Control of the signaling role of PtdIns(4)P at the plasma membrane through H₂O₂-dependent inactivation of synaptojanin 2 during endocytosis

Abstract

Phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P₂] is implicated in various processes, including hormone-induced signal transduction, endocytosis, and exocytosis in the plasma membrane. However, how H₂O₂ accumulation regulates the levels of PtdIns(4,5)P₂ in the plasma membrane in cells stimulated with epidermal growth factors (EGFs) is not known. We show that a plasma membrane PtdIns(4,5)P₂-degrading enzyme, synaptojanin (Synj) phosphatase, is inactivated through oxidation by H₂O₂. Intriguingly, H₂O₂ inhibits the 4-phosphatase activity of Synj but not the 5-phosphatase activity. In EGF-activated cells, the oxidation of Synj dual phosphatase is required for the transient increase in the plasma membrane levels of phosphatidylinositol 4-phosphate [PtdIns(4)P], which can control EGF receptor-mediated endocytosis. These results indicate that intracellular H₂O₂ molecules act as signaling mediators to fine-tune endocytosis by controlling the stability of plasma membrane PtdIns(4)P, an intermediate product of Synj phosphoinositide dual phosphatase.

Graphical abstract

Highlights

• •H₂O₂-dependent oxidation of Synaptojanin increases the level of PtdIns(4)P in the plasma membrane.
• •The 4-phosphatase activity, but not the 5-phosphatase activity, of Synj is inhibited by H₂O₂.
• •The oxidation of Synj is required for EGF receptor-mediated endocytosis.

1. Introduction

Phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P₂] is one of the most abundant phosphoinositides found on the inner surface of the plasma membrane. It plays a crucial role in various cellular processes such as cell signal transduction, endocytosis, exocytosis, phagocytosis, cell movement, and the regulation of certain plasma membrane proteins [[1], [2], [3], [4]]. Plasma membrane PtdIns(4,5)P₂ is primarily synthesized by PtdIns(4)P 5-kinase from PtdIns(4)P [5,6] and is also produced in smaller amounts by PtdIns(5)P 4-kinase from PtdIns(5)P [7]. This is due to the availability of relatively high levels of PtdIns(4)P and very low levels of PtdIns(5)P in cells [8]. The level of plasma membrane PtdIns(4,5)P₂ is also regulated by phosphoinositide (PI) phosphatases and phospholipase C (PLC). Hydrogen peroxide (H₂O₂) molecules are naturally produced during cellular processes in all aerobic organisms. Recent studies have shown that they act as signaling mediators at the plasma membrane in cells stimulated with peptide growth factors, insulin, tumor necrosis factor-α (TNFα), and interleukin-1 (IL-1) [9,10]. However, the precise mechanism by which a transient increase in H₂O₂ levels controls the conversion of PtdIns(4,5)P₂ at the plasma membrane and affects the activities of PI kinases and phosphatases remains poorly understood.

The conversion of plasma membrane PtdIns(4,5)P₂ to PtdIns(3,4,5)P₃ by PI 3-kinase can be activated indirectly by the increased tyrosine phosphorylation of ligand-bound receptors. This activation occurs through the inactivation of protein tyrosine phosphatase (PTP) family proteins, including phosphatase PTP1B [11], by intracellularly produced H₂O₂. A previous study has shown that the tumor suppressor PTEN PI 3-phosphatase is inactivated through direct oxidation by H₂O₂ during peptide growth factor stimulation, resulting in a transient generation of PtdIns(3,4,5)P₃ [12]. The cleavage of plasma membrane PtdIns(4,5)P₂ into inositol 1,4,5-trisphosphate (IP₃) and diacylglycerol (DAG) by PLC-γ family proteins can be enhanced through increased tyrosine phosphorylation of receptors such as platelet-derived growth factor receptor (PDGFR) and epidermal growth factor receptor (EGFR) in cells under oxidative stress [13,14]. However, it remains to be elucidated whether plasma membrane PtdIns(4,5)P₂ phosphatases are direct targets of H₂O₂-dependent oxidation. In addition, the physiological consequences of their oxidations are also unknown.

Synaptojanin (Synj) is a major plasma membrane PtdIns(4,5)P₂ phosphatase among the PI phosphatases that have a well-conserved catalytic cysteine (Cys-X₅-Arg) residue, which may be sensitive to oxidation by H₂O₂. Synj has dual phosphatase activities and contains an amino-terminal Sac1 4-phosphatase domain, a central inositol 5-phosphatase domain, and a carboxyl-terminal proline-rich region [15]. Synj1 plays an essential role in clathrin-mediated endocytosis during synaptic vesicle recycling at the presynaptic nerve terminal [16,17], while Synj2, a ubiquitously expressed homolog to Synj1, functions in the formation of clathrin-coated pits and vesicles at an early stage of clathrin-mediated endocytosis [18].

Here, we show that the catalytic cysteine (Cys-X₅-Arg) within the Sac1 domain of Synj is a target of H₂O₂-induced oxidation, and that the 4-phosphatase activity of Synj is inhibited by H₂O₂, while the 5-phosphatase activity is not. Based on these findings, we propose a mechanism through which the activities of the PI dual phosphatase are differentially regulated by H₂O₂ to control endocytosis through the conversion of PtdIns(4,5)P₂ at the plasma membrane.

2. Materials and methods

2.1. Chemicals and antibodies

N-ethylmaleimide (NEM), iodoacetic acid (IAA), trichloroacetic acid (TCA), catalase, and phenylarsine oxide (PAO) were purchased from Sigma-Aldrich (St. Louis, MO, USA), and diphenylene iodonium (DPI) was obtained from Enzo Life Science (Farmingdale, NY, USA). EZ-Link-Maleimide-PEG2-Biotin and NeutrAvidin UltraLink Resin were obtained from Pierce (Waltham, MA, USA). Epidermal growth factor (EGF) was obtained from Invitrogen, and 4-6-diamidino-2-phenylindole (DAPI) was obtained from Roche (Basel, Switzerland). Chlorpromazine and Dynasore were purchased from Sigma-Aldrich (St. Louis, MO, USA). All antibodies were used in dilutions indicated or otherwise recommended by the manufacturers. Antibodies specific to Synj2 (1:500, ab75880, Abcam), c-Myc (1:1000, SC40, Santa Cruz Biotechnology), β-Actin (1:1000, A2228, Sigma Aldrich), α-Tubulin (1:1000, T9026, Sigma Aldrich), catalase for IF (final concentration 5 μg/ml, 12C2DB9, Abcam; Cambridge, United Kingdom), catalase for IB (1:2000, LF-PA0060, AbFrontier; Seoul, South Korea), and EGFR (1:1000, 2232, Cell Signaling Technology, Danvers, MA, USA) were obtained from commercial sources.

2.2. Cell culture and transfection

AD293 cells (a derivative of HEK293 cells with improved cell adherence; Agilent, Santa Clara, CA, USA) and A549 cells were maintained at 37 °C with 5% CO₂ in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS).

