Contribution of syndecans to cellular uptake and fibrillation of α-synuclein and tau

Scientific evidence suggests that α-synuclein and tau have prion-like properties and that prion-like spreading and seeding of misfolded protein aggregates constitutes a central mechanism for neurodegeneration. Heparan sulfate proteoglycans (HSPGs) in the plasma membrane support this process by attaching misfolded protein fibrils. Despite of intense studies, contribution of specific HSPGs to seeding and spreading of α-synuclein and tau has not been explored yet. Here we report that members of the syndecan family of HSPGs mediate cellular uptake of α-synuclein and tau fibrils via a lipid-raft dependent and clathrin-independent endocytic route. Among syndecans, the neuron predominant syndecan-3 exhibits the highest affinity for both α-synuclein and tau. Syndecan-mediated internalization of α-synuclein and tau depends heavily on conformation as uptake via syndecans start to dominate once fibrils are formed. Overexpression of syndecans, on the other hand, reduces cellular uptake of monomeric α-synuclein and tau, yet exerts a fibril forming effect on both proteins. Data obtained from syndecan overexpressing cellular models presents syndecans, especially the neuron predominant syndecan-3, as important mediators of seeding and spreading of α-synuclein and tau and reveal how syndecans contribute to fundamental molecular events of α-synuclein and tau pathology.


Results
contribution of SDcs to α−syn and tau uptake and fibrillation. Although HSPGs have been already acknowledged as major players in the cellular uptake of α-syn and tau fibrils, evidence on the contribution of specific SDC isoforms to cellular internalization of these misfolded proteins is still waiting to be explored 76 . In order to measure SDCs' contribution to aggregation and cellular spreading of α-syn and tau, with minimal interference with other HSPGs or caveolae-mediated endocytosis, stable SDC transfectants were created in K562 cells, a cell line with reportedly low HSPG expression, along with no detectable levels of caveolin-1, the main component of caveolae 17,[70][71][72][73]77 . The low membrane HSPG expression of K562 cells and their inability to form caveolae -required for caveolar endocytosis -makes K562 an ideal human model cell line to study the effects of SDC overexpression without the interference of other HSPGs or caveolae-mediated endocytosis 17 . Since cell surface HS has been already recognized as a major contributor to cellular binding and uptake of protein aggregates, SDC transfectants were standardized according to their HS expression (as noted in our previous study, contrary to HS, we could not detect any CS on wild-type [WT] K562 cells or any of the SDC transfectants) 17 . Thus SDC transfectants with membrane HS levels alike were selected and together with WT K562 cells, incubated with FITC-labeled α-syn or tau fibrils for 3 h. The state of fibrils added, along with the surface of fibril-treated cells was also examined with electron microscopy (Fig. 1a). Thus scanning electron microscopy revealed that the high number of fibrils on the cell surface of fibril-treated cells at the start (i.e. 10 min) of the incubation period was reduced at 3 h, suggesting the onset of intense internalization. (Previous reports on the very moderate in vitro Detected fluorescence intensities were normalized to fibril (α−syn or tau) or Trf-treated WT K562 cells as standards. The bars represent mean ± SEM of five independent experiments. Statistical significance vs standards was assessed by analysis of variance (ANOVA). *p < 0.05 vs standards; **p < 0.01 vs standards. (d) CLSM visualization of fibril (α−syn, tau) and Trf uptake. (e-g) Colocalization of α−syn or tau fibrils and SDCs. SDC transfectants treated with either of the fluorescent fibrils (α−syn or tau) or Trf for 3 h were permeabilized and treated with the respective APClabeled SDC antibody. Nuclei of cells were stained with DAPI and cellular uptake was then analyzed with CLSM. Representative images of three independent