Introduction

Parkinson’s disease (PD) and many other neurodegenerative disorders are associated with the misfolding and accumulation of a wide range of proteins [1]. A defining pathological characteristic of synucleinopathies, including PD, is the presence of protein aggregates called Lewy bodies and Lewy neurites where the major component is the protein α-synuclein (a-syn) [2, 3]. Although it has been known for almost two decades that duplication [4] and triplication [5] events, as well as mutations in the α-synuclein gene (SNCA) are a cause of familial PD, the precise function of a-syn in neurons remain largely unclear [6–11].

Structurally, a-syn is considered a natively unfolded and disordered ~14-kDa monomer but evidence suggests that a-syn multimerization [12] and post-translational modifications (PTMs) have central roles in PD pathogenesis [13–15]. It is believed that monomeric a-syn is highly disorganized and has a tendency to form fibrils in vitro [16] thus implicating monomeric a-syn as the neurotoxic form of the protein [12]. In support of this, further evidence has suggested that tetrameric a-syn is the most stable form of the protein, possibly representing a neuroprotective a-syn species [15]. However, conflicting data suggests that the a-syn tetramer may simply be a highly disordered a-syn monomeric conformation [16].

Numerous factors have been suggested to influence the transition of monomeric a-syn towards aggregation in neurodegeneration [17]. Importantly, the predominant modification of a-syn in Lewy bodies is phosphorylation at serine-129 (p-S129) suggesting that this phosphorylation event may drive Lewy body formation [18]. Furthermore, phosphorylation of a-syn at serine-87 (p-S87) reduces a-syn oligomerization and influences its interaction with membranes [19]. a-syn interaction with membranes stabilizes the secondary structure where its selective binding to acidic phospholipids leads to a conformational shift from random coil to >70 % α-helix, which in turn may reduce its aggregation potential [20–22]. Interestingly, the a-syn clinical mutation A30P reduces its lipid-binding affinity [23].

A-syn can also bind transition metals, such as copper and iron, and metal binding is thought to trigger aggregation [24–26]. However, metals may also have an indirect effect on a-syn by increasing reactive oxygen species (ROS) levels, which may exacerbate the deleterious effect of a-syn. Interestingly, oxidative stress induced by exogenous H2O2 plays a role in a-syn aggregation [27]. It has also been shown that treatment with dopamine can induce non-dopaminergic neurons to produce dopamine leading to oxidative stress and cell death [28].

It is becoming increasingly clear that a-syn most probably exists as a heterogeneous population of structured and unstructured monomers together with different multimeric a-syn species that may change characteristics as neurons transition from a non-pathological state to a pathological state [29–31]. Although cytosolic and membrane forms of a-syn have been mostly studied, a-syn also localizes to other cellular compartments such as the nucleus where it can promote neurotoxicity by binding to histones and decreasing the level of acetylated H3 [32]. In this study, we combine subcellular fractionation, high-performance liquid chromatography (HPLC) and mass spectrometry (MS) to define subcellular profiles of a-syn species in both healthy and PD neurons, and through this, we have identified a PD-enriched ~36-kDa a-syn species that may represent a prime target for intervention strategies to reduce a-syn pathogenicity.

Materials and Methods

Cell Culture

SH-SY5Y cells (ATCC, Manassas, VA) were cultured in 10 cm tissue culture dishes containing culture media composed of 50/50 ratio of DMEM/F12 media (Invitrogen, Carlsbad CA) with 10 % (v/v) fetal bovine serum (FBS; Atlanta Biologics, Lawrenceville, GA), 1× non-essential amino acids (Invitrogen, Carlsbad, CA), and 1× Glutmax (Invitrogen, Carlsbad, CA). HEK293 cells (ATCC, Manassas, VA) were maintained in a 10-cm tissue culture dish containing DMEM with 10 % (v/v) FBS (Atlanta Biologics, Lawrenceville, GA), 1× non-essential amino acids (Invitrogen, Carlsbad, CA), and 1× Glutmax (Invitrogen, Carlsbad, CA).

Transient and Stable Transfection

The pcDNA3 empty vector (Invitrogen, Carlsbad, CA) and pcDNA3 containing wtSNCA were transfected into SH-SY5Y cells (CRL-2266; ATCC, Manassas, VA) as outlined previously [33]. Transfected cells were analyzed 36–48 h post-transfection unless otherwise stated. Stable cell lines were generated using pcDNA4 (Invitrogen, Carlsbad, CA) empty vector and pcDNA4 containing wtSNCA using 2.5 mg/ml Zeocin selection (Invitrogen, Carlsbad, CA) for 14 days. Cultures were maintained on Zeocin selection throughout all the experimentation.

Antibodies

To ensure consistency within the different experiments, several different primary antibodies were used for all studies performed, showing identical results. a-syn antibodies included rabbit monoclonal anti-a-syn antibody from The Michael J. Fox foundation, rabbit polyclonal anti-a-syn antibody (ABCAM cat. no. ab93432, Cambridge, MA), mouse monoclonal anti-a-syn antibody (Invitrogen cat. no. 180215, Carlsbad, CA), and rat monoclonal anti-a-syn clone 15g7 antibody (ENZO cat. no. ALX-804-258-l001, Farmingdale, NY). For phosphorylation studies, the following antibodies were used: a monoclonal anti-a-syn P87 antibody (gift from Dr. Lashuel), rabbit polyclonal anti-a-syn P87 antibody (Santa Cruz Biotechnology cat. no. SC-19893-R, Dallas, TX), rabbit monoclonal anti-a-syn P129 antibody (ABCAM cat. no. 168381, Cambridge, MA). To determine if ~36-kDa species of a-syn is cross-linked, we used mouse anti-3-nitrotyrosine (Santa Cruz cat. no. sc32757). Antibodies used for quality control are as follows: Mouse monoclonal anti-b-Actin (Sigma, St. Louis, MO), goat polyclonal anti-LDH (Fitzgerald Industries, Acton, MA), rabbit polyclonal anti-EGFR (D38B1) XP (Cell Signaling, Danvers, MA), goat polyclonal anti-lamin B (Santa Cruz Biotechnology, Dallas, TX), rabbit polyclonal anti-Histone H3 (D1H2) XP (Cell Signaling, Danvers, MA), rabbit polyclonal anti-vimentin (D21H3) XP (Cell Signaling, Danvers, MA), rabbit polyclonal anti-calreticulin (Abcam, Cambridge, MA), rabbit polyclonal anti-cytochrome C (Santa Cruz Biotechnology, Dallas, TX), and goat polyclonal anti-Tim23 (Santa Cruz Biotechnology, Dallas, TX).

