Functional Expression of the P2X7 ATP Receptor Requires Eros

Phosphatidylserine (PtdSer), an abundant phospholipid in the plasma membrane, is confined to the inner membrane leaflet by the action of flippases (1–3). In various biological processes, scramblases nonspecifically distribute phospholipids between the outer and inner leaflets, thereby exposing PtdSer. Consequently, cells undergoing apoptosis irreversibly expose PtdSer, whereas activated platelets, capacitated sperm, activated lymphocytes, and transformed cancerous cells reversibly expose PtdSer.

In contrast, flippases specifically translocate PtdSer and phosphatidylethanolamine from the outer to inner leaflet of the plasma membrane in an ATP-dependent manner (2). We previously showed that among 14 P4-ATPases in humans, ATP11A and ATP11C are ubiquitously expressed in various cells, are located at the plasma membrane, and are responsible for maintaining PtdSer at the inner leaflet of the plasma membrane together with their partner, CDC50A (4, 5). Two families of membrane proteins, XK-related (XKR) and transmembrane protein 16 (TMEM16), each carrying 10 putative transmembrane regions, function as scramblases (6–9). XKR8 is a member of the XKR family and is localized to the plasma membrane together with basigin or neuroplasin (10). XKR8 is ubiquitously expressed in various cells and is activated by caspase-mediated cleavage or kinase-mediated phosphorylation to expose PtdSer on the cell surface (11). TMEM16F, also called anoctamin 6, is ubiquitously expressed and functions as a scramblase in response to a high concentration of Ca²⁺ (6). Loss-of-function mutations in TMEM16F cause human and mouse platelets to show delayed PtdSer exposure and microparticle release (12, 13), indicating that TMEM16F is responsible for scrambling the phospholipids in activated platelets.

One of the biological processes in which cells expose PtdSer is inflammation, which can result from tissue injury (14). During this process, a large amount of ATP is released from apoptotic, necrotic, or injured cells, and binds to purinergic receptors (P2X and P2Y) (15). Among the two known families of purinergic receptors, human (h)– and mouse (m)P2X7, also called P2RX7, P2X7R, or P2Z, is an evolutionarily well-conserved membrane protein that carries two transmembrane regions with cytoplasmic short N-terminal and long C-terminal tails (16). P2X receptors, including P2X7, form a homo- or heterotrimer (17, 18), and the binding of ATP or its agonist 2′(3′)-O-(4-benzoylbenzoyl)ATP (BzATP) to P2X7 causes rapid Ca²⁺ uptake and PtdSer exposure by the cell, followed by microvesicle release, pore formation, and pyroptosis (19–22). In addition, P2X7 is essential for the ATP-triggered IL-1β production in endotoxin-primed macrophages (23, 24). How a single protein (P2X7) mediates these various biological processes has been elusive, although one report claims that TMEM16F mediates the downstream signaling of P2X7, including the PtdSer exposure (25).

In this study, we found that the mouse T cell line WR19L responded strongly to ATP by exposing PtdSer. TMEM16F deficiency had no effect on the ATP-induced PtdSer exposure. The screening of a CRISPR/Cas9 knockout library identified essential for reactive oxygen species (Eros) or cytochrome B-245 chaperone 1 (CYBC1) as an essential factor for the P2X7-mediated PtdSer exposure, Ca²⁺ uptake, pore formation, and IL-1β production. Eros was localized to the endoplasmic reticulum and appeared to function as a chaperone for P2X7 by forming a stable homotrimeric complex.

Mouse WR19L cells (TIB-52; American Type Culture Collection [ATCC]) were grown in RPMI 1640 supplemented with 10% FCS. Human THP-1 cells (TIB-202; ATCC) were cultured in RPMI 1640 supplemented with 10% FCS and 55 μM 2-ME. HEK293T cells (CRL-1573; ATCC) were grown in DMEM containing 10% FCS.

The pNEF-BOS vector was derived from pEF-BOS (26), into which an SV40 early promotor-driven neomycin-resistance gene was introduced from pSV2-neo (27). The lentiCas9-Blast vector (28), mouse CRISPR Knockout Pooled Library (genome-scale CRISPR/Cas9 knockout [GeCKO] v2) (28), and pX330 and pX459v2 plasmids (29) were from Addgene. pCMV–vesicular stomatitis virus (VSV)-G–Rous sarcoma virus (RSV)-rev and pCMV-VSV-G (30) were provided by H. Miyoshi (RIKEN BioResource Center). The pMXs-puro retroviral vector (31) and the pGag-pol-IRES-bsr packaging plasmid (32) were from T. Kitamura (Institute of Medical Science, University of Tokyo). The pAdVAntage and pLVSIN–EF-1α lentiviral vectors were purchased from Thermo Fisher Scientific and Takara Bio, respectively.

