MRP8 (rMRP8) (0.78 g/assay) (lane 4), recombinant MRP14 (rMRP14) (0.76 g/assay) (lane 5), or a mixture (1/1) of rMRP8 and rMRP14 (rMRP8/ rMRP14) (0.76 g/assay in total) (lanes 2 and 6). In this case and as previously described, rMRPs were pre-incubated in the presence of 500 nM CaCl 2 at room temperature for 20 min. Oxidase activity was expressed as oxidase turnover (mol of O 2 . ⅐s Ϫ1 ⅐mol of heme b Ϫ1 ). As a reference value, the activity were collected from the mixture before and after adding 100 nmol of arachidonic acid (AA), deposited on a mica surface, and dried before observation with AFM. In some experiments, the medium contained 0.75 g of the mixture of purified MRP8/ MRP14, preincubated or not with 500 nM CaCl 2 for 20 min at room temperature (Ca 2ϩ ). In another experiment (*), purified MRP8/MRP14 was preincubated (ϩAA) or not (ϪAA) with 100 nmol arachidonic acid before cytochrome b 558 was added. The AFM measurements were carried out using a ThermoMicroscopes Explorer AFM in non-contact mode (11). Liposome height was determined on x images (where x represents the number of liposome images acquired for the same preparation; (x Ͼ 3) of three different experiments using the ThermoMicroscopes SPM-Lab software. Results are expressed as the mean value (35 Ͻ n Ͻ 80) Ϯ S.D. B, graph representing the distribution of the measured liposome heights before (Ⅺ) and after (f) adding arachidonic acid to the system containing cytochrome b 558 purified from EBV-B lymphocytes, with 40 M GTP␥S, 5 mM MgCl 2 , and 10 M FAD (left panel), or to the system containing the same medium and purified MRP8/MRP14 (0.75 g) preincubated with 500 nM CaCl 2 , as described for A (right panel). The x axis corresponds to the number of liposomes measured, and each square represents a single measurement of liposome height. ** stands for results significantly different (p Ͻ 0.05) from values obtained before stimulation.
from EBV-B cells. This interaction depends on the presence of calcium and arachidonic acid. The effect of MRP8/MRP14 on the conformational change in cytochrome b 558 , as confirmed by AFM, is related to oxidase measurement in a cell-free assay. The results are presented in Table II. Optimal oxidase turnover was obtained with the system containing neutrophil fractions as previously shown. Oxidase turnover was lower when EBV-B-cell cytosol replaced neutrophil cytosol (56% versus 100%), but unchanged when cytochrome b 558 purified from EBV-B cells was used instead of neutrophil cytochrome b 558 (Table II and Ref. 33). AFM studies showed an increase in liposome height after activation, indicating an assembly process. This assembly occurred, whatever the origin of the cytosol used, and could be fully active (in presence of neutrophil cytosol) or only partially functional (in presence of EBV-B lymphocyte cytosol). Incubating neutrophil cytochrome b 558 with calcium-loaded MRP8/MRP14, before adding cytosol from EBV-B lymphocytes, increased oxidase turnover up to 88% of the control value, suggesting a synergistic effect of MRPs on cytosolic factors in terms of assembly and conformational change of cytochrome b 558 (Fig. 4A, lane 4). Moreover, the interaction of calciumloaded MRP8/MRP14 with cytochrome b 558 , confirmed by AFM, was correlated to an increase in oxidase turnover of purified cytochrome b 558 (46 versus 17%, respectively). This effect was observed only when MRP8/MRP14 was loaded with Ca 2ϩ . The preincubation of MRP8/MRP14 with arachidonic acid in the absence of Ca 2ϩ led to an interaction with cytochrome b 558 , but not to the conformational change of the hemoprotein, which initiates oxidase activity.
MRP8/MRP14 Behaves as a Positive Effector of Oxidase Activation-The data reported in the previous section suggest that MRPs specifically interact with the redox component of NADPH oxidase complex. We previously proposed an allosteric mechanism for the regulation of the phagocyte oxidase and demonstrated that the transition from an inactive to an active oxidase complex initiates the electron transfer from NADPH to oxygen. In vivo, assembling regulatory factors (p67 phox , p47 phox , and Rac) with cytochrome b 558 is necessary to obtain an active enzyme. We have demonstrated that, in vitro, p67 phox is the limiting factor; it is involved in both assembly and activation of oxidase (11). In the present study, we wanted to clarify the role of MRP8/MRP14 on oxidase activity at a molecular level. We therefore assessed the effect of adding rMRP8/ rMRP14 to a semi-recombinant cell-free assay. Oxidase activity was reconstituted in a medium containing purified cytochrome b 558 , rRac1, an increasing concentration of rp67 phox , and an optimal amount of arachidonic acid (Fig. 8). A total reconstitution of oxidase activity was obtained at 200 nM rp67 phox (Fig. 8, open squares). The rp67 phox concentration necessary to reconstitute an optimal oxidase activity was reduced by a factor of ϳ3 and 5, respectively, when rMRP8/rMRP14 (Fig. 8, black squares) or rp47 phox (Fig. 8, open circles) was present in the medium. The shift of the bell-shaped curves was maximum when both rMRP8/rMRP14 and rp47 phox were incorporated in the medium, giving an optimal oxidase activity at 10 nM rp67 phox (Fig. 8, black circles). These data demonstrate that MRP8/MRP14 increased the affinity of p67 phox for cytochrome b 558 , as did p47 phox .

