α-Synuclein Oligomers Impair Neuronal Microtubule-Kinesin Interplay*

Background: α-Synuclein aggregates cause early neurite pathology by as yet unknown mechanisms. Results: α-Synuclein oligomers and seeds decrease microtubule stability, kinesin-microtubule interaction, cellular cargo distribution, and neurite network morphology. Conclusion: Various α-synuclein species interact differently with proteins of axonal transport. Significance: The impairment of the microtubule-kinesin function by α-synuclein oligomers drives early neurite pathology. Early α-synuclein (α-Syn)-induced alterations are neurite pathologies resulting in Lewy neurites. α-Syn oligomers are a toxic species in synucleinopathies and are suspected to cause neuritic pathology. To investigate how α-Syn oligomers may be linked to aberrant neurite pathology, we modeled different stages of α-Syn aggregation in vitro and investigated the interplay of α-Syn aggregates with proteins involved in axonal transport. The interaction of wild type α-Syn (WTS) and α-Syn variants (E57K, A30P, and aSyn(30–110)) with kinesin, tubulin, and the microtubule (MT)-associated proteins, MAP2 and Tau, is stronger for multimers than for monomers. WTS seeds but not α-Syn oligomers significantly and dose-dependently reduced Tau-promoted MT assembly in vitro. In contrast, MT gliding velocity across kinesin-coated surfaces was significantly decreased in the presence of α-Syn oligomers but not WTS seeds or fibrils (aSyn(30–110) multimers). In a human dopaminergic neuronal cell line, mild overexpression of the oligomerizing E57K α-Syn variant significantly impaired neurite network morphology without causing profound cell death. In accordance with these findings, MT stability, neuritic kinesin, and neuritic kinesin-dependent cargoes were significantly reduced by the presence of α-Syn oligomers. In summary, different α-Syn species act divergently on the axonal transport machinery. These findings provide new insights into α-Syn oligomer-driven neuritic pathology as one of the earliest events in synucleinopathies.

␣-Synuclein (␣-Syn) 3 is a 140-amino acid membrane-binding protein widely distributed in brain tissue (1). Its exact physiological role remains unclear, but due to its presynaptic localization, a role in the regulation of neurotransmitter release was shown (2). Multiplications and missense point mutations (A30P, E46K, and A53T) in the ␣-Syn gene cause an early onset of familial forms of Parkinson disease (PD) and strongly link ␣-Syn with the pathogenesis of PD (3)(4)(5)(6). Moreover, deposits mainly composed of ␣-Syn fibrils either in the cytoplasm (Lewy body) or in neurites (Lewy neurite) are the pathologic hallmarks of synucleinopathies, including sporadic PD and dementia with Lewy bodies (7,8). In the process of aggregation of ␣-Syn, soluble oligomers, protofibrils, and insoluble amyloid fibrils are formed (9). Although cytotoxicity in vitro and in vivo is currently attributed to oligomers of ␣-Syn rather than to fibrils (10 -13), the temporal and molecular mechanisms underlying ␣-Syn aggregation-driven neurodegeneration are far from understood.
There is increasing evidence for a direct interplay between ␣-Syn aggregation and axonal transport. Axonal transport relies on the microtubule (MT) network and on the MT-based motor proteins, kinesin and dynein, and is fundamental for the maintenance of neuronal homeostasis (14). A decline in axonal transport motor proteins was reported in sporadic PD brains, and this decline correlated with the presence of ␣-Syn aggregates (15). Furthermore, overexpression of ␣-Syn oligomeric mutants resulted in neurite swellings in rat substantia nigra neurons (13). The interaction of ␣-Syn with tubulin and MTs was shown in vitro and in vivo, and ␣-Syn overexpression and aggregation was associated with disrupted MT networks (16 -22). Together, these data strongly suggest alterations in the intracellular transport machinery in synucleinopathies. Although ␣-Syn oligomers were implicated in decreased tubulin polymerization in a cellular model (23), the specific relation of oligomers/fibrils and different compartments of the neuronal cytoskeletal machinery has not yet been elucidated.
In the present study, we used recombinant ␣-Syn protein (wild type, single point mutants, and the fragment 30 -110) to obtain different aggregate species in vitro and evaluated the direct effect of these species on MT polymerization and kinesin-dependent MT gliding in a cell-free system. Our data showed that ␣-Syn oligomers have a strong influence on kinesin-driven MT motility, whereas seeds of ␣-Syn significantly inhibited MT assembly, promoted by the axon-specific MAP, Tau. In a cellular system, mild overexpression of an oligomerizing ␣-Syn variant and, less potently, ␣-Syn seeds impaired neurite network morphology. This was accompanied by impaired MT stability and reduced presence of kinesin and kinesin-dependent cargoes within neurites. Our results suggest that the MT-kinesin network is the first target of soluble ␣-Syn aggregates.

