Introduction
Parkinson’s disease (PD) is the second most common neurodegenerative disease after Alzheimer’s disease (AD), and its incidence is increasing with the aging of the global population. Clinically, PD is characterized by progressive motor and non-motor symptoms. Pathologically, its hallmarks include progressive neuronal loss mainly involving dopaminergic neurons in the substantia nigra and the appearance of neuronal inclusions called Lewy bodies (LB), which are predominantly composed of aberrant α-synuclein (αSyn) aggregates [
3,
4,
56]. Cumulative evidence has indicated that αSyn aggregation plays a significant role in the pathogenesis of both familial and sporadic forms of PD, and its cell-to-cell propagation is associated with disease progression [
32,
42,
43]. Multiplication and missense mutations in the SNCA gene encoding αSyn that render it prone to aggregation are linked to familial forms of PD, but the initiating event of αSyn aggregation in sporadic PD remains unclear.
Many studies have examined the association between the interaction of αSyn with lipids, especially glucosylceramide, and the propensity of αSyn to aggregate [
1,
2,
10,
22,
29,
37,
54,
57,
58,
63]. Glucosylceramide has been targeted because mutations in
glucocerebrosidase (
GBA) represent the single largest risk factor for the development of PD. These mutations cause a consequent loss of enzymatic activity and excessive buildup of the substrate glucosylceramide. Glucosylceramide converts physiological αSyn conformers into pathogenic species that are prone to aggregate formation [
63]. Interestingly, in our previous study, structural analysis of Lewy bodies in the PD brain revealed that lipids were abundantly distributed in the core of Lewy bodies, even in idiopathic PD patients, indicating the involvement of some lipids in the initiation of αSyn aggregation [
3].
Recently, a transcriptome-wide association study (TWAS) evaluating the protein–protein interaction network connectivity between protein products nominated by PD TWAS and monogenic genes unveiled the link between Synaptojanin 1 (SYNJ1) and SNCA [
31]. SYNJ1 is a brain-enriched phosphatase that dephosphorylates phosphoinositides (PIPs) [
36]. PIPs are a small group of cellular phospholipids composed of two fatty acid chains linked by a glycerol moiety to a water-soluble inositol head group. Phosphorylation at positions 3, 4, and 5 of the inositol rings of phosphatidylinositol yielded seven distinct phosphoinositide derivatives, PI-3-P, PI-4-P, PI-5-P, PI-3,4-P
2, PI-3,5-P
2, PI-4,5-P
2 and PI-3,4,5-P
3 (PIP
3) [
6]. SYNJ1 consists of an amino-terminal Sac1-like phosphatase domain that dephosphorylates PI-3-P and PI-4-P, a central inositol 5′-phosphatase domain that dephosphorylates PI-4,5-P
2 and PIP
3, and a carboxyl-terminal proline-rich region. Mutations identified in the two phosphatase domains of SYNJ1, R258Q, R459P and R839C, have been associated with autosomal recessive, early-onset familial type PD (PARK20), implying involvement of lipid dysregulation in PD pathogenesis [
34,
41,
45]. Importantly, a study assessing genome-wide expression datasets for postmortem sporadic PD brains available in the public domain found that a subset of sporadic PD brains showed significant down-regulation of SYNJ1 transcript in various brain regions including the prefrontal cortex, striatum and substantia nigra [
39].
By integrating genetic evidence, this study aimed to determine the lipid molecule involved in the physiological to pathological shift of αSyn in PD pathogenesis and the underlying mechanism. Towards this goal, we employed genetically engineered cellular and Caenorhabditis elegans (C. elegans) models and demonstrated that depletion of SYNJ1 is associated with higher intracellular αSyn accumulation, accompanied by locomotory defects in C. elegans. Interestingly, lipidomic analysis revealed SYNJ1 depletion elevates the level of its substrate PIP3. Cell-based assays showed that upregulation of intracellular PIP3 itself induces the formation of αSyn inclusions. Subsequent in vitro protein–lipid overlay and aggregation assays further confirmed PIP3 as a lipid molecule that can directly interact with αSyn monomer and initiate aggregation. Notably, PIP3 induces the formation of PD-like fibril polymorphism. Postmortem brain analysis also unveiled the involvement of PIP3 dysregulation in synucleinopathy of PD. Overall, our data demonstrate that the aberrant interaction of αSyn with PIP3 which accumulates upon the loss of SYNJ1 phosphatase activity, is a factor that prompts physiologic αsyn to misfold and form pathologic inclusions, thereby highlighting a substantial role of PIP3 in the context of PD pathogenesis.
Materials and methods
Cell culture
The human cervical adenocarcinoma cell line HeLa (ATCC® CCL-2) and the human neuroblastoma cell line SH-SY5Y (ATCC® CRL-2266) were cultured in Dulbecco’s modified Eagle’s medium (Sigma-Aldrich, D5796) supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37 °C in a 95% air and 5% CO2 humidified incubator. Cells were routinely passaged after reaching 75–90% confluency.