DISCUSSION
The main goal of this study was relative to the change of cytochrome b 558 conformation, which initiates NADPH oxidase activity and can be mediated by other effectors than Phox cytosolic factors. In fact we demonstrated that MRP8/MRP14, belonging to the S100 family, could initiate oxidase activity on their own through a calcium-dependent specific interaction with cytochrome b 558 . Moreover, MRPs are reported as allosteric effectors of phagocyte oxidase regulation.
MRPs Control Phagocyte Oxidase Activation-The properties of MRP8/MRP14 were investigated first in bovine neutrophils; they were shown to bind phenylarsine oxide and synergistically to activate NADPH oxidase (25). Furthermore, it was recently suggested that these proteins participate in NADPH oxidase activation, via a preferential association with the cytosolic activating factors of oxidase, mainly p67 phox (26). In the present TABLE II Effect of MRPs on NADPH oxidase assembly and activation Reconstituted oxidase activity medium was prepared as described under "Experimental Procedures," with cytochrome b 558 (Cyt.b 558 ) purified either from neutrophils or from EBV-B cells (0.2 pmol) in the medium containing MgCl 2 , GTP␥S, and FAD. Activation was initiated by adding an optimal amount of arachidonic acid (AA). In some experiments, the activity was measured in the presence of cytosol isolated from neutrophils or EBV-B cells (300 g) and/or a purified MRP8/ MRP14 fraction (0.78 g) preincubated (Ca 2ϩ ) (or not) with 500 nM CaCl 2 for 20 min at room temperature, or preincubated (or not) with the activator (AA) before addition to the medium. Turnover of reconstituted oxidase was assessed by cytochrome c reduction at 550 nm. Results are expressed as the percentage of the turnover obtained with neutrophil cytochrome b 558 8. Effect of rMRP8 and rMRP14 in a semi-recombinant cell-free system. Reconstituted oxidase activity medium was prepared as described under "Experimental Procedures," with cytochrome b 558 purified from neutrophil (0.215 pmol equivalent to 2.15 nM), increasing concentration of p67 phox (0 -300 nM), rRac1 (100 nM), GTP␥S, and MgCl 2 (squares). In some experiments, rp47 phox (185 nM) was added to the medium (circles). NADPH oxidase activity was measured after addition of arachidonic acid by measuring the cytochrome c reduction sensitive to superoxide dismutase. A mixture (1/1) of human recombinant MRP8 and MRP14 (rMRP8/rMRP14) (0.76 g/assay equivalent to 300 nM/ assay) was added to the medium (black squares and black circles) to evaluate the effect of rMRP8/rMRP14 on the semi-recombinant system. In this case, rMRP8/rMRP14 was incubated in the presence of 500 nM CaCl 2 at room temperature for 20 min, and then cytochrome b 558 and the recombinant cytosolic factors were added. study, MRP8 and MRP14 were co-isolated on an affinity matrix with the soluble factors of neutrophil NADPH oxidase; both components were absent from the cytosol of EBV-B lymphocytes, and coincidentally, these cells displayed a very low oxidase activity compared with neutrophils (13). We demonstrated in vitro and ex vivo, after gene transfection, that complementation of B lymphocytes with MRP8 and MRP14 was sufficient to increase oxidase activity up to control values (Figs. 4 and 6). Moreover, we showed that both MRP8 and MRP14 were necessary in this process, suggesting that at least the dimer MRP8/MRP14 was involved.
There was no oxidase activity even though a normal amount of MRP8 and MRP14 was present (data not shown) in the neutrophils of CGD AR patients (12,42). A preferential interaction between the MRP8/MRP14 complex and p67 phox has been suggested in resting neutrophils (26). This may suggest a co-translocation mechanism. In CGD 47°, the absence of p47 phox prevents translocation of p67 phox (43), whereas in CGD 67°, p47 phox is present and translocates on its own to the plasma membrane, without any MRP8/MRP14.
From these observations, we may conclude that, once translocated to the plasma membrane, the MRP8/MRP14 complex participates in the activation of oxidase as a positive effector of regulation.
