Abstract
Superoxide anion production by the phagocyte NADPH oxidase plays a crucial role in host defenses and inflammatory reaction. The phagocyte NADPH oxidase is composed of cytosolic components (p40phox, p47phox, p67phox, and Rac1/2) and the membrane flavocytochrome b558, which is composed of two proteins: p22phox and gp91phox/NOX2. p22phox plays a crucial role in the stabilization of gp91phox in phagocytes and is also a docking site for p47phox during activation. In the current study, we have used a yeast two-hybrid approach to identify unknown partners of p22phox. Using the cytosolic C-terminal region of p22phox as bait to screen a human spleen cDNA library, we identified the protein interacting with amyloid precursor protein tail 1 (PAT1) as a potential partner of p22phox. The interaction between p22phox and PAT1 was further confirmed by in vitro GST pulldown and overlay assays and in intact neutrophils and COSphox cells by coimmunoprecipitation. We demonstrated that PAT1 is expressed in human neutrophils and monocytes and colocalizes with p22phox, as shown by confocal microscopy. Overexpression of PAT1 in human monocytes and in COSphox cells increased superoxide anion production and depletion of PAT1 by specific small interfering RNA inhibited this process. These data clearly identify PAT1 as a novel regulator of NADPH oxidase activation and superoxide anion production, a key phagocyte function.
This article is featured in In This Issue, p.1313
Introduction
Phagocytes, such as polymorphonuclear neutrophils and monocytes, constitute a major defense line against pathogens, such as bacteria and fungi (1–3). At the site of infection, phagocytes recognize and phagocytize pathogens, a process that includes the formation of intracellular phagosomes and the destruction of the internalized pathogens (3, 4). Activated polymorphonuclear neutrophils and monocytes release antibacterial substances into the phagosome and produce superoxide anion, from which other reactive oxygen species (ROS) derive, which are thought to play an important role in both the direct and indirect killing of pathogens (4–6). The superoxide anion is the precursor of other ROS molecules, such as hydrogen peroxide and hypochlorous acid. In phagocytes, the enzyme responsible for superoxide production is the NADPH oxidase or respiratory burst oxidase (7–11).
The phagocyte NADPH oxidase is a multicomponent enzyme, comprising several subunits: p22phox, p40phox, p47phox, p67phox, gp91phox (also called NOX2), and a small GTPase Rac1 or Rac2 (7–11). In the resting cell, p40phox, p47phox, and p67phox exist in the cytosol as a complex, and Rac1 (in monocytes) or Rac2 (in neutrophils) exist in the GDP-bound form, complexed to Rho/GDI. The other two components, p22phox and gp91phox, form a noncovalently bound complex known as flavocytochrome b558 (7–11). They are located in the membranes of specific granules, gelatinase granules, secretory vesicles, and the plasma membrane. Separating these two groups of components by distributing them between distinct subcellular compartments ensures that the NADPH oxidase is inactive in the resting cell. Upon cell activation by various stimuli, several events take place simultaneously, including phosphorylation and translocation of the cytosolic components p40phox, p47phox, and p67phox to the membranes where they associate with the flavocytochrome b558 (12–18). Rac2 exchanges its GDP for GTP, dissociates from its inhibitor rho/GDI, and migrates to the membrane where it interacts with p67phox. When all components are assembled at the membrane, the flavocytochrome b558 mediates the transfer of electrons from cytosolic NADPH to oxygen to produce the superoxide anion.
Human p22phox is a 195-aa protein; it has been proposed that the p22phox protein sequence consists of a short N-terminal tail, two transmembrane domains, and a long C-terminal tail ranging from aa 130 to 195 (11). In resting cells, p22phox interacts with gp91phox, probably via the transmembrane spanning domains, to stabilize the complex. During activation, the intracellular cytosolic tail of p22phox interacts with p47phox via a proline-rich sequence/SH3 domains docking site (19–21). Indeed, the p22phox C-terminal sequence has a polyproline sequence necessary for this interaction, as shown by site directed mutagenesis and in a chronic granulomatous disease patient who has a mutation at proline 156 (22).
PAT1 is a 585-aa protein; it was first identified as a protein interacting with amyloid precursor protein (APP) tail via the basolateral sorting signal site in the cytoplasmic tail of APP, a cell surface protein implicated in the pathogenesis of Alzheimer disease (23). PAT1 is 99 and 99.6% identical to PAT1a and ARA67, respectively. PAT1 is expressed in several cells, such as epithelial cells and smooth muscle cells (24). PAT1 shares homology with kinesin L chain and was found to bind to microtubules, suggesting its implication in trafficking and protein secretion. Indeed, PAT1 and its isoform PAT1a were found to promote APP processing, resulting in increased secretion of β-amyloid peptides (23, 24). PAT1 was also found in the cytoplasm and the nucleus of Madin–Darby canine kidney cells and interacted and controlled the subcellular localization of the androgen receptor, modulating its function (25, 26). Whether PAT1 interacts with other proteins and regulates other cell functions is not known.
It is not known if gp91phox and p47phox are the only proteins interacting with p22phox under resting and activated conditions. We hypothesized that the cytosolic p22phox region could interact with new, yet-unidentified partners. To identify proteins that interact with p22phox, we performed a yeast two-hybrid screening of a human spleen cDNA library using the cytosolic C-terminal region of p22phox as bait. We have identified PAT1 as a novel p22phox-interacting protein and shown it functions as an enhancer of NADPH oxidase activation and superoxide production in human phagocytes.