2.3. Plasmids and siRNA

The plasmids containing full-length Synj1 or Synj2 cDNA were kindly provided by Pietro De Camilli (Yale University, USA). The mammalian expression plasmids encoding full-length Synj1, 4-phosphatase Synj1, 5-phosphatase Synj1, full-length Synj2, 4-phosphatase Synj2, and 5-phosphatase Synj2 were generated through PCR amplification using the appropriate primers and subcloning into the FseI and AscI sites of the pCS₂ Myc (6 Myc epitope) and pCS₂ GFP or pCS₂ RFP vectors (the pCS₂ parental vector was provided by D. Turner, University of Michigan, Ann Arbor, MI). Various Synj point mutants (Synj1 C383S and Synj2 C386S) were produced using the QuickChange Site-Directed Mutagenesis Kit (Agilent Technologies). The plasmid pHyper-Ras, a plasma membrane-targetable Hyper, was generated by conjugating cHyper cDNA (cytosolic Hyper H₂O₂ sensing reporter, FP941, Evrogen, Moscow, Russia) with cDNA encoding the C-terminal Ha-Ras plasma membrane-target peptide (KLNPPDESGPGCMSCKCVLS). The plasmid for a plasma membrane-targeted catalase (Cat-PM) was generated by conjugating human catalase cDNA with the removal of the 12 nucleotides at the C-terminal encoding KANL residues (peroxisome-targeting sequence) with the same cDNA coding used for the Ha-Ras plasma membrane-targeted peptide. For the generation of the cytosolic catalase (Cat-cyto) plasmid, the cDNA encoding lysine in the C-terminal KQNL peptide of catalase was changed to the stop codon TAA using the QuickChange Site-Directed Mutagenesis Kit (Agilent Technologies). All constructs produced in this study were verified through DNA sequencing. The Synj2 siRNA oligonucleotides (5′-CCA-GGA-UCC-UGA-AAG-CUA-U-3′) and control siRNA oligonucleotides (5-AUG-AAC-GUG-AAU-UGC-UCA-AUU-3′) were obtained from ST Pharm, Korea. The double-stranded siRNA oligonucleotides for Synj2 were previously validated [18]. The plasmid encoding FAPP1-PH(mut)-GFP, in which tyrosine-72 in FAPP1-PH was replaced with arginine was kindly provided by Tamas Balla at NIH, USA.

2.4. Detection of oxidized synj using the biotinyl-cysteine labeling assay

To detect Synj oxidation, a method using Cys biotinylation labeling was carried out as described previously [23] with slight modification. Briefly, H₂O₂-treated or EGF-treated cells (1 × 10⁶ cells per 100-mm dish) were washed with ice-cold phosphate-buffered saline (PBS), collected into 1 ml of PBS in microfuge tubes, quickly centrifuged at 8000×g for 1 min, and rapidly frozen in liquid nitrogen. The frozen cells were incubated for 1 h at room temperature with 1 ml of oxygen-free extraction buffer (50 mM sodium phosphate, pH 7.0, 1 mM EDTA, 10 mM NEM, 10 mM iodoacetic acid, 1% Triton X-100, 5 mM NaF, 50 μg/ml leupeptin, 50 μg/ml aprotinin, 1 mM 4-[2-aminoethyl] benzene-sulfonyl fluoride) in an anaerobic chamber. The supernatant was collected and incubated for 2 h at room temperature in the dark with sodium dodecyl sulfate (SDS) to a final concentration of 1%. The denatured proteins (500 μg) were then precipitated by adding trichloroacetic acid to a final concentration of 10%, followed by further incubation for 1 h at room temperature. The protein precipitate was washed twice with dry ice-chilled acetone and reduced by incubating for 30 min at 50 °C in 0.1 ml of oxygen-free reducing buffer (50 mM Hepes-NaOH, pH 7.7, 1 mM EDTA, 2% SDS, 4 mM DTT) in an anaerobic chamber. The reduced proteins were biotinylated by incubating for 30 min at 50 °C with 0.9 ml of a biotinylating solution, which consisted of 50 mM sodium phosphate at pH 7.0, 1 mM EDTA, and 1 mM biotin conjugated to polyethylene oxide-maleimide (Pierce). The reaction was stopped by the addition of DTT at a final concentration of 1 mM, and the proteins were precipitated by incubation with TCA (final concentration, 10%) for 1 h. The proteins were collected by centrifugation, washed with cold acetone, and solubilized in 0.2 ml of a solution containing 50 mM Hepes-NaOH (pH 7.7), 1 mM EDTA, and 2% SDS with brief ultrasonic treatment at a 15-sec pulse. The samples were then diluted with 0.2 ml of the same solution without SDS, and a 40-μl portion of each resulting mixture was saved for immunoblot analysis. The remaining 360 μl of each mixture was further diluted with the same buffer without SDS to a final SDS concentration of 0.5%. The biotinylated proteins were then precipitated by incubating for 1 h at room temperature with 3 μl of packed UltraLink Immobilized NeutrAvidin (Pierce). The beads were washed five times with washing buffer (20 mM Hepes-NaOH at pH 7.7, 200 mM NaCl, 1 mM EDTA, and 0.5% SDS), and the biotinylated proteins were released from the beads by boiling in sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer and subjected to immunoblot analysis with the antibody to Myc to detect Synj. For quantitative analysis of the protein band intensities, the Multi-Gauge (v3.0, Fuji photo film Co., Minato City, Tokyo, Japan) software was used.

2.5. Expression and purification of Synj2

The bacterial expression plasmids containing cDNA encoding the 4-phosphatase domain protein (1–506) or 5-phosphatase domain protein (507–904) of Synj2 were constructed by cloning into the pGEX-4T-1 vector. The expression and purification of GST-Synj2 4-phosphatase and 5-phosphatase proteins were performed according to the manufacturer's instructions (GE Healthcare, Boston, MA, USA). Both proteins were purified using Glutathione-Sepharose 4B beads (GE Healthcare), and the purified recombinant proteins were concentrated and dialyzed against a storage buffer (10 mM Tris at pH 7.7, 100 mM KCl, and 1 mM EGTA). The proteins were stored at −80 °C.

2.6. Malachite green assay

Each recombinant GST-Synj2 4-phosphatase protein (3 μg) or 5-phosphatase protein (3 μg) was incubated in a storage buffer containing 0, 10, 50, or 200 μM H₂O₂ for 5 min at room temperature. The sample was then incubated with 2 μg of catalase for 10 min at 37 °C to remove any remaining H₂O₂. The GST-Synj2 4-phosphatase or 5-phosphatase protein mixture was incubated with diC₈-PtdIns(4)P (Echelon, Salt Lake City, UT, USA, 3000 pmol) or diC8-PtdIns(4,5)P₂ (Echelon, 3000 pmol), respectively, in 15 μl of phosphatase reaction buffer containing 30 mM Hepes (pH 7.7), 100 mM KCl, 1 mM EGTA, and 2 mM MgCl₂ for 10 min at 37 °C. The free phosphate generated during the reaction was measured by incubation with 100 μl of malachite green assay solution (Echelon) for 15 min at room temperature, and analysis of the absorbance was performed at a wavelength of 620 nm in a microplate reader. Conversion (%) = [(Free phosphate in reaction, pmol) – (Background phosphate, pmol)] × 100/3000 pmol. Phosphatase activity (%) = (Conversion value)/(Highest conversion value) × 100.