Plasmid and virus production
The expression vectors for αSyn-GFP and αSyn-mRFP were created using the Multisite Gateway system, according to the manufacturer’s protocol (12537100, Invitrogen). Briefly, complementary DNA (cDNA) of αSyn with a stop codon was inserted between the attL1 and attR5 sites of the entry vector. Then, this was recombined with pENTR-L5-EGFP-L2 or pENTR-L5-mRFP-L2 (gift from Professor H. Kuroyanagi, University of the Ryukyus Graduate School of Medicine) into the destination vector with the CMV promoter pcDNA™-DEST40 (12274015, Invitrogen). To generate pMRX-ib-αSyn-mRFP for preparation of recombinant retroviruses, cDNA corresponding to αSyn-mRFP was subcloned into the pENTR1A plasmid. The sequences inserted into the pENTR1A plasmid were then transferred into pMRX-ib-DEST using Gateway LR Clonase. Recombinant retroviruses were prepared as previously described [
46].
gRNA design to target SYNJ1
Targeting sequences of sgRNA in the CRISPR/Cas9 system were determined using CRISPRdirect (
https://crispr.dbcls.jp/). The targeting sgRNA sequence used in this study was 5′-GGGACCAGGTTTAATGTCCG-3′. According to the Zhang Lab General cloning protocol (
http://www.addgene.org/crispr/zhang/), synthesized and annealed sgRNA targeting human
SYNJ1 was inserted into the modified pX330 plasmid (pX330; Addgene #42230) in which a P2A-puromycin resistance gene was conjugated to Cas9 (hereafter referred to as pCas9-puroR).
Establishment of SYNJ1-KO stable SH-SY5Y cell line
SH-SY5Y cells were transfected with pCas9-puroR vectors, and the cells were selected using 1 µg/ml puromycin (P8833, Sigma-Aldrich) for 7 days. Single colonies were manually isolated under a microscope. Genomic DNA of isolated clones was extracted using the PureLink™ Genomic DNA Mini Kit (K1820-01, Invitrogen), and sequences were confirmed. Out of 16 clones, 5 clones contained the homozygous frameshift SYNJ1 mutation. Depletion of SYNJ1 was confirmed by western blotting and immunostaining, and the established SYNJ1-KO lines were used for further experiments.
Transfection by electroporation
For each electroporation reaction of SH-SY5Y, 1.5 × 106 cells/100 μL optiMEM (Gibco, 31-985-062) was used. Ten μg of expression vectors encoding eGFP or αSyn-eGFP were added to the cell suspension. Cells/plasmid DNA suspensions were then transferred into 2-mm gap cuvettes (Nepa Gene Co., Ltd, EC-002S) and electroporated using a NEPA21 Super Electroporator (Nepa Gene Co., Ltd, Chiba, Japan). Immediately after electroporation, 500 μL of the pre-equilibrated culture medium was added to the cuvette, and the cell suspension was transferred to a 6-well plate. At 24 h post-transfection, media were replaced with fresh media; cells were harvested 72 h post-transfection for protein immunoblot analysis and immunostaining.
Western blotting
Cells were harvested and lysed in CelLytic MT Cell Lysis Reagent (C2978, Sigma-Aldrich) with a protease inhibitor (539131, Calbiochem) and phosphatase inhibitor (07574-61, Nacalai Tesque) mixture. The protein concentration of cell lysates was determined using a Pierce™ BCA protein assay kit (23225, Thermo Scientific). Lysates (10 μg/lane) were resolved on 10–20% sodium dodecyl sulfate (SDS)–polyacrylamide gel (2331840, ATTO Corporation) and transferred to polyvinylidene difluoride membranes (1620177, Bio-Rad Laboratories, Inc.). The membranes were incubated with blocking buffer and probed with the following primary antibodies: anti-Synj1 (1:1000) (ATL ATLAS antibodies AB, HPA011916); anti-phosphorylated αSyn (1:1000) (014-20281, Wako); αSyn antibody (Syn211) (1:1000) (32-8100, Invitrogen) and ACTB/β-actin antibody, clone C4 (1:10,000) (MAB1501, Merck Millipore). Following incubation with Amersham ECL horseradish peroxidase-conjugated anti-rabbit or anti-mouse secondary antibody (1:20,000) (NA934, NA931, GE Healthcare), the bound antibody was visualized with Amersham ECL Prime western blotting Detection Reagent (RPN2236, GE Healthcare) using the ChemiDoc Touch Imaging System (BioRad, Berkeley, CA, USA).
Immunocytochemistry
Cells were fixed with 4% PFA (09154-85, Nacalai Tesque) for 30 min at room temperature. After permeabilization with 0.1% Triton X-100 solution for 10 min and washing in tris-buffered saline (TBS), cells were blocked with 0.2% gelatin-TBS for 30 min and subsequently with primary antibodies, purified anti-PtdIns (3,4,5) P3 IgG (1:200) (Z-P345B, Echelon Biosciences), anti-Synj1 (1:500) (HPA011916, ATL ATLAS antibodies AB), anti-LAMP1 (1:100) (#9091, Cell Signaling Technology), and anti-LAMP-2 (sc-18822, Santa Cruz) overnight at 4 °C. After three washes in TBS, the cells were incubated with fluorescein AffiniPure donkey anti-mouse antibody (715-095-150, Jackson ImmunoResearch), Cy™3 AffiniPure donkey anti-Rabbit antibody (111-165-003, Jackson ImmunoResearch), or Alexa Fluor 647 donkey anti-rabbit antibody (A31573, Molecular Probes) (1:1000) for 1 h at room temperature. After washing in TBS, the cells were counterstained with Hoechst (H21486, Thermo Scientific) and observed microscopically using SpinSR10 (Olympus, Tokyo, Japan).
LC–MS/MS analysis
Lipidome analysis was conducted according to the Lipidome lab Multiphospholipid Scan package (Lipidome lab, Akita, Japan), using liquid chromatography triple quadrupole mass spectrometry (LC–TQMS) based on the methods described previously [
27,
53]. Fifty mg of cell pellets and brain samples were prepared and kept at − 80 ℃ until analysis. Briefly, the sample was dissolved with methanol, and then homogenized using a glass homogenizer. Total lipids were extracted using the liquid–liquid extraction as Bligh and Dyer methods [
8]. In addition, to analyze acidic phospholipids (PLs) such as PI and PIP, another aliquot of the same lipid extract was added with an equal volume of methanol before being loaded onto a diethylaminoethyl-cellulose column (Santa Cruz Biotechnology) pre-equilibrated with chloroform. After successive washes with chloroform/methanol (1:1, v/v), the acidic PLs were eluted with chloroform/methanol/HCl/water (12:12:1:1, v/v). The resultant fraction was subjected to a methylation reaction with TMS-diazomethane, followed by evaporation to dryness to give a residue, which was re-dissolved in methanol.
