JFC1, a novel tandem C2 domain-containing protein associated with the leukocyte NADPH oxidase.

We have employed a yeast two-hybrid system to screen a B lymphoblast-derived cDNA library, searching for regulatory components of the NADPH oxidase. Using as bait the C-terminal half of p67(phox), which contains both Src homology 3 domains, we have cloned JFC1, a novel human 62-kDa protein. JFC1 possesses two C2 domains in tandem. The C2A domain shows homology with the C2B domain of synaptotagmins. JFC1 mRNA was abundantly expressed in bone marrow and leukocytes. The expression of JFC1 in neutrophils was restricted to the plasma membrane/secretory vesicle fraction. We confirmed JFC1-p67(phox) association by affinity chromatography. JFC1-containing beads pulled down both p67(phox) and p47(phox) subunits from neutrophil cytosol, but when the recombinant proteins were used, only p67(phox) bound to JFC1, indicating that JFC1 binds to the cytosolic complex via p67(phox) without affecting the interaction between p67(phox) and p47(phox). In contrast to synaptotagmins, JFC1 was unable to bind to inositol 1,3,4,5-tetrakisphosphate but did bind to phosphatidylinositol 3,4,5-trisphosphate and to a lesser extent to phosphatidylinositol 3,4-diphosphate. From the data presented here, it is proposed that JFC1 is acting as an adaptor protein between phosphatidylinositol 3-kinase products and the oxidase cytosolic complex.

The NADPH oxidase, a multisubunit enzymatic complex that is present in neutrophils and B lymphocytes, is responsible for the monoelectronic reduction of oxygen to produce superoxide anion (O 2 . ) at the expense of NADPH (1,2). Free radical production is directly related to the bactericidal capacity of the cell, since patients with chronic granulomatous disease, whose NADPH oxidase is inactive (3,4), suffer recurrent bacterial infections. In resting cells, the components of the oxidase are distributed between different subcellular compartments and thus re-main unassembled while the oxidase is inactive. The main membrane component is cytochrome b 558 , an integral membrane protein containing one subunit of gp91 phox and one of p22 phox that is located in the secretory vesicles and specific granules (5). Meanwhile, the p47 phox and p67 phox , components that are known to be essential for the oxidase activation in vivo, remain in the cytosol in a complex that also includes p40 phox (6,7), a protein that is reported to regulate the activity of the oxidase. Two other factors, the small GTPases Rac2 (8) and Rap1a (9,10), are also known to participate in the regulation of the oxidase. The exact mechanism of the activation process, however, remains obscure. In the presence of adequate stimuli, the cytosolic factor p47 phox is phosphorylated and translocated, together with p67 phox (11,12), to the particulate fraction, where the cytosolic complex interacts with cytochrome b 558 . The subunit gp91 phox is a flavohemoprotein that contains two hemes necessary to transfer electrons from NADPH to molecular oxygen (13,14). The C terminus of p22 phox binds to tandem SH3 1 domains present in p47 phox (15). Because p47 phox binds to p67 phox (16,17), it is considered to be responsible for assembling the oxidase in vivo. On the other hand, recent studies showed that p47 phox is not required for the reconstitution of NADPH oxidase in a cell-free system when high concentrations of p67 phox and Rac are present (18,19). Therefore, these studies suggest that p67 phox is the essential cytosolic factor for enzyme activity (19,20).
pression. Moreover deletions of either SH3 domain dramatically reduced NADPH oxidase activity in this system, findings that correlated with decreased membrane binding (23). It is well known that SH3 domains play an important role in protein-protein interactions regulating cellular localization of interacting factors, and, although the importance of these domains in the p67 phox -p47 phox interaction has been described (16,17), it is not unlikely that p67 phox -SH3 domains interact with other accessory proteins that could take part in the still unclear process of NADPH oxidase assembly or activation.
We used the yeast two-hybrid system to search for additional components of the phagocyte antimicrobial machinery. Using as bait the C-terminal half of p67 phox , including both SH3 domains, we isolated a protein that interacts with oxidase components. In this communication, we describe the cloning and some of the properties of this protein, which we have designated JFC1.

EXPERIMENTAL PROCEDURES
Materials-Reagents for the yeast two-hybrid assay, including vectors, yeast strains, the yeast two-hybrid cDNA library, and control vectors, were the generous gifts of Stephen Elledge (Baylor College of Medicine, Houston, TX). The -ZAP cDNA leukocyte library was kindly provided by Dr. Ernest Beutler (The Scripps Research Institute, La Jolla, CA).