The Alexa 488–conjugated rat anti-mouse P2X7 mAb (clone Hano43) was from Bio-Rad Laboratories. The HRP mouse anti-FLAG M2 mAb, anti-FLAG M2-conjugated magnetic beads, and 3× FLAG peptide were from Merck. The HRP rabbit anti-GFP Ab was from Medical & Biological Laboratories. The rabbit anti-rat P2X7 Ab was from Alomone Labs. The rabbit anti-human Eros Ab was from Atlas Antibodies. The mouse anti–Na⁺/K⁺-ATPase mAb (clone 464.6) was from Abcam. The HRP–goat anti-rabbit Igs Ab and the HRP–goat anti-mouse Igs Ab were from Agilent Technologies. The BzATP and ATP were purchased from Wako Pure Chemical and Nacalai Tesque, respectively. The BAPTA-AM and 1-[2-amino-5-(2,7-difluoro-6-acetoxymethoxy-3-oxo-9-xanthenyl)phenoxy]-2-(2-amino-5-methylphenoxy)ethane-N,N,N′,N′-tetraacetic acid, tetra(acetoxymethyl) ester (Fluo-4 AM) were from Dojindo Molecular Technologies. Cy5-labeled annexin V was from BioVision. Ionomycin and propidium iodide (PI) were from Merck. SYTOX Blue, YO-PRO-1, ER-Tracker Red and Hoechst 33342 were from Thermo Fisher Scientific. The 1-oleoyl-2-{6-[(7-nitro-2-1,3-benzoxadiazol-4-yl)amino]hexanoyl}-sn-glycero-3-phosphocholine (NBD-PC) was from Avanti Polar Lipids.

The TMEM16F, P2X7, and Eros genes were knocked out using the CRISPR/Cas9 system and the pX330 or pX459v2 plasmid, as described previously (5, 29). Complementary oligonucleotides carrying the sgRNA target sequence were the following: mTMEM16F, 5′-GGATGAAGTCGATTCGCCTC-3′; mP2X7, 5′-TGAGCGATAAGCTGTACCAG-3′; mEros, 5′-ATTTGTGGCTGTACAGAACT-3′; hP2X7, 5′-TGATGACAGGCTCTTTCCGC-3′; and hEros, 5′-TGGAAGCTCTTCTACGTCAC-3′. The oligonucleotides were ligated into pX330 or pX459v2, and the resultant plasmid DNA was introduced into WR19L or THP-1 cells by electroporation using an NEPA21 Super Electroporator (Nepa Gene). In some cases, the transfection was carried out twice with a 3-d interval. At 20–30 h after transfection with the pX459v2 vector, the cells were treated with 1 μg/ml puromycin for 30 h. Single clones were isolated by limiting dilution and were genotyped by sequencing the sgRNA target region of the corresponding chromosomal gene.

To express P2X7 and Eros, the coding sequences for mP2X7k (FJ436444), mEros (NM_144832), hP2X7a (NM_002562), and hEros (NM_001100407) were prepared by RT-PCR using RNA from WR19L (mP2X7k and mEros), or THP-1 (hP2X7a and hEros) cells. These cDNAs were tagged with FLAG, hemagglutinin (HA), enhanced GFP (EGFP), or mCherry at the C terminus and inserted into pNEF-BOS (mP2X7k), pMXs-puro (mP2X7k and mEros), or pLVSIN–EF-1α (hP2X7a and hEros). The authenticity of the expression plasmids was confirmed by DNA sequencing. The expression plasmids constructed with the pNEF-BOS vector were digested singly with a restriction enzyme and introduced into WR19L cells by electroporation. Stable transformants were selected by culturing in the presence of 2 mg/ml G418. For viral transformation, pMXs-puro–based vectors were introduced into HEK293T cells together with pGag-pol-IRES-bsr, pCMV-VSV-G, and pAdVAntage, whereas pLVSIN–EF-1α–based vectors were introduced with pCAG-HIVgp and pCMV-VSV-G-RSV-rev. Retroviruses and lentiviruses in the culture supernatant were concentrated at 4°C by centrifugation at 6000 × g for 16 h and used to infect WR19L and THP-1 cells, respectively. Transformants were selected in the presence of 1 μg/ml puromycin. If necessary, mP2X7-, EGFP-, or mCherry-positive cells were sorted using a FACSAria II (BD Biosciences).

The Eros^−/− mice (Cybc1^{tm1a(KOMP)/Mmucd}) were obtained from the Mutant Mouse Resource and Research Centers and maintained on a C57BL/6N genetic background. All mice were housed in a specific pathogen-free facility at Research Institute for Microbial Diseases, Osaka University, and all mouse studies were approved by the Ethics Review Committee for Animal Experimentation of Research Institute for Microbial Diseases, Osaka University.

To detect mP2X7 on WR19L cell transformants, 5 × 10⁵ cells were washed with PBS containing 2% FCS (PBS/FCS), and incubated on ice for 30 min in 100 μl of PBS/FCS containing 40-fold-diluted Alexa 488–anti-mP2X7 mAb. The cells were then washed with PBS/FCS, suspended in 250 μl of PBS/FCS containing 250 nM SYTOX Blue, and analyzed by flow cytometry with a FACSCanto II (BD Biosciences). The data were analyzed by FlowJo software (BD Biosciences).