MRP8/MRP14 Interaction with Cytochrome b 558 Is Mediated by Arachidonic Acid-MRPs need calcium to be active. MRPs are EF-hand molecules, which bind calcium selectively and with high affinity (17). The transitory elevation of cytosolic calcium, in response to inflammatory stimuli, initiates reaction cascades beginning in cytosol with phosphorylation of oxidase cytosolic factors, mainly p47 phox , but also MRP14. After calcium binding, MRPs associate in a non-covalent oligomeric MRP8/MRP14 structure (heterodimer or heterotetramer) and translocate to the plasma membrane (44). In myelomonocytic cells, phosphorylation of MRP8 and MRP14 has been suggested as the main parameter of subcellular compartmentation (27,45). The results tend to show that MRP8 and MRP14 must be assembled in an oligomeric structure to have a biological activity. A preferential translocation of phosphorylated MRP14 isoforms was reported (45) in response to a calcium signal.
Arachidonic acid has been suggested to mediate MRP association with membrane structures (46). In fact, a conformational change in the tertiary structure of the MRP8/MRP14 dimer, caused by an interaction with arachidonic acid, may improve its binding on membrane or extracellular surfaces. Arachidonic acid has been shown to enhance oxidase activity in vitro (47) and to induce a significant structural change in cytochrome b 558 (48). In vivo, among its multiple functions in neutrophils, arachidonic acid has been reported to have a role in NADPH oxidase activation through phospholipase A 2 (PLA 2 ). Cytosolic PLA 2 was recently introduced as the major PLA 2 activated in stimulated phagocytic-like cell (49,50). Moreover, transfection of antisense cytosolic PLA 2 oligonucleotide completely abolished arachidonic acid release and NADPH oxidase activation in stimulated PLB985 cells, but the translocation of p47 phox and p67 phox was unaffected (49). These data suggest that arachidonic acid participates in the oxidase complex assembly in vivo. Biological concentrations of arachidonic acid require micromolar levels to elicit most biological activities (51). The binding of MRPs to arachidonic acid may raise its local concentration in the membrane and facilitate MRP interaction with cytochrome b 558 .
We have seen that both MRP8 and MRP14 were necessary to induce NADPH oxidase activation in our experiments, and that this effect was dependent on calcium and arachidonic acid (Figs. 4 and 6). Most of the S100 proteins formed homodimers, but when both MRP8 and MRP14 were expressed, the heterodimer MRP8/MRP14 was preferred (17,52). More sensitive immunochemical labeling will be necessary to determine which MRP oligomeric structure is involved, but we can reasonably suggest that the heterodimer MRP8/MRP14 is indeed implicated.
Conformation Change of Cytochrome b 558 Initiates Oxidase Activation-Other teams, as well as ours, have demonstrated in vitro that cytochrome b 558 is the sole redox component and that p67 phox is the limiting cytosolic factor (9, 10) that initiates both assembly of the complex and activation (11). An allosteric regulation mechanism was suggested in which p47 phox behaves as a positive effector that increases the affinity of p67 phox for cytochrome b 558 (53,54,11). In vivo, phosphorylated p47 phox serves as an adapter protein bringing p67 phox into proximity with flavocytochrome b 558 (55).
Successfully applying AFM to the structural analysis of oxidase complex assembly opened the way to exploring the structure-function relationships of cytochrome b 558 incorporated into liposomes in the presence of MRP8/MRP14. The results of AFM cytochrome b 558 topography assessment were correlated to oxidase turnover measured at the molecular level. Purified and relipidated cytochrome b 558 displays a low but measurable level of NADPH oxidase activity in vitro in the absence of cytosolic factors: this activity depends on the nature of the phospholipid environment (56,11). The data presented here show that, in vitro, the sole Ca 2ϩ -loaded MRP8/MRP14 modulates the transition from an inactive to an active conformation state of cytochrome b 558 in the presence of arachidonic acid. Moreover, we found that adding MRP8/MRP14 to the cytosol of p67 phox -deficient EBV-B lymphocytes partly restores oxidase activity, highlighting MRP8/MRP14 capacity to induce the transition between the resting and activated states of oxidase (data not shown).
When the reconstitution assay was performed with physiological effectors such as p67 phox and p47 phox , MRP8/MRP14 increased the affinity of p67 phox for cytochrome b 558 and potentiated the effect of p47 phox . These findings suggest that MRPs directly interact at the molecular level with cytochrome b 558 ; this association, which depends on calcium and arachidonic acid, induces reorganization in its structural conformation (Figs. 4 and 6, lanes 2) resulting in oxidase activation. These data suggest that MRP8/MRP14 will positively regulate NADPH oxidase activity in cells when cytosolic subunits of oxidase are in limited quantity.
In conclusion, conformation change of cytochrome b 558 has been shown to initiate the electron transfer from NADPH to oxygen, and to generate superoxide anions. MRP8/MRP14 is suggested to mediate the transition from an inactive to an active conformation state of cytochrome b 558 and to behave physiologically as positive mediators of oxidase regulation. The respective function and modality of binding of both MRPs in the allosteric process should be further investigated.