Alpha-synuclein inside cell nuclei is found in the human brain and changes in dementia with Lewy bodies

Post-mortem human brain tissue from clinico-pathologically confirmed cases of DLB (n = 22) and non-neurodegenerative diseases controls (n = 24) was obtained from the Newcastle Brain Tissue Resource. For histology, cingulate or lateral temporal cortex sections were prepared from the right hemisphere of the brain (4% paraformaldehyde immersion fixed, 6-weeks). Frozen temporal grey matter (Broadman area 21/22, middle and superior temporal gyrus, 500 mg) for subcellular fractionation, biochemical analysis, and MS was obtained from the contralateral left hemisphere of the brain, (snap frozen between copper plates at − 120 °C).

Forebrain samples from three month-old aSyn knockout mice, generated as previously outlined [57], were snap frozen in liquid nitrogen, and stored at − 80 °C prior to use.

Quantification of nuclear pS129, was performed using a Lecia SP8 confocal microscope and slides imaged via 63 × oil immersion lens (n = 12 DLB and 12 controls). For each case, 5 images from layers V-VI of the lateral temporal cortex were selected at random and sampled via Z-Stack imaging (0.3 µm step thickness). For analysis, Z-stacks of 21 images were summated using Z-project maximum intensity function of ImageJ (NIH), DAPI fluorescence was used to manually trace individual nuclei as regions of interest (ROI), which were grouped according to NeuN reactivity enabling the quantification of pS129 per ROI as per NeuN +ve and NeuN −ve nuclei. Mean values of fluorescence intensity, without threshold application, were calculated per image and then further averaged to provide a mean value of pS129 immunoreactivity for NeuN +ve and NeuN −ve cells per case, before being pooled according to disease (e.g. Control Cf. DLB). Comparisons between control and DLB cases were made via Mann–Whitney tests (GraphPad Prism, Ver 5) and correlative analysis between NeuN +ve and NeuN −ve within individuals performed via Spearman’s correlation (Prism), p < 0.05 was deemed significant. Additionally, frequency plots of pS129 immunoreactivity of all NeuN +ve and NeuN −ve nuclei were generated for control and DLB cases as a visual representation of nuclear changes at a cellular population level, using Prism.

Purified nuclear extracts were generated as per previously reported method [40]. Tissue blocks (250 mg) and individual cerebral hemispheres from aSyn knockout mice were lysed 1:16 (W:V) in nuclear extraction buffer (0.32 mM sucrose, 5 mM CaCl_2, 3 mM Mg(Ac)₂ 10 mM Tris–HCl, 0.1% NP-40, pH 8, containing cOmplete protease inhibitor cocktail and phostop tablets, 1/10 ml, Sigma) via dounce homogenisation to generate crude whole tissue lysates. Cytoplasmic fractions were obtained from the centrifugation (800 rcf × 40 min, 4 °C) of crude lysates (500 µl) and retention of the resulting supernatant. Nuclei were isolated from the remaining lysates via sucrose gradient centrifugation (1.8 M Sucrose, 3 mM Mg(Ac)₂, 10 mM Tris–HCl, pH 8, 107,000 rcf × 2.5 h, 4 °C). Pelleted nuclei were washed twice in PBS and resuspended in 500 µl 0.01 M phosphate buffer saline. Fractions were directly applied to microscope slides for immunochemical staining (as above) or frozen at − 80 °C.