The P2X7-mediated PtdSer exposure was analyzed by annexin V binding, followed by flow cytometry, as described previously with slight modifications (6). In brief, 1.8 × 10⁶ cells were washed with PBS, resuspended in annexin buffer (10 mM HEPES–NaOH [pH 7.5], 140 mM NaCl, and 2.5 mM CaCl₂) containing 1000-fold-diluted, Cy5-labeled annexin V and 2.5 μg/ml PI, and preincubated at 20°C for 5 min or at 4°C for 10 min. The cells were then stimulated at 20°C or 4°C with ionomycin, BzATP, or ATP and analyzed by flow cytometry using a FACSCanto II. The P2X7-dependent internalization of phosphatidylcholine (PtdCho) was assayed by the internalization of NBD-PC as described previously (6). In brief, 1.8 × 10⁶ cells were suspended in 300 μl of annexin buffer and preincubated at 4°C for 10 min. Aliquots of 250 μl of the cell suspension were mixed with an equal volume of annexin buffer containing 500 nM NBD-PC and 1 mM ATP and incubated at 4°C. A 90-μl aliquot of this mixture was then mixed with 150 μl of annexin buffer containing 5 mg/ml fatty acid–free BSA, incubated on ice for 1 min, and analyzed with a FACSCanto II. The incorporation of YO-PRO-1 was similarly assayed by flow cytometry (33). In brief, 1.3 × 10⁶ cells were preincubated at 4°C for 10 min in 650 μl of annexin buffer, mixed with 650 μl of annexin buffer containing 4 μM YO-PRO-1, 500 nM SYTOX Blue, and 1 mM ATP, incubated at 4°C, and analyzed with a FACSCanto II.

To monitor intracellular Ca²⁺, 7 × 10⁵ cells were incubated with 1.4 ml of 4 μM Fluo-4 AM in HEPES/NaCl buffer (10 mM HEPES–NaOH [pH 7.5] and 140 mM NaCl) at 25°C for 15 min, washed, incubated at 25°C for 15 min in 350 μl of HEPES/NaCl buffer, and chilled at 4°C for 10 min. A 150-μl aliquot of the cell suspension was then mixed with an equal volume of HEPES/NaCl buffer containing 5 mM CaCl₂, 500 nM SYTOX Blue, and 1 mM ATP, incubated at 4°C for 5 min, and analyzed by flow cytometry with a FACSCanto II.

Cells stably expressing mP2X7k-EGFP, mEros-EGFP, and mEros-mCherry were washed with HBSS supplemented with 2% FCS (HBSS/FCS). The cells were suspended in HBSS/FCS containing 5 μg/ml Hoechst 33342, seeded into a glass-bottom dish (Matsunami), and observed by an IX81 confocal fluorescence microscope (Olympus). In some cases, cells were stained at 37°C for 15 min with 1 μM ER-Tracker Red before observing by microscope.

Several mouse tissues express two splice variants of P2X7, P2X7a, and P2X7k, which use a different exon 1 (34). To determine which variant is expressed in WR19L cells, the total RNA was prepared from the WR19L cells and the resident peritoneal macrophages of C57BL/6J mice using an RNeasy Mini Kit (QIAGEN), and reverse transcribed using SuperScript III Reverse Transcriptase (Thermo Fisher Scientific) or a High-Capacity RNA-to-cDNA Kit (Thermo Fisher Scientific). cDNA was subjected to PCR with PrimeSTAR GXL DNA Polymerase (Takara Bio) and the following primers: specific forward primers, 5′-TTTTTAATTAAGCCACCATGCCGGCTTGCTGCAGCTG-3′ on exon 1 for mP2X7a and 5′-TTTTTAATTAAGCCACCATGCTGCCCGTGAGCCAC-3′ on exon 1′ for mP2X7k, and a common reverse primer, 5′-AAAGAATTCGTAGGGATACTTGAAGCCAC-3′ on exon 13.

Genome-wide CRISPR screening was carried out essentially as described previously (35, 36). In brief, TMEM16F^−/−P2X7^−/−(DKO) WR19L cells established by the CRISPR/Cas9 system were stably transformed with pNEF-BOS–mP2X7k to generate DKO-WR19L–mP2X7k cells. A lentivirus-carrying Cas9–FLAG was produced by transfecting HEK293T cells with lentiCas9-Blast, pCAG-HIVgp, and pCMV-VSV-G-RSV-rev, and was used to infect DKO-WR19L–mP2X7k cells. Transformants were selected in the presence of 10 μg/ml blasticidin S, and the clone expressing Cas9–FLAG was identified by Western blotting with anti-FLAG. The lentivirus for the sgRNA library was produced by transfecting HEK293T cells (1 × 10⁷) with 17 μg of GeCKO v2 library (A and B) DNA, 8 μg pCAG-HIVgp, and 5 μg pCMV-VSV-G-RSV-rev and concentrated at 4°C by centrifugation at 6000 × g for 16 h. This virus was used to infect DKO-WR19L–mP2X7k/Cas9 cells at a multiplicity of infection of 0.3. In brief, 2 × 10⁷ cells in 48-well plates at 4 × 10⁵ cells per well were spin infected at 700 × g for 1 h at 30°C in the presence of 10 μg/ml polybrene. The cells were then cultured for 4 d in medium containing 1 μg/ml puromycin and 800 μg/ml G418, with one passage with a 3-fold dilution. The cells were then cultured without the antibiotics for 2 d, during which they were passaged once with a 4-fold dilution.