2.7. Subcellular fractionation

Cells (150-mm dish) expressing Cat-PM or cytosol catalase (Cat-cyto) were suspended in 0.25 ml of a hypotonic buffer containing 10 mM Hepes-NaOH (pH 7.5), 2 mM MgCl₂, 25 mM KCl, 0.5 mM EDTA, 0.5 mM EGTA, a mixture of protease inhibitors (1 mM DTT, 5 mM NaF, 10 μg/ml aprotinin, 10 μg/ml leupeptin), and a phosphatase inhibitor cocktail (Sigma-Aldrich). The cell suspension was then incubated on ice for 30 min. The cells were gently homogenized by forcing them 10 times through a 30-gauge syringe needle. The homogenate was mixed with an equal volume (0.25 ml) of hypertonic buffer containing 0.5 M sucrose to generate an isotonic condition. The homogenate was centrifuged at 1000×g for 10 min to remove nuclei and debris, and the resulting supernatant was then centrifuged again at 15,000×g for 15 min. The final supernatant was used as the clear cytosolic fraction. The pellet contained the heavy membrane fraction, which included the mitochondria, peroxisome, and other organelles. The pellet was resuspended in 0.5 ml of isotonic buffer containing 0.25 M sucrose and was centrifuged at 15,000×g for 30 min; the final pellet was used as the heavy membrane fraction. The protein concentration was measured using the Bradford assay (Bio-Rad, Hercules, CA, USA), and equal amounts of protein were subjected to SDS-PAGE for immunoblot analysis.

2.8. Confocal microscopy

Both immunofluorescence (IF) staining and live cell imaging experiments were performed as described previously [19]. Cells were cultured in a 12-well dish containing cover slips (diameter, 12 mm; Marienfeld, Lauda-Königshofen, Germany) coated with 0.01% poly-l-lysine (Sigma-Aldrich) for both experiments. To stain catalase, the cells were fixed in 4% paraformaldehyde (Electron Microscopy Sciences, Hatfield, PA, USA) in PBS for 10 min at room temperature, followed by incubation in PBS containing 5% horse serum (Gibco-BRL, Grand Island, NY, USA) and then in 0.1% Triton X-100 for 30 min to block non-specific antibody binding. The cells were incubated with antibodies to catalase for 30 min at room temperature. After washing with PBS three times, the cells were stained with Alexa Fluor® 488-conjugated goat anti-rabbit IgG (1:1000, Invitrogen). Following three additional washes with PBS, 4′6′-diamidino-2-phenylindole (DAPI, 0.5 μg/ml) was used to stain the DNA. Finally, the samples were mounted onto glass slides using Fluoromount-G solution (Southern Biotech, Birmingham, AL, USA). For live-cell imaging, cells were observed in an incubation chamber (Chamlide TC; Live Cell Instrument, Namyangju, South Korea) at 37 °C under an atmosphere of 5% CO₂ in air. Confocal images were obtained using a 60× Plan Apochromat VC objective, NA 1.40, by illuminating with a 405-nm laser, a 488-nm multi-Ar laser, or a 561-nm diode-pumped solid-state laser under a Nikon A1R confocal microscope. The emission (Em_500–530nm) intensity ratios of Ex_488nm vs Ex_405nm were measured to assess the change in H₂O₂ levels in live cells expressing plasma membrane-targeted Hyper (Hyper-Ras). All images were processed and analyzed using the Nikon NIS elements software.

2.9. Analysis of EGFR endocytosis

Serum-starved A549 cells, with or without Cat-PM, were washed with PBS twice and incubated with DMEM containing 500 ng/ml Alexa fluor 555-conjugated EGF (E35350, Molecular Probes, Eugene, OR, USA) for 10 min at 37 °C to facilitate internalization. The cells were then washed with cold PBS three times to stop the reaction and incubated with an acidic buffer (0.2 M acetic acid, 0.5 M NaCl) for 5 min at room temperature to remove excess Alexa fluor 555-EGF on the cell surface. The cells were washed with cold PBS twice, fixed in 4% paraformaldehyde in PBS for 10 min, and stained with DAPI for confocal imaging.

2.10. Delivery of PtdIns(4)P

The PtdIns(4)P mixture was prepared by combining 100 μM GloPIPs BODIPY® FL-PI(4)P (C–04F6a, Echelon) with Carrier 3 (P–9C3, Echelon) in PBS at a 1-to-1 molar ratio. The mixture was briefly treated with ultrasonic waves for 1 min to ensure complete dissolution of the complex. Serum-starved A549 cells were incubated with DMEM containing 10 μM of the PtdIns(4)P mixture for 30 min at 37 °C. The endocytosed EGF-555 and BODIPY® FL-PI(4)P were observed using a confocal microscope.

2.11. Endocytosis inhibition

Endocytosis inhibition was performed using chlorpromazine, a clathrin-mediated endocytosis inhibitor and Dynasore, a dynamin inhibitor respectively as described previously [19,20] with minor modifications. For inhibition with chlorpromazine, serum-starved A549 cells (2 × 10⁵ per 12 well) were pretreated with chlorpromazine (2.5 μg/ml) for 8 min and incubated with EGF (200 ng/ml) in the presence or absence of chlorpromazine for 5 min. For inhibition with Dynasore, serum-starved A549 cells (2 × 10⁵ per 12 well) were pretreated with Dynasore (40 μM) for 2 h and incubated with EGF (200 ng/ml) in the presence or absence of Dynasore for 5 min. The cells were washed three times with PBS and then, immunofluorescence staining was carried out as described previously [21]. Cells were fixed for 20 min in 2% paraformaldehyde and permeabilized with PBS containing 0.5% saponin (Sigma-Aldrich) and 1% bovine serum albumin (BSA, Bovogen, East Keilor, Australia) for 30 min at room temperature. Cells were incubated with antibodies to PtdIns(3,4)P₂ (Z-P034b, Echelon Biosciences, Salt Lake City, UT, USA; 1:200) for 2 h in PBS containing 1% BSA. After three 5-min washes with PBS, the cells were incubated with a fluorescent secondary antibody in PBS containing 1% BSA and washed three times for 5 min each with PBS. Confocal images were acquired using an LSM 880 Airy (Carl Zeiss) equipped at Fluorescence Core Imaging Center, Ewha Womans University and were processed and analyzed using Imaris software (Oxford Instruments, Abingdon, United Kingdom).

2.12. Statistical analysis

Quantitative data are presented as the mean ± standard error of the mean (SEM) from multiple determinations obtained from at least three independent experiments. The data were analyzed using Student's t-test with Sigma Plot 10.0 software, and the p value was calculated to assess the statistical significance.

2.13. Molecular modeling and docking

The three-dimensional structural models of the Sac1 domain, as well as the combined Sac1 and 5-phosphatase domains, were predicted using the ColabFold pipeline, which integrates the capabilities of AlphaFold2 with the MMseqs2 sequence search tool [22]. A molecular model of PtdIns(4)P was constructed and subsequently optimized utilizing the Builder toolbox in PyMOL (The PyMOL Molecular Graphics System, Version 2.4.0, Schrödinger, LLC). To explore potential binding interactions, PtdIns(4)P was docked into the active site of the Sac1 domain structure using the AutoDock Vina software [23].