EXPERIMENTAL PROCEDURES
␣-Syn Preparation-␣-Syn wild type (WTS) and the ␣-Syn variants E57K (glutamic acid to lysine substitution at amino acid 57 of the ␣-Syn protein), A30P (alanine to proline substitution at amino acid 30 of the ␣-Syn protein), and aSyn (30 -110) (core part of the ␣-Syn protein consisting of amino acids 30 -110) were used. WTS and all ␣-Syn variants were expressed in Escherichia coli. ␣-Syn protein purification was performed with a modified version of previously published protocols taking advantage of several protein precipitation steps (13,24). Briefly, the pellet obtained in the 50% (v/v) ethanol precipitation step was resuspended in 9 ml/liter of culture of 20 mM Tris, pH 8, and dialyzed overnight at 4°C against the same buffer with a 3.5-kDa cut-off membrane (Spectrum Laboratories Inc., Rancho Dominguez, CA). The sample was applied onto a HiTrap Q HP column (GE Healthcare), equilibrated with 20 mM Tris, pH 8, and eluted with a linear gradient from 0 to 500 mM NaCl over 10 column volumes. The collected fractions were analyzed by SDS-PAGE and Coomassie Brilliant Blue staining, and ␣-Syn-containing fractions were merged. For aSyn (30 -110), the flow-through was collected, analyzed by SDS-PAGE, and concentrated with 3-kDa cut-off Amicon concentrators (Merck Millipore, Darmstadt, Germany). All preparations were finally dialyzed against 20 mM ammonium acetate, lyophilized, and stored at Ϫ20°C. The lyophilized samples were resuspended in MT assembly buffer (20 mM PIPES, 80 mM NaCl, 1 mM EGTA, 0.5 mM MgCl 2 , pH 6.8) at a concentration of 400 M and stored at Ϫ20°C until usage. The purity of ␣-Syn proteins was 98 -100%, as determined by SDS-PAGE followed by Coomassie Brilliant Blue staining (data not shown).
␣-Syn Aggregation-400 M ␣-Syn solutions were diluted to 100 M solutions with MT assembly buffer before aggregation. One part of each 100 M ␣-Syn protein solution was left untouched and was considered as a monomer fraction. The rest was transferred to a multimer fraction, applying a previously published protocol (13) with some modifications: incubation at 37°C for 12 h followed by 56°C for 16 h with slight agitation.
Electron Microscopy (EM) of ␣-Syn Multimers-The samples for negative stain transmission EM were prepared at the completion of the aggregation protocol. 3 l of the 100 M ␣-Syn multimer solutions were adsorbed onto 100-mesh Formvar/ carbon-coated copper grids (Plano GmbH, Wetzlar, Germany), briefly washed with distilled water, and negatively stained with 1 g/100 ml of aqueous uranyl acetate for 1 min. Grids were analyzed on a transmission electron microscope (JEM 1400, Jeol, Tokyo, Japan). Digital images were taken using a CCD-Camera Orius SC1000 (Gatan Inc., Pleasanton, CA).
Preparation of Microtubule Proteins and Kinesin-In the present study, full-length human neuron-specific kinesin-1 (KIF5A full ) and its truncated form comprising 560 amino acids from the N terminus and containing the motor domain, neck, and part of the stalk (KIF5A 560 ) were used. Kinesins were expressed in E. coli as fusion proteins with a C-terminal chitinbinding intein tag and purified by binding to chitin beads followed by removal of the intein tag by 50 mM DTT-induced specific self-cleavage (25). The kinesins were stored in motility buffer (50 mM imidazole, 0.5 mM MgCl 2 , 0.5 mM EGTA, 0.5 mM DTT, pH 6.8) supplemented with 150 mM NaCl and 1 M glycerol at Ϫ80°C.
Microtubule protein (MTP) consisting of 80% tubulin and 20% complete MAP set (MAP1, MAP2, and Tau) was purified from porcine brain homogenates by two cycles of temperaturedependent disassembly (0°C)/reassembly (37°C) (26). To obtain pure tubulin, co-purified MAPs were removed by phosphocellulose ion exchange chromatography, according to a previously published protocol (27). MAP2 and Tau were eluted from the phosphocellulose, treated for 5 min at 95°C (28), and separated from denatured material (MAP1 and residual tubulin) by ultracentrifugation (45 min, 40,000 rpm, Beckman Coulter Ti50.2 rotor, Beckman Coulter, Krefeld, Germany). Tau and the high molecular weight MAP2 were separated by passing the clear supernatant through a 100-kDa ultrafilter. All proteins (i.e. MTP, tubulin, MAP2, and Tau) were frozen in liquid nitrogen and stored at Ϫ80°C until usage.