experiments are shown. Scale bar = 10 μm. (h) SDS-PAGE showing FITC-labeled α-syn and tau fibrils immunoprecipitated with either SDC3 or SDC4 antibodies from extracts of SDC3 and SDC4 transfectants. Lane 1: 0.5 ug of FITC-α-syn; Lane 2: immunoprecipitate of FITCα-syn-treated stable SDC3 transfectants; Lane 3: immunoprecipitate of untreated SDC3 transfectants; Lane 4: immunoprecipitate of FITC-α-syn-treated SDC4 transfectants; Lane 5: immunoprecipitate of untreated SDC4 transfectants; Lane 6: Molecular weight (MW) marker; Lane 7: 0.5 µg of FITC-tau; Lane 8: immunoprecipitate of FITC-tau-treated SDC3 transfectants; Lane 9: immunoprecipitate of untreated SDC3 transfectants. Lane 10: immunoprecipitate of FITC-tau-treated SDC4 transfectants; Lane 11: immunoprecipitate of untreated SDC4 transfectants. Standard protein size markers are indicated on the left. degradation of α-syn or tau fibrils in the extracellular space also suggest that internalization is the major cellular process responsible for the clearance of extracellular α-syn or tau aggregates [78][79][80][81]. Thus the intracellular fate of the fibrils was then studied with quantitative flow cytofluorometric and microscopic assays 17 . Flow cytometric measurement of uptake was conducted by adding trypan blue (dissolved at a concentration of 0.25% in ice-cold 0.1 M citrate buffer) 1 min before the analyses, thus extracellular fluorescence of surface bound fluorescent proteins was quenched, hence enabling the exact assessment of the internalized proteins 17,36,82 . The rate of classical endocytic pathways was simultaneously detected by measuring the uptake of fluorescently labeled transferrin (Trf), the marker of clathrin-mediated endocytosis 83 . As Fig. 1b,c show, SDCs -especially the neuron predominant SDC3increased the cellular uptake of fibrils, while internalization of Trf, the marker of clathrin-mediated endocytosis was reduced by SDC1-3 (and left unaffected by SDC4) suggesting that SDC mediated uptake of the fibrils occur through clathrin-independent routes. Microscopic studies revealed similar pattern as flow cytometry: namely that compared to WT K562 cells, fibril-treated SDC transfectants exhibited higher intracellular fluorescent signals (Fig. 1d). CLSM (confocal laser scanning microscopy) colocalization studies then revealed apparent intracellular colocalization of SDCs with α-syn and tau fibrils (the Mander's overlap coefficients [MOC] for SDCs with α-syn and tau exceeded 0.7, an indicator of significant colocalization [ Supplementary Fig. S1]), suggesting the common intracellular pathway SDCs and α-syn or tau fibrils follow during internalization (Fig. 1e,f). Unlike α-syn or tau, Trf -demonstrating the characteristic features clathrin-mediated endocytosis (i.e. vesicle-like intracytoplasmic structures) -exhibited very weak colocalization with any of the SDCs after 3 h of incubation (i.e. MOCs < 0.4 for all SDCs), indicating that SDC-mediated pathways occur independent of clathrin ( Fig. 1g and Supplementary  Fig. S1). Co-immunoprecipitation of fibril treated SDC transfectants also confirmed that α-syn or tau fibrils bind SDC3 and SDC4 (Fig. 1h).
HSPGs are internalized via a lipid raft-dependent, yet clathrin-and caveolin-independent pathway 17,52,84 . The flotillin family of membrane proteins (i.e. flotillin 1 [FLOT1] and flotillin 2 [FLOT2]) have been acknowledged as markers of lipid rafts 85 . The involvement of lipid rafts in the SDC-mediated internalization of α-syn and tau fibrils was justified with both microscopic and Co-IP studies ( Fig. 2a-f): CLSM studies demonstrated apparent intracellular colocalization of both flotillins with SDCs, along with α-syn and tau fibril in SDC transfectants (MOCs of FLOT1 and 2 with SDCs ~0.8), while Co-IP studies confirmed the colocalization of SDC3 and 4 with flotillins and α-syn and tau fibrils ( Fig. 2a-f). It is worth noting that in our recent affinity-based proteomics study, FLOT1 and -2 were detected in a pull-down experiment of SDC4, further supporting the findings of the CLSM colocalization studies 17 .