3. Results

3.1. The catalytic cysteine in the Sac1 domain of synj is susceptible to oxidation by H₂O₂

Synj family proteins are responsible for converting or degrading plasma membrane PtdIns(4,5)P₂. Synj1 is enriched in the presynaptic region of the nerve terminal [17], while Synj2 has a broader expression pattern [24]. Both Synj1 and Synj2 have an NH₂-terminal Sac1 phosphatase domain that can dephosphorylate PtdIns(4)P in vitro [25], a central 5-phosphatase domain that can dephosphorylate PtdIns(4,5)P₂ [16,18], and a COOH-terminal domain that can interact with plasma membrane-recruited proteins (Fig. 1A). To determine whether oxidative stress affects the PtdIns(4,5)P₂ degradation process and Synj activity, we investigated the presence of oxidized Cys using the Sac1 domain, 5-phosphatase domain, and full-length Synj1 or Synj2. This was achieved by alkylating reduced cysteines, followed by reducing the oxidized cysteines with DTT, and subsequently biotinylating the newly revealed thiolated Cys residues [26]. The oxidized forms of the Sac1 4-phosphatase domain of Synj1 and Synj2 were observed with Myc-tagged forms after exposure to extracellular H₂O₂, but the oxidized forms of the 5-phosphatase domain were not present in cells (Fig. 1B and C). The catalytic Cys residues (Cys-383 in Synj1 and Cys-386 in Synj2) at the Sac1 domain of Synj were targets of H₂O₂-induced oxidation, similar to the low-pK_a Cys residue at the Cys-X₅-Arg motif of protein tyrosine phosphatase (PTP) family proteins, such as PTP1B [11], LMW-PTP [27], PTEN [12], and Cdc14B [26]. However, the central 5-phosphatase domain, which contains a conserved histidine in its active site, was not sensitive to oxidization by H₂O₂ (Fig. 1B and C). The oxidized form was also observed in Myc-tagged full-length Synj2 containing the COOH-terminal interacting motif (Fig. 1D).

Fig. 1.

The catalytic cysteine in the 4-phosphatase domain of Synj1 and Synj2 is oxidized by H₂O₂. (A) Schematic representation of Synj1 and Synj2, where the 4-phosphatase domains contain a cysteine residue and the 5-phosphatase domains contain a histidine residue at the catalytic sites. Each proline-rich region is responsible for binding with other proteins. (B) Immunoblot analysis for the detection of oxidized Synj1 using the cysteine biotinylation assay in HEK293 (AD293) cells expressing Myc-tagged Synj1 4-phosphatase domain, Synj1 4-phosphatase mutant (C383S) domain, or Synj1 5-phosphatase domain. (C) Immunoblot analysis for the detection of oxidized Myc-tagged Synj2 4-phosphatase domain using the cysteine biotinylation assay. (D) Immunoblot analysis for the detection of oxidized Synj2 using the cysteine biotinylation assay in 293T cells expressing Myc-tagged full-length Synj2.

3.2. H₂O₂-dependent inhibition of 4-phosphatase activity of synj

While the 5-phosphatase activity of Synj on PtdIns(4,5)P₂ is well characterized in vitro and in vivo [16,18,28,29], the 4-phosphatase activity needs to be clarified regardless of the requirement of the Sac1-phosphatase activity of Synj1 for efficient endocytosis after short stimuli in neurons [28]. In addition, the direct correlation of Sac1 activity loss of Synj1 with early-onset parkinsonism needs to be examined [[30], [31], [32]]. We purified recombinant GST-conjugated 4-phosphatase domain protein and 5-phosphatase domain protein of Synj2 using an E. coli expression system (Fig. 2A). The activity of the Sac1-like domain of Synj2 was approximately 70% of that of the 5-phosphatase domain of Synj2 when the purified Sac1-like protein was incubated with diC₈-PtdIns(4)P and the 5-phosphatase protein with diC₈-PtdIns(4,5)P₂ as a substrate, respectively (Fig. 2B) indicating that Synj2 has both 4-phosphatase and 5-phosphatase activity. The vicinal phosphate at the 5’ carbon position on an inositol ring of diC₈-PtdIns(4,5)P₂ makes it an informal substrate for Sac1-like proteins. Next, we investigated whether the phosphatase activity was inhibited by H₂O₂. The 4-phosphatase, which contains a catalytic cysteine residue, was inactivated by H₂O₂ in a concentration-dependent manner. However, the 5-phosphatase activity remained constant during treatment with H₂O₂ below concentrations of 200 μM (Fig. 2C and D). Together with the oxidation results in Fig. 1, the data in Fig. 2 demonstrate that H₂O₂ can work as a regulator of Synj dual phosphatase by inhibiting 4-phosphatase activity without affecting 5-phosphatase activity.

Fig. 2.

Inhibition of Synj2 4-phosphatase activity by H₂O₂. (A) Coomassie blue staining of purified GST-tagged Synj2 4-phosphatase and 5-phosphatase domains. Each lane contained 1 μg of protein. (B) Phosphatase assay of Synj2 4-phosphatase and 5-phosphatase domains using the malachite green method. The purified Synj2 4-phosphatase and 5-phosphatase proteins were incubated with diC₈-PtdIns(4)P (echelon) and diC₈-PtdIns(4,5)P₂ (echelon), respectively. (C–D) Control of Synj2 phosphatase activity by H₂O₂. (C) Purified GST-4-phosphatase domain (3 μg) and (D) purified GST-5-phosphatase domain (3 μg) were incubated with 0, 10, 50, and 200 μM H₂O₂ for 5 min at room temperature, and the samples were then incubated with 2 μg of catalase for 10 min at 37 °C to remove any remaining H₂O₂ in a tube. The activity was measured using the malachite green method. The results are shown as the mean ± SEM of triplicates (Student's t-test; *, p < 0.05; **, p < 0.01; n = 3 for each experiment).