Yeast Two-hybrid Assay and Retrieval of a Full-length Clone-For the yeast two-hybrid assay "bait" constructs, the C-terminal portion of p67 phox , including the two SH3 domains and the intervening sequence (residues 245-512), was amplified from cDNA by polymerase chain reaction using a 5Ј primer (GAATTCGCTCACCGTGTGCTATTT) that contained an EcoRI site (underlined) and a sequence that annealed to nucleotides 730 -747 and a 3Ј antisense primer (GAATTCCCCTTCAA-CAAAAACTTTGGGGAA) that also contained an EcoRI site (underlined) and a sequence that annealed to nucleotides 1516 -1536. The resulting fragment was then ligated into pAS1 so as to be in frame with the upstream GAL4 coding sequences. This construct was sequenced to verify that it was in frame as planned and that no mutations had been introduced during polymerase chain reaction. The construct was used to screen a yeast two-hybrid cDNA library derived from Epstein-Barr virus-transformed B lymphoblasts cloned in the vector pACTII, whose inserts are expressed as fusion proteins consisting of the activation domain of GAL4 attached to the polypeptide encoded in the insert. The constructs were transfected by the method of Gietz et al. (24) in Saccharomyces cerevisiae strain Y190, and the screening was performed following procedures and using negative controls that have been described previously (25,26). From 4.2 million transformants, 46 positive colonies were detected. Twenty-four of these were selected for sequencing. Of these 24 clones, four corresponded to a single protein that showed a strong homology to C2 domains in synaptotagmins (27) and rabphilin 3 (28), two proteins that are involved in lipid binding, secretion, and protein-protein interactions. One of these clones was chosen for further study.
To obtain a full-length clone, a peripheral blood leukocyte cDNA library in -ZAP was screened by standard filter hybridization methods (29), using the [␣-32 P]dCTP-labeled PstI fragment of the yeast twohybrid clone as a probe. After isolation, the -ZAP clone containing the full-length JFC1 insert was circularized to the plasmid pBK-CMV-JFC1 by Cre-lox-mediated recombination using the ExAssist helper phage and the Escherichia coli strain XLOLR (Stratagene, La Jolla, CA).
Sequencing-The sequences of both the yeast two-hybrid cDNA fragment in pACTII and the full-length JFC1 cDNA in -ZAP were determined in The Scripps Research Institute molecular biology facility, using an automated fluorescent dye terminator sequencer. Overlapping oligonucleotide primers (Life Technologies, Inc.) were designed so as to obtain complete sequence information from both strands of the cDNA.
RNA Blot Analysis-Leukocyte RNA was obtained as previously described (29). The probe was labeled with [␣-32 P]dCTP (PerkinElmer Life Sciences) using the Prime-It RmT random primer labeling kit from Stratagene (La Jolla, CA) and probed as described elsewhere (29). Multiple Tissue Northern blots (CLONTECH, Palo Alto, CA) were probed with the full-length EcoRI fragment of the JFC1 cDNA, and the multiple tissue Northern blots were probed as recommended by the manufacturer.
Preparation and Fractionation of Neutrophils-Whole blood was obtained from the anonymous donor program at The Scripps Research Institute General Clinical Research Center. Neutrophils (polymorphonuclear leukocytes (PMNs)) were isolated by dextran sedimentation, hypotonic lysis, and Ficoll density centrifugation as previously described (30). Fractionation was carried out by a published protocol (31). Briefly, PMNs were resuspended in PIPES buffer (10 mM), pH 7.3 containing 100 mM KCl, 3 mM NaCl, 3.5 mM MgCl 2 , and 1 mM ATP (buffer A) supplemented with protease inhibitors (diisopropylfluorophosphate, 0.5 mM (Sigma) and Complete protease inhibitor mixture (Roche Molecular Biochemicals)) and then disrupted by nitrogen cavitation at 400 p.s.i. EGTA was added to a final concentration of 1.25 mM, and nuclei and unbroken cells were removed by centrifugation at 500 ϫ g. The supernatant from this spin (postnuclear supernatant (PNS)) was either centrifuged at 200,000 ϫ g for 30 min to separate cytosol from particulate matter or fractionated on a two-step discontinuous Percoll gradient (step 1, ␦ ϭ 1.076 g/ml; step 2, ␦ ϭ 1.11 g/ml, with tonicity adjusted with buffer A containing 1.25 mM EGTA and protease inhibitors) as previously described (31). All particulate fractions were washed with buffer A containing 1.25 mM EGTA to remove Percoll.