Nuclear fractions, pre-treated with DNAse (50u/ml, 15 min RT, RQ1 DNAse, Promega) and cytoplasmic fractions, were adjusted to 80 ng/µl with LDS sample buffer (Fisher Scientific), reducing agent (Fisher Scientific) and dH₂O and heated at 70 °C for 10 min. Samples (4 µg/lane for phospho-aSyn blots and 1.2 µg/lane for all other blots) were separated via SDS page electrophoresis (200 V, 35 min in MES buffer) and transferred to nitrocellulose membranes (0.2 µm) via Iblot (7 min, 20 mV; Fisher Scientific). Initial blots were optimised for aSyn retention by post-transfer heat treatment in 10 mM phosphate buffered saline (PBS; microwaved, 800 W, 5 min in boiling PBS) or via 30 min chemical fixation (4% paraformaldehyde, 30 min, RT). All subsequent aSyn blots were pre-treated via microwave-heat assisted protein retention (800 W, 5 min in boiling in 10 mM PBS), prior to being blocked in 5% milk powder containing Tris buffered saline with Tween-20 (TBST; 5 mM Tris, 145 mM NaCl, 0.01% Tween-20, pH 7.4, 1 h, RT). Membranes were stained for aSyn (pS129-aSyn, EP1536Y, 1:1000, Syn-1, 1:5000 and C-terminal directed MJRF1 [aa119-123, Cat# ab138501, Abcam], 1:5000) in TBST with 0.05% sodium azide overnight at 4 °C and labelled with secondary antibodies (goat anti-mouse HRP or goat anti-rabbit-HRP, 1:5000, in 5% milk powder containing TBST, 1 h, RT). Between incubation steps, blots were washed 3 × 5 min in TBST). Immunoreactivity was visualised via enhanced chemiluminescence (1.25 mM Luminol, 30 μM coumaric acid and 0.015% H₂O₂), images captured with a Fuji LAS 4000 with imaging software (Fuji LAS Image, Raytek, Sheffield, UK). Protein loading and fractionation purity was established via reprobing membranes with the cytoplasmic marker GAPDH (1:5,000, 14C10, Cell signalling) and the nuclear marker Histone H3 (1:5000). Western blots were quantified via area under the curve of immunoreactive bands (AUC; ImageJ) adjusted to the AUC of corresponding loading controls. Samples were processed in batches, each blot contained ≥ 3 control and DLB cases, allowing normalisation of measurements to control values within blot prior to pooling between blots. Comparisons were made via Mann–Whitney tests, p < 0.05 was deemed significant.

To maximise the sensitivity/detectability of aSyn in the sub cellular fractions, a peptide level fractionation approach was used. A pooled sample (combined for controls and DLB cases) for each subcellular fraction was created by taking the volume for 20 µg of peptide solution from each sample, adding 20 µg of each sample to the same microcentrifuge tube. Peptides were dried by centrifugal evaporation. The near dried peptide pellet was dissolved in 50 µl, 20 mM ammonium formate, pH 8.0, sonicated for 5 min in a sonicating water bath then the solution was clarified by centrifugation at 13,000×g for 10 min at room temperature. The supernatant was transferred to a liquid chromatography (LC) vial with 200 µl insert. Alkaline reverse phase LC was carried out on a Thermo Ultimate 3000 UHPLC, with a 25 cm Phenomenex C18 column, 5 µM particle size, 100 Å pore size. The buffers used were A: ammonium formate, pH 8 and B: 100% (v/v) acetonitrile. 50 µl of peptide solution was injected and separated over a 40 min linear gradient from 2% buffer B to 40% buffer B at a flow rate of 300 µl/min. Fractions were collected every 1 min, with fraction 1 and 13, 2 and 14, 3 and 15 etc. collected into the same vial. This created 12 fractions. These samples were dried in a centrifugal evaporator to remove acetonitrile and ammonium formate. The dried peptides were dissolved in 15 µl of 3% Acetonitrile, 0.5% TFA, sonicated for 5 min in a sonicating water bath then the solution was clarified by centrifugation at 13,000 xg for 10 min at room temperature then transferred to an LCMS vial. All 12 fractions were acquired with the same LCMS parameters as below with ~ 1 µg of peptides loaded on column per fraction.