3.3. Increased levels of plasma membrane PtdIns(4)P following H₂O₂ accumulation during EGF activation

Increased levels of extracellular H₂O₂ induced the oxidation of Synj proteins (Fig. 1). Synj1 is a highly abundant protein in presynaptic neural cells [33]. A recent study in the nematode Caenorhabditis elegans proposed that low concentrations of H₂O₂ promote signal transduction by sensory neurons [34]. However, whether the changes in the reactive oxygen species levels at synaptic clefts are involved in neural signal transduction has not been clear until now. Synj2 is broadly expressed in various organs, including the lung and heart [24]. The production of intracellular H₂O₂ is required for signal propagation in cultured cells stimulated with growth factors such as epidermal growth factor (EGF) or platelet-derived growth factor (PDGF) [10,13]. Therefore, we investigated whether EGF-induced H₂O₂ production results in the oxidation of Synj2 in A-549 human lung carcinoma cells, where Synj2 was previously reported to regulate EGF receptor-mediated endocytosis [18]. No substantial amount of oxidized Synj2 was observed in serum-starved cells. Immunoblots showed a significant increase in the levels of oxidized Synj2 1–3 min after EGF activation (Fig. 3A). The amount of H₂O₂ at the plasma membrane is regulated by the local inactivation of the PrxI antioxidant protein [35] and activation of membrane-integrated NADPH oxidase (Nox) [36,37] during receptor activation. To monitor the increase in H₂O₂ levels at the plasma membrane, we generated a plasma membrane-targeted Hyper with a small targeting peptide. The C-terminal 20-amino-acid peptide of Ha-Ras, a plasma membrane-targeted peptide containing cysteine residues modified with palmitoylation and farnesylation [38], was attached to the C-terminus of Hyper, a cytosolic H₂O₂-sensing ratiometric fluorescent reporter [39] to target the plasma membrane microdomain, and the targetable sensor was called Hyper-Ras (Supplementary Fig. 1A). A transient increase in H₂O₂ levels at the plasma membrane was observed in cells expressing Hyper-Ras, and quantification of ratiometric fluorescence intensity showed a significant accumulation of plasma membrane H₂O₂ (Supplementary Figs. 1B–D). The amount of plasma membrane H₂O₂ peaked at approximately 100 s after EGF treatment and decreased slowly over 10 min. At the peak time, the plasma membrane H₂O₂ level in EGF-activated cells was about 70% higher than that in EGF-activated cells treated with diphenyliodonium (DPI), a Nox inhibitor, or catalase (Supplementary Fig. 1D). Considering that Synj2 is a 5- and 4- phosphatase that can convert PtdIns(4,5)P₂ to PtdIns(4)P and sequentially to PtdIns at the plasma membrane, we tested the possibility that H₂O₂-dependent inactivation of only the 4-phosphatase activity of Synj2, without affecting the 5-phosphatase activity of Synj2 (Fig. 2), induces an increase in the amount of plasma membrane PtdIns(4)P. We used the FAPP1-PH(mut)-GFP reporter (from T Balla, NIH, USA), which can effectively monitor PtdIns(4)P at the plasma membrane by reducing its binding to the Golgi ARF1 GTPase through a mutation (Y72R) in the ARF-binding domain of FAPP1-PH [40] (personal communication with T. Balla). Live fluorescence imaging and quantitative analysis of the fluorescence intensity around the plasma membrane showed a statistically significant transient increase in PtdIns(4)P around the plasma membrane in A549 cells activated with EGF compared with cells activated with EGF and DPI or catalase (Fig. 3B and C). A previous study analyzing the total amount of PIs reported that the total phosphatidylinositol bisphosphate level remains consistent during EGF activation in A431 epidermoid carcinoma cells [41]. Live cell imaging studies using a fluorescent PtdIns(4,5)P₂ indicator (GFP-PLC1_δ1 PH) demonstrated that the transient reduction in plasma membrane PtdIns(4,5)P₂ concentration in receptor-activated cells was associated with the degradation of PtdIns(4,5)P₂ by activated PLC [42] and with the increase in calcium concentration [43]. Considering the various enzymes such as PIP 5 kinases [6], PLC [44], PI3 kinases [45] and several lipid phosphatases [46] that can affect PtdIns(4,5)P₂ concentration, the plasma membrane PtdIns(4,5)P₂ level is thought to vary in cells during receptor activation.

Fig. 3.

The transient increase in plasma membrane PtdIns(4)P levels in response to H₂O₂ generated in EGF-activated cells. (A) Western blot analysis for the detection of oxidized Synj2 using cysteine biotinylation assay in A549 cells expressing Myc-tagged full-length Synj2 during EGF (200 ng/ml) activation (Left). Quantification of oxidized Synj2 bands. Error bars represent the SEM from three independent experiments (Student's t-test; *, p < 0.05; **, p < 0.01) (Right). (B) A transient increase in the plasma membrane PtdIns(4)P levels. Selected snapshot images of live A549 cells expressing FAPP1-PH (mut)-GFP. Serum-starved A549 cells were preincubated with DPI (5 μM), a Nox inhibitor, or catalase (2 mg/ml). Scale bars, 20 μm; 4 μm in magnification. (C) Quantification of relative plasma membrane PtdIns(4)P fluorescence in live A549 cells during EGF stimulation. The bar graph shows the relative fluorescence intensity at 2 min in EGF-activated A549 cells. The error bars represent the SEM (Student's t-test; *, p < 0.05; n = 19).

3.4. The EGF-dependent increase in plasma membrane PtdIns(4)P is mediated through Synj2 oxidation

We evaluated the level of plasma membrane PtdIns(4)P in Synj2 knock-down cells compared to control cells to determine whether Synj2 oxidation results in PtdIns(4)P accumulation. Quantification of plasma membrane PtdIns(4)P through plasma membrane fluorescence imaging analysis showed that a transient increase in plasma membrane PtdIns(4)P requires Synj2 (Fig. 4A and B). The EGF-induced increase in PtdIns(4)P levels was approximately 20% at 2 min in activated cells, but PtdIns(4)P accumulation at the plasma membrane was not observed in Synj2 knock-down cells. The expression of Myc-Synj2 in siSynj2-treated cells restored the increase in plasma-membrane PtdIns(4)P levels (Fig. 4B). Endogenous Synj2 protein levels were reduced by 70–80% in Synj2 knock-down cells two days after transfection with siSynj2 oligos (Fig. 4C). Synj is suggested to be recruited to the plasma membrane by curvature generators, such as the BAR protein endophilin [47], and is co-localized with clathrin-coated pits in activated cells to degrade plasma membrane PtdIns(4,5)P₂ [48]. Taken together with the increase in the amount of PtdIns(4)P around the plasma membrane due to H₂O₂ production during EGF activation (Fig. 3, Supplementary Fig. 1), these findings demonstrate that increased plasma membrane H₂O₂ results in Synj oxidation and inactivation, leading to an increase in PtdIns(4)P levels around the plasma membrane. PtdIns(4)P is a major precursor of plasma-membrane PtdIns(4,5)P₂, but it also plays an independent role in determining membrane identity at the plasma membrane [49]. We focused on the physiological consequences of the transient increase in PtdIns(4)P levels during growth factor activation.

Fig. 4.

Control of the plasma-membrane PtdIns(4)P levels by Synj2. (A) Transient increase in the plasma membrane PtdIns(4)P levels induced by Synj2. Selected snapshot images of live A549 cells expressing FAPP1-PH (mut)-GFP. siRNA oligo against Synj2 and/or siRNA-resistant Myc-Synj2 was transfected into the cells. Scale bars, 20 μm; 4 μm in magnification. (B) Quantification of the relative plasma-membrane PtdIns(4)P fluorescence in (A). The bar graph shows the relative fluorescence intensity at 2 min in EGF-activated A549 cells. The error bars represent the SEM (Student's t-test; *, p < 0.05; **, p < 0.01; n = 45). (C) Immunoblot of Synj2 in the lysates of control and Synj2 knock-down cells.