Treatments

SH-SY5Y cells were cultured to ~85–90 % confluence on the day of treatment, divided into four treatment groups, and subjected to the appropriate treatments for 6 h. Treatments were performed in culture without FBS. Group 1 was treated with treatment media only, group 2 was treated with 100 μM H2O2 in treatment media, and group 3 was treated with 100 μM l-Dopa in treatment media. After 6 h, cells were washed with phosphate-buffered saline (PBS), trypsinized, and collected into 15 ml conical tubes containing cell culture medium to neutralize the trypsin. Cells were harvested by centrifugation at 500×g for 5 min at 4 °C and washed with 5 ml of PBS and pelleted again at 500×g at 4 °C. Cells were resuspended in 1 ml of PBS, transferred to a 1.5-ml pre-chilled centrifuge tube, and centrifuged at 500×g for 5 min at 4 °C. The cell pellets were saved for subcellular fractionation or whole cell lysate preparations.

Whole Cell Lysate and Subcellular Fractionation of Cultured Cells

Whole cell lysates were prepared using Radio Immuno-Precipitation Assay Buffer (RIPA, 0.15 mM NaCl, 1 % (v/v) NP-40, 0.50 mM Tris, pH 8.0, 0.5 % (w/v) Sodium deoxycholate and 0.01 % (v/v) SDS). Seventy microliters of RIPA buffer, containing 1× protease inhibiters per 10 μl of cell pellet, was used to lyse the cells. Cells were vortexed for 10 s, allowed to lyse for 10 min at 4 °C, and centrifuged at 16,000×g for 10 min at 4 °C to pellet the insoluble material. The supernatant was removed and stored at −80 °C. Subcellular fractionations were obtained using the Pierce subcellular fractionation kit (Pierce, Rockford, IL).

Whole Cell Lysate and Subcellular Fractionation of Brain Tissue

Age-matched (66–86 years old), post-mortem unfixed, frozen cerebellum and frontal cortex tissue was obtained from patient with confirmed PD and healthy control individuals with no history of neurological disease (New York Brain Bank and Center for Mitochondrial Medicine and Neurogenetics, Department of Neurology, Haukeland University Hospital, Bergen, Norway). See Supplementary Table 1 for detailed information on brain samples used for this study. Tissue morphology was confirmed by routine histological techniques. Approximately 100 mg of frozen post-mortem cerebellum were placed into separate 3 ml polypropylene tubes (VWR, Radnor, PA), which contained 1.5 ml of ice-cold PBS/1× protease inhibitors (Pierce, Rockford, IL). Once thawed, the tissue was homogenized in a Polytron homogenizer (Kinematica, Bohemia, NY) for 10 s. Homogenized tissues were then put through a tissue strainer provided in the fractionation kit. Particulates in the strainer were discarded. The strained cells were transferred to pre-chilled 1.5 ml centrifuge tubes and centrifuged at 500×g at 4 °C. The cells were gently resuspended in 1.5 ml of PBS containing 1× protease inhibitors. Cells were collected by centrifugation at 500×g at 4 °C. Samples were prepared for whole cell lysates as described above or for subsequent fractionation steps using the Subcellular Fractionation Kit for Tissues (Pierce, Rockford, IL).

Protein Concentration Estimation

Protein concentrations were determined using Pierce detergent-compatible protein assay (Pierce, Rockford, IL). Approximately 20 μg of total protein extract for each fraction was used for SDS-PAGE and subsequent Western blot analysis.

Cross-linking

Cells growing at 85–90 % confluency were trypsinized, removed from the plates, and washed in ice-cold PBS. Cells were collected by centrifugation at 500×g at 4 °C, the supernatant discarded followed by addition of ice-cold PBS to a final volume of 8 ml. The cells were placed onto fresh 10 cm tissue culture dishes and subjected to 150 mJ of energy using a UV cross-linker (VWR, Radnor, PA) and collected by centrifugation. Cells were subjected to lysis for whole cell extract or for fractionation analysis as outlined above.

Phosphatase Treatment

In a 1.5-ml microcentrifuge tube, samples containing approximately 20 μg of protein were precipitated in 1 ml of ice-cold 10 % (v/v) tri-chloro-acetic acid (TCA) in 90 % (v/v) acetone containing 0.07 % (v/v) beta-mercaptoethanol. The samples were stored at −20 °C for 1 h to complete precipitation. Tubes were centrifuged at 18,000×g for 15 min to pellet. The pellets were then washed in triplicate with 100 % acetone containing 0.07 % (v/v) beta-mercaptoethanol. Pellets were air dried at −20 °C until all the acetone had evaporated. Pellets were resuspended in 200 μl (10 μl of buffer/1 μg protein) of calf intestinal alkaline phosphotase (CIP; New England Biolabs, Ipswich, MA) buffer composed of 100 mM NaCl, 50 mM Tris-HCL, 10 mM MgCl2, 1 mM dithiothreitol (DTT), and 1× protease inhibitor at pH 7.9. CIP was added at a concentration of 1 unit/1 μg of total protein. Treated samples were incubated at 37 °C for 2 h and prepared for SDS-PAGE and Western blotting analysis.