Materials and Methods
Reagents
Buffers, PMA, fMLF, C5a, PMSF, diisopropylfluorophosphates (DFP), diphenyleneiodonium, the anti-actin mAb and other chemicals were purchased from Sigma-Aldrich (St. Louis, MO). NaDodSO4-PAGE (SDS-PAGE) and Western blotting reagents were purchased from Bio-Rad Laboratories (Richmond, CA). Dextran T500, Ficoll, G-Sepharose beads, pGEX-6p1, and glutathione Sepharose were purchased from GE Healthcare (Little Chalfont, U.K.). MACSxpress Neutrophil Isolation Kit was from Miltenyi Biotec (Paris, France) and EasySep immunomagnetic negative selection kit was from STEMCELL Technologies (Grenoble, France). A mouse polyclonal anti-PAT1 Ab raised against the last 100 aa (486 aa–586 aa) was from Abnova (Taipei, Taiwan). Anti-p22phox and anti-gp91phox Abs were from Santa Cruz Biotechnology (Santa Cruz, CA). Anti–cytochrome b558 mAb (7D5) was from MBL International (CliniSciences, France). Secondary Abs were from The Jackson Laboratory.
Two-hybrid screening of human spleen cDNA library
Two-hybrid screening was performed using the method described previously (27, 28). The cDNA encoding the C-terminal tail of p22phox (residues 132–195) was cloned into the pLEX10 vector in frame with the LexA DNA-binding protein. The yeast strain L40, established with the LexA/p22phox (residues 132–195) pLEX10 plasmid, was transformed further with pACT2 plasmids containing cDNAs from a human spleen cDNA library (Clontech Laboratories, Basingstoke, U.K.) and fused with the sequence of the activation domain of Gal4. The transactivation of the two reporter genes His3 and LacZ was monitored first by growth of transformed yeasts on selective medium lacking histidine, but also tryptophan and leucine. His+ colonies were then assayed for β-galactosidase activity (LacZ+) by a qualitative color filter assay (27, 28). Recombinant pACT2 plasmids were recovered from His+/LacZ+ phenotype yeasts, then amplified in Escherichia coli and finally sequenced.
Cloning and expression of human PAT1a and expression of the p22phox cytosolic domain (residues 132–195) and the gp91cytosolic domain (residues 291–570)
PAT1a cDNA contained in PKH3 plasmid (a generous gift from Dr. S. Kins, University of Kaiserslautern, Kaiserslautern, Germany) was amplified using Pfu polymerase (StrataGene) with flanking ECOR1 and XhoI restriction sites open reading frame-flanking primers (sense, 5′-ccg gaa ttc atg gcg gcc gtg gaa cta-3′; antisense, 5′-ccg ctc gag tca gca gct cgg tcc ctc-3′) and cloned into the pCRII-TOPO vector (Invitrogen). For recombinant expression of PAT1a, the encoding cDNA was subcloned into pGEX-6P1 vector (Pharmacia, Piscataway, NJ) and sequenced to rule out unexpected mutations and to confirm the sequence. It was then transformed in BL21-DE3 (pLysS) E. coli strain and expressed as follows. An overnight culture was diluted 10-fold in fresh Terrific Broth medium containing 100 μg/ml ampicillin and grown for one additional hour at 37°C. The culture was then induced with 0.2 mM isopropyl β-d-thiogalactoside for 18 h at 15°C. Bacteria were harvested by centrifugation (4000 g, 20 min, 4°C), and the pellet was resuspended in lysis buffer (50 mM Tris-HCl [pH 7.5], 50 mM NaCl, 5 mM MgCl2, 1 mM DTT, 1% [v/v] Triton X-100, and protease inhibitors). Cells were lysed by sonication (6 × 30 s), lysates were centrifuged (15,000 × g, 30 min, 4°C), GST-recombinant proteins were affinity precipitated from supernatant by overnight incubation at 4°C with glutathione/Sepharose 4B beads (Pharmacia). Beads were then washed in lysis buffer, and the fusion protein was cleaved by incubation with PreScission Protease (Amersham Pharmacia Biotech) for 4–6 h at 4°C in 150 mM NaCl, 50 mM Tris-HCl (pH 7), 1 mM DTT, and 1 mM EDTA and subjected to another round of glutathione bead adsorption. PAT1a (aa 1–585) migrates at 67 kDa as a monomeric protein in a nondenaturating gel. p22phox cDNA contained in a plasmid (a generous gift from Dr. M. Dinauer, University of Saint Louis). Cloning and expression of p22phox (aa 132–195) cytoplasmic tail and gp91phox (aa 291–570) cytoplasmic tail in the pGEX-6P1 expression plasmid was performed as previously described (18, 21).
GST pulldown assay
The assay was performed as described previously (18, 29). Briefly, 80 pmol of GST-p22phox (aa 132–195), GST/gp91phox (aa 291–570), and GST alone were incubated in the presence of 5 pmol of recombinant PAT1 and glutathione/Sepharose beads in interaction buffer (20 mM Hepes [pH 7.5], 1% Nonidet P-40, 50 mM NaCl, and 1 mM EGTA) for 1 h. After washing, the complex was eluted with 10 mM glutathione and analyzed by SDS-PAGE and Western blots using protein-specific Abs.