Stimulation of PMN with Phorbol 12-Myristate 13-Acetate (PMA)-PMNs prepared as described above were washed once in Ca 2ϩ -and Mg 2ϩ -free phosphate-buffered saline (Life Technologies, Inc.) and then resuspended in PBS containing 0.5 mM CaCl 2 and 1.5 mM MgCl 2 and warmed to 37°C for 10 min. PMA (Sigma) was added to a final concentration of 1 g/ml and the cells were incubated for an additional 6 min at 37°C. The cells were then washed once in ice-cold PBS and subjected to the cavitation/fractionation protocol.
In Vitro Binding Studies-Recombinant JFC1 was expressed in E. coli as the glutathione S-transferase fusion protein and purified using glutathione-Sepharose beads (Amersham Pharmacia Biotech) as recommended by the manufacturer. Sepharose beads (100 l) containing 10 g of the recombinant fusion protein or GST alone were washed twice in RIPA buffer (10 mM Tris/HCl, pH 7.5, 140 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 0.025% NaN 3 ) and incubated overnight at 4°C in the presence of 1.8 ϫ 10 7 or 9 ϫ 10 7 cell equivalent of neutrophil-derived cytosol or buffer. Beads were washed five times during 10 min with buffer RIPA at 4°C, spun down, and boiled in Laemmli SDS sample buffer (31). Samples were resolved by SDS-PAGE, and proteins were transferred to nitrocellulose and probed with primary antibodies directed against p47 phox or p67 phox . Detection was performed using alkaline phosphatase-conjugated goat anti-rabbit secondary antibody. The bound alkaline phosphatase activity was detected using the 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium colorimetric reagent (Bio-Rad).
In Vitro Translated JFC1 and p67 phox -Wild type JFC1 cDNA and p67 phox cDNA were cloned into pBK-CMV under the control of the T7 promoter. In vitro translated proteins labeled with [ 35 S]methionine were produced using the TNT coupled transcription and translation system from Promega (Madison, WI) as recommended by the manufacturer. Translational grade [ 35 S]methionine was purchased from Amersham Pharmacia Biotech. In order to normalize the amount of radioactive protein added to the experiments, serial dilutions of each in vitro translation reaction were electrophoresed on SDS-PAGE, and the specific protein bands were quantified using a PhosphorImager (Molecular Dynamics, Inc., Sunnyvale, CA).
Binding of [ 35 S]JFC1 to p47 phox or p67 phox in Vitro-The binding of [ 35 S]JFC1 to GST-p47 phox or GST-p67 phox was evaluated as described above under "In Vitro Binding Studies" except that 5 g of GST-p47 phox , GST-p67 phox , or GST were used in the assays. The reactions were performed either in RIPA buffer or in a buffer that contained 50 mM Tris/HCl, pH 7.5, 50 mM NaCl, and 0.1% Triton X-100. Serial dilutions of the in vitro translation reactions containing [ 35 S]JFC1, [ 35 S]p67 phox as a positive control, or [ 35 S]luciferase as a negative control were used in the assay with a 200-l final volume. Samples were rotated at 4°C during 1 h, washed four times for 15 min with the reaction buffer, spun down, and boiled in Laemmli SDS sample buffer. Samples were resolved by SDS-PAGE, and the specific protein bands were quantified using a PhosphorImager (Molecular Dynamics).
Binding of JFC1 to Phosphoinositides: Dot Blot Assay-The binding of in vitro translated [ 35 S]JFC1 to several phospholipids and phospho-inositides was evaluated by a dot blot assay as previously described (32) with minor modifications. Phospholipids at 2 mg/ml in 1:1 chloroform/ methanol solution containing 0.1% HCl were spotted (4 g) onto nitrocellulose sheets. After drying, nitrocellulose was blocked overnight at 4°C in Tris-buffered saline plus 3% bovine serum albumin. In vitro translated [ 35 S]JFC1 (0.1-0.5 Ci) in Tris-buffered saline containing 1.5% bovine serum albumin and 300 M L-methionine was then used to probe the phosphoinositide-containing nitrocellulose for 30 min at room temperature. Filters were washed five times with Tris-buffered saline and dried, and bound radioactivity was visualized by autoradiography. Samples were eluted from the nitrocellulose spots containing the radioactive probe and resolved by SDS-PAGE. Radioactivity was visualized using a PhosphorImager (Molecular Dynamics). Negative controls using in vitro translated [ 35 S]luciferase were performed in parallel.