The resulting DKO-WR19L–mP2X7k/Cas9/GeCKO cells (3 × 10⁷) were washed with PBS, resuspended in 2 ml of annexin buffer, and chilled at 4°C for 10 min. One milliliter of chilled annexin buffer containing 180 μM BzATP was then added to the cells, and the mixture was incubated at 4°C for 5 min. After a 15-fold dilution with chilled annexin buffer, the cells were collected by centrifugation at 4°C at 300 × g for 3 min, suspended in 3 ml of annexin buffer containing 1000-fold-diluted, Cy5-labeled annexin V and 2.5 μg/ml PI, and subjected to cell sorting with a FACSAria II. The cell population with the lowest annexin V signal (∼0.8%) was collected and cultured for 3 d in the presence of 800 μg/ml G418 and for 2 d in the absence of G418. This procedure (stimulation with BzATP, sorting of the cells expressing low levels of PtdSer, and expansion) was performed three times.

Genomic DNA was prepared from the third-sorted cells using the QIAamp DNA Mini Kit (QIAGEN), and a portion of the integrated lentiviral DNA carrying the sgRNA sequence was amplified by PCR using PrimeSTAR GXL DNA Polymerase and primers (5′-AATGGACTATCATATGCTTACCGTAACTTGAAAGTATTTCG-3′ and 5′-CTTTAGTTTGTATGTCTGTTGCTATTATGTCTACTATTCTTTCC-3′). The PCR product was then subjected to PCR using PrimeSTAR HS DNA polymerase (Takara Bio) to attach adaptor sequences for next-generation sequencing (NGS) with a mixture of forward primers (NGS-Lib-Fwd-1–10) (36) and a common reverse primer (5′-CAAGCAGAAGACGGCATACGAGATGTGACTGGAGT TCAGACGTGTGCTCTTCCGATCTTCTACTATTCTTTCCCCTGCACTGT-3′). The resultant PCR product was purified using spin columns (Promega), quantified with a Quant-iT PicoGreen dsDNA Assay Kit (Thermo Fisher Scientific), and analyzed by NGS with MiSeq (Illumina) using the MiSeq Reagent Kit v3 (Illumina). The obtained reads were linked to the corresponding sgRNA sequence, and the abundance of each unique sgRNA was calculated using software that was custom prepared at Amelieff.

Whole-cell lysates were prepared by rotating cells at 4°C for 1 h in 50 mM Tris-HCl buffer (pH 7.5) containing 1% NP-40, 150 mM NaCl, and a protease inhibitor mixture (cOmplete, EDTA-free; Roche). The insoluble materials were removed by centrifugation at 4°C, 20,000 × g, for 20 min.

To prepare the membrane fractions from WR19L cells, 2.5 × 10⁷ cells were washed with PBS and then homogenized on ice in 2.7 ml of solution A [10 mM Tris-HCl (pH 7.5) and 1 mM (p-amidinophenyl)methanesulfonyl fluoride hydrochloride (p-APMSF)] using a Dounce homogenizer. After adding 2.7 ml of solution B (10 mM Tris-HCl [pH 7.5], 500 mM sucrose, 100 mM KCl, 10 mM MgCl₂, and 1 mM p-APMSF) to the homogenate, the nuclei and mitochondria were removed at 4°C by sequential centrifugations at 800 × g for 10 min and at 8000 × g for 10 min. The membrane fractions were then precipitated by centrifugation at 100,000 × g for 60 min and suspended in 600 μl of 20 mM Tris-HCl buffer (pH 7.5) containing 1% n-dodecyl-β-d-maltopyranoside (DDM), 50 mM KCl, 1 mM MgCl₂, 10% glycerol, 1 mM p-APMSF, and a protease inhibitor mixture (cOmplete, EDTA-free). The suspensions were homogenized by passing them through a 29G needle and solubilized by rotation at 4°C for 2 h. The insoluble materials were removed by centrifugation at 20,000 × g for 20 min, and the supernatant was used as the crude membrane fraction.