Overlay assay
Human recombinant PAT1 was subjected to SDS-PAGE and transferred to nitrocellulose membranes. Overlay with recombinant p22phox cytosolic tail was performed as we previously described (21).
Ethics statement
Cells were isolated from the venous blood of healthy volunteers with their written informed consent in accordance with the Declaration of Helsinki. All experiments were approved by the INSERM Institutional Review Board and ethics committee. Data collection and analyses were performed anonymously.
Human neutrophil preparation and fractionation
Human neutrophils were isolated from the blood of healthy volunteers by dextran sedimentation and Ficoll centrifugation, as described previously (30, 31). The isolated cells were resuspended in PBS at a concentration of 108 cells/ml in the presence or absence of DFP (2.7 mM) for 20 min at 15°C and washed in the same buffer. Nuclei and cytoplasms were prepared using the Nuclear and Cytoplasmic Protein Extraction Reagents Kit from Pierce (Rockford, IL). For mRNA expression experiments, highly purified neutrophils were obtained using the MACSxpress Neutrophil Isolation Kit, and the EasySep immunomagnetic negative selection kit. These two kits yielded more than 99% purified neutrophils required for mRNA studies (32).
Human monocytes and lymphocytes isolation
PBMC were freshly isolated from the whole blood of healthy volunteers by Ficoll-Paque separation (Pharmacia, Uppsala, Sweden) as previously described (30, 31). PBMC were subjected to an additional separation step to isolate monocytes and lymphocytes (STEMCELL isolation kit). The purity of the resulting cell suspensions was randomly tested by a Sysmex XE-2100 automated blood cell counter (Sysmex, Kobe, Japan) and yielded more than 99% monocytes and 98% lymphocytes, respectively.
mRNA expression of PAT1 in human blood leukocytes
Total RNAs were isolated by lysing the cells with TRIzol Reagent (Life Technologies) according to the manufacturer's instructions. Afterwards, 1 μg of total RNAs were reverse transcribed, and PAT1a mRNA was quantified by real-time PCR using the Light Cycler Technology (Roche, Mannheim, Germany) and PAT1 (covering a common part of the three isoforms) specific forward 5′-tgc aaa aag tca cta tga tga gg-3′ and reverse 5′-tcc acc aca act ttc act gg-3′ primers (Eurogentec). Detection of PCR product was based on SYBR Green fluorescence signal. GAPDH housekeeping gene expression was used to normalize the results.
Cell lysate preparation for Western blotting analysis
To detect PAT1 by Western blotting in leukocytes, the cells were pretreated as following: cells were lysed in a lysis buffer (50 mM Tris-HCl buffer [pH 7.4], 1% Triton X-100, 2 mM β-glycerophosphate, 25 mM NaF, 150 mM NaCl, 10 μg/ml each of Leupeptine, Pepstatine, and Aprotinine), then diluted twice with the same lysis buffer without Triton X-100, sonicated three times for 3 s each, and centrifugated for 5 min at 4000 rpm at 4°C. Protein concentration in the supernatant was determined prior to denaturation by addition of Laemmli sample buffer (2×) (31) and boiling at 100°C for 3 min.
Transfection of PAT1 in COSphox cells
COSphox cells, which had been transfected with NADPH oxidase components (gp91phox, p22phox, p47phox, and p67phox), were used to study the effect of PAT1 on NADPH oxidase activity in the resting state and following stimulation with fMLF, C5a, or PMA (33, 34). The cDNA for PAT1a fused to a hemagglutinin tag was subcloned into pRK5 vector. Lipofectamine 2000 reagent (Invitrogen) was used for transient transfection of expression vectors for human formyl peptide receptor 1 (FPR1) (2.5 μg), C5aR (2.5 μg), P-Rex1 (0.5 μg), and pRK5-PAT1a (1 μg) into COSphox cells in a 100-mm culture dish (0.5–1 × 106 cells). Cells were analyzed 24 h later for superoxide anion production based on isoluminol enhanced chemiluminescence (ECL) as previous described (33, 34). The ECL counts per second was continually recorded for 5–10 min before and 20–30 min after stimulation with fMLF (1 μM), C5a (100 nM), or PMA (200 ng/ml). The relative level of superoxide produced was calculated based on the integrated ECL during the first 10 min after agonist stimulation. Whole-cell lysate was prepared for Western blotting.
Immunoprecipitation
The technique used for cell lysis and cytochrome b558 immunoprecipitation was adapted from previously described protocols (15, 18). Cells were lysed by resuspending (5 × 107 cells/ml) in lysis buffer (50 mM Tris-HCl buffer [pH 7.4], 1% Triton X-100, 25 mM NaF, 2 mM β-glycerophosphate, P8340 protease inhibitor mixture [1:1000 dilution], 150 mM NaCl, and 1 mM DFP). Following sonication on ice and mixing by rotation, the lysates were centrifuged at 114,000 × g for 30 min at 4°C. The supernatants were diluted twice in the same buffer without Triton X-100. Immunoprecipitation was performed by addition of anti-gp91phox or anti-p22phox or anti–cytochrome b558 or control IgG (1:200 dilution) mAbs and Protein A/G Beads (Santa Cruz Biotechnology, Santa Cruz, CA) saturated with BSA and incubated for 4 h. Beads were then washed four times, and proteins were denaturated in sample buffer by boiling at 100°C for 3 min.