Effect of SDC domains on cellular internalization of α-syn and tau fibrils. The involvement of SDC
domains in the interaction with α-syn and tau fibrils were then explored in studies affecting various parts of the SDC ectodomain. Inducing undersulfation of SDCs' HS chains with sodium chlorate (NaClO 3 ), an inhibitor of protein sulfation 86 , significantly (p < 0.01) reduced on α-syn and tau fibril uptake in all cells lines, highlighting the importance of polyanionic HS chains in interaction with the proteins (Fig. 3a-d).
The HS chains of SDCs offer a binding site for a large number of macromolecular ligands. Besides HS chains, SDC4 also possess another binding site, the so-called cell-binding domain (CBD) mediating cell-to-cell attachment 17,36,68 . To study whether CBD influence the uptake of α-syn and tau fibrils, we created several structural mutants of SDC4, one of the isoform that induced significant increase in the uptake of α-syn and tau fibrils ( Fig. 1b-d). Among the SDC4 structural mutants created (Fig. 3e), the CBD.pSi4 is made of the CBD and the secretion signal sequence (Si), but lacking any HS chains 17,36,68 . HSA.pSi4 mutants on the other hand lack the CBD, but contain the HS attachment site (HSA) with the HS chains, while pSi4 mutants possess a truncated extracellular domain made of only the Si of SDC4 (Fig. 3e) 17,36,68 . All of the above mentioned SDC4 mutants, along with one coding WT SDC4, were tagged with GFP at the juxtamembrane region and expressed in K562 cells 17 . Clones with equal extent of GFP, hence SDC expression were then selected with flow cytometry and -along with WT K562 cells as controls -treated with α-syn and tau fibrils for 3 h at 37 °C. Cellular uptake and attachment were then analyzed with flow cytometry, as well as confocal microscopy. Compared to WT K562 cells, WT SDC4 transfectants exhibited the most significant (p < 0.05) increase in α-syn and tau fibril uptake (Fig. 3f,g). Mutants without HS chains (i.e. pSi4 or CBD.pSi4) did not increas α-syn and tau uptake, showing the minor involvement of the signal sequence or the CBD domain in the interactions with α-syn or tau fibrils. On the other hand, mutants with a truncated ectodomain containing only the HSA with the HS chains increased fibril internalization at an extent similar to SDC4 transfectants, thus confirming the significance of polyanionic HS chains in the interactions with α-syn or tau fibrils. Microscopic studies also showed apparent intracellular colocalization of HSA.pSi4 with both α-syn and tau fibrils (MOC = 0.76 and 0.8, respectively), proposing that α-syn and tau fibrils are internalized bound to the HS chains of SDC4 (Fig. 3h,i).