LC–MS/MS analysis was performed using the Xevo TQ-XS mass spectrometer with an ACQUITY UPLC H-Class (Waters). The lipids were separated on a Waters X-Bridge C18 column (3.5 μm, 150 mm × 1.0 mm internal diameter) at 40 °C using a gradient solvent system as follows: mobile phase A was isopropanol/methanol/water (5/1/4 v/v/v) supplemented with 5 mM ammonium formate and 0.05% ammonium hydroxide (28% in water); mobile phase B was isopropanol supplemented with 5 mM ammonium formate and 0.05% ammonium hydroxide (28% in water) with flow rate was 80 μL/min. Lipid species were measured using multiple reaction monitoring (MRM) in positive ion mode. Peak areas of individual species were normalized with those of the internal/surrogate standards PI 15:0/18:1-d7, PI(4)P 17:0/20:4, PI(4,5)P2 17:0/20:4 and PI(3,4,5)P3 17:0/20:4 (Avanti Polar Lipids), which were added to the samples before lipid extraction. The LC–MS/MS raw data were processed using analytical software (MassLynx4.2; Waters).
Caenorhabditis elegans culture and strains
Standard methods were used to culture
C. elegans on nematode growth medium (NGM) agar seeded with OP50 Escherichia coli (
E. coli) [
9]. The worms were maintained at 20 °C unless otherwise indicated.
The following strains, obtained from the
C. elegans Genetics Center, were used: N2 wild-type (Bristol), NL5901 pkIs2386 [unc-54p::alpha-synuclein::YFP] [
20], and EG3027 (mutant unc-26:s1710) [
47]. To create NL5901/N2 and NL5901/unc-26 heterozygote worms, NL5901 males were mated with N2 and unc-26 hermaphrodites, and F1 worms were used for the analysis. To create N2/N2 and N2/unc-26 heterozygote worms, N2 males were mated with N2 and unc-26 hermaphrodites.
Quantification of αSyn aggregates formed in C. elegans
Aggregates were quantified as previously described, with some modifications [
20,
23,
44]. Briefly, NL5901/N2 and NL5901/unc-26 heterozygote worms were created by mating. Synchronized nematodes were cultured until they reached the young adult stage (day 3). Animals were then transferred onto new NGM plates and cultured until adult day 2 stage, and αSyn aggregates were counted under a Zeiss LSM 700 confocal microscope. For each independent experiment, 14 worms from each group were examined. Aggregates were defined as discrete bright structures with boundaries distinguishable from the surrounding fluorescence. The aggregates were measured visually on all aggregates observed in the head region of the worms. The experiments were performed by an experimenter blinded to the quantification grouping of worms.
Locomotion assay
The locomotion speed of worms was analyzed using a multi-worm tracker (MWT), as previously described [
24,
60]. Briefly, age-synchronized adult day 2 worms (
n = 50 for each group) were washed three times in NG buffer and transferred from the NGM culture plate onto the assay plate. The assay plate was a 13 × 10-cm plate filled with agar, which was divided into equal four regions. Regions were surrounded with glycerol, an aversive stimulus for
C. elegans, to prevent the animals from moving to other regions. The locomotion of the worms was captured using the MWT system, and the images were binarized to calculate the locomotion speed of individual worms.
Analysis of the MWT data
Analysis of the recordings was performed using Choreography (part of the MWT software) and custom-written scripts to organize and summarize the data. Animal tracks were collected as previously described [
24]. Using the MWT software, we drew the trajectory of the worm's movement and measured the average speed of the locomotion.
Establishment of HeLa-αSyn-mRFP stable line
To generate a stable HeLa-αSyn-mRFP line (gift from Professor T. Yoshimori, Osaka University), HeLa cells were first infected with recombinant viruses prepared from pMRX-ib-αSyn-mRFP using polybrene (TR-1003, Sigma-Aldrich). At 48 h after infection, cells were cultured in selection-medium containing 5 µg/ml blasticidin (A1113902, Gibco). The selection process was conducted for a series of passages by introducing fresh medium and antibiotics to the cells. After 10–14 days, polyclonal populations of blasticidin-selected cells were pooled, expanded, and subjected to fluorescence-activated cell sorting (FACS) using BD FACS AriaIIIu (Becton, Dickinson) to isolate single clones from the top 0.5% RFP-signal cells, followed by manual colony pickup.
PIP3 delivery to Hela αSyn-mRFP cells
Intracellular delivery of PIP3 was performed using the PIP3 Shuttle PIP™ Kit (P-9039, Echelon Biosciences), with slight modifications to the manufacturer’s protocol. Briefly, HeLa cells stably overexpressing αSyn-mRFP were seeded on a 4-well glass bottom dish (Matsunami, D141400) and incubated overnight at 37 °C. Shuttle PIP carrier 2 (Histone H1) and Bodipy®-FL-PIP3 were incubated in a 0.2 ml tube in a 1:1 molar ratio for 10 min at room temperature. The complex was diluted with Opti-MEM and added to media covering HeLa cells with a final carrier and PIP3 concentration of 5 µM. The following day, the dye-containing media was removed, and cells were washed with phosphate-buffered saline (PBS) before live imaging using SpinSR10 (Olympus) or fixation with 4% paraformaldehyde (PFA) for immunostaining. Confocal images were taken randomly across the entire well at 60× magnification, and the percentage of αSyn puncta-positive cells was quantified. For transient gene overexpression, HeLa cells were seeded and transfected with pcDNA-αSyn-mRFP 24 h prior to delivery of PIP3 using Fugene HD transfection reagent (E2311, Promega Corporation), according to the manufacturer’s protocol. For live imaging of lysosomes, cells were incubated with 300 nM LysoTracker™ Blue DND-22 (L7525, Invitrogen) for 1 h and washed three times with PBS before observation.