Urea/Thiourea Treatment

Samples were precipitated as described above followed by addition of 200 μl of a solution containing 7 M urea and 2 M thiourea, 4 % (w/v) CHAPS, and 30 mM Tris, at pH 8.5. Samples were resuspended and placed on a vortex overnight at room temperature (RT) and prepared for SDS-PAGE and Western blotting analysis.

SDS-PAGE and Western Blot Analysis

Five times loading buffer was added to samples to generate final concentrations of 50 mM Tris buffer pH 6.8, 4 % (w/v) SDS, 10 % (v/v) glycerol, 0.2 % (w/v) Bromophenol Blue, and 100 mM beta-mercapthoethonal in de-ionized water. Samples were boiled for 5 min, then loaded onto an SDS-polyacrylamide gel consisting of a 4 % (v/v) acrylamide stacking gel and a 12 % (v/v) acrylamide resolving gel. Samples were run at 100 V for 94 min in running buffer (192 mM glycine, 25 mM Tris base, and 2 % (w/v) SDS on a Mini-PROTEAN® system (Bio-rad, Hercules, CA). After electrophoresis, proteins were transferred to a PVDF membrane (Pall Life Sciences, Ann Arbor, MI) using a transfer buffer composed of 192 mM Glycine, 25 mM Tris base, 20 % (v/v) methanol, and 0.005 % (w/v) SDS at 200 mA for 1.5 h using a Mini Trans-blot system (Bio-rad, Hercules, CA). After the transfer, membranes were incubated in 5 % (w/v) non-fat dry milk in TBS-T (137 mM NaCl, 15.4 mM Trizma HCl, 0.1 % (v/v) Tween 20, pH 7.6) for 1 h at RT. Following blocking, membranes were incubated with the appropriate primary antibody at 1:1000 in 2.5 % (w/v) non-fat dry milk in TBS-T overnight at 4 °C. The next day, membranes were washed in triplicate with TBS-T for 10 min at RT. The membranes were then incubated with the appropriate secondary antibody at a 1:10,000 dilution in 2.5 % (w/v) non-fat dry milk in TBS-T for 2 h at RT. The membranes were then washed four times with TBS-T for 10 min at RT. The membranes were developed using ECL (Pierce, Rockford, IL) and signals were recorded using a Bio-Rad Molecular Imager Chemi Doc XRS+ imaging system (Bio-Rad, Hercules, CA).

Scanning and Analysis of the Images

Following de-staining steps, gels were scanned using the EPSON scan perfection V750 PRO software (Digital ICE technologies, Long Beach, CA). The gels were scanned in professional mode at 600 dpi and in 16-bit grayscale.

Mass Spectrometry

The protein bands of interest were manually excised from the gel and processed as described in refs. [33, 34] with some modifications. The gel slices were reduced, alkylated, and digested with trypsin. Tryptic peptides were extracted and the pooled solution was dried under vacuum and resuspended in 10 μl 0.1 % (v/v) formic acid. Three microliters of each sample were loaded onto a 75 μm × 15 cm column self-packed with 3 μm ReproSil-Pur C18-AQ beads (Dr. Maisch GmbH, Germany). Peptides were eluted with a gradient of 3–30 % acetonitrile in 0.1 % formic acid over 60 min at a flow rate of 250 nl/min at 45 °C using a Thermo Scientific EASY-nLC 1000 coupled to a Q Exactive HF mass spectrometer (Thermo Fisher Scientific, Waltham, MA). The Q Exactive was operated in data-dependent analysis mode with survey scans acquired at a resolution of 120,000. Up to the top 15, most abundant precursors from the survey scan were selected with an isolation window of 1.6 Thompsons and fragmented by higher-energy collisional dissociation with normalized collision energies of 27. The maximum ion injection times for the survey scan and the MS/MS scans were 20 and 60 ms, respectively, and the ion target value for the survey scan and the MS/MS scans was set to 3,000,000 and 1,000,000, respectively. The raw files were processed using the MaxQuant [35] computational proteomics platform (version 1.2.7.0) for peptide identification and quantitation. The fragmentation spectra were searched against the UniProt human protein database (downloaded 27 June 2014) containing 87,938 protein sequences and allowing up to two missed tryptic cleavages. Carbamidomethylation of cysteine was set as a fixed modification and oxidation of methionine and protein N-terminal acetylation were used as variable modifications for database searching. The precursor and fragment mass tolerances were set to 7 and 20 ppm, respectively. Both peptide and protein identifications were filtered at 1 % false discovery rate (FDR) and thus were not dependent on the peptide score.

A-Syn Purification

The protocol was adapted from Bartels et al.: “nondenaturing purification of α-synuclein from erythrocytes” [30]. SHSY5Y cells were trypsinized and removed from plates, washed 2× with ice-cold PBS. Cells were lysed under native conditions by sonication in column equilibration buffer (20 mM Tris buffer, pH 8.0, 25 mM NaCl) containing protease inhibitors. Lysate was spun at 18,000×g for 15 min at 4 °C. Supernatant was removed then injected onto a 5-ml HiTrap Q HP anion exchange column under native conditions on an Akta pure HPLC (GE Healthcare, Pittsburg, PA) equilibrated with 20 mM Tris buffer, pH 8.0, 25 mM NaCl, 5 mM EDTA. a-syn was eluted from the column with a 25 mM-1 M NaCl gradient in 20 mM Tris buffer, pH 8.0. Pool and concentrated fractions were injected onto a Superdex 200 (10/300 GL; GE Healthcare) that had been equilibrated with 20 mm Tris buffer containing 150 mM NaCl and 5 mM EDTA at pH 8.3 and eluted with 20 mm Tris buffer containing 150 mM NaCl and 5 mM EDTA at pH 8.3. Fractions were collected and concentrated, then run on 12 % SDS-PAGE gels as described above for Western blot analysis or stained with Blue Silver stain [36] and prepared for mass spectrometry analysis.