Confocal microscopy
Following treatment with fMLF or PMA, neutrophils were spotted using round, gaped filter paper on poly-l-lysine–coated glass slides and allowed to dry in a humid chamber by gravity sedimentation, then fixed with 2% paraformaldehyde for 10 min, permeabilized with 0.2% Tween 20 for 15 min at 37°C, and blocked with 5% BSA in PBS. Cells were then incubated overnight at 4°C with rabbit anti-p22phox polyclonal Ab (1:200) and mouse anti-PAT1 mAb (1:200) diluted in 1% BSA/PBS. After washing, cells were incubated with Alexa Fluor 488 (green)–conjugated goat anti-rabbit Ab (1:200) and Alexa Fluor 568 (red)–conjugated goat anti-mouse (1:200) for 1 h at room temperature in the dark. Nuclei were stained with TO-PRO-3 iodide (Invitrogen). Stained cells were examined with a Zeiss LSM 510 confocal microscope (63/1.4 numerical aperture objective), and the images were imported into a laser-scanning microscope image browser for analysis. The designation “Merge” corresponds to colocalization of PAT1 and p22phox.
Pat1 silencing by small interfering RNA in human monocytes
To achieve PAT1 silencing, small interfering RNAs (siRNAs) (Santa Cruz, CA) were used. We used a liposome-based technology to deliver the nucleotides (35). Freshly isolated monocytes resuspended in RPMI 1640 without serum and antibiotics at the density of 1.5 × 106 cells. Lipids (HiPerFect; Qiagen) and 200 nM PAT1 siRNA were allowed to bind at the ratio of 1:1 (v/v) for 20 min at room temperature. The complex was then added to a new six-well plate and then harvested cells were added, dripping directly to the complex. Transfection was carried out for 4 h, followed by addition of new complete RPMI 1640 media, and cells were analyzed 1–2 d later (35).
PAT1a overexpression in monocytes
Transfection of human blood monocytes was performed in the Amaxa (Lonza, Cologne, Germany) system, as recommended by the company. In brief, after centrifugation (1800 rpm for 8 min), monocytes were resuspended in supplemented Human Monocyte Nucleofector solution (Amaxa) to a final concentration of 5 × 106 cells/100 μl; 100 μl of cell suspension was mixed with 2 μg of either pcDNA3.1-PAT1a plasmid or empty vector. After transfection with the program Y-001, monocytes were seeded in 24-well plates at a density of 5 × 106 cells per well in 2-ml supplemented medium and analyzed after 24–36 h of culture. Transfection efficiency was between 40 and 60%, as verified by immunofluorescence.
Measurement of superoxide anion production
Monocytes (5 × 105) were suspended in 0.5 ml HBSS containing 1 mg/ml cytochrome C at 37°C. cytochrome C reduction at 550 nm was then measured before (resting) and after the addition of stimuli (10−6 M fMLF or 100 nM PMA) to the cells. OD was recorded with a spectrophotometer at 550 nm during 15 min. Superoxide dismutase (2.5 U) was used in each experiment to inhibit cytochrome C reduction, ensuring superoxide production. The quantification of superoxide anions was calculated using ε = 21.1 mM−1 cm−1. Total superoxide anion production was calculated and expressed as nanomoles/15 min/1 million cells or as the initial rate nanomoles/min/1 million cells.
Measurement of luminol-amplified chemiluminescence
Monocytes (5 × 105) were suspended in 0.5 ml HBSS containing 10 μM luminol at 37°C. ROS production was then measured before (resting) and after the addition of stimuli (10−6 M fMLF or 100 nM PMA) to the cells. Luminol-amplified chemiluminescence was recorded with a luminometer (Berthold-Biolumat LB937) (36, 37).
SDS-PAGE and Western blotting
Gel electrophoresis and protein transfer to nitrocellulose membranes were performed with classical techniques (31). Nitrocellulose membranes were blocked with 5% nonfat dry milk in borate-buffered saline (pH 8.4) (100 mM boric acid, 25 mM borax, and 75 mM NaCl) for 1 h at room temperature and then incubated with 1:1000 mouse anti-gp91phox, -p22phox, or -PAT1 mAb overnight. The membranes were then washed extensively and incubated with HRP-conjugated 1:5000 anti-mouse IgG for 1 h at room temperature. Blots were visualized by using ECL Western blotting reagents (Amersham Pharmacia).
Statistical analysis
Data were analyzed with the GraphPad Prism 5 software. Differences between groups were analyzed by the one-way ANOVA test with Tukey multiple comparison posttest. The *p < 0.05, **p < 0.01, and ***p < 0.001 values were considered as significant.