For the determination of the abundance of aSyn and subcellular compartment purity controls, samples from controls and DLB were processed separately (giving 6 nuclear and 6 cytoplasmic subcellular fractions, from the 3 controls and 3 DLB cases) as per the s-trap method, as described in the previous section. On elution of peptides from the S-trap cartridge, samples were frozen and dried with centrifugal evaporator until at a volume of ~ 1 µl. The sample was dissolved in 0.1% Formic acid at a concentration of 1 µg/µl.

Peptide samples were loaded with 1 µg per LCMS run, peptides separated with a 125 min nonlinear gradient (3–40% B, 0.1% formic acid (Line A) and 80% acetonitrile, 0.1% formic acid (Line B)) using an UltiMate 3000 RSLCnano HPLC. Samples were loaded onto a 300 μm × 5 mm C18 PepMap C18 trap cartridge in 0.1% formic acid at 10 µl/min for 3 min and further separated on a 75μmx50cm C18 column (Thermo EasySpray -C18 2 µm) with integrated emitter at 250 nl/min. The eluent was directed into a Thermo Orbitrap Exploris 480 mass spectrometer through the EasySpray source at a temperature of 320 °C, spray voltage 1900 V. The total LCMS run time was 150 min. Orbitrap full scan resolution was 60,000, RF lens 50%, Normalised ACG Target 300%. Precursors for MSMS were selected via a top 20 method. MIPS set to peptide, Intensity threshold 5.0 e3, charge state 2–7 and dynamic exclusion after 1 times for 35 s 10 ppm mass tolerance. ddMS2 scans were performed at 15,000 resolution, HCD collision energy 27%, first mass 110 m/z, ACG Target Standard.

The acquired data was searched against the human protein sequence database, available from (https://www.uniprot.org/uniprot/?query=proteome:UP000005640↗), concatenated to the Common Repository for Adventitious Proteins v.2012.01.01 (cRAP, ftp://ftp.thegpm.org/fasta/cRAP↗), using MaxQuant v1.6.43. Fractions were assigned as appropriate. Parameters used: cysteine alkylation: iodoacetamide, digestion enzyme: trypsin, Parent Mass Error of 5 ppm, fragment mass error of 10 ppm. The confidence cut-off representative to FDR < 0.01 was applied to the search result file. Data processing was performed in Perseus [66]. All the common contaminants, reversed database hits and proteins quantified with less than 2 unique peptides were removed from the dataset. For quantitative proteomics, per subcellular fraction per case, intensity values for proteins were transformed to log2 and technical replicates averaged. Median was then subtracted within each sample to account for unequal loading and the width of the distribution adjusted. The dataset was filtered, keeping proteins with > 2 valid values in at least one experimental group. Remaining missing values were imputed from the left tail of the normal distribution (2StDev away from the mean, ± 0.3 StDev). These compositional values were used for principle component analysis of the two fractions and were used to demonstrate relative abundance of selected proteins known to be enriched with various cellular compartments as per The Universal Protein Resource Knowledgebase (UniProtKB [11]).

aSyn^Nuc was further confirmed in a larger number of cases, following the boiling of membranes in PBS and aSyn detection via pan-Syn1 antibody (Fig. 3c). In each sample the purity of the nuclear and cytoplasmic fractions was determined via Histone H3 and GAPDH controls (Fig. 3c). The levels of monomeric aSyn^Nuc were not significantly different between control and DLBs (Fig. 3d, p > 0.05), but were ~ tenfold lower than those in the cytoplasmic fractions, independent of disease status (Fig. 3c, 0.09 ± 0.04 cf. 0.11 ± 0.03 in controls cf. DLB respectively, p > 0.05). Longer exposures of Syn-1 blots again revealed the occurrence of higher molecular weight aSyn species of ~ 28 kDa, 42–45 kDa and 54–56 kDa consistent with the presence of nuclear aSyn oligomers in DLB cases (Fig. 3c, arrows). Similar results were also obtained using MJRF1 pan-aSyn antibody (Additional file 4: Fig. S5). The comparable levels of aSyn^Nuc monomer in nuclear fractions derived from control and DLB cases, argues against the detection of aSyn^Nuc as a consequence of the erroneous enrichment of aggregated aSyn due to fractionation. Both monomeric and oligomeric aSyn^Nuc were pS129 positive and the levels of pS129 reactive monomers were elevated in DLB cases compared to controls (14.3 ± 3.2-fold increase, p < 0.01, Fig. 2d). Critically, no immunoreactivity towards Syn-1, MJFR1 or pS129 was detected in similarly processed tissue from aSyn knockout mice (Fig. 3 and Additional file 4: Fig. S5). To further rule out the potential of cytoplasmic contamination of the nuclear fraction, analysis of GAPDH expression across cytoplasmic and nuclear fraction was conducted. Of the 13 cases investigated, quantifiable GAPDH signals within the nuclear fraction were detected in only 6 samples (~ 46%), at ~ 300 fold lower than that of cytoplasmic GAPDH (Additional file 4: Fig. S6), thus the magnitude of cytoplasmic contamination cannot account for the presence of aSyn within the nuclear fraction.