3.5. The accumulation of H₂O₂ in the plasma membrane is required for the elevation of PtdIns(4)P levels and receptor-mediated endocytosis

To identify the role of local H₂O₂ around the plasma membrane in the inactivation of Synj2, we produced Cat-PM to specifically reduce H₂O₂ levels locally at the plasma membrane without affecting H₂O₂ levels in other parts of the cell. Cat-PM was generated through the C-terminal conjugation of catalase without the KANL peroxisomal targeting sequence with 20 amino acid peptides of Ha-Ras, similar to the production of Hyper-Ras (plasma membrane targeting Hyper) as shown in Supplementary Fig. 1. Cytosolic catalase (Cat-cyto) was produced with a C-terminal peroxisomal targeting 4 amino acid (KANL) deletion. The proper localization of Cat-PM and Cat-cyto was confirmed through immunostaining and immunoblotting following fractionation (Supplementary Figs. 1A and B). The localization of catalase in peroxisomes was confirmed with immunofluorescence (Supplementary Fig. 2C). The accumulation of H₂O₂ at the plasma membrane in cells expressing Cat-PM was significantly reduced, with a 92% decrease in Hyper-Ras fluorescence compared to non-transfected cells and a 90% decrease compared to cells expressing Cat-cyto (Fig. 5A). Cells expressing Cat-PM also showed a significant reduction in plasma membrane PtdIns(4)P levels compared to control cells and cells expressing Cat or Cat-cyto (Fig. 5B and Supplementary Fig. 3), indicating that plasma membrane PtdIns(4)P levels are specifically regulated by plasma membrane H₂O₂ and not by cytosolic or peroxisomal H₂O₂ in cells activated with EGF. Next, we examined whether the increase in PtdIns(4)P levels caused by plasma-membrane H₂O₂ can regulate EGF receptor-mediated endocytosis. We compared the internalization of fluorescent-tagged EGF between Cat-PM-expressing cells and control cells. The fluorescence intensity of endocytosed EGF spots in control cells was 3 times higher than that in Cat-PM-expressing cells (Fig. 5C). This result demonstrates that the transient accumulation of H₂O₂ at the plasma membrane is required for receptor-mediated endocytosis.

Fig. 5.

Control of receptor-mediated endocytosis by plasma-membrane H₂O₂. (A) Analysis of plasma membrane H₂O₂ levels in live A549 cells expressing plasma membrane-targeted Hyper (Hyper-Ras) and plasma membrane-targeted catalase (Cat-PM), catalase (Cat), or cytosolic catalase (Cat-cyto). The H₂O₂ levels were measured using relative emission activated by 488-nm and 405-nm lasers (Em₄₈₈/Em₄₀₅). The bar graph shows the relative fluorescence intensity at 2 min in EGF-activated A549 cells. The error bars represent the SEM (Student's t-test; *, p < 0.05; n = 18). (B) Analysis of the plasma membrane PtdIns(4)P levels in live A549 cells expressing FAPP1-PH (mut)-GFP and plasma membrane-targeted catalase (Cat-PM), catalase (Cat), or cytosolic catalase (Cat-cyto). The bar graph shows the relative fluorescence intensity at 2 min in EGF-activated A549 cells. The error bars represent the SEM (Student's t-test; *, p < 0.05; n = 23). (C) Inhibition of EGFR endocytosis through the ectopic expression of Cat-PM. Representative confocal images of fluorescent EGF-555 (EGF conjugated with Alexa 555 dye) endocytosed in control or Cat-PM-expressing cells. Scale bars, 20 μm; 2 μm in magnification. The graph shows the relative intensity of EGF-555 spots 30 min after EGF treatment. The error bars represent the SEM (Student's t-test; *, p < 0.05; n = 19). Western blot analysis shows Cat-PM expression. (D) Increase in EGFR endocytosis induced by PtdIns(4)P. The amount of endocytosed EGF is increased in cells treated with soluble fluorescent PtdIns(4)P (10 μM). Representative confocal images show PtdIns(4)P (green) and EGF-555 (red) 30 min after EGF treatment. Scale bars, 20 μm; 2 μm in magnification. The error bars represent the SEM (Student's t-test; *, p < 0.05; n = 13).

The presence of local PtdIns(4,5)P₂ at the plasma membrane is critical for the formation of clathrin-coated pits (CCPs) during clathrin-mediated endocytosis [50]. Its conversion by the 5-inositol phosphatases, such as Synj, is required for CCP initiation and stabilization [48]. As PtdIns(4)P is an intermediate product of Synj with PtdIns(4,5)P₂, we investigated whether PtdIns(4)P directly controls receptor-mediated endocytosis by monitoring the endocytosis EGFs during receptor activation in cells with or without excess PtdIns(4)P. We administered water-soluble fluorescent PtdIns(4)P (Echelon) and shuttle PIP™ carrier (Echelon) to cells and monitored EGF receptor endocytosis. We found a 61% increase in endocytosis in cells containing extra PtdIns(4)P compared to control cells (Fig. 5D). Our results, combined with a previous study demonstrating the requirement of Synj2 for clathrin-mediated endocytosis [18], show the consequence of the temporary accumulation of plasma membrane PtdIns(4)P, a dephosphorylated intermediate of PtdIns(4,5)P₂, through the action of the Synj2 dual phosphatase. Because the PtdIns(4)P level at the plasma membrane is also regulated by PI4 kinase III alpha [43,51], we investigated whether the observed transient increase in plasma membrane PtdIns(4)P levels is a result of PI4 kinase activation. We utilized an inhibitor of PI4 kinase activity, phenylarsine oxide (PAO), which depletes cellular PtdIns(4)P with minimal effects on the total amount of PtdIns(4,5)P₂ [43,49,52]. Treatment of recombinant Synj2 with PAO for 10 min did not affect the phosphatase activity of the Synj2 4-phosphatase domain or 5-phosphatase domain (Supplementary Fig. 4A). Cells treated with PAO showed a gradual reduction in plasma membrane PtdIns(4)P levels (Supplementary Figs. 4B and C), indicating the importance of PI4 Kinase in maintaining the PtdIns(4)P pool at the plasma membrane [51]. Cells activated with EGF in the presence of PAO exhibited a transient increase in plasma membrane PtdIns(4)P levels (Supplementary Figs. 4B and C), which may be attributed to the H₂O₂-dependent inactivation of PI phosphatases, such as Synj2, that contain redox-sensitive cysteine at the plasma membrane. A previous study [21] demonstrated that PtdIns(3,4)P₂ production from PtdIns(4)P by phosphatidylinositol 3-kinase C2α [PI(3)K C2α] is required for clathrin-mediated endocytosis. We investigated whether an increase in PtdIns(3,4)P₂ levels in the plasma membrane precedes invagination of endocytic CCPs. The administration of chlorpromazine [53], an inhibitor of coated pit assembly at the plasma membrane, or Dynasore [54], a dynamin inhibitor, enhanced plasma membrane PtdIns(3,4)P₂ amount in A549 cells during EGF stimulation (Supplementary Fig. 5), indicating that PtdIns(3,4)P₂ synthesis via PI(3)K C2α occurs before fission of CCPs by dynamin.

3.6. Structural insights into the H₂O₂-dependent accumulation of PtdIns(4)P and the successive cleavage of 5-P and 4-P phosphates in PtdIns(4.5)P₂ by synaptojanin

According to the in-silico model structure (Fig. 6A), the Sac1 domain is composed of two subdomains: an N-terminal subdomain and a catalytic subdomain with the Cys-X₅-Arg catalytic motif [55,56]. The catalytic subdomain adopts a three-layer αβα fold, characterized by a central β-sheet consisting of three parallel and five anti-parallel β-strands. The central sheet is surrounded by seven α-helices, with two located on one side and five on the opposite side. Notably, the P-segment of residues 385–392 spans the C-terminal end of β16, a connecting loop, and the N-terminal end of α10. This segment, which contains the Cys-X₅-Arg motif, is known to be responsible for phosphate binding.