Stabilized MTs were formed by taxol-promoted self-assembly at 37°C from 2 mg/ml pure tubulin in the MT assembly buffer supplemented with 0.2 mM GTP and 20 M taxol. MTs were usually prepared at the day of usage at least 30 min prior to an experiment. The purity of kinesins was 75-80%, and the purities of MTP, tubulin, MAP2, and Tau were all 95-98%, as determined by SDS-PAGE followed by Coomassie Brilliant Blue staining.
Ligand Blot-Cytoskeleton proteins (total amount of 1 g/spot) were dotted onto nitrocellulose membrane stripes in the order shown in Fig. 1. BSA was used to control for a nonspecific binding of ␣-Syn. As positive control, 3 l of 100 M ␣-Syn solution were applied. Proteins were allowed to bind to the nitrocellulose for 5 min at room temperature. Unbound proteins were removed by washing three times with a blocking buffer (0.5 g/100 ml BSA, 0.5% Tween 20 in PBS). All further washing steps and antibody dilutions for ligand blot experiments were performed in the blocking buffer. The stripes were incubated in 10 M solutions of monomeric or multimeric WTS or ␣-Syn variants overnight at 4°C. To control for nonspecific binding, one stripe was incubated in the blocking buffer in the absence of ␣-Syn. Unbound ␣-Syn was washed out three times. ␣-Syn binding was detected after incubation with monoclonal mouse anti-␣-Syn antibody (1:1000; clone 42, BD Biosci-ences) for 1 h at room temperature followed by three times washing and incubation with goat anti-mouse antibody labeled with an infrared dye, IRDye 800CW (1:2000; LI-COR Biosciences, Lincoln, NE) for 1 h at room temperature. After extensive washing, the stripes were developed on the Odyssey infrared imager using the Odyssey application software version 3.0 (both from LI-COR Biosciences). Densitometry analysis was performed by estimating an integrated intensity value of each spot that is proportional to the amount of dye-labeled antibodies on the membrane using the Odyssey application software. Three independent ligand blots were performed, and mean values of the intensities Ϯ S.D. were calculated.
Microtubule Assembly Assay-MT assembly was recorded at 37°C by a Cary 100 spectrophotometer equipped with a temperature-controlled multicell holder (Agilent Technologies, Böblingen, Germany) using 2-mm OS glass cuvettes (Hellma GmbH & Co. KG, Müllheim, Germany). For this, the proteins were adjusted to a total volume of 275 l in the MT assembly buffer containing 0.5 mM GTP.
To analyze the influence of ␣-Syn on MT polymerization in the presence of MAPs, assembly was performed 1) from MTP As a control, MTs were assembled at the respective conditions in the absence of ␣-Syn. As a negative control for MT polymerization, MT assembly was performed from purified tubulin (16 M) with or without ␣-Syns (6.3 M) and in the absence of any MAPs. The MT assembly measurements with MAP2 and MTP were performed with an additional cooling step to 10°C. The rationale for the cooling step was to test for a possible stabilizing effect of ␣-Syn on established MT in the absence of the ␣-Syn inhibitory effect on MT polymerization. Measurements of Tau-promoted MT assemblies were performed with an addition of the MT assembly-promoting drug taxol to a final concentration of 10 M. The rationale for the taxol step in Tau assemblies was to prove that tubulin had not lost its assembling capacity under ␣-Syn-inhibitory conditions. MT polymerization was measured as the turbidity (A 360 nm ) of the solution recorded every 4 s at 360 nm over 30 -60 min. Data are shown as representative assembly curves and mean assembly levels at 30 min (plateau level, representing the assembly/disassembly steady state) of three independent experiments for each condition. To quantitatively analyze the effect of ␣-Syns on MT assembly, the turbidity signals after 30 min in each assay were compared with that of the control condition.
In Vitro Gliding Assay-Gliding assays were performed on glass slides (for principles, see Ref. 29) covered by casein (5 mg/ml). MT were assembled and stabilized as described above. All gliding assays were performed with the KIF5A 560 in the motility buffer supplemented with 20 M taxol. The gliding assay mixture containing KIF5A 560 (0.4 mg/ml), stabilized MTs (40 g/ml), 100 mM NaCl, 0.5 mM ATP, and 10 M ␣-Syn was incubated for 5 min at room temperature prior to transfer onto the casein-coated glass slide. The sample was covered by a coverslip, and MT gliding was observed as described below. The control gliding assay was performed in the absence of ␣-Syn. MT gliding velocities were measured during observation by following the tracks of individual MTs using the Argus-20 software (Hamamatsu Photonics, Herrsching am Ammersee, Germany). All gliding assays were carried out at 36°C. The mean gliding velocities were calculated from the gliding velocities of at least 10 MTs in each sample.