Phosphatase inhibitor treatment
SF1670 (B-0350, Echelon Biosciences), dissolved in dimethylsulfoxide (DMSO), was added to the culture medium 24 h after cell seeding. DMSO was used as the vehicle control. After 24 h of treatment, the cells were fixed for immunostaining and imaging.
Primary neuronal cultures and treatment
Tissue culture plates were coated with poly-l-ornithine (0.2 mg/ml) for 1 h at 37 °C and washed 3 times with autoclaved milli-Q water. C57BL/6 mouse E15.5 pup brains were collected for primary neuronal cultures. Neurobasal media supplemented with B-27, glutamax, penicillin and streptomycin was used for cultures. Glial inhibitor, 5-fluoro-2-deoxyuridine, was added at 3 days in vitro (DIV). A mixture of shuttle PIP carrier 3 and Bodipy®-FL-PIP3 was added to the primary neuronal cultures at DIV10. Cells were fixed with 4% paraformaldehyde 4 days after Bodipy®-FL-PIP3 treatment, followed by immunostaining of anti-phosphorylated S129 α-syn (1:1000) (EP1536Y, ab51253, abcam), anti-NeuN antibody (1:500) (A60, MAB377, EMD Millipore) and anti-microtubule-associated protein 2 (MAP2) (1:5000) (NB300-213, Novus Biologicals) before microscopic observation using IN Cell Analyzer 6000 (GE healthcare, Chicago, IL, USA). Sixty-four images covering the entire area of culture well were taken at 20× magnification. Automated quantification of the intensity, area and count of phosphorylated α-syn, total area of MAP2 and number of NeuN-positive cells was performed using the software IN Cell Developer Toolbox. For inhibitor experiment, SF1670 (0.5 µM) was added to the primary neuronal cultures at DIV10, and cell fixation was performed at DIV14 followed by immunostaining with purified anti-PtdIns (3,4,5) P3 IgG (1:200) (Z-P345B, Echelon Biosciences), anti-phosphorylated S129 α-syn (1:1000) (EP1536Y, ab51253, abcam) and anti-MAP2 (1:5000) (NB300-213, Novus Biologicals). Confocal images were acquired using SpinSR10 microscope (Olympus, Tokyo, Japan). Following image acquisition, fluorescence intensity of PIP3 and pSyn per unit of total MAP2 area were analyzed using ImageJ software. For immunofluorescence assessment of the synaptic localization of PIP3 and pSyn, primary antibodies anti-PtdIns (3,4,5) P3 IgG (1:200) (Z-P345B, Echelon Biosciences), anti-phosphorylated α-syn (pSyn#64) (1:500) (015-25191, Wako), anti-SNAP25 (1:250) (1113839, GeneTex) and PSD95 (D27E11) antibodies (1:100) (3450, Cell Signaling Technology) were used.
Protein purification
Human wild-type (WT) αSyn was purified from
E. coli as described previously [
26,
62]. Briefly, a plasmid containing WT human αSyn was expressed in
E. coli BL21 (DE3) (69450, Novagen, Merck, San Diego, CA, USA).
E. coli were suspended in buffer, crushed by sonication, and centrifuged at 8000 rotations per minute (rpm). Streptomycin sulfate (06339-52, Nacalai Tesque, Kyoto, Japan) (final 2.5% [w/w]) was added to the supernatant and centrifuged at 8000 rpm. The supernatant was then heated at 90 °C in a water bath and centrifuged at 20,000 rpm. The supernatant was then (1) precipitated with solid ammonium sulfate (02620-75, Nacalai Tesque, Kyoto, Japan) to 70% saturation, (2) centrifuged at 20,000 rpm, (3) dialyzed overnight, (4) applied onto a Resource-Q column (GE Healthcare, Little Chalfont, UK) with 50 mM Tris–HCl buffer (pH 7.5) containing 0.1 mM dithiothreitol (14112-52, Nacalai Tesque, Kyoto, Japan) and 0.1 mM phenylmethylsulfonyl fluoride (022-15371, FujiFilm Wako Pure Chemical Corporation, Osaka, Japan) as the running buffer, and (5) eluted with a linear gradient of 0–1 M NaCl. αSyn-enriched fractions were pooled and further purified by size exclusion chromatography using a HiLoad Superdex 200 26/600 pg column (GE Healthcare) equilibrated with 50 mM Tris–HCl (pH 7.5) and 150 mM NaCl. The purified fractions were combined and dialyzed against deionized water at 4 °C overnight. Sample solutions were flash-frozen in liquid nitrogen, lyophilized, and stored at − 80 °C until use. The fractions containing αSyn (as determined by SDS-PAGE/Coomasie blue staining) were joint, dialyzed versus deionized water, acidified with 5 mM HCl and loaded onto a Reverse Phase Cosmosil Protein R × 250 mm Preparative Column (Nacalai-Tesque, Kyoto, Japan) and eluted with a linear gradient of 30–90% acetonitrile. The pure fractions were combined and flash-frozen in liquid nitrogen, lyophilized and stored at − 80 ºC until use.