Although we are the first to systematically characterise aSyn^Nuc in human tissue, aSyn^Nuc has been reported at endogenous levels in brain tissue from mice, primates and originally in electric rays [20, 29, 37, 41, 49, 59, 71]. Likewise, aSyn^Nuc is evident in transgenic aSyn overexpression animals [39] and cell models [21, 44, 54, 63, 72], some studies of which have implemented reporter fusion proteins to negate potential non-specific antibody cross-reactivity [21, 44]. Combined the evidence supports a conserved physiological occurrence of aSyn^Nuc in the brain. Nevertheless, given that several aSyn antibodies demonstrate non-aSyn specific cross-reactivity [26], and that many neuropathological reports have failed to observe aSyn^Nuc in human tissue [4, 8, 33, 43], this issue has remained controversial. Here, we demonstrate the specificity of the EP1536Y pS129 antibody in human brain tissue, as per the abolition of staining in the presence of a blocking peptide and that the specific antigen retrieval method employed upon fixed tissue prior to immunohistochemical protocols is a key factor in aSyn^Nuc detection. As antigen retrieval methods typically employed for the visualisation of LB pathology, namely formic acid and proteinase K [5], were suboptimal for aSyn^Nuc, the occurrence of aSyn^Nuc in the human brain may have been largely overlooked. Although, we have previously identified subtle intranuclear aSyn aggregates in DLB temporal cortex with conformational dependent antibodies KM51 and 5G4 staining as following an EDTA + formic acid pre-treatment [54] in line with a modified protocol as per BrainNet Europe guidelines [1]. Accordingly, aSyn^Nuc may be detectable with modest adaptation of existing diagnostic protocols. In relation to the biochemical detection of aSyn^Nuc, the low levels of endogenous aSyn^Nuc in human brain tissue, may have previously precluded immunoblot quantification of aSyn^Nuc. The loss of low molecular weight proteins, including aSyn, from transfer membranes can lead to target proteins falling below detectable concentrations without post-transfer fixation [32]. Accordingly, we evaluated a paraformaldehyde fixation method, previously evaluated for increased aSyn retention [32] as well as boiling membranes in PBS, a method more commonly used for β-amyloid detection [30]. As optimal conditions for the membrane fixation for any protein are likely dependent on its chemical composition [69], it should be noted that heating in PBS was more favourable for optimising aSyn^Nuc whilst paraformaldehyde was more favourable for monomeric cytoplasmic aSyn, suggesting a difference in chemical composition between cytoplasmic and nuclear pools of aSyn. Independent from antigen binding and retention, aSyn was detected in nuclear enriched fractions of human brain tissue, in all cases investigated, by label free MS. Despite the nuclear preparation enviably containing non-nuclear contamination, comparison of prominent cytoplasmic, mitochondrial and membrane proteins, clearly demonstrated a much-reduced presence within the nuclear fraction compared to the abundance of aSyn within this nuclear fraction. Thus, aSyn detection within nuclear fractions is unlikely to be a consequence of erroneous non-nuclear compartmental enrichment. Indeed, our data not only argues against a cytoplasmic source of aSyn, but also specifically against pre-synaptic contamination, where aSyn is most prominently enriched [6] and also mitochondria [53], where aSyn has also been reported to be present.