SDC3's contribution to cellular entry of α-syn and tau fibrils were also analyzed in the more complex SH-SY5Y cell line. The effect of SDC3 overexpression -in both differentiated and undifferentiated SH-SY5Y cells -was thus studied on α-syn and tau fibril uptake. SDC3 overexpression -either in differentiated or undifferentiated SH-SY5Y cells -induced increased HS expression, along with increased internalization of both fibrils (Fig. 4a,b). Colocalization studies also confirmed the common intracellular pathway SDC3 and α-syn or tau fibrils follow during cellular internalization (Fig. 4c,d), with Mander's overlap coefficients (MOCs) all (i.e. either between SDC3 and α-syn or SDC3 and tau) exceeding 0.7, indicating strong colocalization ( Supplementary Fig. S2). The SDC3-mediated uptake of α-syn or tau fibrils in SH-SY5Y cells occur through lipid rafts, as shown by the apparent colocalization of the aggregates with flotillins ( Fig. 4e-h with relevant MOCs shown in Supplementary  Fig. S2b,c). Further Co-IP experiments also confirmed the colocalization of α-syn or tau fibrils with either SDC3 or flotillins, as both α-syn or tau could be immunoprecipitated with either SDC3 or flotillin 1 or 2 ( Fig. 4i-k).  www.nature.com/scientificreports www.nature.com/scientificreports/ Conformation dependence of SDC-mediated uptake. Previous studies suggested that HSPG-mediated α-syn and tau uptake is heavily dependent on conformation, namely that fibrils enter the cells via HSPG-dependent pathways, while monomers and smaller oligomers enter the cells through other routes 87 . Thus we investigated SDCs' effect on α-syn and tau fibrillation and uptake by incubating (freshly prepared) monomeric of α-syn and tau with SDC transfectants (established in K562 cells) and monitoring the fate of the monomers with fibrillation and uptake studies. ThT fluorescence studies quantitatively showed the fibrillation triggering effect of SDCs on α-syn and tau (Fig. 5). Thus SDC3 and to a lesser extent, SDC4 and triggered fibrillation of α-syn and tau over time (Fig. 5a,b). Simultaneous scanning electron microscopy studies also revealed the aggregation inducing effect of SDCs on surface attached α-syn or tau (Fig. 5c). Scanning electron microscopy showed the growing number of fibrillar α-syn or tau assemblies on the surface of SDC overexpressing cells. (Fig. 5c). The extent of fibril formation due to SDC overexpression, especially in the case of SDC2-4 is very spectacular: 18 h after addition of α-syn and tau monomers to SDC transfectants, scanning electron microscopy revealed widespread appearance of mature fibrils (i.e. larger than 2 µm) expanding transcellularly. Electron microscopic visualization of the structure of fibrils induced by SDC3, shown in Fig. 5d, reveals that in the very brief time of 18 h the overexpression of SDC3, along with SDC4 (and in case of tau, SDC2) triggered the formation of mature fibrils, thus highlighting the importance of fibrillation triggering effects of SDCs. It is also evident, that SDCs can induce the formation of neurodegeneration related α-syn or tau fibrils overexpressed in the membrane of a non-neuronal cell line (i.e. K562).
Uptake studies showed that while at 1 h of incubation the cellular uptake of monomeric α-syn and tau were lower in the SDC overexpressing lines, at 18 h, when fibril are already formed due to SDCs, SDC overexpression also triggered higher amount of α-syn or tau uptake (Fig. 6a-f). On the other hand, SDC transfectants -except for SDC4 -internalized less Trf at both time points, referring to affected clathrin-mediated endocytosis. Confocal microscopy also showed ThT-labeled aggregates intracellularly 18 h after treatment with monomeric α-syn or tau, especially in SDC transfectants, suggesting once SDCs trigger the formation of α-syn or tau fibrils, they also mediate the intracellular translocation of the fibrils (Fig. 6h). Further colocalization studies showed minimal overlap of SDCs with monomeric α-syn or tau at 1 h (Fig. 7a,b,d). However once in fibrillar state, α-syn or tau colocalize with SDCs, as shown by the increased MOCs for SDCs with α-syn or tau at 18 h of incubation (Fig. 7e,f,h). The strong colocalization of SDCs with α-syn or tau at 18 h of incubation suggests the common pathway SDCs and fibrillar α-syn or tau enter the cells. The overlap of SDCs with Trf remained quite minimal at both (i.e. 1 and 18 h) time points (Fig. 7c,d,g,h).