Lipid binding assay
To assess the direct binding of αSyn with various lipids, a protein–lipid overlay assay was performed using membrane lipid strips (P-6002), PIP strips (P-6001), and Sphingo strips (S-6000) purchased from Echelon Biosciences (Salt Lake City, USA). First, lipid membranes were blocked with chemical-blocking buffer (EzBlock Chemi, AE-1475, ATTO, Tokyo, Japan) for 30 min, and incubated with 0.5 µg/ml of αSyn protein in blocking buffer for 1 h at room temperature with gentle agitation. The strips were washed with TBS-T three times for 10 min with shaking, and then incubated with αSyn antibody (Syn211) (32-8100, Invitrogen) or α/β-synuclein (F-11) antibody (sc-514908, Santa Cruz Biotechnology) for 1 h. After three washes, the strips were treated with anti-mouse IgG HRP-conjugated antibody for 1 h and washed three times. Finally, αSyn binding to each lipid was evaluated by chemiluminescence detection with ECL prime (RPN2232, Cytiva, Tokyo, Japan).
SUV preparation
SUVs were prepared according to the method described by Suzuki et al. [
58]. POPC (L-1618), PI (P-0016), PI-3-P (P-3016), PI-4-P (P-4016), PI-5-P (P-5016), PI-3,4-P
2 (P-3416), PI-3,5-P
2 (P-3516), PI-4,5-P
2 (P-4516), and PI-3,4,5-P
3 (P-3916) were purchased from Echelon Biosciences. Briefly, lipid mixtures of POPC with respective PIPs (90:10 molar ratio, total concentration at 10 mM) in chloroform were dried under an atmosphere of N2 and lyophilized overnight to remove any trace of organic solvent. The thin lipid film obtained was hydrated in fibrillation buffer (50 mM Tris–HCl pH 7.4, 150 mM NaCl) and sonicated using Bioruptor II (BM Equipment, Japan), with 5 cycles of 30 s of sonication and 30 s of quiescence. The size and integrity of the vesicles were verified using TEM and dynamic light scattering.
Amyloid fibrils were formed by dissolving lyophilized αSyn in fibrillation buffer, filtering through 0.22 µm membrane, and adjusting to 0.5 mg/mL supplemented with 10 µL ThT (202-01002, FujiFilm Wako Pure Chemical Corporation, Osaka, Japan); all experiments were performed in the presence or absence of 1 mM of POPC-PIPs SUVs. Each reaction mixture (100 µL) was transferred to a 96-well sealed plate (Costar Assay Plate, Corning, USA), with each well containing 35 mg ZrO2 beads (YTZ-0.5, Nikkato Corporation) to facilitate the fibril formation. The microplate with the reaction mixtures was subjected to cyclic agitation with a 3 min orbital shaking period at 2000 rpm, followed by a 12 min quiescent period at 37 °C. The kinetics of fibril formation was monitored according to ThT intensity fluorescence (excitation at 450 nm and emission at 485 nm) every 15 min in an MTP-900 microplate reader (Corona Electric Co., Tokyo, Japan). All the reaction conditions were evaluated with at least 15 replicates, and the kinetics was characterized according to the lag time (i.e., the time required to reach a fluorescence value of 500 A.U.) and the maximum ThT intensity (i.e., the highest intensity value in the measuring period).
Brain lysate preparation
Amygdala sections from the frozen side and amygdala sections from the formalin-fixed side were prepared. Amygdala slices (100 mg) from the brains of patients with PD and MSA were placed into Precellys Lysing tubes (P000912-LYSK 0-A, M&S Instruments, Osaka, Japan), resuspended in 1 mL of fibrillation buffer, and subjected to two cycles of high-speed shaking for 20 s in a lysis and homogenization system (Bertin Instruments, France). Then, the homogenates were transferred into Eppendorf tubes and centrifuged at 2000×g for 2 min at room temperature. The concentration of total protein in the supernatant fractions was quantified using the MicroBCA Protein Assay Reagent Kit (23235, Thermo Pierce), and the homogenates were aliquoted and stored at − 80 °C until use.
Amplification of ɑSyn aggregates from amygdala brain homogenates
A volume of amygdala brain homogenate was added to solutions of ɑSyn to reach a final concentration of 20 µg/mL total protein. Then, 200 µL of 0.5 mg/mL monomeric ɑSyn in the fibrillation buffer containing the amygdala lysate was added into a 96-multiplate (675096, Greiner Bio-One). This was subjected to ultrasonication to accelerate the amyloid formation at an optimized frequency of 30 kHz in cycles of 300 ms of irradiation and 800 ms of quiescence using a Handai Amyloid Burst Inducer (HANABI) equipment (CORONA ELECTRIC, Ibaraki, Japan). ThT fluorescence intensity was recorded as a function of time.
Transmission electron microscopy
Fibrils were adsorbed onto 400-mesh grids (Nisshin EM Co., Ltd., Tokyo) and negatively stained with 1% phosphotungstic acid (27807-62, Nacalai Tesque, Kyoto, Japan), and their structures were observed using an H-7650 TEM (Hitachi High Technologies Corporation, Tokyo, Japan) operated at 80 kV.
Proteinase K resistance assay
ɑSyn fibrils (0.5 mg/mL) in the fibrillation buffer were digested using proteinase K (03115887001, Sigma) (1 µg/mL) at 37 °C and agitation at 400 rpm for different time intervals. To stop the reaction, the samples were incubated at 95 °C for 5 min, mixed with loading buffer (1610747, Bio-Rad) (50 mM Tris–HCl, pH 6.8, 4% SDS, 2% β-mercaptoethanol, 12% glycerol, and 0.01% bromophenol blue) and incubated at 95 °C for an additional 10 min. The digestion patterns were analyzed using SDS-polyacrylamide gel electrophoresis, followed by Coomassie Brilliant Blue (11642-31, Nacalai Tesque, Kyoto, Japan) staining. The first five digestion products, B1–B5, were used for analysis. The proteinase K resistance (PKR) score was established as the band intensity ratio between bands B2 and B1 (B2/B1).