In Situ Immunofluorescence Cells

SH-SY5Y were seeded onto a single well chamber slide (Fisher Scientific, Waltham, MA) and grown in culture media until approximately 85–90 % confluency. Cells were then treated for 6 h according to the treatment protocols outlined above. Untreated and H2O2-treated cells were rinsed with PBS and fixed with 4 % (w/v) paraformaldehyde in PBS for 1 h at RT and then washed 2× with PBST. Cells were permeablized with acetone for 2 min at −20 °C and washed in triplicate (5 min each) with PBS. The cells were then blocked in 10 % (v/v) normal goat serum in PBS for 1 h at RT and incubated with 5 μg/ml rabbit polyclonal anti-a-syn polyclonal antibody (Abcam, Cambridge, MA) at 4 °C overnight. The cells were then washed in triplicate (5 min each) with PBS and incubated with the secondary antibody (Alexa-Flour goat anti-rabbit, Invitrogen, Carlsbad, CA) at 1:200 dilution for 2 h in the dark at RT, washed in triplicate with PBS, cover slipped with mounting medium containing DAPI, and imaged using a Nikon T2000S fluorescence microscope.

Neutral Red Assay

Cells were seeded into 96- well plate and allowed to grow to 90 % confluency. The cells were then treated with H2O2 and l-Dopa as previously described in “Treatments.” Assay was performed as described by Repetto et al. [37].

Statistical Analysis

Protein bands were analyzed using ImageJ (NIH, Bethesda, MD) software. Statistical significance was determined using the t test normalized to untreated cells.

Results

A-Syn Exists as Subcellular-Specific Species

Since a-syn is present as SDS-resistant higher molecular weight species in PD brains [38, 39], which are different from fibrillar a-syn [40], we initially analyzed a-syn in total SH-SY5Y neuroblastoma cell extracts using SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting employing four different a-syn monoclonal and polyclonal antibodies. In SH-SY5Y cells, a-syn was present, irrespective of the antibody used, as an ~17-kDa monomer in combination with higher molecular weight a-syn species ranging from ~50 to ~150 kDa in size (Fig. 1a, untreated). Interestingly, neither H2O2 nor l-Dopa treatment appeared to have any effect on the a-syn profiles in total cell extracts (Fig. 1a, H2O2 and l-Dopa) although ROS-dependent a-syn oligomerization has been reported [41].

Fig. 1
figure 1

Subcellular a-syn multimers in SH-SY5Y cells. a A-syn profiles in untreated, H2O2-treated, and l-Dopa-treated whole cell extracts (n = 6). b Quality control of the fractionation process (n = 8). c A-syn profiles in different subcellular locations in SH-SY5Y cells (n = 8). d Untreated SH-SY5Y cells showing no a-syn aggregation and SH-SY5Y cells treated with H2O2 showing a-syn aggregation (n = 3). e Untreated and urea/thiourea-treated SH-SY5Y cells showing an increase in specific a-syn species in response to urea/thiourea (arrows; n = 3). Quantification of monomeric a-syn in response to urea/thiourea treatment. f Twenty micrograms of purified human a-syn without urea treatment and after 12 h of 8 M urea treatment showing a decrease in the total amount of very high molecular weight (MW) a-syn aggregates (high MW doublet) in response to urea treatment whilst an ~36-kDa species appears stable (n = 3). Fold change is calculated from control value of 0 (for untreated cells). Negative change denotes a decrease in the amount of protein. *p < 0.05; ***p < 0.001

Full size image

We next fractionated SH-SY5Y cells using buffer-based protocols, showing <10 % cross-contamination of marker proteins between subcellular fractions (Fig. 1b). Rapid (<3 h) buffer-based fractionation was chosen instead of conventional ultra-centrifugation-based protocols, in order to minimize disruption of a-syn species during the procedure. Western blotting showed that a-syn was present as a monomeric form and as higher molecular weight species in the cytosolic, membrane, and nuclear-soluble fractions (Fig. 1c). No monomeric a-syn was detected in the chromatin-bound fraction, and no a-syn was detectable in the cytoskeletal fraction (Fig. 1c). Interestingly, the higher molecular weight a-syn species show distinctive profiles in the different subcellular fractions suggesting that a-syn forms defined multimeric species dependent on its cellular localization (Fig. 1c). The subcellular-specific profiles were not due to the fraction protocol itself, as the different species combined from each individual fraction, equalled the species observed in total cell extracts (Fig. 1c, whole cell lysate).

To ensure that the subcellular-specific, higher molecular weight a-syn species were not fibrillar a-syn, untreated and H2O2-treated SH-SY5Y cells were immune-stained for a-syn showing no aggregation in untreated cells (Fig. 1d, upper panels) compared with aggregation in H2O2-treated cells (Fig. 1d, lower panels). The inherent stability of the a-syn species was also tested by subjecting cell extracts to urea/thiourea treatment. Although lower molecular weight species of a-syn (~68, ~50, ~36 kDa) were remarkably stable in response to the treatment, monomeric a-syn showed increased abundance in response to urea/thiourea treatment (Fig. 1e). Similar stability of a-syn was also observed for human a-syn purified from Escherichia coli (Fig. 1f).