Whilst each method used in this study have limitations, such as antibody specificity, addressed via blocking peptides as well as the absence of immunoreactivity within aSyn KO mouse tissue and the potential for fractional contamination, addressed by comparative analysis of equally abundant non-nuclear proteins, the consistent detection of aSyn^Nuc across methods collectively negates overall concerns of artificial aSyn^Nuc detection. Furthermore, the visual and fractional sub-cellular localisation of aSyn immunoreactivity, with both pan and phosphorylation specific antibodies, alongside, the apparent disease dependent alteration in aSyn^Nuc phosphorylation and oligomerisation, strongly supports its genuine occurrence.

aSyn positive nuclei were present in both immunohistochemistry and immunoblot experiments, in all cases examined, independent of disease status. Moreover, pS129 aSyn^Nuc was detected in both NeuN positive neurons and NeuN negative glial lineage cells, consistent with the low aSyn expression in astrocytes and oligodendrocytes [23, 48, 56, 64]. Such ubiquity of aSyn^Nuc implies its involvement in physiological processes. Accordingly, aSyn^Nuc is associated with genomic integrity, participating in DNA double strand break repair [59], and/or enhancing transcription of genes downstream from retinoic acid [13]. Critically, however, when overexpressed, aSyn^Nuc downregulates gene transcription as a consequence of inhibited histone acetylation [29, 52], and/or via direct interactions with gene promotors [62]. Genes associated with DNA repair [52, 54] and mitochondrial regulation [62], are among the affected transcripts which highlights the potential for dysfunctional aSyn^Nuc to engage in established neurodegenerative pathways.

The disease dependent alteration of aSyn^Nuc, comprising of hyperphosphorylated pS129 aSyn and higher molecular weight oligomeric aSyn species was evident in DLB cases. Phosphorylation [2, 19] and oligomerisation [22] of aSyn are widely associated with LB pathology and, therefore, the nuclear alteration is consistent with speculated pathological mechanisms [51]. However, comparable molecular weight banding of aSyn was not detected in the cytoplasm of matched cases of DLBs, suggesting the changes in aSyn^Nuc composition are unlikely to reflect changes in whole cell aSyn composition and instead represent novel aSyn conformers. Methodologically, it remains possible that the presence of aSyn oligomers within the nuclear fraction, may in part be due to the sedimentation of large insoluble aggregates from the brain tissue of DLB cases. However, such oligomers are typically observed within denaturing western blot preparations in the protein denaturing anionic SDS detergent or urea fractions which are known to solubilise larger aggregates [3]. Here, tissue lysis and nuclear isolation, was performed in the absence of SDS or Urea, employing only the non-protein denaturing non-anionic detergent NP-40 [27] at a very low concentration (0.1%), thus such solubilisation of larger aggregates seems unlikely. In any case, the potential for larger aggregate contamination of the nuclear fraction, does not detract from the robust detection of monomeric aSyn^Nuc in both controls and DLBs and nor the serine 129 phosphorylation of such monomers.

Both control and DLB cases demonstrated pS129 aSyn^Nuc, albeit this was ~ 14-fold higher in DLB cases in western blots and apparent in controls when quantified via immunohistochemistry, although this was elevated in DLB cases. Thus, aSyn^Nuc phosphorylation, at least at low levels, is not overtly pathological. Indeed, cellular studies imply pS129aSyn^Nuc can be regulated endogenously by Polo-like 2 kinase activity [17] and promotes its nuclear import [54, 60]. aSyn^Nuc phosphorylation may further facilitate the protein’s recruitment to DNA damage, modulate transcriptional effects and protect against cellular stress [54, 59], whilst also preventing intranuclear oligomer formation [31].