Binding of JFC1 to [ 3 H]InsP 4 -[ 3 H]InsP 4 binding to JFC1 was evaluated as previously described (33) with minor modifications. Briefly, the reactions, containing 3 g of GST-JFC1 and 2 nM [ 3 H]InsP 4 (10,000 cpm/assay) were carried out in a medium containing 50 mM Tris-HCl, pH 7.5, and 2 mM EDTA for 10, 20, or 30 min at 4, 30, or 37°C. The final volume was 200 l. Free and bound InsP 4 were resolved either by protein precipitation or by filtration through nitrocellulose. In the first case, protein was precipitated (15 min on ice) by adding 300 g of ␥-globulin and 18% (final concentration) polyethylene glycol. The reactions were centrifuged at 12,000 rpm at 4°C for 5 min, and the supernatant was removed by aspiration. The pellet was solubilized with 500 l of tissue solubilizer; 35 l of glacial acetic acid was added, the suspension was transferred to 4.0 ml of liquid scintillation fluid, and radioactivity was determined by scintillation counting. For filtration assays, samples were vacuum-filtered on 0.45-m pore size cellulose nitrate membrane filters (Whatman, Maidstone, England). For this purpose, a sampling vacuum manifold (Millipore Corp., Bedford, MA) was used. The filters were immediately washed four times with 2 ml of ice-cold 50 mM Tris-HCl, pH 7.4, containing 2 mM EDTA in less than 20 s. Filters were counted using a scintillation counter (Beckman LS 6000SC). Negative control assays were run in parallel using either bovine serum albumin or GST. Nonspecific binding was determined in the presence of 10 M InsP 4 .
Antibodies-The antibody against p67 phox was raised against recombinant p67 phox purified from baculovirus and was the generous gift of Robert M. Smith (University of California, San Diego, La Jolla, CA). The anti-JFC1 antibody was raised by inoculating rabbits with the N-terminal peptide NH 2 -MAHGPKPETEGLLDLS-COOH conjugated to keyhole limpet hemocyanin (Chiron Mimotopes, San Diego, CA). The antibody was affinity-purified over a column of Sulfolink-activated Sepharose (Amersham Pharmacia Biotech) on which a peptide identical except for the addition of a C-terminal cysteine had been immobilized. The antibody raised against p47 phox has been previously described (12).

A C2 Domain-containing Protein
Interacts with the C-terminal Half of p67 phox -We were interested in identifying proteins that might be involved in the modification and regulation of the cytosolic subunits of the NADPH oxidase. For this purpose, we constructed a yeast two-hybrid system "bait" vector containing a fragment of p67 phox comprising residues 245-512, spanning both SH3 domains and the intervening sequence, and used it to screen a cDNA library derived from Epstein-Barr virus-transformed human B lymphoblasts, cells known to contain a fully functional NADPH oxidase. Among the library clones that were found to interact strongly and specifically with the C-terminal half of p67 phox was a sequence encoding p47 phox , which served as a convenient internal control, and four copies of a sequence that encoded the C-terminal part of a previously undescribed protein. This protein fragment contained a region of homology to the C2 domains of a number of proteins. C2 domains have been found in proteins that function in protein phosphorylation, lipid modification, GTPase regulation, and membrane trafficking (34). These domains also serve as lipid-binding domains in various isoforms of protein kinase C (35), the synaptotagmins (36,37), rabphilin 3 (38), and the class II PtdIns 3-kinase (39). Because all of these processes occur in concert with NADPH oxidase activation, we decided to pursue this clone as a potential regulator of oxidase activity.
A full-length cDNA was retrieved from a -ZAP human pe-ripheral blood leukocyte library by conventional plaque screening. The sequence of this clone is shown in Fig. 1. Translation of the open reading frame disclosed a 562-amino acid protein (starting from the first methionine) that contained a second C2 domain just upstream of the one found in the original yeast library clone (Fig. 1). According to a previously described classification (40), both C2 domains present in JFC1 correspond to topology I. Comparison of the C2 domains from JFC1 with other C2 domain-containing proteins by sequence alignment identified that the C2A domain has structural homology with the synaptotagmin-C2B domains and with rabphilin 3-C2B domain (27,41).