To prepare the membrane fraction from THP-1 cells, 1.3 × 10⁷ cells were treated with 120 ng/ml PMA for 3 d, and cultured for another 3 d without PMA. After washing with PBS, the cells were collected with a cell scraper and suspended in 2.5 ml of solution A. The cells were lysed with an ultrasonic liquid processor (Q55; Qsonica). After adding 2.5 ml of solution B, the membrane fraction was prepared as described above, and solubilized in 200 μl of 50 mM Tris-HCl buffer (pH 7.5) containing 1% NP-40, 150 mM NaCl, and a protease inhibitor mixture (cOmplete, EDTA-free).

To immunoprecipitate the mP2X7-mEros complex, 500 μl of crude membrane lysate (60 μg protein) from WR19L cell transformants expressing mP2X7-FLAG and mEros-EGFP was incubated at 4°C overnight with 10 μl (bed volume) of anti-FLAG M2 magnetic beads. After washing three times with 1 ml of 20 mM Tris-HCl buffer (pH 7.5) containing 0.05% DDM, 50 mM KCl, 1 mM MgCl₂, and 10% glycerol, proteins bound to the beads were eluted with 150 μg/ml 3× FLAG peptide in 100 μl of 20 mM Tris-HCl buffer (pH 7.5) containing 1% DDM, 50 mM KCl, 1 mM MgCl₂, and 10% glycerol.

For SDS-PAGE, whole-cell lysates or crude membrane lysates were incubated for 1 h at room temperature in SDS sample buffer (62.5 mM Tris-HCl [pH 6.8], 2% SDS, 10% glycerol, 2.5% 2-ME, and 0.005% bromophenol blue) and separated by electrophoresis on a 7.5% or 15% polyacrylamide gel (Nacalai Tesque). Precision Plus Protein Dual Color Standards (Bio-Rad Laboratories) were used as m.w. markers. For Blue native PAGE (BN-PAGE), samples were mixed with NativePAGE Sample Buffer (4×) and NativePAGE 5% G-250 Sample Additive (20×) and loaded onto a NativePAGE Novex 4–16% BisTris gel (Thermo Fisher Scientific). After electrophoresis at 150 V for 35 min at 4°C, the concentration of Coomassie Brilliant Blue (CBB) G-250 in the running buffer was changed from 0.02 to 0.002%, and the samples were subjected to further electrophoresis successively at 150 V for 25 min, 250 V for 30 min, and 350 V for 20 min. The NativeMark Unstained Protein Standard (Thermo Fisher Scientific) was used as m.w. markers.

For Western blotting, separated proteins in a gel were transferred to a polyvinylidene difluoride membrane (MilliporeSigma) immediately after SDS-PAGE or after incubating a BN-PAGE gel in SDS-PAGE running buffer (25 mM Tris-HCl [pH 8.3], 192 mM glycine, and 0.1% SDS) for 15 min at room temperature. The membranes were blocked with 5% skim milk and probed with HRP-labeled Abs, followed by detection with the Immobilon Western Chemiluminescent HRP substrate (Merck). Proteins on the polyvinylidene difluoride membrane were stained with CBB R-250 as a loading control.

THP-1 cells were treated with PMA to promote their differentiation to macrophages, and ATP-induced IL-1β secretion was performed as described previously (37, 38) with slight modification. In brief, THP-1 cells at a density of 2.5 × 10⁵ cells per well in 24-well plates were cultured in RPMI 1640 containing 10% FCS and 120 ng/ml PMA for 3 d. The PMA was removed, and the cells were further cultured successively for 2 d in RPMI 1640–10% FCS, and overnight in the same medium containing 100 ng/ml LPS. The cells were then stimulated for 5 h with 1–3 mM ATP in RPMI 1640–10% FCS. The culture medium was centrifuged at 4°C, 20,000 × g for 15 min, and IL-1β in the supernatant was quantified using an ELISA kit (Quantikine ELISA Human IL-1β/IL-1F2; R&D Systems).

For bone marrow–derived macrophages (BMDMs), bone marrow was prepared from the femurs of 3-wk-old, wild-type Eros^+/+ and Eros^−/− littermate mice and treated with RBC Lysis Buffer (Buffer EL; QIAGEN). The cells were cultured for 4 d in RPMI 1640 supplemented with 10% FCS, 55 μM 2-ME, and 10% of medium conditioned with CMG14-12 (39) (Medium M) by changing the medium every other day. The macrophages were suspended in PBS containing 1 mM EDTA using cell scrapers and collected by centrifugation at 300 × g for 3 min. The cells were suspended in Medium M, and 7.5 × 10⁴ cells per well were cultured in 48-well plates. One day later, the cells were primed with 100 ng/ml LPS for 3 h and stimulated for 6 h with ATP. The culture medium was centrifuged at 20,000 × g for 15 min at 4°C, and IL-1β in the supernatant was quantified using an ELISA kit (Quantikine ELISA Mouse IL-1β/IL-1F2; R&D Systems).