Oxidative Stress Alters Subcellular a-Syn Profiles

To investigate possible dynamic a-syn changes, we next analyzed subcellular a-syn profiles in SH-SY5Y cells in response to neurotoxic levels of H2O2 and l-Dopa. In contrast to cytosolic a-syn, which showed no significant profile changes in response to either treatment, membrane-localized a-syn showed a reduction of an ~90-kDa a-syn species in response to H2O2 and l-Dopa treatment whilst a nuclear-soluble ~70-kDa a-syn species was more abundant in response to both treatments (Fig. 2a). Quantification was performed based on equal loading of total protein from each fraction from untreated and treated cells (Fig. 2a). These data suggest that subcellular-specific higher molecular weight a-syn species are dynamic in nature dependent on the oxidative environment. We also analyzed monomeric a-syn profiles in different subcellular localizations in response to H2O2 and l-Dopa exposure and found a significant decrease in response to both treatments in the cytosol and membrane fractions whilst monomeric a-syn increased in abundance in the nuclear-soluble fractions (Fig. 2b). Included in the membrane fraction are membranes of mitochondrial origin. Mitochondria represent the main cellular site for ROS production, which by default constantly exposes mitochondrial components to oxidative damage [42]. Mitochondrial dysfunction is intimately linked to PD pathogenesis and several PD-associated proteins, such as Parkin, PINK1, and PARK13 (Htra2), are coupled with mitochondrial functions. PINK1 is the second most common genetic cause of early-onset PD after Parkin [43], and many studies have shown that the PINK1 and Parkin mutations result in mitochondrial dysfunction [44, 45]. Treatment of cells with H2O2 and l-Dopa does not significantly alter their viability as measured by neutral red assay (Fig. 2c).

Fig. 2
figure 2

Subcellular a-syn profile changes in response to oxidative stress. a A-syn profiles in the cytosol, membrane, and nuclear-soluble fractions in untreated, H2O2-treated, and l-Dopa-treated SH-SY5Y cells. Open arrow, an ~90–100-kDa a-syn species that decreases in abundance in response to both treatments. Arrowheads, an ~70-kDa a-syn species that increases in abundance in response to both treatments (n = 6). b Changes in subcellular monomeric a-syn in response to H2O2 and l-Dopa (n = 6). c There is no statistical significance in SH-SY5Y cell viability with and without treatments for each clone as measured by neutral red assay (n = 8, for each). Changes in a-syn multimeric species levels in treated fractions are calculated from similar size multimeric species in untreated fractions (n = 4 for each treatment, *p < 0.05; **p < 0.01; ***p < 0.001). A negative number indicates a decrease in abundance

Full size image

Most studies analyzing higher-order a-syn profiles have employed in vivo cross-linking arguing for the capture of a-syn oligomers in their physiological state [12, 46]. We UV cross-linked untreated, H2O2-treated, and l-Dopa-treated SH-SY5Y cells followed by fractionation into cytosolic, membrane and nuclear-soluble fractions (Fig. 3a) followed by analysis of a-syn profiles. In fractions from cross-linked SH-SY5Y cells, a-syn is present as a monomeric structure and as three major species (~50, ~68, and ~90 kDa) (Fig. 3a). Although the ~50- and ~90-kDa a-syn species are common to both in vivo cross-linked and non-cross-linked SH-SY5Y fractions, both species are significantly more abundant in the cross-linked fractions (Figs. 1c and 3a). In contrast to non-cross-linked cells (Figs. 1c and 2a), the a-syn profiles in cross-linked cells do not differ significantly between the different cellular fractions or in response to H2O2 or l-Dopa treatments (Fig. 3a). This suggests that cross-linking is trapping transient a-syn species. It has been previously shown that a-syn can be cross-linked at tyrosine residues [47], yet when whole cell lysate of SH-SY5Y cells overexpressing a-syn are probed with anti-3-tyrosine monoclonal antibodies, the multimeric species of a-syn are not found to be cross-linked under normal conditions (Fig. 3c).

Fig. 3
figure 3

A-syn profiles of UV cross-linked SHSY cells and HEK293 cells. a Untransfected SH-SHY5Y UV cross-linked fractionation profile of a-syn shows no difference between treatment groups. b QC of fractionated SH-SY5Y UV cross-inked cells. c Western blot using an anti-3-nitrotyrosine demonstrating that SHSY species found in whole cell lysate are not naturally cross-linked at tyrosine. d QC of fractionated HEK293 cells. e Total a-syn profiles in whole cell extracts and from cytosolic, membrane, nuclear-soluble, and chromatin-bound fractions from untreated, H2O2-treated, and l-Dopa-treated HEK293 cells demonstrates that a-syn profiles are cell-type specific

Full size image

To determine if the fractionation profile of a-syn observed in SH-SY5Y is representative of all cells, we fractionated HEK293 cells (Fig. 3d, e). The a-syn multimeric profile observed is somewhat similar yet there exist distinctly different multimeric species of a-syn that are unique to HEK293 cells.

A-Syn Phosphorylation Is Species- and Subcellular-Specific and Influenced by the Oxidative Environment

Phosphorylation of a-syn has implications for PD pathogenesis. In Lewy bodies, a-syn is predominantly p-S129 [18] whilst p-S87 is suggested to have a neuroprotective effect by preventing oligomerization [19]. In light of this, we analyzed p-S129 and p-S87 phosphorylation patterns of a-syn species in SH-SY5Y cells in different subcellular localizations and in response to H2O2 and l-Dopa treatments using p-S129- and p-S87-specific monoclonal antibodies. In the cytosolic and nuclear fractions from untreated cells, a predominant ~150-kDa p-S129 a-syn species was observed which was almost absent in the membrane fraction (Fig. 4a). However, in response to H2O2 and l-Dopa treatment, the ~150-kDa a-syn species in the membrane fraction was significantly phosphorylated on S129 (Fig. 4a). The ~90–100-kDa cytosolic a-syn species present in untreated, H2O2-treated, and l-Dopa-treated SH-SY5Y cells (Figs. 1c and 2a) also showed clear S129 phosphorylation in untreated cells, but phosphorylation was significantly reduced in response to H2O2 and l-Dopa treatment (Fig. 4a). Conversely, the ~70-kDa nuclear-soluble a-syn species present in untreated cells (Figs. 1c and 2a), which increases in abundance in response to H2O2 and l-Dopa treatment (Fig. 2a), showed low level of S129 phosphorylation in untreated and H2O2-treated cells but significant phosphorylation in response to l-Dopa (Fig. 4a).