Gliding Assay Microscopy and Data Analysis-MT gliding was visualized by video-enhanced differential interference contrast microscopy. All analyses were performed using an Zeiss Axiophot microscope (Carl Zeiss, Munich, Germany) equipped with a Plan-Neofluar ϫ63/1.25 oil immersion objective (Carl Zeiss), a Chalnicon video camera type C2400-0.1 (Hamamatsu Photonics), and the image processing system Argus-20 (Hamamatsu Photonics), which enables background subtraction and electronic contrast enhancement. MT movement was documented on a digital hard disk recorder (JVC SR-DVM70EU, JVC, Wayne, NJ). Video sequences were digitized using Adobe Photoshop, resulting in image sequences of 30 frames/s.
Lund Human Mesencephalic (LUHMES) Cell Culture-LUHMES cells (gift of Dr. M. Leist, Konstanz, Germany; described in Ref. 30) were cultured at 37°C in a humidified 95% air, 5% CO 2 atmosphere. Cells were cultured in plastic flasks or in plastic plates coated with poly-L-ornithine (50 g/ml) and fibronectin (1 g/ml) (both from Sigma-Aldrich) overnight at 37°C and washed once with PBS prior to seeding the cells. LUHMES cells were maintained and expanded in advanced DMEM/F-12/Glutamax medium supplemented with N2 supplement (both from Invitrogen) and 40 ng/ml FGF2 (R&D Systems, Wiesbaden-Nordenstadt, Germany) in flasks. Cells were split using TrypLE Express (Invitrogen) at 80% confluence. Neuronal differentiation of LUHMES cells was performed for 6 days in 24-well plates (150,000 cells in 500 l/well) in advanced DMEM/F-12/Glutamax/N2 medium supplemented with 1 g/ml tetracycline, 1 mM cAMP (both from Sigma-Aldrich), and 2 ng/ml glial cell-derived neurotrophic factor (R&D Systems). Lentiviral (LV) constructs encoding for either WTS or the E57K ␣-Syn variant containing an internal ribosome entry site GFP sequence or LV encoding only an internal ribosome entry site GFP (referred to hereafter as WTS, E57K, and GFP, respectively) were used to infect LUHMES cells. LUHMES cells were predifferentiated for 2 days before infection at a multiplicity of infection (MOI) of 2 (mild) or 4 (high overexpression). LV infection was performed for 48 h.

␣-Syn Oligomers Alter Microtubule-based Cytoskeleton
Germany), mouse monoclonal anti-␣-Syn antibody (1:250; LB509) or rabbit polyclonal anti-Tau antibody (1:50) (both from Abcam, Cambridge, MA), or mouse monoclonal antiacetylated tubulin antibody (1:10,000; clone 6-11B-1) or rabbit anti-KIF5 (1:100) (both from Sigma-Aldrich) and were counterstained with DAPI (1 g/ml). The following secondary antibodies were used: donkey anti-rabbit AlexaFluor546 and donkey anti-mouse AlexaFluor649 (both from Invitrogen). Cells were visualized on a Zeiss AxioVert 200 inverted microscope with the Axiovision Software (Carl Zeiss). Cell death was visualized by the addition of 0.4 g/100 ml trypan blue and was determined as a percentage of trypan blue-positive cells within a total number of cells on a particular image. Network morphology was characterized by the number of neurites crossing three unified random horizontal parallel lines over the image. Acetylated tubulin was quantified by the fluorescence intensities over the picture and normalized to the number of DAPI-positive nuclei in this picture. KIF5 signals within neurites were determined by the mean intensity along the tracing of a single neurite. At least 100 neurites were analyzed for each condition. Distribution of mitochondria in neurites was defined as the number of mitochondria/m of a neurite length. Colocalization of ␣-Syn with ␤3-tubulin and Tau was quantified by the determination of Pearson's overlap coefficient. Each condition was performed in duplicates. Image analyses were performed using ImageJ software version 1.47f (National Institutes of Health, Bethesda, MD).