Fig. 6.

Model structures of Synj and the Sac1 domain/PtdIns(4)P complex. (A) Overall structure of the Sac1 domain. The N-terminal and catalytic subdomains are colored in gray and pale cyan, respectively. The P-segment is highlighted in magenta with labeled secondary structural elements (β16 and α10). (B–C) Binding mode of PtdIns(4)P into the active site of Sac1 domain. (B) PtdIns(4)P (yellow) and the interacting residues of both the Sac1 domain (pale cyan) and the P-segment (magenta) are represented as sticks. The 4-P phosphate, 4’-, and 5′-carbon atoms of the inositol ring are marked as 4-P, 4′C, and 5′C, respectively. Gray dotted lines indicate polar interactions. Oxygen, nitrogen, sulfur, and phosphorus atoms are colored in red, blue, yellow, and orange, respectively. (C) Electrostatic potential surfaces of the Sac1 domain. Blue, red, and white colors indicate positively-charged, negatively-charged, and hydrophobic surfaces, respectively. The P-segment is highlighted with a white dotted circle. (D) Overall structure of Synj. For clarity, only the 5-phosphatase domain (green) and the Sac1 domain (pale cyan) are depicted, with emphasis on the active sites of each domain highlighted in gray shading. The PtdIns(3,4,5)P₃ (blue stick) and a magnesium ion (blue sphere) in the 5-phosphatase domain are derived from the superposed structure of the Synj1/PtdIns(3,4,5)P₃ complex (PDB code: 7A17). The PtdIns(4)P (yellow stick) in the Sac1 domain is derived from the docked complex structure.

To examine the substrate specificity of the Sac1 domain towards PtdIns(4)P compared to PtdIns(4,5)P₂, we docked PtdIns(4)P into the active site (Fig. 6B). The resulting complex model revealed that the 4-P phosphate on the inositol ring fits snugly into the positively charged pocket formed by the P-segment (Fig. 6C). The phosphate is electrostatically stabilized by the side chain guanidium of Arg-392 and the main chain –NH groups of Asp-388, and Asp-391. In addition to phosphate binding, it is worth noting that the –OH group on the 5′ carbon of PtdIns(4)P forms hydrogen bonds with both the guanidium moiety of Arg-392 and the imidazole ring of His-338. This molecular arrangement is related to the Sac1 domain's preference for PtdIns(4)P over PtdIns(4,5)P₂. In the case of PtdIns(4,5)P₂, the 5-P phosphate at the 5′ carbon could cause steric clash with the side chains of His-338 and Arg-392, consequently hindering the molecule's binding to the active site.

In the P-segment, the Cα carbons of Cys-386 and Cys-389 are positioned 4.8 Å apart, indicating a potential for disulfide bond formation in oxidative conditions (Fig. 6B). Such a bond would likely disrupt the conformation of the P-segment that is responsible for binding phosphates. Consequently, the catalytic activity of the Sac1 domain may be reduced due to the formation of disulfide bonds, which aligns with the H₂O₂-induced accumulation of PtdIns(4)P.

To gain insight into the successive cleavage of the 5-P and 4-P phosphates in PtdIns(4,5)P₂ by the 5-phosphatase and Sac1 domains, respectively, we predicted a model structure that contains both domains of Synj (Fig. 6D). Notably, the two domains were found to be closely positioned, with both orienting their active sites in the same direction. This structural arrangement is suitable for the transfer of PtdIns(4)P, which is the product of the 5-phosphatase domain, to the active site of the Sac1 domain. Remarkably, the substrate specificities of the 5-phosphatase and Sac1 domains towards PtdIns(4,5)P₂ and PtdIns(4)P, respectively, are reflected in their distinct structures. As described above, the Sac1 domain has a three-layer αβα fold with a single phosphate-binding site formed by the P-segment (Fig. 6A). In contrast, the crystal structure of the 5-phosphatase domain in a complex with PtdIns(3,4,5)P₃ revealed a four-layer αββα fold that allows the metal-mediated accommodation of two phosphates on the 4′ and 5′ carbons of the inositol ring (Fig. 6D).

4. Discussion

In this study, we demonstrated that the synaptojanin dual PtdIns(4,5)P₂ phosphatase is deactivated through cysteine oxidation of the 4-phosphatase domain by H₂O₂ molecules in the plasma membrane. These H₂O₂ molecules are produced in cells that are activated with growth factors. Since the 5-phosphatase domain of Synj is insensitive to high levels of H₂O₂, an intermediate product called PtdIns(4)P, derived from PtdIns(4,5)P₂, transiently accumulates at the plasma membrane. We observed that a high concentration of PtdIns(4)P molecules promotes receptor-mediated endocytosis. Considering that the production of PtdIns(3,4)P₂ through PtdIns(4)P is required for clathrin-mediated endocytosis driven by PI(3)K C2α [21], we believe that the transient deactivation of Synj by H₂O₂ is critical for PtdIns(3,4)P₂ production during endocytosis. In cells activated with EGF, the increased levels of H₂O₂ in the plasma membrane enhanced PtdIns(4)P through the oxidation of the 4-phosphatase domain of Synj, making them signaling mediators for receptor-mediated endocytosis. Plasma membrane PtdIns(4)P has distinct roles, including targeting specific proteins and serving as a substrate for PtdIns(4,5)P₂ synthesis [49]. During clathrin-mediated endocytosis, the temporary accumulation of PtdIns(4)P from PtdIns(4,5)P₂ by Synj at clathrin-coated pits (CCPs), followed by the timed synthesis of PtdIns(3,4)P₂ by PI(3)K C2α, is required for the formation of clathrin-coated vesicles [21,57]. Considering that Synj has both 4-phosphatase and 5-phosphatase activity (Fig. 2B) [25,28], the inhibition of the 4-phosphatase activity without interfering with the 5-phosphatase activity is critical for the temporary accumulation of PtdIns(4)P by Synj. Here, we showed that H₂O₂ molecules in the plasma membrane are transiently increased and selectively inhibit the 4-phosphatase activity of Synj2 through Cys oxidation in cells activated with EGF.

Based on the results of Synj oxidation and dynamic analysis of plasma membrane PtdIns(4)P levels, we propose a mechanism in which the activities of PI dual phosphatases are regulated differentially by H₂O₂ to regulate endocytosis through the conversion of PI around the plasma membrane (Fig. 7). EGF-mediated endocytosis takes place in a submembrane compartment containing EGF receptors and PtdIns(4,5)P₂ molecules. Subsequently, an activated EGF receptor promotes H₂O₂ production by activating NADPH oxidase. Increased H₂O₂ deactivates the 4-phosphatase domain of Synj phosphatase associated with the plasma membrane and induces the accumulation of PtdIns(4)P, an intermediate product of the enzyme, from PtdIns(4,5)P₂. Subsequently, PI(3)K C2α protein in clathrin-coated pits (CCPs) produces PtdIns(3,4)P₂ from PtdIns(4)P [21,57]. Increased PtdIns(3,4)P₂ is required for the recruitment of sorting nexin-9 (SNX9) BAR domain protein and SNX-18 protein during CCP maturation [21]. After CCP invagination, PtdIns(4)P is degraded to PtdIns by reduced Synj, and the depletion of PtdIns(4)P results in a decrease in PtdIns(3,4)P₂ levels in the endocytosed vesicle. The remaining PtdIns(3,4)P₂ molecules are converted to PtdIns(3)P by inositol polyphosphate 4-phosphatase (INPP4) [58] and ultimately converted to other PI species during endosomal maturation [59].