Tissue-specific expression of JFC1 mRNA-We examined the tissue-specific expression of JFC1 by probing a semiquantitative mRNA dot blot with the radiolabeled full-length cDNA JFC-1 fragment (Fig. 2). As expected, the clone was abundantly expressed in bone marrow and lymphoid tissues, which have a high leukocyte content. There was also significant expression of JFC1 in pancreas, trachea, stomach, salivary gland, and prostate. Northern blot analysis of leukocyte mRNA using the library cDNA as a probe identified a single band at ϳ1.8 kilobases, consistent with the size of the full-length cDNA (Fig. 3).
Subcellular Localization of JFC1-Many C2 domain-containing proteins are associated with membranous subcellular structures. We examined the subcellular localization of JFC1 by immunoblotting (Fig. 4). An antibody raised against an N-terminal peptide from JFC1 recognized a specific doublet in whole neutrophils that had been lysed by boiling in SDS-PAGE sample buffer (Fig. 4A), probably representing either different translation starting points or just phosphorylated and unphosphorylated forms of JFC1. JFC1 has been successfully phosphorylated in vitro by protein kinase C, mitogen-activated protein kinase, and Ca 2ϩ /calmodulin kinase II. 2 When neutrophils were lysed by nitrogen cavitation, JFC1 segregated to the PNS. Ultracentrifugation of the PNS yielded soluble (cytosol) and particulate (membranes/organelles) fractions. JFC1 was detected exclusively to the particulate fraction of neutrophils by the method employed.
A more detailed fractionation was carried out using a method previously described (31). Briefly, the PNS was layered onto a two-step discontinuous Percoll gradient and centrifuged, resulting in the separation of azurophil granules (␣), the specific granules (␤), the plasma membranes and secretory vesicles (␥), and a layer of cytosol. As can be seen in Fig. 4B (upper panel), JFC1 was detected only in the ␥ fraction of neutrophils that contain the plasma membranes and secretory vesicles. JFC1 was not detected in azurophilic or specific granules. Furthermore, stimulation of the neutrophils by PMA did not have an effect on the subcellular distribution (Fig. 4B, upper panel). In contrast, p47 phox and p67 phox were found almost exclusively in the cytosol in resting neutrophils, and PMA stimulation induced some of each protein to become associated with the ␥ fraction (Fig. 4B, lower panel), consistent with earlier findings (42). Despite the protein's tight association with the membranous structures, a hydropathy plot of the amino acid sequence did not indicate the presence of any large hydrophobic regions, suggesting that JFC1 is peripherally associated with the membrane (Fig. 4C). This was confirmed by the finding that the protein could be partially eluted from the membranes by 500 mM NaCl.
JFC1 Interacts with NADPH Oxidase Components in Vitro-Before functional studies were carried out, we confirmed that JFC1 was able to interact with the p67 phox NADPH oxidase component independent of the yeast two-hybrid system. This was achieved using an affinity adsorption technique that tests the ability of proteins from neutrophil cytosol to bind to immobilized GST-JFC1. As described earlier, JFC1 was identified through the ability of its C-terminal region to associate with p67 phox in a yeast two-hybrid system. Fig. 5 demonstrates that affinity chromatography confirms this association. p67 phox was pulled down from neutrophil cytosol by JFC1-containing beads (Fig. 5A, lower panel, lanes 3 and 5), while GST alone was not effective (Fig. 5A, lower panel, lanes 1 and 2). It is noteworthy that p47 phox , another cytosolic factor essential for the oxidase activity in vivo, was also pulled down by JFC1 in the same reaction (Fig. 5A, upper panel, lanes 3 and 5). It is very well established that p67 phox exists as a stable complex with p47 phox in the cytosol of nonstimulated neutrophils (6), raising the question whether JFC1 associated independently with both p47 phox and p67 phox or whether it bound to p67 phox and took up p47 phox pari passu as a component of the complex. The fact that we were unable to detect binding of GST-p47 phox attached to glutathione Sepharose beads to in vitro translated [ 35 S]JFC1 (Fig. 5B, upper panel) although [ 35 S]JFC1 was pulled down by GST-p67 phox (Fig. 5B, lower panel) supports the idea that JFC1 binds to p67 phox but not to p47 phox . Binding of the latter occurs because it is complexed to p67 phox . Moreover, increasing the concentration of cytosol in the reaction augmented the binding of p67 phox and p47 phox to JFC1 (Fig. 5A, lanes 3-5), again suggesting that they bind as a complex and indicating that JFC1 binds to a p67 phox site that is different from that involved in the recognition of p47 phox . JFC1 Binds to PtdIns(3,4,5)P 3 -As described above, JFC1 contains two tandem C2 domains that resemble those present in synaptotagmins and rabphilin 3. The C2A domain of JFC1 is highly homologous to the synaptotagmin C2B domain (Fig. 6), suggesting that a functional correlation may exist. The synaptotagmin C2B domain is known to bind inositol polyphosphates, mainly InsP 4 , in the absence of calcium (43). In Fig. 6, we show a comparison of putative inositol polyphosphate binding domains of various proteins. It is of special interest that, despite the homology observed between the JFC1-C2A domain and the polyinositol binding domain of synaptotagmins, the former was unable to bind [ 3 H]InsP 4 when evaluated either by a previously described precipitation assay (33) or by nitrocel-lulose filter binding assays (Fig. 7B).