Results
Identification of PAT1 as a p22phox-interacting protein
To identify potential p22phox-interacting proteins, we performed a yeast two-hybrid screen of a human spleen cDNA library using the cytosolic C-terminal region of p22phox (residues 132–195) (Fig. 1A). This construct does not induce background as manifested by the absence of transactivation of the two reporter genes, His3 and LacZ. Among the histidine and β-galactosidase–positive clones, one clone has passed all the nonspecific interaction controls; in the absence of the bait, the pACT2 isolated from this clone does not induce the transactivation of both reporter genes (Fig. 1B). This clone contained a 886-kb insert whose sequence matched the human PAT1 sequence, encoding aa 1–277 (Supplemental Fig. 1). To confirm the p22phox/PAT1 interaction detected in the two-hybrid experiments, we performed a GST pulldown assay using purified recombinant human PAT1a (99% identical to PAT1) and GST-p22phox (residues 132–195). Results obtained by Western blots and protein quantification show that GST-p22phox was able to pulldown PAT1a, whereas GST alone was not (Fig. 1C, left and right). This interaction was also confirmed by an overlay assay, Western blots, and quantification of the proteins (Fig. 1D, left and right). To investigate if PAT1 also associates with gp91phox or not, GST pulldown assay was also performed using the GST fusion protein containing the cytoplasmic C-terminal domain of gp91phox (aa 291–570). Results show that GST-gp91phox-Cter was not able to interact with PAT1 (Supplemental Fig. 2). These results demonstrate the specific interaction between PAT1 and p22phox, but also show a direct association without the need for any other protein or agent.
Schematic representation of p22phox cytosolic tail and its interaction with PAT1 in the yeast two-hybrid system, in the GST pulldown assay and overlay assay. (A) Human p22phox is a transmembrane protein with two transmembrane domains and a long cytosolic carboxy terminal sequence predicted from aa 132–195. (B) Summary of results obtained from the two-hybrid system. The p22phox cytosolic C-terminal region (aa 132–195) was fused to the LexA protein (pLex10 vector) as bait to screen a human spleen cDNA library in pACT2 vector. The histidine- and β-galactosidase–positive clone was designed as (+), and negative result was designed as (−). Specificity of p22phox (aa 132–195)/PAT1 interaction: the pACT2-PAT1 clone was used to verify its interaction with the empty pLex vector and with pLex10-p22phox (aa 132–195). The histidine- and β-galactosidase–positive clones were designed as (+), and negative result was designed as (−). (C) Recombinant human PAT1 (PAT1) was incubated with GST-p22phox (aa 132–195) or with GST protein and then with glutathione (GSH)–agarose beads for 1 h. After this incubation, the beads were washed three times, and proteins were analyzed by SDS-PAGE and Western blot using anti-GST or polyclonal anti-PAT1 Abs (left). Bound PAT1 and GST were quantified using ImageJ 1.43u software (Wayne Rasband, National Institutes of Health), their ratio was calculated, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 GST-p22phox as compared with GST alone (right). (D) Recombinant human PAT1 was analyzed by SDS-PAGE and overlay assay using GST alone or GST-p22phox, and the proteins were detected using anti-GST Ab (left). Bound PAT1 were quantified using ImageJ 1.43u software, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 GST-p22phox as compared with GST alone (right).
To study the PAT1/p22phox interaction in intact cells, we used the COSphox cell line system (i.e., COS cells transfected with all the components of the phagocytes NADPH oxidase) (33, 34). Transfection of COSphox cells with hemagglutinin-tagged PAT1-pRK5 plasmid resulted in a clear overexpression of PAT1, as shown by Western blots and protein quantifications data (Fig. 2A, left and right). Immunoprecipitation of cytochrome b558 using an anti-gp91phox or anti-p22phox Ab showed that PAT1 coimmunoprecipitated with the complex (Fig. 2B), whereas the use of control IgG did not demonstrate the presence of PAT1. These results clearly demonstrate that PAT1 interacts with p22phox in intact cells using a reconstituted cell system.
Overexpression of PAT1 in COSphox and coimmunoprecipitation with p22phox and gp91phox. (A) COSphox cells were transfected with a plasmid-expressing PAT1a (pRK5-PAT1a-hemagglutinin [HA]) or control plasmid without PAT1a and lysed in lysis buffer. Proteins were analyzed by SDS-PAGE and Western blot with anti-HA, anti-PAT1, anti-p22phox, and anti-p47phox Abs (left). PAT1 and p22phox were quantified using ImageJ 1.43u software, their ratio was calculated, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 pRK5-PAT1 as compared with pRK5 (right). (B) Proteins were immunoprecipitated from resting COSphox cells lysates with anti-gp91phox or anti-p22phox Ab or control IgG and analyzed by SDS-PAGE and Western blot with anti-PAT1a, anti-p22phox, or anti-gp91phox Abs (left). Bound and total PAT1 were quantified using ImageJ 1.43u software, their ratio was calculated, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 anti-gp91phox and anti-p22phox as compared with IgG (right). IP, immunoprecipitation; I.B, immunoblot.