1H-15N heterogeneous single-quantum coherence NMR spectroscopy
1H-
15N heterogenous single-quantum coherence (HSQC) NMR measurements were performed using 100 µM
15N-labeled ɑSyn dissolved in fibrillation buffer prepared in H
2O/D
2O (9:1, v/v). The
15N-labeled ɑSyn was expressed in M9 minimal medium containing
15NH
4Cl and purified as described for the unlabeled protein. NMR spectra were acquired at 37 °C on a Bruker AVANCE III HD 600 MHz NMR spectrometer equipped with a 5 mm quadruple resonance cryogenic probe (Bruker Biospin). The data size and spectral width were 256 (t1) × 2048 (t2) and 1338 Hz (ω1,
15N) × 9,615 Hz (ω2,
1H), respectively. The carrier frequencies of
1H and
15N were 4.7 and 118 ppm, respectively. The number of scans/FID was 32. The repetition time was 1 s. The peak assignment at pH 7.4 was achieved employing the assignment data reported by El Turk et al. [
61]. The chemical shift perturbation (CSP) is calculated as follows:
$$\Delta \delta =\sqrt{{\Delta \delta }_{\mathrm{H}}^{2}+{\left(\frac{1}{8}{\Delta \delta }_{\mathrm{N}}\right)}^{2}},$$
where Δ
δH and Δ
δN are the chemical shift changes (in ppm) with respect to the H and N axes, respectively. All NMR spectra were processed with Topspin (Bruker Biospin), NMRPipe [
15] and NMRFAM-sparky [
30].
Tissue preparation and chromogenic immunohistochemistry (IHC) staining
Clinical profiles of human autopsy cases (disease control patients,
n = 3; PD patients,
n = 3) used for chromogenic IHC are shown in Table
1. The midbrains, including the substantia nigra, from the patients were fixed overnight in 4% PFA and then immersed in PBS containing 30% sucrose until sinking. The brain samples were cut into 40-μm-thick sections using a cryostat (CM1850; Leica Microsystems). Free-floating sections were washed in TBS and immersed in a solution of 3% H
2O
2 to quench endogenous peroxidase activity. Then, they were incubated with the primary antibody against PtdIns(3,4,5)P3 (1:100) (Z-P345B, Echelon Biosciences) in TBS containing 10% blockace (UKB80, KAC Co., Ltd.) overnight at 4 °C with continuous shaking. The sections were then washed three times in TBS-T and incubated with biotinylated anti-mouse secondary antibody (BA-9200, Vector Laboratories) in TBS-T for 2 h at room temperature. The sections were then incubated with avidin–biotin peroxidase complex (PK-6100, Vector Laboratories) for 1 h. Thereafter, the reaction products were visualized with 3,3-diaminobenzidine tetrahydrochloride. All sections were then washed in TBS, mounted on amino propyltriethoxysilan-coated slides, dried, stained with crystal violet, dehydrated in a graded series of ethanol, cleared in xylene, and coverslipped. Images were obtained using an Eclipse Ni-E microscope (Nikon, Tokyo, Japan).
Table 1
Clinical information on the postmortem human brain samples
Control |
#1 | 69 | M | 24 | Multiple cerebral infarction, aspiration pneumonia | Respiratory failure | None | | | 〇 |
#2 | 67 | M | 58 | Intracerebral hemorrhage, subarachnoid hemorrhage | Cerebral herniation | None | | | 〇 |
#3 | 74 | M | 4 | Intracerebral hemorrhage, pneumonia | Respiratory failure | None | 〇 | | 〇 |
#4 | 84 | M | 11 | Multiple cerebral infarction, Alzheimer's disease, pulmonary aspergillosis | Septic shock | None | 〇 | | 〇 |
#5 | 83 | M | 19 | Intracerebral hemorrhage, aspiration pneumonia | Respiratory failure | None | 〇 | 〇 | 〇 |
#6 | 88 | M | 15 | Metabolic encephalopathy | Multiple organ failure | None | | 〇 | |
#7 | 65 | M | 2.5 | Chronic liver failure, chronic renal failure | Multiple organ failure | None | | 〇 | |
#8 | 84 | F | 21 | Intracerebral hemorrhage, aspiration pneumonia | Respiratory failure | None | | 〇 | |
PD |
#1 | 83 | M | 10 | PD, pneumonia | Respiratory failure | Limbic | | | 〇 |
#2 | 66 | M | 17 | PD, aspiration pneumonia | Respiratory failure | Limbic | 〇 | | 〇 |
#3 | 86 | F | 9.5 | PD | Suffocation | Limbic | 〇 | | 〇 |
#4 | 88 | M | 3 | PD, aspiration pneumonia | Respiratory failure | Limbic | 〇 | 〇 | |
#5 | 85 | M | 8 | PD, pneumonia | Respiratory failure | Limbic | | 〇 | |
#6 | 84 | F | 3 | PD | Chronic heart failure | Limbic | | 〇 | |
#7 | 79 | F | 19.5 | PD, acute myelocytic leukemia, fungal pneumonia | Multiple organ failure | Brainstem | | 〇 | |
Tissue preparation and immunofluorescence (IF) staining
Clinical profiles of human autopsy cases (disease control patients,
n = 4; PD patients,
n = 4) used for IF staining are shown in Table
1. For IF analyses, sections of the midbrains were cut at 40-μm thickness using cryostat (CM1850; Leica Microsystems) and placed onto glass slide. Sections were stored in a sealed slide box at − 80 °C until use. Before immunostaining, the tissue sections were dried at room temperature for 20 min and immersed in pre-cooled acetone (− 20 °C) for 5 min. Fixative was poured off and the tissue sections was dried for 30 min to allow acetone to evaporate from the tissue sections. The slides were rinsed in TBS for 3 changes, 5 min each. The sections were then soaked with blocking agent 10% normal goat serum (NGS) (S-1000 Vector Laboratories) and incubated with the primary antibodies against PtdIns(3,4,5)P3 (1:200) (Z-P345B, Echelon Biosciences), phosphorylated S129 α-syn (1:1000) (EP1536Y, ab51253, abcam) and MAP2 (1:5000) (NB300-213, Novus Biologicals) in 2% NGS a humidified chamber at 4 ℃ overnight. Sections were washed three times before incubation with secondary antibodies for 1 h at room temperature. Alexa Fluor