Overexpressing SDC3 in SH-SY5Y cells, either differentiated or undifferentiated, induced lower uptake of α-syn or tau monomers at 1 h (Fig. 8a,c), but significantly (p < 0.05) higher at 18 h of incubation, when fibrils are already formed, as revealed by both ThT fluorescence and electron microscopic studies ( Fig. 8a-f). Thus 18 h after addition of α-syn or tau monomers, overexpressed SDC3 -either in undifferentiated or differentiated SH-SY5Y cells -triggered the fibrillation of both proteins. Electron microscopy also revealed widespread presence of mature fibrils on SDC3 overexpressing SH-SY5Y cells at 18 h ( Fig. 8c-e), while confocal microscopic studies showed higher number of ThT-labeled aggregates in SDC3 transfectants at the same time point (Fig. 8f), suggesting that while inducing fibrillation, SDC3 also facilitate the uptake of fibrillar α-syn or tau.

Discussion
The "protein-only hypothesis" proposed that small proteinaceous particles, termed prions, could be the infectious agents responsible for the transmission of spongiform encephalopathies, such as scrapie 88,89 . Despite initial skepticism, it is now widely accepted that prions are formed by a misfolded version of the cell-membrane prion protein that accumulates as aggregates 90,91 . Since other, more common neurodegenerative diseases, such as AD and PD, are similarly characterized by the accumulation of aggregated misfolded proteins, "prion-like" behavior of these pathologic aggregates has been hypothesized and investigated [92][93][94] . It is increasingly emerging that exogenous aggregates of α-syn or tau "infect" neighboring cells and propagate in experimental in vitro and in vivo settings 10 . Experimental studies have shown that the injection of tau and α-syn into animals induces neurons to form intracellular inclusions at the injection sites, from where they can also spread to distant brain regions [95][96][97][98][99][100] . These and other findings suggest that tau and other misfolded proteins have prion-like properties, and that spreading and seeding constitutes a central pathological mechanism for AD and other neurodegenerative diseases 94 . It is therefore important to understand the molecular mechanisms underlying the formation and cellular spreading of misfolded proteins 101 . Exploration of the cellular seeding and spreading phenomena of misfolded proteins in neurodegenerative diseases requires a strong understanding of protein uptake and endocytosis. In recent years, our research group has been exploring fundamental biological processes governing the SDC-mediated uptake of proteins 63,64 . First we described the SDC-mediated macropinocytic entry of peptidic drug delivery systems (DDSs) into the cells 36,63,68 . Utilizing our SDC overexpressing cellular models later we also revealed that SDCs, especially the neuron predominant SDC3, induce fibrillation and subsequent intracellular uptake of amyloid-β(1-42) 17 .  www.nature.com/scientificreports www.nature.com/scientificreports/ Considering the evidence backing up the HSPG-mediated entry of misfolded proteins, along with the increased expression of SDCs in human AD brains, we studied the uptake of α-syn and tau in our SDC cellular assays. K562 cells, a cell line expressing a minimal amount of membrane HSPGs and no caveolin-1, the major component of caveolae, are ideal cellular models to create stable SDC transfectants and assess the exact contribution of SDCs to the cellular uptake of α-syn and tau, without the interference of other HSPGs and caveolae-mediated endocytosis. As several observations suggest that the HS sulfation pattern is the same for every HS chain that a given cell synthesizes [27][28][29][30] , exploration of SDCs' interactions with tau or α-syn in a given cell type (i.e. K562 cells) helped us to study whether the SDC type would also influence interactions with the misfolded proteins.
Cellular uptake assays in our SDC overexpressing cell lines showed that SDCs mediate the uptake of already fibrillar form of α-syn or tau, while α-syn or tau monomers enter the cells primarily through SDC independent routes. Besides their contribution to the uptake of fibrils, SDCs also induce the fibrillation of α-syn and tau. Once fibrils are formed, SDCs, especially the neuron predominant SDC3, facilitate the uptake of fibrillar α-syn and tau. It has become evident that the effect of SDC3 on fibrillation and cellular translocation α-syn and tau is independent of cell type and could be demonstrated in both neuronal and non-neuronal cell lines. Thus, the amount and type of SDC expressed in the membrane mattered more, then the cell type itself. Increased number of SDCs on the cell surface offer an ideal environment for the seeding and subsequent spreading of aggregation-prone proteins. Since HS fine structure appears to reflect the cellular source of the SDC and not the SDC type 31,102 , evidence gathered in multiple cell lines confirmed the importance of the SDC3 isoform type on interactions with tau or α-syn. Namely, the observed difference in SDCs' interactions with α-syn and tau might go beyond the fine structure HS chains and should be also influenced by the other isoform specific regions of the SDC3 core protein.