To evaluate the ability of JFC1 to bind phospholipids and polyphosphoinositides, we performed dot blot assays as described under "Experimental Procedures" (SDS-PAGE analyses of samples eluted from the nitrocellulose spots are shown here). Fig. 7A shows that in vitro translated [ 35 S]JFC1 bound specifically to PtdIns(3,4,5)P 3 in these assays. JFC1 binding to PtdIns(3,4)P 2, although still significant, was 48% lower than that detected with PtdIns(3,4,5)P 3, indicating that the fully   phosphorylated head group is necessary for maximum binding. JFC1 binding to PtdIns(4,5)P 2 was only 16.6% of the maximum observed with PtdIns(3,4,5)P 3 , suggesting that 3-phosphorylated phosphoinositides play an important role in JFC1 recognition. In the same way, JFC1 bound to PtdIns(3)P to the same extent as that detected for PtdIns(4,5)P 2 . No significant binding of JFC1 to phosphatidylcholine, phosphatidylserine, or PtdIns was detected under these assay conditions. DISCUSSION Using a yeast two-hybrid system, we have identified a protein that interacts with p67 phox , a cytosolic factor of the NADPH oxidase, opening a new chapter in the regulation of this enzymatic complex. We confirm that JFC1 specifically binds to p67 phox in an in vitro study, using neutrophil cytosol as a source of the cytosolic factors. The polypeptide p67 phox is generally thought to be complexed with p47 phox in the cytosol of resting PMNs (6). Binding between these two factors has been shown to take place through interaction between the p67 phox N-terminal domain and the p47 phox SH3 domain (17), although association between C-terminal SH3 domain of p67 phox and a proline-rich C-terminal sequence in p47 phox has also been shown by the same group (16). We observed that JFC1 did not bind to p47 phox , although p47 phox and JFC1 interacted in the presence of cytosol, which implies that their association is through an intermediary factor, presumably p67 phox . The fact that JFC1 binds to p67 phox without altering the association of this cytosolic factor with p47 phox has physiological significance and suggests that JFC1 could have an important role in vivo. It should be taken into account that p47 phox is present in molar excess over p67 phox in the cytoplasm and that a large percentage of the protein actually exists in a dissociated form (6). In this work, we show that augmenting the cytosol concentration in the reaction medium increased the binding of p67 phox to JFC1 and, consequently, the detection of p47 phox , supporting the idea that free p47 phox does not interfere with the recognition of the complex by JFC1 and that JFC1 recognizes p67 phox at a molecular site different from that involved in the p67 phox -p47 phox interaction.