Fig. 4
figure 4

Phosphorylation status of subcellular a-syn species in response to oxidative stress. a Western blots showing phosphorylation of S129 in subcellular fractions in response to H2O2 and l-Dopa treatments. Closed arrows, increase in abundance of a membrane-specific phosphorylated S129 ~150 kDa a-syn species in response to both treatments. Open arrow, decreased abundance of a cytosolic phosphorylated S129 ~90–100-kDa a-syn species in response to both treatments. Arrowhead, increased abundance of a nuclear soluble phosphorylated S129 ~70-kDa a-syn species in response to l-Dopa treatment (n = 6). b Western blots showing phosphorylation of S87 in subcellular fractions in response to H2O2 and l-Dopa treatments. Arrowhead, increased abundance of a cytosolic-specific phosphorylated S87 ~70-kDa a-syn species in response to H2O2 and l-Dopa treatment. c Monomeric a-syn is not phosphorylated at S129 nor at S87. Bottom, blots re-probed for the presence of a-syn. *p < 0.05; **p < 0.01; ***p < 0.001

Full size image

The phosphorylation profiles observed were verified by treating the different fractions with alkaline phosphatase, which abolished the p-S129 signals (Supplementary Fig. 1). Conflicting evidence suggests that S129 phosphorylation may or may not be necessary for a-syn-induced neurotoxicity [48, 49]. Our data suggest that both may be the case depending on the subcellular localization, as the cytosolic ~90–100-kDa a-syn species is not phosphorylated in response to oxidative stress whilst the membrane ~150-kDa and the nuclear-soluble ~70-kDa a-syn species are S129 phosphorylated in response to stress treatments (Fig. 4a).

We next analyzed phosphorylation of a-syn S87 in the cytosolic, membrane, and nuclear-soluble fractions from SH-SY5Y cells. In the membrane fraction, the ~150- and ~70-kDa a-syn species showed significant S87 phosphorylation (Fig. 4b). In contrast, only the ~70-kDa a-syn species was S87 phosphorylated in the nuclear-soluble fraction (Fig. 4b). S87 phosphorylation patterns in the membrane and nuclear-soluble fractions were less abundant in response to H2O2 and l-Dopa treatments whilst the cytosolic ~70-kDa a-syn species was only phosphorylated in the cytosol in response to H2O2 and l-Dopa treatments (Fig. 4b). We also analyzed S129 and S87 phosphorylation of monomeric a-syn in the cytosolic, membrane, and nuclear-soluble fractions from untreated, H2O2-treated, and l-Dopa-treated SY-SY5Y cells showing no evidence of phosphorylation [50] (Fig. 4c). These results suggest that S129 and S87 phosphorylation events are both a-syn-species and location specific and that phosphorylation patterns change dependent on the oxidative environment.

Neurodegeneration Results in Distinct PD-Enriched a-Syn Species

To determine whether human brain tissue also harbor defined subcellular-specific a-syn profiles, we fractionated post-mortem tissue from the cerebellum of a control subject and from an age-matched patient with pathologically confirmed PD (Fig. 5a–g). The ~90–100-kDa a-syn species was present in the cytosolic fraction from both subjects (Fig. 5b). By contrast, the PD subject harbored a significant ~36-kDa a-syn species which was far less abundant in the control subject (Fig. 5b). Furthermore, the control subject harbored an ~70-kDa a-syn species which was absent in the PD subject (Fig. 5b). In the membrane fraction, from the same cerebral brain tissue, the PD subject harbored an ~50- and ~70-kDa a-syn species which is absent in the control subject (Fig. 5c). Equal protein loading was verified using cytosolic (LDH) and membrane (EGFR) marker antibodies (Fig. 5b, c). We also analyzed monomeric a-syn profiles in the cytosol, membrane, and nuclear-soluble factions from both subjects showing no significant differences (Fig. 5d). The finding that a-syn profiles differ in brain tissue from a healthy individual as compared with an individual with PD suggests that some a-syn species may play a role in PD pathogenesis. Furthermore, it is possible that the cytosolic PD-enriched ~36 kDa and the membrane ~70- and ~50-kDa species (Fig. 5b, c) represent neurotoxic drivers.

Fig. 5
figure 5

Subcellular-specific a-syn profile changes in human brain tissue. Total of n = 4 PD positive brains (PD1–PD4) were fractionated and n = 4 PD-negative brains (N-PD1–N-PD4) were fractionated. Ages range from 66- to 86-year-old individuals. a Quality control of fractionations of human cerebellum and frontal Cortex. b Cytosolic a-syn profiles in cerebellum from non-PD and PD subjects. Arrowheads, ~36- and ~70-kDa a-syn species in PD and non-PD cerebellum, respectively. LDH was used as a loading control. c Membrane a-syn profiles in cerebellum from non-PD and PD subjects. Arrowheads, ~50- and ~70-kDa a-syn species in PD cerebellum. EGFR was used as a loading control. d Monomeric a-syn in non-PD and PD cerebellum. e S129 phosphorylation of a-syn in non-PD and PD cerebellum. f S87 phosphorylation of a-syn in non-PD and PD cerebellum. g Representative Western blots of additional fractionated human frontal cortex brain samples demonstrating the enrichment of the cytosolic ~36-kDa a-syn species in PD samples as compared with non-PD samples

Full size image

As we observed both S129 and S87 phosphorylation profile differences in SH-SY5Y cells in response to oxidative stress, we analyzed phosphorylation events in the cytosolic and membrane fractions from both subjects. We found no differences in S129 or S87 phosphorylation profiles between the control and PD subjects in either the cytosolic or membrane fractions (Fig. 5e, f). Indeed, the same ~36-kDa a-syn species was phosphorylated on both S129 and S87 in both fractions from both subjects. Most S129 and S87 phosphorylation studies have been performed in animal models showing conflicting evidence [48, 49, 51]. This together with our phosphorylation data in SH-SY5Y cells and in human brain tissue suggests that a-syn phosphorylation patterns are cell-type specific. Additional brain samples (frontal cortex) were fractionated similarly showing the ~36-kDa speices of a-syn is present in the cytosolic fraction of PD-patient brain samples but not as abundant in the non-PD controls (Fig. 5g).