PAT1 is expressed in human neutrophils, monocytes, and lymphocytes and it colocalizes with p22phox
The expression of PAT1 has been observed in some cells, such as epithelial cells (23–26), but its presence in human neutrophils and other leukocytes had not been yet investigated. To look for the expression of PAT1 in these human blood cells, the total RNAs were extracted from highly purified neutrophils, monocytes, and lymphocytes, and PAT1 mRNA expression was analyzed using specific primers and real-time quantitative PCR cDNA quantification. Results show that PAT1 mRNA was expressed in the three types of blood cells at comparable levels (Fig. 3A). To validate the expression of PAT1 protein, equal amounts of cells were analyzed by SDS-PAGE and Western blot using two different specific Abs against human PAT1. Results with both Abs show that PAT1 protein is expressed in the three cell types (Fig. 3B). Confocal microscopy analysis confirmed PAT1 expression in human neutrophils (Fig. 3C, PAT1). To investigate PAT1/p22phox interaction in neutrophils, we first explored the possibility of colocalization in intact neutrophils using confocal microscopy. Interestingly, in resting neutrophils, PAT1 showed a cytoplasmic, nuclear, and perinuclear localization, whereas p22phox showed granular and perinuclear localization (Fig. 3C). In resting neutrophils, PAT1 colocalized partially with p22phox, especially in the peri-nuclear region. We found that, upon stimulation with fMLF (10−6 M) or PMA (100 ng/ml), the distribution of PAT1 is modified; it was mainly recruited at the plasma membrane, where it displays colocalization with p22phox (Fig. 3C). A PAT1/p22phox colocalization was also clearly found in resting and PMA-stimulated monocytes (Supplemental Fig. 3). These results suggest that translocation of the PAT1 to the plasma membrane takes place following stimulation. To confirm this result, cytoplasmic, nuclear, cytosolic, and membrane extracts were prepared from resting and stimulated neutrophils. Protein analysis shows that, in resting cells, PAT1 is mainly located in the nucleus and in cytosol at a lower extent and that stimulation of neutrophils induced a decrease of the PAT1 level in the nuclear fraction with an increase in the cytosolic or cytoplasmic fractions as well as a translocation to the membrane (Fig. 3D, 3E).
Expression of PAT1 in human neutrophils, monocytes, and lymphocytes and subcellular localization in neutrophils. (A) PAT1 mRNA expression: freshly isolated neutrophils, lymphocytes, and monocytes were used to purify mRNA. One microgram of mRNA was retrotranscribed into cDNA, which was quantified by real-time PCR using specific primers amplifying common regions of the PAT1 isoforms. Arbitrary units were attributed based on EΔCT calculation (mean ± SEM; n = 4). (B) PAT1 protein expression: recombinant human PAT1 (rhPAT1) (5 ng) and resting human neutrophils, monocytes, and lymphocytes were lysed (1 × 106 cells) and analyzed by SDS-PAGE and Western blot with a polyclonal or an anti-PAT1 mAb. (C) Neutrophils in the resting state or treated with fMLF (10−6 M) or PMA (100 ng/ml) were fixed and permeabilized as described in the Materials and Methods section. Cells were incubated with a rabbit anti-p22phox Ab and a mouse anti-PAT1 Ab, followed by incubation with Alexa Fluor 488 (green)–conjugated goat anti-rabbit and Alexa Fluor 568 (red)–conjugated goat anti-mouse. Nuclei were stained with TO-PRO-3. Stained cells were examined with a confocal microscope, and the images were analyzed. The overlap of PAT1 and p22phox is clearly seen as the white signal (arbitrary color) in the merged image and is enriched in the perinuclear region of resting neutrophils and at the membrane periphery in the activated cells. (D) Neutrophils in the resting state or treated with fMLF (10−6 M) or PMA (100 ng/ml) were lysed, and cytoplasm and nuclei were prepared. Proteins were analyzed by SDS-PAGE and Western blot with anti-PAT1, anti-actin, and anti-histone Abs. (E) Neutrophils in the resting state or treated with fMLF (10−6 M) or PMA (100 ng/ml) were lysed. Cytosol and membranes were prepared. Proteins were analyzed by SDS-PAGE and Western blot with anti-PAT1, anti-p22phox, and anti-p47phox Abs. (F) Membrane PAT1 and p22phox were quantified using ImageJ 1.43u software, their ratio was calculated, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 fMLF and PMA as compared with resting cells (Rest.) (right).
PAT1 interacts with cytochrome b558 in human neutrophils
To further investigate the PAT1/p22phox interaction in human neutrophils, we used a coimmunoprecipitation assay from intact neutrophils. Cell lysates from resting or fMLF- or PMA-stimulated neutrophils were immunoprecipitated with anti–cytochrome b558 7D5 Ab, which recognizes extracellular gp91phox/p22phox complex (also known as the cytochrome b558), without interfering with intracellular interactions. Results show that PAT1 was detected in the cytochrome b558 complex containing gp91phox and p22phox of resting and stimulated neutrophils, whereas it was less detected in beads with control IgG isotype (Fig. 4A). The PAT1/cytochrome b558 interaction was further confirmed by coimmunoprecipitation assay using the specific NOX2 Ab in COSphox cells transfected with the PAT1 and the FPR plasmids (Fig. 4B), whereas the use of control IgG did not demonstrate the presence of PAT1. Protein quantification analysis of several Western blots showed that fMLF induced a significant increase of PAT1/cytochrome b558 interaction in both cells, whereas PMA was less effective.