Ousingsawat et al. (25) reported that P2X7-induced PtdSer exposure is mediated by a Ca²⁺-dependent scramblase, TMEM16F. We examined this possibility using a mouse cell line derived from WR19L T cell leukemia. WR19L cells that were stimulated at 20°C with 3 μM ionomycin exposed PtdSer within 5 min (Fig. 1A, 1B). As we previously reported (8), this PtdSer exposure was completely inhibited by chelating the intracellular Ca²⁺ with 25 μM BAPTA-AM (Fig. 1A), and it was not observed in TMEM16F^−/− (16F^−/−) WR19L cells (Fig. 1B) prepared by the CRISPR/Cas9 system (Supplemental Fig. 1A). Treating WR19L cells at 20°C for 5 min with 300 μM BzATP, an agonist for P2X7 (40), strongly increased the PtdSer exposure (Fig. 1A). As expected, BzATP did not induce the PtdSer exposure in P2X7^−/−WR19L cells (Fig. 1B, Supplemental Fig. 1B). However, in contrast to the proposal by Ousingsawat et al. (25), the null mutation of TMEM16F had little effect on the BzATP-induced PtdSer exposure (Fig. 1B). Furthermore, preincubating WR19L cells with BAPTA-AM had no effect on the BzATP-induced PtdSer exposure (Fig. 1A), indicating that other TMEM16 family members with Ca²⁺-dependent scrambling activity (8) were probably not involved in this process. From these results, we concluded that the P2X7-induced PtdSer exposure in WR19L cells proceeds independently from the TMEM16F-mediated phospholipid scrambling.

There are several alternatively spliced forms of mP2X7 mRNA (41), two of which (P2X7a and P2X7k) are found in the thymus and spleen (34, 42). To examine which splice variants were expressed in WR19L cells, their RNA was subjected to RT-PCR for the P2X7a and P2X7k isoforms. As reported previously (42), the mouse macrophages exclusively expressed P2X7a mRNA, whereas WR19L cells predominantly expressed the P2X7k isoform (Fig. 2A). To confirm the contribution of mP2X7k to the BzATP-induced PtdSer exposure, TMEM16F^−/−P2X7^−/−(DKO)-WR19L cells were established using the CRISPR/Cas9 system (Supplemental Fig. 1B). As expected, the DKO-WR19L cells did not respond to ionomycin or BzATP (Fig. 1B). Transforming the DKO-WR19L cells with mP2X7k (DKO–mP2X7k) resulted in a high expression of mP2X7k on their cell surface (Fig. 2B) and strongly sensitized the cells to the BzATP-induced PtdSer exposure. That is, three times less concentration of BzATP than that for the wild-type cells was sufficient to activate the mP2X7k-transformed cells to expose PtdSer at 20°C for 5 min in the absence of TMEM16F, confirming that TMEM16F was not required for the P2X7-mediated PtdSer exposure.

To identify molecules that were required for the P2X7-mediated PtdSer exposure, DKO-mP2X7k cells were subjected to a CRISPR screen (35) (Fig. 2C). In this screen, a clone of the mP2X7k transformants that strongly and universally exposed PtdSer in response to 60 μM BzATP at 4°C for 5 min was identified by limiting dilution and cell sorting. The cloned cells were transformed with Cas9, and were infected with a lentivirus carrying a GeCKO library (28) at a multiplicity of infection of 0.3 to achieve <1 sgRNA-carrying lentivirus per host cell. The cells were then stimulated with 60 μM BzATP at 4°C for 5 min, and a 0.8% population that severely lost the ability to expose PtdSer was isolated by cell sorting (Fig. 2C). This sorting of cells that lost the ability to expose PtdSer in response to BzATP was repeated two more times. Flow cytometry analysis indicated that after the third sorting, some cell population almost fully lost the ability to expose PtdSer in response to BzATP (Fig. 2C). A deep sequencing analysis of the chromosomal DNA of the sorted cell population indicated that the sgRNAs for Eros (CYBC1) were the most enriched, suggesting that Eros was required for the P2X7-mediated PtdSer exposure.

To confirm that Eros was involved in the P2X7–mediated PtdSer exposure, the Eros in DKO-mP2X7k was knocked out by the CRISPR/Cas9 system (Supplemental Fig. 1C). The ability of the resultant cell line (TKO-mP2X7k) to expose PtdSer in response to BzATP or ATP was severely impaired, and this impairment was fully rescued by transforming the cells with mEros (TKO-mP2X7k/mEros) (Fig. 3A). P2X7 mediates not only PtdSer exposure, but also ATP-induced phospholipid scrambling (43), dye uptake (44), Ca²⁺ influx (16, 45), and IL-1β production (46). Accordingly, stimulating DKO-mP2X7k, but not DKO, cells with ATP promoted phospholipid scrambling, assayed by the incorporation of NBD-PC (Fig. 3B). This activity was not observed in TKO-mP2X7k cells and was rescued by transforming these cells with Eros. Similarly, ATP promoted the incorporation of YO-PRO-1 in DKO-mP2X7k cells. This effect was decreased in TKO-mP2X7k, and transforming TKO cells with mEros together with mP2X7 strongly increased the incorporation of YO-PRO-1. Furthermore, DKO cells did not respond to ATP for Ca²⁺ internalization, but DKO-mP2X7k cells responded well to ATP, and the intracellular Ca²⁺, detected by the Fluo-4 Ca²⁺ sensor, increased within 5 min after the ATP treatment (Fig. 3D). This response was not observed in TKO-mP2X7k cells, but was rescued by expressing mEros in TKO-mP2X7k/mEros cells.