The observation that overexpression of the neuron predominant SDC3 can indeed interfere with the uptake of α-syn and tau monomers, suggests that SDC3 overexpression can also support the propagation and cellular spread of fibrillar aggregates by affecting the clearance of monomeric α-syn and tau via classical endocytic routes. Attachment to SDCs (especially SDC3) therefore facilitates fibril formation as the local concentration increase of aggregation-prone α-syn and tau, the acidic molecular surrounding of HS chains and SDCs' propensity to induce oligomerization of their ligands provide ideal conditions for aggregation 47,103-108 . The observed fibril triggering effect of SDCs, especially of SDC3, suggests that SDC binding could also enhance fibril maturation, thus creating a vicious circle of seeding and spreading. The neuronal predominance of SDC3 could also account for neurons being the main cellular targets of pathological protein fibrils.
Considering our data on the striking similarities on the interaction of neurodegeneration-related misfolded proteins (α-syn, tau and amyloid-β ) with SDCs, along with clinical findings on the increased expression of SDCs in human AD brains suggest that neurodegenerative disorders, especially AD, could be considered as "syndecanopathies", namely the elevated level SDCs, especially the neuron predominant SDC3, creates favorable environment for fibrillation of aggregation-prone proteins (α-syn and tau or amyloid-β), besides facilitating their cellular translocation into neurons. As macromolecules or parasites entering the cells via SDCs can exert their biological activity intracellularly, cellular entry of pathological fibrillar assemblies via SDCs could be indeed a harmful uptake route that becomes dominant once fibrils are formed 64 .
In summary our paper reveals how SDCs, especially the neuron predominant SDC3, can mediate -regardless of cell type -the seeding and spreading of tau and α-syn. Our cellular data further supports recent clinical reports on the increased expression of SDCs in AD brains, providing fundamental biological evidence on the contribution of SDC overexpression to central molecular events in neurodegenerative disorders.

Materials and Methods
Fluorescent labeling aggregation of proteins. Recombinant tau-441 (2N4R isoform, purchased from rPeptide) was incubated for 5 days to form fibrils as described by Holmes et al. 13 . Briefly, 10 μM protein, in PBS 1 mM DTT pH 7.4, was mixed with low molecular weight heparin (0.05 mg/ml) and incubated with rotary agitation (400 rpm) at 37 °C for 5 days, while confirming fibrils formation was with Thiofavin T (ThT) assays and electron microscopy. Before use, fibrillization mixture was centrifuged and the resultant pellet resuspended in the PBS 1 mM DTT pH 7.4 without heparin to a stock solution. Formation of α-syn fibrils was induced as described by Ihse et al 87 . Briefly, α-syn (purchased from rPeptide) was dissolved in PBS to a concentration of 140 µM (~2 mg/ml), and incubated at 37 °C with rotary agitation (400 rpm) for 10 days, while monitoring fibrils formation with ThT assays and electron microscopy. After 10 days of incubation, the fibril solution was centrifuged at 20,000 × g for 30 min to separate the insoluble fibrils from smaller soluble aggregates and/or any monomers. The pellet was re-dissolved in PBS to a stock solution. Concentrations were determined by measuring absorbance using NanoDrop. Labeling of α-syn or tau fibrils was performed with the FITC or Alexa Fluor 633 labeling kits according to the manufacturer's instructions (Thermo Fisher Scientific). The calculated yield of the labelling was 3.4 mol dye/tau, and 2.2 mol dye/α-syn. For all the experiments only freshly prepared sonicated fibrils (8 pulses of 30% amplitude) or freshly prepared monomers were used. were created as described previously 17,36,68 . Differentiation protocol of SH-SY5Y cells. SH-SY5Y cells maintained at 37 °C in a humified 5% CO 2 containing air environment were seeded at an initial density of 10 4 cells/cm2 on 24 or 8 well plates in culture dishes (Corning) previously coated with 0.05 mg/ml collagen (Merck). All-trans-retinoic acid (RA, Sigma) was added the day after plating at a final concentration of 10 μM in Gibco ™ Advanced MEM (Thermo Fischer   82 . In the case of the SDC4 mutants treated with Alexa Fluor 633-labeled proteins, extracellular fluorescence of surface attached α-syn or tau was removed by trypsinization according to the method described by Nakase et al. 110 . Cellular uptake was then measured by flow cytometry using a FACScan (Becton Dickinson). A minimum of 10,000 events per sample was analyzed. Viability of cells was determined by using propidium iodide in the cell suspension (10 µg/ml) and appropriate gating in a forward-scatter-against-side-scatter plot to exclude dead cells, debris, and aggregates.