Coimmunoprecipitation of PAT1 with the cytochrome b558 in resting and activated neutrophils and COSphox cells. (A) Neutrophils in the resting state or treated with fMLF (10−6 M) or PMA (100 ng/ml) were lysed. Cytochrome b558 was immunoprecipitated using the 7D5 Ab (+). Control IgG was used as control (−). The beads were washed, and proteins were denaturated and analyzed by SDS-PAGE and Western blot using an anti-PAT1 or an anti-gp91phox or an anti-p22phox Ab (left). Bound and total PAT1 were quantified using ImageJ 1.43u software (Wayne Rasband, National Institutes of Health), their ratio was calculated, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 fMLF and PMA compared with resting cells (right). (B) The same experiment was performed with the COSphox cells transfected with the FPR and PAT1 plasmids (left). *p < 0.05 fMLF and PMA compared with resting cells (right). REST., resting.
PAT1 promotes superoxide anion production in intact cells
We further investigated the functional effects of the PAT1/p22phox interaction in living cells overexpressing PAT1. Because neutrophils have a short life span and are resistant to transfection procedures, we used alternative cells for PAT1 overexpression or PAT1 knockdown. First, we used COSphox cell lines, which are transfected with either the fMLF receptor (FPR) or the C5a receptor (34), with or without PAT1 plasmid. Transfection of PAT1 in COSphox cells clearly increased fMLF- and C5a-induced ROS production (Fig. 5A, 5B). Interestingly, constitutive and PMA-induced ROS productions were only slightly enhanced (Fig. 5C).
Overexpression of PAT1 in COSphox cells enhanced ROS production. COSphox cells were transfected with human FPR1 or hC5aR and the PAT1a expression constructs (PAT1a) or the empty vector (control) and stimulated with fMLF (1 μM), C5a (100 ng/ml), or PMA (200 ng/ml). ROS production was measured using isoluminol-amplified chemiluminescence (Chemilum.) in fMLF (A)–, C5a (B)–, and PMA (C)–stimulated cells and in nonstimulated (-NS) conditions. An example of each condition is shown in the left panel. Integrated Chemilum. (area under the curve) of five experiments was quantified in counts per second (cps). Mean ± SEM were calculated and presented in the right panels. The data were analyzed by two-way ANOVA using GraphPad Prism 5 software and shown as mean ± SEM from four independent experiments. *p < 0.05.
Second, we used the siRNA approach to inhibit endogenous PAT1 expression in monocytes. Western blots and quantification data show that monocyte PAT1 protein level was downregulated by PAT1 siRNA, but not by scramble control siRNA (Fig. 6A, left and right). We used the cytochrome C reduction assay to precisely monitor superoxide anion production and NADPH oxidase activation. Interestingly, inhibition of PAT1 expression resulted in total superoxide anion production (15-min period) inhibition in fMLF- and PMA-stimulated neutrophils (Fig. 6B); however, the initial rate of superoxide anion production was not affected (Fig. 6C). These results were further confirmed by luminol-amplified chemiluminescence technique (Supplemental Fig. 4). Furthermore, we overexpressed PAT1 in human monocytes using an Amaxa transfection system. As shown in Fig. 6D, transfection of human monocytes with a plasmid-encoding PAT1a resulted in overexpression of the protein. GFP expression was also used to check the transfection efficiency of monocytes (data not shown). Interestingly, monocytes overexpressing PAT1 showed enhanced superoxide anions production, as measured by the cytochrome C reduction assay (Fig. 6E), as well as ROS production, as measured by luminol-amplified chemiluminescence of fMLF- and PMA-stimulated monocytes (data not shown).
Effect of inhibition of PAT1 expression and overexpression on human monocytes ROS production. (A) Human monocytes were freshly prepared and transfected with either scramble primers (Scr) or siRNA/PAT1 (si-PAT1) primers for 48 h. Proteins were analyzed by SDS-PAGE and Western blot with anti-PAT1 and anti-actin Abs (left). PAT1 and actin bands were quantified using ImageJ 1.43u software (Wayne Rasband, National Institutes of Health), their ratio was calculated, and the values are expressed as means ± SEM of three independent experiments. *p < 0.05 si-PAT1 compared with Scr (right). Superoxide anion production was measured using the cytochrome C reduction assay at 550 nm by fMLF (10−6 M)– and PMA (100 ng/ml)–stimulated cells treated with Scr and si-PAT1. Total superoxide anion production was calculated and expressed as nanomoles/15 min/1 million cells (B) or as the initial rate nanomoles/min/1 million cells (C) as total. *p < 0.05 si-PAT1 compared with Scr. (D) Human monocytes were freshly prepared and transfected with either a plasmid-expressing PAT1a (pcDNA3.1-PAT1a; pcDNA-PAT1) or with an empty pcDNA3.1 plasmid (pcDNA). Proteins were analyzed by SDS-PAGE and Western blot with anti-PAT1 and anti-actin Abs and quantification analysis. (E) Superoxide anion production was measured using the cytochrome C reduction assay at 550 nm by resting (REST.), fMLF (10−6 M)–, and PMA (100 ng/ml)–stimulated cells. Superoxide anion production was calculated and expressed as the total nanomoles/15 min/1 million cells. *p < 0.05 pcDNA-PAT1 compared with pcDNA.