hP2X7 has 81.3% aa sequence identity with mP2X7. As previously reported (37), human THP-1–derived macrophages primed with LPS dose dependently responded to ATP to produce IL-1β (Fig. 3E). To examine the contribution of Eros in this process, the Eros and P2X7 genes in THP-1 cells were individually knocked out by the CRISPR/Cas9 system (Supplemental Fig. 2). As shown in Fig. 3E, not only the P2X7-null mutation, but also the Eros-null mutation completely blocked the IL-1β production, and this defect was rescued by expressing hP2X7 or hEros, respectively.

We then prepared BMDMs from wild-type and Eros^−/− mice. Consistent with previous reports (23, 24), the LPS-primed, wild-type BMDMs produced IL-1β in response to ATP (Fig. 3F). In contrast, Eros^−/− BMDMs almost completely lost the ability to produce IL-1β in response to ATP (Fig. 3F). From these results, we concluded that various P2X7-mediated biological processes in mouse and human cells require Eros.

Eros was recently shown to be required for the stable expression of the NADPH oxidase subunits gp91^phox and p22^phox by acting as a chaperone (47). To examine whether Eros acts as a chaperone for P2X7, FLAG-tagged mP2X7k was introduced into DKO, TKO, and TKO-mEros cells, and the mP2X7k expression was evaluated by Western blotting with anti-P2X7 Ab. As shown in Fig. 4A, mP2X7k was stably expressed in DKO cells. However, its expression was severely reduced in TKO cells, and this decrease was rescued by transforming the cells with mEros. Similar results were obtained for endogenous hP2X7 in THP-1 cells. Western blotting with anti-P2X7 Ab detected bands of 75 and 200 kDa in the membrane fraction from the wild-type, but not P2X7^−/−THP-1, cells. The intensity of these bands increased by transforming the cells with hP2X7-FLAG (Fig. 4B), indicating that these bands corresponded to the monomer and trimer forms of hP2X7. The expression of hP2X7 was severely decreased by the null mutation of Eros, and this reduction was fully rescued by expressing hEros. Similarly, the wild-type (Eros^+/+) BMDMs expressed a high level of P2X7 (Fig. 4C). This expression was not reduced by the heterozygous null mutation of Eros (Eros^+/−) but severely reduced by the homozygous null mutation (Eros^−/−).

The requirement of Eros for the cell-surface expression of P2X7 was then examined with WR19L cells. Flow cytometry analysis with anti-mP2X7 Ab showed that mP2X7 was highly expressed in the DKO-mP2X7k cells. In contrast, mP2X7 on the surface of TKO-mP2X7k cells was severely decreased, but it was fully rescued in TKO-mP2X7k/mEros cells (Fig. 4D). To examine the localization of mP2X7 and mEros, mP2X7k and mEros were fused with EGFP, and expressed in TKO cells together with mEros-HA or mP2X7-FLAG, respectively. Observation by confocal microscopy indicated that a substantial portion of the mP2X7k was localized to the plasma membrane (Fig. 4E). In contrast, in agreement with a previous report (47), the mEros-EGFP in TKO-mP2X7k cells was exclusively observed intracellularly, and merged well with the ER-Tracker, indicating that mEros is localized at endoplasmic reticulum. When TKO cells were transformed to coexpress mP2X7k-EGFP and mEros-mCherry, the intracellular EGFP signal was colocalized with mCherry signals (Fig. 4F), suggesting that mEros assisted the folding or maturation of mP2X7 at the endoplasmic reticulum and that the mature mP2X7 moved to the plasma membrane. To test this possibility, DKO, TKO, and TKO-mEros-EGFP cells were transformed with FLAG-tagged mP2X7k, and their membrane fractions were analyzed by BN-PAGE. As shown in Fig. 4G, Western blotting with anti-FLAG Ab showed two bands at around 800 and 550 kDa in the DKO and TKO-mEros, but not in the TKO cells. The lower band at 550 kDa appeared to correspond to a trimeric form of mP2X7k, as reported previously (34). The anti-GFP Ab for mEros-EGFP detected two bands, of which the upper one was similar to the upper band observed with the anti-FLAG Ab for mP2X7k. The membrane fraction was then subjected to immunoprecipitation with anti-FLAG beads for the mP2X7k-FLAG. As shown in Fig. 4H, the precipitates carried the upper large complex, which could be recognized not only with the anti-FLAG for mP2X7k, but also with the anti-GFP for mEros. These results indicated that the large band of 800 kDa was a complex between mP2X7k and mEros in the endoplasmic reticulum. Mature mP2X7k with an apparent M_r of 550 kDa was present on the plasma membrane, whereas mEros, free or complexed with other proteins, was present in the endoplasmic reticulum. We concluded that mEros transiently interacts with mP2X7 at the endoplasmic reticulum and assists its folding. mP2X7 then moves to the plasma membrane, where it is present as a homotrimeric complex.