Microscopic visualization of uptake. Internalization of the fluorescently labeled (either FITC or Alexa Fluor 633) α-syn or tau fibrils or monomers, along with FITC-Trf, was visualized by confocal laser scanning microscopy (CLSM). WT SH-SY5Y and WT K562 cells, along with SDC transfectants and SDC4 mutants were grown on poly-D-lysine-coated glass-bottom 35-mm culture dishes (MatTek Corp.). After 24 h, the cells were preincubated in DMEM/F12 medium (supplemented with 10% FCS) at 37 °C for 30 min before incubation with the fluorescently labeled fibrils or monomers at a concentration of 5 μM monomer equivalent (or 25 μg/ml in case of Trf). After incubation, the cells were rinsed two times with ice-cold PBS, fixed in 4% paraformaldehyde (Sigma) and nuclei were stained with DAPI (1:5000, Sigma) for 5 min. For colocalization studies, after fixation, the cell membranes were permeabilized (1% Triton X-100), followed by 1 h of treatment with APC-labeled SDC antibodies (1:100) with or without either of the Alexa Fluor 546-labeled flotillin antibodies (flotillin-1 or 2, all Santa Cruz Biotech, cat.no. sc-74566 AF546; sc-28320 AF546). The samples were then rinsed three times with PBS containing 1% goat serum and 0.1% Triton X-100, then stained with DAPI (1:5000) for 5 min, washed three times with PBS and embedded in Fluoromount G (SouthernBiotech). Distribution of fluorescence was then analyzed on an Olympus FV1000 confocal laser scanning microscope equipped with three lasers. A laser diode (excitation, 405 nm) and a band-pass filter (420-480 nm) were used to capture the signal recorded as blue; an argon laser (excitation, 488 nm) and a bandpass filter (505-530 nm) were used to capture the signal recorded as green; and finally, a helium/neon laser (excitation, 543 nm) and a band-pass filter (550-625 nm) were used to capture the signal recorded as red. Sections presented were taken approximately at the mid-height level of the cells. Photomultiplier gain and laser power were identical within each experiment. The Olympus Fluoview software (version 4.2b) was used for image acquisition and analysis. For visualizing internalization of ThT-labeled fibrils, WT K562 cells and SDC transfectants grown on poly-D-lysine-coated glass-bottom 35-mm culture dishes were incubated with α-syn, tau at a concentration of 5 µM monomer equivalent (in DMEM/F-12 without Phenol Red) at 37 °C for 18 h, then treated with Thioflavin T (ThT, Sigma) at a concentration of 25 µM for 10 min at 37 °C and rinsed two times with ice-cold PBS. After fixation in 4% paraformaldehyde (Sigma), nuclei were stained with DAPI (1:5000) for 5 min, the after three washing with PBS, the samples were embedded in Fluoromount G and distribution of fluorescence was analyzed on an Olympus FV1000 confocal laser scanning as described above. For colocalization analyses (SDCs with α-syn, tau or Trf; SDCs with flotillins, SDC3 with α-syn, tau or flotillins), the Mander's overlap coefficient (MOC) was calculated by analyzing 21 cellular images (7 images per sample, experiments performed in triplicate) with the built-in colocalization module of Olympus Fluoview software (version 4.2b). www.nature.com/scientificreports www.nature.com/scientificreports/ Co-immunoprecipitation experiments. Stable SDC3 transfectants or WT SH-SY5Y cells were incubated with or without FITC-labeled α−syn or tau fibrils at a concentration of 5 µM monomer equivalent for 3 h at 37 °C. After incubation the cells were washed twice with ice cold PBS and treated with cold Pierce IP lysis buffer. Then the cells were scrapped off to clean Eppendorf tubes, put on a low-speed rotating shaker for 15 min and centrifuged at 14,000 g for 15 min at 4 °C. The supernatants were then transferred to new tubes and combined with 5 µg of the human SDC3 or SDC4 antibody (R&D Systems; cat.no. AF3539 and AF2918F) or 5 μg of flotillin-1 and 2 (FLOT1 and 2) antibody (Santa Cruz Biotechnology; cat.no. sc-28320, sc-74566). The antigen sample/SDC3, SDC4 antibody or antigen sample/FLOT1, FLOT2 mixture was then incubated for overnight at 4 °C with mixing. The antigen sample/SDC3, SDC4 antibody and antigen sample/FLOT1 or 2 antibody mixture then were added to a 1,5 ml microcentrifuge tube containing pre-washed Pierce Protein A/G Magnetic Beads (Thermo Fisher Scientific, cat.no. 88802) and after incubation at room temperature for 1 hour with mixing, the beads were then collected with a MagJET Separation Rack magnetic stand (Thermo Fisher Scientific) and supernatants were discarded. To elute the antigen, 100 µl of SDS-PAGE reducing sample buffer was then added to the tubes and samples were heated at 96 °C for 10 minutes in 1% SDS and the samples were proceeded to SDS-PAGE.
Thioflavin T binding assays. WT K562, SH-SY5Y cells and SDC transfectants seeded into black-sided, clear bottom 96-well microplates (Corning) at a density of 1,5 × 10 5 cells/well in 100 µl of DMEM/F-12 (without Phenol Red) were exposed to monomeric α−syn and tau at a concentration of 5 µM for various amounts of time Scanning electron microscopy of surface attachment and fibrillation. WT SH-SY5Y and WT K562 cells, along with SDC transfectants were grown on poly-D-lysine-coated glass-bottom 35-mm culture dishes. After 24 h, the cells were preincubated in DMEM/F12 medium (supplemented with 10% FCS) at 37 °C for 30 min before incubation with either α−syn or tau fibrils or monomers (at a concentration of 5 μM monomer equivalent) for various amounts of time (10 min and 3 h for the fibrils and 1, 6 and 18 h for the monomers). The cells were then rinsed two times with ice-cold PBS, then fixed in 2.5% glutaraldehyde and 0.15% alcian blue 8GX (Sigma) for 1 hour. After post-fixation in 1% OsO 4 (Sigma) for 1 hour, the samples were dehydrated in aqueous solutions of increasing ethanol concentrations, critical point dried, covered with 10 nm gold by a Quorum Q150T ES sputter and observed in a JEOL JSM-7100F/LV scanning electron microscope.

Promoting undersulfation.
To study the effect of proteoglycan sulfation on fibril uptake, cells were incubated with 60 mM sodium chlorate (NaClO 3 ; Sigma) for 48 h, the processed for the flow cytometry and microscopy studies as described above.
Statistical analysis. Results are expressed as means ± standard error of the mean (SEM). Differences between experimental groups were evaluated by using one-way analysis of variance (ANOVA). Values of p < 0.05 were accepted as significant.