The ~36-kDa PD-Enriched a-Syn Species Is Recapitulated in PD Cell Models

To test whether we could recapitulate, in cultured cells, the PD-enriched a-syn profiles observed in human brains, we overexpressed wild-type human a-syn (wt-a-syn) in SH-SY5Y cells mimicking a pathogenic a-syn gene multiplication event. As compared with empty vector control transfected cells, wt-a-syn overexpression in SH-SY5Y cells resulted in the appearance of the ~36-kDa a-syn species in whole cell extracts (Fig. 6a). Importantly, the ~36-kDa a-syn species was specific to the cytosolic fraction mirroring the situation observed in the PD brain cytosolic fraction (Fig. 5b). Furthermore, we found that the a-syn monomer was phosphorylated in whole cell extracts at S129 in SH-SY5Y cells over expressing wt-a-syn (Fig. 6a).

Fig. 6
figure 6

Transient overexpression of a-syn in pCDNA3 expression vector containing wtSNCA transfected into SH-SY5Y cells recapitulates the ~36-kDa a-syn species enriched in human PD brain. a Whole cell lysate of SH-SY5Y cells overexpressing SNCA (SH-SY5Y/wt-a-syn) showing the ~36-kDa a-syn species (arrowhead), increased monomeric a-syn, and S129 phosphorylated a-syn as compared with pcDNA3-control-transfected cells. b Cytosolic fraction of SH-SY5Y/wt-a-syn cells showing the ~36-kDa a-syn species in overexpressing cells as compared with pcDNA3-control-transfected cells. c Membrane fraction of SH-SY5Y/wt-a-syn showing lack of the ~36-kDa a-syn species. d Nuclear soluble fraction of SH-SY5Y/wt-a-syn showing lack of the ~36-kDa a-syn species

Full size image

To further validate these findings, we also overexpressed the clinical A53T a-syn variant in SH-SY5Y cells followed by subcellular fractionation. As for overexpressed wt-a-syn, the A53T a-syn variant resulted in the presence of a cytosolic-specific ~36-kDa a-syn species (Fig. 7).

Fig. 7
figure 7

Transient overexpression of the clinical A53T a-syn variant in pCDNA3 expression vector transfected into SH-SY5Y cells recapitulates the ~36-kDa a-syn species found in human PD brain. a Whole cell lysate of SH-SY5Y cells overexpressing A53T a-syn showing the ~36-kDa a-syn species compared with the pcDNA3-control-transfected cells. b Cytosolic fraction of SH-SY5Y/A53T a-syn cells showing the ~36-kDa a-syn species in overexpressing cells as compared with pcDNA3-control-transfected cells. c Membrane fraction of SH-SY5Y/A53T a-syn showing lack of the ~36-kDa a-syn species. d Nuclear soluble fraction of SH-SY5Y/A53T a-syn showing lack of the ~36-kDa a-syn species

Full size image

Purification and Identification of the ~36-kDa a-Syn Species

To ensure that the ~36-kDa a-syn species, detected by Western blotting, was indeed a-syn, we purified this species by HPLC from SH-SY5Y under native conditions from cells overexpressing wt-a-syn (Fig. 8). Following previous HPLC-based methods with slight modifications [30], anion exchange chromatography resulted in the enrichment of the ~36-kDa a-syn species and monomeric a-syn (Fig. 8a; peak 1—fractions 2–5). Fractions 2–5 were pooled and concentrated and subjected to size exclusion chromatography resulting in sufficiently pure monomeric a-syn and ~36-kDa a-syn for mass spectrometry analysis (Fig. 8b, c,; arrowheads 1 and 2). Mass spectrometry verified the identity of monomeric a-syn (Fig. 8b, c; arrowhead 1), the ~36-kDa a-syn species (Fig. 8b, c; arrowhead 2), and a smaller ~30-kDa a-syn species, possibly a degradation product of the ~36-kDa species (Fig. 8b, c; arrowhead 3).

Fig. 8
figure 8

Purification and mass spectrometry analysis of the ~36-kDa a-syn species from stable SH-SY5Y cell line overexpressing wt-a-syn in pCDNA 4 expression vector. a Cell lysis with RIPA and sonication under native conditions yield similar results (left). Whole cell lysate of SH-SY5Y/wt-a-syn purified by HPLC anion exchange and a-syn species detected by Western blotting. Peak 1 (fractions 2–5) was subjected to HPLC size exclusion chromatography. b Enrichment of monomeric a-syn (arrow 1), the ~36-kDa a-syn species (arrow 2), and an ~30-kDa a-syn species by size exclusion chromatography shown by Coomassie stain and Western blotting. Arrows 4 and 5 represent monomeric a-syn and the ~36-kDa a-syn species from purified a-syn, respectively. c Table showing the identity of the a-syn species by mass spectrometry. PEP posterior error probability

Full size image

Discussion

Although a-syn has been studied for almost two decades, it is poorly understood how this small protein shifts from a monomeric form to higher-order species as neurons transition towards a pathological state in neurodegeneration [7]. Similarly, it is unclear whether neurons possess PD-enriched a-syn species and whether some of these influence neurodegeneration. Our data shows that a-syn exists physiologically as defined, subcellular-specific SDS-resistant species that change characteristics in response to oxidative stress and neurodegeneration and we identify a cytosolic PD-enriched ~36-kDa a-syn species.