Herein we have demonstrated that JFC1 is associated with a subset of membranes in the phagocyte, although by our method we were not able to distinguish whether it localized to the secretory vesicles, the plasma membrane, or both. Therefore, it is unlikely that JFC1 interacts with the cytosolic complex in unstimulated cells, the components remaining in different intracellular compartments. Although little is known about the molecular basis for the up-regulation of the oxidase in the phagocytic vesicles or at the phagocyte surface, it is general knowledge that after stimulation, the cytosolic factors translocate to the particular fraction, suggesting that, in vivo, the interaction between JFC1 and the cytosolic factors would take place at some point after cell activation. In the presence of an appropriate stimulus, the cytosolic component p47 phox becomes sequentially phosphorylated on serines Ser 359 and Ser 370 followed by Ser 303 and Ser 304 , and the cytosolic complex migrates to the particulate fraction where it is known to interact with cytochrome b 558 (44). p67 phox also becomes phosphorylated (45), although the significance of this phosphorylation in the activation of the oxidase remains unknown. Although there are still aspects of the oxidase assembly that are unclear, several studies have shown that p47 phox interacts with the p22 phox subunit of the cytochrome b 558 (16). Moreover, it has been previously described that the cytosolic factors fail to translocate in the absence of cytochrome b 558 (46), suggesting that the latter would serve as a docking site. However, a recent immunofluorescence study examining phagocytosis indicated that the cytochrome b 558 , although required for stable membrane binding, is insufficient for targeting the cytosolic complex to the cell periphery and suggested that several other adaptor proteins may be involved in the process (47). Since JFC1 localizes in the particular fraction where the assembly of the oxidase takes place and interacts with the cytosolic subunits of the oxidase, it is conceivable that this protein could play such a role in vivo.
As discussed above, JFC1 possesses tandem type I C2 domains in its C-terminal end. Comparison of the C2A domain of JFC1 with the sequence alignment of other C2-containing proteins showed that it is highly homologous to the C2B domains present in synaptotagmins (Fig. 6). Several functions have been attributed to the synaptotagmin C2B domains including dimerization, interaction with clathrin assembly protein-2 (48), a process involved in endocytosis of synaptic vesicles, and binding to ␤ SNAP (49) and SNAP25 (50). Notably, several members of the synaptotagmin family have been shown to bind to the second messenger InsP 4 by an InsP 4 -binding site present in their C2B domains (43). Despite possessing a close match to a putative InsP 4 -binding site, K(K/R)KTXXK(K/R), in its C2A domain, JFC1 did not bind to the second messenger in an in vitro binding assay (Fig. 7B). A similar lack of binding to InsP 4 has been demonstrated for synaptotagmins III, V, and X, although their C-terminal truncated forms have been shown to bind InsP 4 . In this case, an inhibitory effect has been proposed for the C-terminal end of these proteins (43). JFC1, on the other hand, lacks an inhibitory domain homologous to those found in synaptotagmins III, V, and X. Another example of a protein that possesses a synaptotagmin-like C2 domain unable to bind InsP 4 is the GTPase-activating protein GAP1m. GAP1m binds InsP 4 by its pleckstrin homology domain (51) but not by its "lysine stretch-like region" located in its C2B domain, which is homologous to those present in synaptotagmins (52,53).
The strong binding of JFC1 to PtdIns(3,4,5)P 3 , a phosphoinositide whose polar inositol group is InsP 4 with the variation that the 1-phosphate is attached to the diacylglycerol presented here, appears to be an important observation. The fact that JFC1 binds to PtdIns(3,4,5)P 3 but not to InsP 4 indicates that the lipid moiety of the molecule plays an essential role in the recognition of the phosphoinositide by JFC1. This establishes a difference between JFC1 and other putative PtdIns(3,4,5)P 3 receptors like Arf-nucleotide-binding site opener (54) and Bruton's tyrosine kinase (55), since these have been described to bind not only to PtdIns(3,4,5)P 3 but also to InsP 4 (for a review, see Ref. 52). Moreover, synaptotagmin I should also be included in this group, since it has recently been shown to possess C2B-mediated phosphoinositide binding activity (37). This supports the idea that the "lysine stretch" moiety in JFC1 could be responsible for its PtdIns(3,4,5)P 3 binding capacity.