TMEM16F mediates Ca²⁺-induced PtdSer exposure in platelets, lymphocytes, and myoblasts (48, 49). Because the P2X7-mediated PtdSer exposure is accompanied by Ca²⁺ influx, TMEM16F was proposed to be involved in this process (25). However, we showed in this study that a TMEM16F-null mutation in mouse T cell lines had little effect on the P2X7-induced PtdSer exposure. Our preliminary results also showed that TMEM16F-null BMDMs responded to BzATP to expose PtdSer as efficiently as wild-type macrophages to expose PtdSer. Although we cannot rule out the possibility that Ca²⁺ internalized by a P2X7-mediated ion channel activates TMEM16F to expose PtdSer, it is clear that a TMEM16F-independent mechanism also exists for the P2X7-mediated PtdSer exposure. Two families of membrane proteins (TMEM16F and XKR) have phospholipid scramblase activity (48). Treatment with an intracellular Ca²⁺ chelator did not inhibit the P2X7-induced PtdSer exposure, suggesting that TMEM16 family members are not involved in this process. Whether a member of the XKR family or another unidentified molecule is responsible for the P2X7-mediated PtdSer exposure remains to be studied.

Screening of the GeCKO-CRISPR library (28) identified Eros as the molecule essential for the high, stable expression of P2X7. Eros is a membrane protein carrying two transmembrane regions. Thomas et al. (47) previously showed that Eros is localized to the endoplasmic reticulum and functions as a chaperone for the components of NADPH oxidase, a heterodimeric complex of gp91^phox and p22^phox, present in the plasma membrane. Similarly, we found that Eros was located at the endoplasmic reticulum, in WR19L cells. BN-PAGE and immunoprecipitation analyses indicated that P2X7 was present in two forms: associated or not associated with Eros. The intracellular colocalization of P2X7 with Eros suggested that Eros binds to P2X7 at the endoplasmic reticulum and helps its folding. Once P2X7 is correctly folded, it leaves the endoplasmic reticulum and moves to the plasma membrane, whereas Eros is recycled to act as a chaperone for newly synthesized P2X7 molecules. The recycling of chaperones requires energy or ATP (50). Eros does not appear to carry an ATPase or nucleotide-binding domain, indicating that Eros may act as a cochaperone for specific client proteins (e.g., NADPH oxidase and P2X7). As is found for many other membrane proteins (51), other classical chaperones (Hsp90 and Hsp70) with ATPase activity may be involved in the folding of P2X7. In this regard, it may be relevant that the NADPH oxidase activity is regulated by Hsp90 (52).

Similar to mutations in NADPH oxidase, loss-of-function mutations in Eros cause chronic granulomatous disease in humans (53, 54). There are two antimicrobial systems in phagocytes: reactive oxygen species (ROS) in the early host response, and reactive nitrogen species (RNS) in the late or chronic phase of infection (55–57). Accordingly, mice lacking both NADPH oxidase and inducible NO synthase exhibit a much stronger immunocompromised phenotype, such as granuloma formation, than mice lacking only one of these systems (58). NADPH oxidase is responsible for ROS generation, whereas RNS are generated by the action of inducible NO synthase, the expression of which is upregulated by signaling from P2X7 (59). These results suggest that Eros, acting as a chaperone for NADPH oxidase and P2X7, regulates the production of both ROS and RNS. It would be interesting to examine whether the RNS system is defective in Eros-null macrophages. Although the inflammasomes for IL-1β production are activated, not only by the P2X7 pathway, but also by other noncanonical pathways (60), P2X7 was shown to be necessary for the ATP-induced IL-1β production in LPS-treated mice (24). It will be interesting to examine whether Eros-null mice or patients produce IL-1β and, if so, to correlate its level with the formation of granuloma.

P2X7 mutants including a spliced form that lacks the C-terminal region, or a form that carries point mutations in the C-terminal region, maintain their ion channel activity but cannot mediate the pore formation or cell death (19, 61, 62), suggesting that these signaling pathways are independent of the Ca²⁺ internalization or that they require an additional signal from the C-terminal cytoplasmic region of the P2X7 receptor. Screening a CRISPR library for certain respective outcomes, such as YO-PRO-1 uptake, IL-1β release, or PtdCho uptake, may identify molecules responsible for their specific signaling pathways.

We thank M. Fujii for secretarial assistance.

Y.R. is on a leave of absence from Otsuka Pharmaceutical Co. to pursue a Ph.D. at the Graduate School of Frontier Biosciences, Osaka University. The other authors have no financial conflicts of interest.