The biology of a-syn has long been controversial. Most studies have focused on monomeric a-syn and species larger than ~50 kDa have traditionally been labeled as a-syn aggregates [15, 16]. Interestingly, recent data has suggested that a-syn exists physiologically as a stable ~68-kDa helical tetramer arguing that higher molecular weight a-syn species may play a role in healthy neurons [12, 15]. However, conflicting evidence has proposed that a-syn predominantly exists as a monomer in neurons [16]. The discrepancy between these findings may be because a-syn characteristics are both cell type and tissue specific [7]. Indeed, we find that subcellular a-syn profiles in HEK293 cells differ substantially from both SH-SY5Y neuroblastoma cells and brain tissue (Figs. 1, 3, and 5).

The subcellular distribution of total a-syn has been reported in both cell culture and in post-mortem PD brains, however, the distribution of different a-syn species across subcellular locations has not been reported [52, 53]. Our data show that a-syn profiles are unique to different subcellular regions where the cytosol, the membrane fraction and the nuclear fraction all contain monomeric a-syn but also distinct SDS-resistant higher molecular weight species, ranging from ~50 to ~150 kDa in size (Fig. 1c). Although urea treatment results in a marginal increase in monomeric a-syn, in addition to ~40-, ~50-, and ~70-kDa a-syn species, the overall resistance towards 7 M/2 M urea/thiourea suggests that the different a-syn species are either alternate forms of monomeric a-syn or a-syn covalently coupled to itself or other proteins (Fig. 1e). Paramagnetic relaxation enhancement NMR spectroscopy has demonstrated that a-syn harbors long-range intramolecular interactions, which may explain the lower molecular weight species [54]. Further, small-angle X-ray scattering techniques have shown that the native a-syn populations consist of multiple conformers, including SDS- and 6 M guanidine hydrochloride-resistant a-syn dimers and trimmers [55], suggesting covalent linkages between individual monomeric units.

Oxidative stress can induce a-syn oligomerization through ROS-mediated protein oxidation [41]. Indeed, exogenous H2O2 seems to effect a-syn aggregation to a larger extent than endogenously secreted H2O2 [27]. The finding that monomeric and higher molecular weight a-syn species change in abundance in response to H2O2 and ROS-inducing l-Dopa levels (Fig. 2a, b), suggests oxidative stress-induced dynamic behavior within the overall a-syn species profile. Studies have demonstrated that oxidative stress increases nuclear a-syn accumulation [56] and our data suggests that this observed accumulation is enriched in monomeric a-syn and the ~68-kDa a-syn species (Fig. 2a, b).

The propensity of a-syn to aggregate appears to also depend on phosphorylation status as serine-129 phosphorylated a-syn is highly enriched within Lewy bodies [18] and serine-87 phosphorylation is proposed to reduce a-syn aggregation [50]. Our findings that the nuclear ~68-kDa a-syn species and the membrane-associated ~150-kDa a-syn species show increased serine-129 phosphorylation in response to stress (Fig. 4a) suggests that these two species may represent aggregation drivers. We also show that the cytosolic ~68-kDa a-syn species becomes phosphorylated at serine-87 in response to oxidative stress suggesting that this cytosolic species may be neuroprotective in decreasing the overall aggregation potential of the a-syn pool [50]. This is in agreement with data indicating that a possible tetramer may represent a stable neuroprotective a-syn species [15].

Little is known regarding subcellular a-syn species in human brain and how the neuronal a-syn pool transitions towards neurotoxic species in neurodegeneration. By extending our findings in SH-SY5Y cells, we show that post-mortem PD brain cerebellum contain an ~68- and ~50-kDa a-syn species, associated with membranes, which are absent in age- and gender-matched healthy cerebral tissue (Fig. 5c). We further identify a cytosolic, PD-enriched ~36-kDa a-syn species in human brain cerebellum (Fig. 5b) that can be recapitulated in SH-SY5Y cells overexpressing human wt-a-syn and the clinical A53T variant (Figs. 6 and 7). This cytosolic PD-enriched ~36-kDa a-syn species was purified and subjected to mass spectrometry showing the presence of a-syn peptides (Fig. 8c).

According to the Braak hypothesis, the pathogenic process begins at or near the brainstem and progresses upwards to other regions of the brain [57, 58]. Examining tissue from the cerebellum region of the brain presents an opportunity to study a-syn behavior in the terminal stages of PD pathogenesis in a region near to where the pathogenic process began. In addition, analysis of brain tissue from the frontal cortex demonstrates that both regions of the brain show the similar ~36-kDa-enriched a-syn species suggesting that this species may play a widespread role in PD pathogenesis.

Whether this PD-enriched ~36-kDa a-syn species represents a conformationally extended monomeric form or a dimeric structure remains unknown. Despite this, our data suggests the presence of pathologically relevant a-syn species and it is possible that this ~36-kDa a-syn species modulates neurotoxicity or represents a conformational intermediate in the a-syn aggregation pathway. If the ~36-kDa a-syn species indeed modulates neurotoxicity, it would be reasonable to assume that this a-syn variant may act as a seed for further aggregation into higher molecular weight species in neurons. Conversely, the ~36-kDa a-syn species may represent a neuroprotective variant, accumulating as a response to oxidative stress and/or neuronal damage, and through this accumulation minimizing further aggregation into higher molecular weight species.

This study demonstrates that a-syn exists as multiple subcellular-specific species within neurons that change profile characteristics in response to oxidative stress and neurodegeneration. The study further identifies PD-enriched a-syn species in human brain tissue. Although only two brain regions from PD patients were used, we do see similar a-syn profiles in both the cerebellum and frontal cortex. Further studies utilizing multiple brain regions would be necessary to determine if the reported a-syn profile is also observed in other brain regions. These findings lay a solid foundation to further dissect their possible neuroprotective and/or neurotoxic nature and the targeted manipulation of these PD enriched a-syn species may aid in the design of intervention strategies for individuals suffering from PD and other synucleinopathies.