JFC1 showed preference to PtdIns(3,4,5)P 3 over PtdIns(3,4)P 2 , bound relatively weakly to PtdIns(3)P and PtdIns(4,5)P 2 , and showed no binding to PtdIns(4)P in vitro (Fig. 7A), suggesting a central role for 3-OH-phosphorylated inositol lipids in JFC1 function. It is well established that phosphorylation of the D-3 position of phosphoinositides is catalyzed by members of the PtdIns 3-kinase family (52). Furthermore, there is clear evidence that PtdIns 3-kinase is involved in the activation of the NADPH oxidase, although the details of this mechanism remain to be elucidated. It has been previously shown that PtdIns(3,4,5)P 3 accumulates in fMLP-stimulated neutrophils (56). Moreover, PtdIns 3-kinase inhibitors markedly inhibited the NADPH oxidase activation in fMLP-stimulated neutrophils (57)(58)(59). On the other hand, extracellular superoxide anion production by PMA-stimulated neutrophils has been shown to be unaffected by the same PtdIns 3-kinase inhibitors (58,59); thus, the suggestion that PtdIns 3-kinase is involved in the oxidase activation pathway upstream of the activation of protein kinase C has been proposed (57). It is well established that p47 phox is a substrate for protein kinase C (11,60,61), and as described above, its phosphorylation is an essential event for the activation of the oxidase in vivo. It has been demonstrated that the phosphorylation of this cytosolic factor is inhibited by the PtdIns 3-kinase inhibitor wortmannin in fMLP-stimulated PMNs (58), which also correlates with the abolition of the oxidase activity. However, a recent study showed that p47 phox phosphorylation is only moderately reduced in the presence of wortmannin in opsonized zymosanactivated PMNs, although complete inhibition of the oxidase activity was observed at identical inhibitor concentrations (62). These data suggest that mechanisms other than the protein kinase C-associated p47 phox phosphorylation take place downstream of PtdIns(3,4,5)P 3 formation in association with oxidase activation. Moreover, it was recently observed that wortmannin inhibits intracellular production of O 2 . in FIG. 6. Sequences alignments of C2 domains containing a "lysine stretch." Sequence similarity of the C2B domains of synaptotagmin I (residues 303-352) (SytI; NP-005630); GTPase-activating protein 1m (residues 181-233) (GAP1m; AAD09821); Doble C2␣ (residues 281-330) (DocC2␣; NP-003577); rabphilin 3A (residues 590 -639) (Rab3A, A8097); and the C2A domain of JFC1 (residues 295-342). Lysines described to be essential for InsP 4 in synaptotagmins (43) are indicated by asterisks. Residues that are identical in the five sequences are darkly shaded; three or more matches are lightly shaded.

FIG. 7.
Binding of JFC1 to phosphoinositides. A, the binding of in vitro translated [ 35 S]JFC1 to several phospholipids and phosphoinositides was evaluated by dot blot assay. Phospholipids were spotted (4 g) onto nitrocellulose sheets. After drying, nitrocellulose was blocked with bovine serum albumin. In vitro translated [ 35 S]JFC1 (0.1-0.5 Ci) was then used to probe the phosphoinositide-containing nitrocellulose for 30 min at room temperature. Filters were washed and dried, and bound radioactivity was visualized by autoradiography. Samples were eluted from the nitrocellulose spots containing the radioactive probe and were further resolved by SDS-PAGE. Radioactivity was visualized using a PhosphorImager (Molecular Dynamics). PC, phosphatidylcholine; PS, phosphatidylserine; PI, phosphatidylinositol. Results shown are representative of three different experiments. B, [ 3 H]InsP 4 binding to JFC1 was evaluated as described under "Experimental Procedures." The reactions, containing 3 g of GST-JFC1 or an equimolar amount of GST and 2 nM [ 3 H]InsP 4 (10,000 cpm/assay) were carried out during 30 min at 30°C. The final volume was 200 l. Free and bound InsP 4 were resolved by protein precipitation.
PMA-activated neutrophils, leading to the suggestion that PtdIns 3-kinase is not necessarily upstream of protein kinase C in the oxidase signaling pathway (63). Thus, other events might be involved downstream of PtdIns 3-kinase in the signaling pathway that lead to the activation of the oxidase. In this way, it has been indicated that members of the unidentified "renaturable kinases" take action downstream of PtdIns 3-kinase during human neutrophil oxidase activation, while mitogen-activated protein kinase and mitogenactivated protein kinase/extracellular signal-regulated kinase kinase would be involved in the secretory pathway, also downstream of PtdIns(3,4,5)P 3 formation (59). In conclusion, while there is no doubt about the essential participation of PtdIns(3,4,5)P 3 in the activation of the oxidase, the signaling pathways downstream of PtdIns 3-kinase that direct the activation remain uncertain. We speculate that JFC1 could act as an accessory factor in the regulation of the oxidase, since it seems directly to establish a connection between an early event associated with oxidase activation (i.e. PtdIns(3,4,5)P 3 generation) and an oxidase subunit itself.
Finally, it is worth mentioning that although NADPH oxidase is restricted to leukocytes, PtdIns 3-kinase-related events are ubiquitous, and also JFC1 is expressed in several tissues, mostly with secretory function. Therefore, it is not unlikely that JFC1 plays a more general role, namely as an adaptor protein, presumably interacting with other SH3 domain-containing proteins. This function is currently under investigation in our laboratory.