® 405 goat anti-mouse IgG (H + L) antibody (A31553, Thermo Fisher Scientific), Alexa Fluor
® 488 donkey anti-mouse IgG (H + L) antibody (A21202, Thermo Fisher Scientific), Alexa Fluor
® 594 donkey anti-rabbit IgG (H + L) antibody (A21207, Thermo Fisher Scientific) and Alexa Fluor
® Plus 647 goat anti-chicken (A32933, Thermo Fisher Scientific) were used as secondary antibodies. Slides were washed 3 times, 5 min each, in TBS before applying mounting media and coverslip. Confocal mages were obtained using SpinSR10 microscope (Olympus, Tokyo, Japan). Following image acquisition, fluorescence intensity of protein of interest per unit of total MAP2 area were analyzed using ImageJ software.
Sample preparation for lipidomic analysis
Clinical profiles of human autopsy cases (disease control patients,
n = 5; PD patients,
n = 3) used for lipidomic analysis are shown in Table
1. The medulla oblongata and cerebellar cortex of the patients were frozen in powdered dry ice. Frozen samples from the tegmentum of medulla oblongata including the dorsal nucleus of vagus nerve and cerebellar cortex were used for lipidomic analysis.
Statistical analysis
For statistical comparisons, all data were checked for normality using D’Agostino and Pearson normality test before parametric and non-parametric tests were performed. For parametric test, data were analyzed by one-way and two-way ANOVA followed by Dunnett’s and Tukey’s multiple post hoc test for comparing more than three samples, and two-tailed Student’s t test and multiple t test for comparing two samples with 95% confidence. For non-parametric analysis, Kruskal–Wallis test with Dunn’s multiple comparisons was performed. Significance was accepted when p < 0.05. p values are presented as *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. All statistical tests were performed using GraphPad Prism 7 (GraphPad Software, Inc, San Diego, CA, USA).
Ethics statement
This study was approved by the Ethics Committee of Osaka University Hospital (no. 12148) and was conducted in accordance with the Declaration of Helsinki and the Ethical Guidelines for Medical and Health Research Involving Human Subjects endorsed by the Japanese government. All subjects provided informed consent.
Discussion
Genetic and pathological studies have revealed that lipid dysregulation plays a key role in the pathomechanisms of αSyn aggregation in patients with PD [
28,
50,
59]. In this study, we demonstrate that functional loss of lipid phosphatase SYNJ1 promotes the pathological aggregation of αSyn via the dysregulation of its substrate PIP
3. Concomitantly, we identify PIP
3 as a novel αSyn interactor and aggregation inducer and relate its dysregulation to PD.
Mutations in SYNJ1 have been identified in an early-onset autosomal recessive form of PD (PARK20) [
45]. The pathomechanisms of PARK20 have been related to the dysfunction of synaptic vesicle recycling and the autophagy system [
12,
16,
40]. However, the direct interplay between SYNJ1 and αSyn aggregation is not fully understood. Our results indicate that SYNJ1 deficiency causes intracellular αSyn aggregation mediated by the accumulation of PIP
3 and locomotor decline in
C. elegans models. Indeed, Synj1 haploinsufficiency mice have been reported to exhibit PD-like pathologies comprising αSyn accumulation, impaired autophagy and dopaminergic terminal degeneration as well as age-dependent motor function abnormalities. In the study, the authors showed the elevation of 5′-phosphatase substrate, PI-4,5-P
2, in the Synj1 haploinsufficiency mice but did not investigate the level of PIP
3. Importantly, down-regulation of SYNJ1 transcript could be observed in a subset of sporadic PD brains, implicating involvement of SYNJ1 deficiency and the upregulation of its PIP substrates in αSyn accumulation in sporadic PD [
39]. Starting off with genetic approach, we demonstrate PIP
3 is one of the lipid candidates that can initiate αSyn aggregation.
PIPs make up only a small fraction of cellular phospholipid and yet they play important roles in a wide range of cellular processes, including membrane dynamics, trafficking, and intracellular signaling, with each PIP exhibiting distinct subcellular localization and function [
6,
14]. For example, PI-4,5-P
2 and PIP
3 are found at the plasma membrane, PI-3,4-P
2 is largely localized at the plasma membrane and in the early endocytic pathway, PI-3-P is concentrated in early endosomes, PI-3,5-P
2 exists in late compartments of the endosomal pathway, and PI-4-P is enriched at the Golgi complex but also present at the plasma membrane[
41]. More studies have focused on the function of PI-4,5-P
2 rather than that of PIP
3 as the steady-state abundance of PIP
3 at the plasma membrane is low. However, local levels of PIP
3 change dynamically following stimulation [
7]. Indeed, αSyn has been shown to form discrete foci at the cellular plasma membrane, whereby the abundance and localization of these foci correlate with pools of PIP
3 and PI-4,5-P
2 [
25].