Discussion
ROS production by the phagocyte NADPH oxidase NOX2 is an essential process for host defense against pathogens (3–5). However, excessive NADPH oxidase activation is believed to be involved in inflammatory reactions (38, 39). Tight regulation of NADPH oxidase activation is important to ensure that ROS are produced only when and where required. Several mechanisms, such as protein phosphorylation, GTPase activation, and protein/protein interactions, regulate NADPH oxidase activation (8–12). Results presented in this study uncover a novel mechanism of regulating NOX2 and ROS production based on p22phox/PAT1 interaction. We used different approaches to show the interaction between p22phox and PAT1; the yeast two-hybrid system, the GST pulldown assay, the overlay assay, confocal microscopy, and coimmunoprecipitation techniques. Interestingly, PAT1 does not interact with the gp91phox cytosolic region, suggesting that PAT1 interacts specifically with p22phox. We showed that this interaction is functional, PAT1 overexpression increased NADPH oxidase-derived ROS production in human monocytes and in the COSphox cells. Furthermore, we have shown that inhibition of PAT1 expression in human monocytes resulted in an inhibition of total superoxide anion production, but not the initial rate of superoxide production. This result suggests that PAT1 is not involved in the initiation of NADPH oxidase activation, but rather, it is required to sustain the activation in human monocytes.
PAT1 mRNA is expressed in different tissues, such as heart, brain, placenta, skeletal muscle, kidney, and pancreas (23–26). At least three isoforms of PAT1 have been described in the literature; they are almost identical, differing by only a few amino acids. These isoforms compose PAT1 itself, ARA67, which is 99.6% identical to PAT1 and PAT1a, which is 99% identical to PAT1. The sequence identified in this study is shared by the three isoforms; however, the isoforms expressed in human neutrophils remain undetermined. Because the three isoforms are 99% identical, and the yeast two-hybrid identified the common N-terminal domain as interacting with p22phox, all three isoforms could potentially interact with the p22phox cytosolic tail.
Interestingly, PAT1 enhances fMLF- and C5a-induced NADPH oxidase activation and, to a lesser extent, the PMA-induced activation. Because fMLF and C5a are physiological agonists that activate two different G-protein–coupled receptors, and PMA is a strong pharmacological activator of conventional protein kinase C, these results suggest that PAT1 could play a major role in G-protein–coupled, receptor-mediated NADPH oxidase activation. The conventional protein kinase C–mediated NADPH oxidase pathway may be regulated to a lesser extent by PAT1. More studies are required to try to understand this process.
PAT1 binds microtubules and is involved in intracellular trafficking of several proteins, including β-APP (23, 24), Us11 virus protein (40), and androgen receptor (26). It is not excluded that PAT1 could play a role in p22phox production or in the trafficking of p22phox-rich granules in phagocytes. We have confirmed that PAT1 binds to microtubules in vitro and was able to mediate p22phox binding (data not shown). Interestingly, cytochrome b558 was shown to undergo conformational changes and to bind to cytoskeleton with activated NADPH oxidase (41–43). PAT1 may facilitate the conformational changes of cytochrome b558 and its binding to cytoskeleton. Thus, PAT1 could control localization of NADPH oxidase activation in the cell.
p22phox expression is not restricted to phagocytes; it is expressed in several tissues and cells such as vascular smooth muscle cells, epithelial cells, endothelial cells, and neurons. In addition, p22phox is associated with other NOX enzymes such as NOX1, NOX3, and NOX4 (10, 44). We have recently shown that p22phox could bind to NOX5 in monocyte-derived DC35. Using the yeast two-hybrid system and a vascular smooth muscle cell cDNA library, Lyle et al. (45) have identified Poldip2 protein as a partner of p22phox. Poldip2 increased NOX4 enzymatic activity in vascular smooth muscle cells and modulated cytoskeletal remodeling and cell migration. Our results and those reported by Lyle et al. (45) suggest that p22phox could have several partners that probably depend on p22phox tissue expression.
In conclusion, we have shown that PAT1 is able to bind to the region of the p22phox cytosolic tail composed of aa 132–195 and is capable of enhancing NADPH oxidase activation in vitro and in intact cells. PAT1 was first identified in human cells as binding to APP and to microtubules. In this study, we describe a novel function of PAT1 in human phagocytic cells. Several other cells express p22phox in association with other NOX enzymes; therefore, PAT1 could be a novel regulator of superoxide anion and ROS production in these cells.
Disclosures
The authors have no financial conflicts of interest.
Acknowledgments
We thank Dr. Mary Dinauer (Riley Hospital for Children, Indiana University School of Medicine, Indianapolis, IN) for the generous gift of the p22phox plasmid, Dr. Marie-Christine Lecomte (INSERM U1134) for help on the two-hybrid system, and Dr. Stefan Kins (University of Kaiserslautern, Kaiserslautern, Germany) for the generous gift of the PAT1a cDNA–containing plasmid.
Footnotes
This work was supported by grants from the Agence Nationale de la Recherche, Arthritis Fondation Courtin, Vaincre La Mucoviscidose, INSERM, CNRS, Labex Inflamex, and University Denis-Diderot Paris7.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- APP
- amyloid precursor protein
- DFP
- diisopropylfluorophosphate
- ECL
- enhanced chemiluminescence
- FPR
- formyl peptide receptor
- ROS
- reactive oxygen species
- siRNA
- small interfering RNA.
- Received April 30, 2018.
- Accepted December 22, 2018.
- Copyright © 2019 by The American Association of Immunologists, Inc.