Nevertheless, an overabundance of phosphorylated aSyn, may impede the proteins interaction with DNA, with potential consequences for DNA stabilization and repair [15]. In addition to the facilitation of nuclear import, further studies point to the increased nuclear retention of pS129 aSyn [21], suggesting that an elevation such phosphorylated species, without regulation may lead to excessive intranuclear aSyn accumulation. Indeed, it is possible that physiological phosphorylation of aSyn may be required for its nuclear export. In disease conditions where phosphorylated aSyn is excessively high, such regulatory mechanisms may be overwhelmed and thus subcellular intranuclear accumulation beyond a physiological range may in turn be detrimental to nuclear homeostatic processes. The determinantal consequences of altered nuclear aSyn content is particularly evident from in-vivo studies, where enhanced aSyn^Nuc levels results in the loss of dopaminergic neurons and motor impairments [29, 35] and by the toxicity associated with the nuclear translocation of aSyn in response to oxidative stressors [20, 62, 68, 73]. It is interesting to note that the increased pS129 aSyn^Nuc as per immunohistochemical quantification is seen not in a selective subpopulation of cells but is clearly seen in a rightward (increasing) shift in distribution of pS129 aSyn^Nuc levels of control tissue. Such alteration in distribution suggests that changes in pS129 aSyn^Nuc are not limited to neurons containing LB pathology, but instead are evident in many cells of the cortex in DLB and thus potentially indicates a disruption of nuclear homeostasis throughout the cellular population. Nevertheless we did not find robust evidence of an overall accumulation of total aSyn and thus perhaps associated nuclear dysfunction may closer relate to the altered ratio between phosphorylated and non-phosphorylated aSyn^Nuc.

As a consequence of altered post-translational modification, the risk of self-aggregation for aSyn^Nuc in disease conditions is likely altered/increased [38]. The observed pS129 positive oligomeric aSyn species would support this and is consistent with our previous observations of intranuclear reactivity towards oligomer/aggregation specific aSyn antibodies in DLB cases and the formation of aSyn^Nuc oligomers in cellular overexpression systems [54]. Certainly, the nuclear environment would appear prone to aggregation, as DNA [10, 12, 24] and histones [20] accelerate aSyn aggregation. Although we did not detect overt signs of mature fibril inclusions, neuronal and oligodendrocytic intranuclear inclusions of aSyn filaments are a pathological feature of MSA, alongside aSyn cytoplasmic aggregates [34]. As part of the MSA disease process elevated pS129 aSyn^Nuc is considered as an early event and is assumed to precede the formation of mature aggregates [25, 70]. In any case, the occurrence of neuronal nuclear inclusions in the brain stem nuclei of MSA cases correlates with surviving neuron numbers, suggestive of a protective role for mature inclusions [50], which is in line with cellular toxicity being largely independent of overt aSyn aggregates in cell models [68] and that instead, at least within the nucleus, toxicity may be driven by intermediate aSyn oligomers [67].