Expanding on these previous findings, we demonstrated that increased cellular PIP
3 levels resulted in the formation of αSyn inclusions in an overexpression cultured cell line. Following cellular uptake, Bodipy-FL-PIP
3 can be observed in small round vesicular structures resembling endosomes/lysosomes. PIP
3 localization at endosomes has been previously reported [
17]. The current study confirmed that the PIP
3- and αSyn-positive inclusions were localized in the lysosomes, suggesting that this could be the site of their encounter. We speculate that the association with PIP
3 in lysosomes initiates αSyn aggregation, which leads to disruption of the lysosomal membrane and, in turn, cascading aggregation of cytoplasmic αSyn [
18,
49]. Importantly, the lysosomal localization of these inclusions closely resembles the pathological feature of LB as it has been reported that lysosome-like vesicles could be detected inside or at the edge of LB inclusions [
33,
51]. We also present evidence that PIP
3 hastened the formation of aggregates from endogenously expressed αSyn using primary neurons. In primary cells, PIP
3 and αSyn aggregates were predominantly localized in the presynaptic region in the neuronal processes. Presynaptic localization PIP
3 observed here is consistent with a previous study that reported developing cortical neurons exhibited intense PIP
3 levels in their axon and growth cone during the period of rapid axon growth [
38]. Our result also agrees with the previous studies using mouse models showing αSyn being enriched in the presynaptic region [
11] and phosphorylation of endogenous αSyn initiates at the presynaptic region and spreads through the axon to the cell body [
5]. Intriguingly, the phenotype observed in the neuronal processes of primary neurons fittingly recapitulates the initial stage of the formation of Lewy neurites found in diseased synucleinopathy brains.
A more thorough assessment on how distinct PIPs affect αSyn aggregation was performed using in vitro-based assays. In vitro protein–lipid interaction studies and seeding assays revealed that mono-, di-, and triphosphorylated PIPs bind to αSyn and promote their aggregation to different extents, with PIP3 ranked top on both assessments. The higher affinity of PIP3 for αSyn than of other PIPs could be attributed to the stereochemistry of the phosphorylation of the inositol moiety. The results of PIPs–αSyn binding indicated that single phosphorylation of PI promotes the interaction with αSyn, probably due to electrostatic contacts between the N-terminal domain of αSyn and the negative phosphate group. Among the mono-phosphorylated PIPs, PI-3-P showed the highest affinity for αSyn, which increased if a second phosphate group was added at position 5 or was affected if the position was 4. Considering the stereochemistry of the inositol moiety, a double phosphorylation at positions 3 and 5 would enhance the interaction because both phosphate groups are oriented in the same direction. Alternately, if double phosphorylation takes place at positions 3 and 4, the phosphate groups are oriented in different directions, negatively affecting the interaction. Finally, PIP3 showed the strongest interaction, probably due to the additive effect of simultaneous phosphorylation at positions 3, 4, and 5, while maintaining favorable stereochemistry.
This is the first detailed report showing a direct association between PIP
3 and αSyn which is attributed to specific interaction with the N-terminal and NAC regions of αSyn, as revealed by NMR spectra. Collectively, in vitro experiments help clarify the mechanism of how PIP
3 initiates αSyn misfolding. Notably, the interaction with PIP
3 induces the formation of fibrils exhibiting structural and biochemical properties similar to those amplified from PD brains. Recent studies have evaluated the relationship between structural polymorphism and disease diversity to understand the mechanisms by which a single amyloidogenic protein causes different diseases. For example, αSyn causes both PD and multiple system atrophy (MSA). Ultrastructural analysis of αSyn amyloid-like fibrils extracted from patients’ brains showed that brain αSyn fibrils differ between those in MSA and in PD/DLB, with the former being predominantly twisted and the latter being mostly straight rod-like [
13,
48,
55]. The current study demonstrated that the PIP
3-derived fibrils showed rod-like morphology similar to that of the PD brain-derived fibrils. To the best of our knowledge, this is also the first study to connect the accumulation of a specific lipid to the creation of fibrils with morphological and biochemical similarities to those of PD-related fibrils.
Previous studies on the interplay between αSyn and other lipids were mainly based on in vitro and in vivo preclinical studies with little direct investigation of PD brain samples [
19]. The key strengths of this study are that we provide evidence showing PIP
3 induces the pathological aggregation of αSyn to form fibrils showing structural and biochemical resemblance to those derived from PD brains and, importantly, the accumulation of PIP
3 colocalized with pSyn in postmortem PD brain samples. Regarding the expression levels of SYNJ1, we did not observe a significant increase in the PD brains of our study population, indicating that PIP
3 accumulation is a common pathological change in both familial (PARK20) and sporadic PD and the reduced SYNJ1 activity is not the sole reason for the accumulation of PIP
3. It would be an important theme to investigate the prevalence of SYNJ1 deficiency among the sporadic PD patients and to explore the upstream causative mechanism of PIP
3 accumulation in the remaining populations in future studies. We also acknowledge the limitation of the small sample size of postmortem cohorts used in our study which may not sufficiently reflect the heterogeneity of PD. Larger sample size may be required in future studies.
In conclusion, aberrant interaction of αSyn with PIP3 which accumulates upon the loss of function of lipid phosphatase SYNJ1, promotes the transition of physiological αSyn to pathological assemblies showing structural and biochemical similarities to those derived from PD brains. This study thereby highlights an emerging role of PIP3 in the context of PD pathogenesis and opens new therapeutic perspectives targeting PIP3 to improve PD pathology.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.