Fig. 7.

Proposed explanation for the consequence of H₂O₂-dependent inactivation of Synj during receptor-mediated endocytosis. Conversion of PtdIns(4,5)P₂ to PtdIns(4)P is achieved through the H₂O₂-dependent oxidation of the 4-phosphatase domain of the Synj dual phosphatase and is required for receptor-mediated endocytosis (RME) during EGF activation. PI(3)K C2α uses PtdIns(4)P as a substrate to produce PtdIns(3,4)P₂, which is critical for the formation and maturation of clathrin-coated pits. Complete degradation of PtdIns(4)P to PtdIns by Synj occurs in the endosome with a low level of H₂O₂. The membrane in red shows a PI-rich submembrane compartment. See the discussion for details.

The Synj structural prediction model (Fig. 6) supports the fact that the 5-phosphatase domain of Synj converts PtdIns(4,5)P₂ to PtdIns(4)P and releases it from the membrane in the presence of H₂O₂. In the absence of H₂O₂, the 4-phosphatase domain of Synj binds to PtdIns(4)P and converts it to PtdIns. Considering the sequential 5-phosphatase and 4-phosphatase activity of Synj, the selective regulation of 4-phosphatase activity in Synj is indispensable for the accumulation of plasma membrane PtdIns(4)P. The selective inhibition of Synj's 4-phosphatase activity by H₂O₂ increases during receptor activation and can drive cells to form endocytic vesicles through the spatiotemporal increase in PtdIns(4)P levels and the synthesis of PtdIns(3,4)P₂ by PI(3)K C2α. Without the accumulation of plasma membrane H₂O₂, Synj completely degrades PtdIns(4,5)P₂ into PtdIns, and, as a result, PtdIns(3,4)P₂ is not produced, leading to defective receptor-mediated endocytosis in cells expressing Cat-PM (Fig. 5). In addition to the presence of several PI 4-phosphatases or PI 5-phosphatases in the plasma membrane [60], we propose that the role of Synj dual phosphatase in receptor-mediated endocytosis is to regulate the temporal and spatial conversion of PtdIns(4,5)P₂ to PtdIns or PtdIns(4)P.

Synj phosphatase is a unique protein that contains both 4-phosphatase and 5-phosphatase activity. In contrast, most PtdIns(4,5)P₂ phosphatases, including OCRL (oculocerebrorenal syndrome of Lowe), INPP5B, and INPP5E, have only 5-phosphatase activity. Synj1 is highly abundant in presynaptic cells which need active clathrin-mediated endocytosis for signal transduction through exocytosis and endocytosis of synaptic vesicles. Synj converts PtdIns(4,5)P₂ to PtdIns(4)P and, subsequently, PtdIns(4)P to PtdIns. The rapid conversion of PtdIns(4,5)P₂ to PtdIns(3,4)P₂ is important for the formation and maturation of clathrin-coated pits during endocytosis. Our study demonstrates that an increase in PtdIns(4)P levels depend on Synj expression and the inactivation of Synj's 4-phosphatase activity by H₂O₂. Synj is also required for the uncoating of clathrin [16], and a defect in the 4-phosphatase domain of Synj1 impairs clathrin uncoating in presynaptic nerve terminals [32]. Synj not only plays a role in the formation of clathrin-coated pits through the accumulation of PtdIns(3,4)P₂, together with PI(3)K C2α, by supplying PtdIns(4)P, but also induces clathrin uncoating of vesicles by converting PtdIns(4)P to PtdIns. This precise temporal control of Synj phosphatase activity can be achieved through the H₂O₂ molecule, a short-lived signaling mediator at the plasma membrane. Synj-enriched clathrin-coated pits can generate PtdIns(3,4)P₂ in the presence of H₂O₂ and induce endocytosis. In endocytosed vesicles, Synj-associated vesicles induce clathrin uncoating and further endosomal maturation.

H₂O₂ molecules act as signaling mediators by oxidizing various physiological substrates such as phosphatases and kinases, and controlling their activities around the plasma membrane of cells that are activated by extracellular stimuli, including growth factors [61]. The concentration of H₂O₂ above 10–100 μM is temporary and local in the cell [62] because cells have developed an antioxidant system to protect themselves from unwanted damage. Here, we show that a transient increase in H₂O₂ is involved in the dynamic change from PtdIns(4,5)P₂ to PtdIns(4)P and subsequently to PtdIns(3,4)P₂ during endocytosis. The Synaptojanin dual phosphatase is responsible for the production of PtdIns(4)P through the temporary inactivation of its 4-phosphatase domain. H₂O₂ can control the half-life of PtdIns(4)P, an intermediate of Synj, through cysteine oxidation. Synj1 is known to participate in synaptic vesicle recycling in presynaptic cells [17], and the 4-phosphatase domain of Synj is critical for endocytosis in response to small stimuli [28]. Our findings provide evidence that H₂O₂ molecules control the ratio of the two products of Synj (PtdIns(4)P vs PtdIns) derived from PtdIns(4,5)P₂. During clathrin-mediated endocytosis, PtdIns(4,5)P₂ is required for the initiation of endocytic clathrin-coated pits and the production of PtdIns(3,4)P₂ by PI(3)K C2α is critical for the maturation of late-stage clathrin-coated pits [21]. We demonstrate a previously unknown interplay between PtdIns(4,5)P₂ and PtdIns(3,4)P₂ that depends on PtdIns(4)P, an intermediate product of Synj. This study proposes that the temporal and spatial accumulation of PtdIns(4)P through the inactivation of the 4-phosphatase domain of Synj by H₂O₂ at the clathrin-coated pits is a key process. The present study on Synj2 oxidation and its inhibition by H₂O₂ during endocytosis will facilitate future investigations into the possible oxidation of Synj1, an abundant PtdIns(4,5)P₂ phosphatase in the presynaptic cells of the brain, and its impact on synaptic transmission.

CRediT authorship contribution statement

Su In Jo: Writing – original draft, Visualization, Methodology, Investigation, Formal analysis. Suree Kim: Visualization, Methodology, Investigation, Data curation. Jung Mi Lim: Methodology, Investigation. Sue Goo Rhee: Writing – original draft, Resources. Bo-Gyeong Jeong: Visualization, Investigation. Sun-Shin Cha: Visualization, Validation, Investigation, Funding acquisition. Jae-Byum Chang: Methodology, Data curation. Dongmin Kang: Writing – review & editing, Writing – original draft, Visualization, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no conflict of interest.

Appendix A. Supplementary data

The following is the Supplementary data to this article.

Data availability

Data will be made available on request.