Additional file 1: Table S1. Individual Human cases. Diagnosis (Diag), age (in years), sex, post mortem interval (PMI, in hrs) and neuropathological assessment scores for neurofibrillary tangle (NFT) Braak stage, Thal phase, Consortium to Establish a Registry for Alzheimer’s Disease (CERAD), the National Institute of Ageing – Alzheimer’s Association (NIA-AA) criteria, Lewy body (LB) Braak stage and McKeith criteria are provided. For McKeith criteria, absence of Lewy pathology (No LB), Limbic predominate and Neocortical (neoctrx) predominate are indicated. Additionally, the use of each case in western blots (WB), multi-channel fluorescence histochemistry (His) and/or mass spectrometry (MS) is also listed. N.A = not available.Additional file 2: Table S2. Fraction enrichment. Top 30 cellular component categories identified from proteomic analysis of nuclear and cytoplasmic tissue fractions as per ShinyGo database. Assigned category name, number of proteins associated genes detected out of total list are provided as well as false discovery rate (FDR).Additional file 3: Table S3. Correlative analysis of age and post-mortem delay with nuclear pS129 aSyn intensity. Table reports Spearman’s correlation (r) and associated significance of nuclear pS129 aSyn for NeuN +ve and NeuN −ve cell types when correlated with age or post-mortem delay. Data is presented for total cohort and when spilt into control and DLB cases. N.S = not significant, p > 0.05.Additional file 4: Fig. S1. Nuclear antigen accessibility following antigen retrieval methods. Representative images captured via a 40 × objective lens from cingulate cortex from control cases stained for Histone H3 with DAPI nuclear stained. Comparison of staining with citrate, formic acid and EDTA + formic acid clearly demonstrates optimum nuclear labelling following a combined treatment of EDTA and formic acid. Scale bar = 10 µm. Fig. S2. Comparison of nuclear aSyn detection under different antigen retrieval method. Example micrograph images (× 40) from dementia with Lewy body (DLB) cases of a) phospho-serine 129 positive aSyn (pS129) mouse IgG2 immunoreactivity. Sections were pre-treated prior to antibody staining with Citrate buffer (a), Formic acid (b), Protienase K (c) and EDTA + Formic acid (d) methods of antigen retrieval. Expanded area inserts (ii, dotted line boxes in i) are shown for both pS129 alone and in combination with DAPI nuclear stain, where the nuclear outline is highlighted (dotted outline). Note robust detection of punctate intranuclear staining of pS129 following EDTA + Formic acid-based antigen retrieval. Scale bar in i = 10 µm and ii = 5 µm. Fig. S3. Specificity of EP1536Y pS129 antibody immunoreactivity. Demonstration of control temporal cortex sections, stained either with EP1526Y phospho-antibody, with or without blocking peptide or with no primary, following EDTA + formic acid antigen retrieval pre-treatment, all sections stained with secondary antibody and nuclei co-labelled with DAPI. Scale bar = 10 µm. Fig. S4. Nuclear phosphorylated s129aSyn in control and Dementia with lewy body cases. Micrographs captured via 20 × objective of temporal cortex tissue from control (Con) and Dementia with Lewy body (DLB) cases. Sections stained for pS129 aSyn (EP1536Y) and neurons (NeuN) with a DAPI nuclear stain. Scale bar = 50 µm. Fig. S5. Quantification of cytoplasmic proteins within nuclear fractions. Example western blots of GAPDH immunoreactivity of cytoplasmic (C) and nuclear (N) fractionates (a) without (i) and with (ii) enhanced contrast to allow for visualisation of GAPDH within the nuclear fraction. Comparative analysis between GAPDH immunoreactivity from cytoplasmic and those sample in which GAPDH was above detection threshold (~ 46%) indicated ~ a 300 fold dilution of cytoplasmic components in the nuclear fraction (b). Fig. S6. Detection of nuclear aSyn via pan-aSyn antibody MJRF1. Example western blots of pan-aSyn MJFR1 immunoreactivity of cytoplasmic (c) and nuclear (N) fractionates from control (con) and cases of dementia with Lewy bodies (DLB) (1.8 µg/lane). Monomeric aSyn is shown under optimised exposure conditions, with large panel depicting monomeric and oligomeric aSyn species captured following overexposure of the blot. Antibody specificity was confirmed by means of similar probing of tissue fractionates generated from aSyn knockout mice (aSyn−/−). Cytoplasmic and nuclear loading controls GAPDH and Histone H3 are also shown. Note faint appearance of high molecular weight aSyn species in the nuclear fraction only in DLB cases (arrows) only in the nuclear fraction. Fig. S7. Mass spectrometry detection of aSyn in the nuclear fraction. Annotated spectra for each of the recovered peptides, b and y ions are highlighted accordingly.