Statistical Analysis-Ligand blot densitometry data were analyzed by repeated measures analysis of variance with a Dunnett post hoc test separately for ␣-Syn monomers and multimers. MT assembly levels and MT gliding velocities were analyzed by one-way analysis of variance using the Dunnett post hoc test for monomers and multimers separately versus control condition. Neuronal network morphology, cell viability, imunohistochemical results, and mitochondria distribution were analyzed by one-way analysis of variance with Tukey's multiple comparison post hoc test. Dose dependence of the WTS effect on MT assembly was analyzed by estimating a linear regression coefficient (R 2 ) between assembly levels at 30 min and WTS concentrations. Differences were considered statistically significant at p Ͻ 0.05. Statistical tests were performed using the GraphPad Prism 5 software (GraphPad Software, San Diego, CA).

WTS Directly
Interacts with KIF5A, Tubulin, and MAPs-Previously, ␣-Syn oligomers were shown to be more toxic than ␣-Syn fibrils and were observed to cause prominent axonal swellings in dopaminergic neurons in the rat substantia nigra in vivo (13,31). We therefore aimed to investigate the interaction of different aggregated states of ␣-Syn with proteins involved in axonal anterograde transport.
Using a ligand blot technique, WTS, E57K, A30P, and aSyn (30 -110) were tested for their interaction capacity with the following: the full-length neuron-specific motor protein kinesin KIF5A (KIF5A full ), a truncated form of a KIF5A molecule (containing the first 560 amino acids (KIF5A 560 )), tubulin, assembled MTs, MAP2, and Tau (Fig. 1, B-D). The rational for using WTS was to mimic sporadic PD. E57K is an artificial mutant known to form oligomers but rare fibrils (13). A30P is a familial mutant, also known to form oligomers, but to a lesser extent than E57K. aSyn(30 -110) represents a fibril-forming artificial mutant (32). For each ␣-Syn protein, monomeric and multimeric forms were analyzed separately. Multimers of the ␣-Syn proteins formed in vitro were represented either by oligomers (multimers of E57K and A30P variants) or fibrils (multimers of aSyn(30 -110) variant) or a mixture of oligomers and protofibrils of WTS multimers referred to hereafter as WTS seeds (Fig. 1A).
WTS monomers revealed binding to KIF5A (both full-length and truncated), tubulin, MTs, MAP2, and Tau. Compared with WTS, aSyn(30 -110) and two oligomerizing ␣-Syn variants showed weaker binding to these cytoskeleton proteins (Fig. 1C). Notably, WTS seeds demonstrated a prominent interaction with all tested cytoskeleton proteins compared with weaker binding of oligomers and fibrils to these proteins (Fig. 1, B and  D). In summary, direct binding of oligomer and fibril-forming ␣-Syn variants to the proteins of MT-based cytoskeleton is significantly lower compared with WTS and WTS seeds.
␣-Syn Seeds Reduce Tau-promoted MT Formation-We first tested the possibility that WTS and ␣-Syn variants might promote MT assembly from tubulin dimers ( Fig. 2A). Purified tubulin was not able to assemble into MTs either in the absence of MAPs (Control in Fig. 2B) or with the addition of any of the ␣-Syn species (Fig. 2B). This indicates that ␣-Syn does not have an intrinsic MAP-like activity. Because WTS and, less potently, ␣-Syn variants were found to bind MAP2, Tau, and tubulin, we checked whether ␣-Syn proteins and its different multimeric forms are able to affect MT assembly in the presence of MAPs. None of the ␣-Syn variants (either as monomers or as multimers) had an impact on MT assembly promoted by the complete MAP set or MAP2 alone; nor did they exert an MT-stabilizing activity (Fig. 2, C and D).
This led us to investigate the influence of WTS, ␣-Syn variants, and their multimers on MT formation in the presence of the axon-specific MAP, Tau (Fig. 3A). Monomeric forms of WTS and of ␣-Syn variants did not influence Tau-promoted MT assembly (Fig. 3, B and E; p Ͼ 0.05 for the assembly levels at 30 min). ␣-Syn seeds significantly reduced Tau-promoted MT assembly (Fig. 3, C and F, p Ͻ 0.05 for the assembly levels at 30 min in the presence of WTS seeds versus control). Interestingly, neither oligomers alone nor ␣-Syn fibrils were able to influence Tau-promoted MT assembly significantly (Fig. 3, C and F). To test whether the inhibitory effect of WTS seeds on Tau-promoted MT polymerization is accompanied by irreversible changes of tubulin, we used the MT assembly-promoting drug, taxol, to trigger the formation of MTs. The addition of taxol enabled MT formation in the presence of ␣-Syn monomers and multimers (40 min time point; Fig. 3B and data not shown), suggesting that the inhibition of MT assembly by WTS seeds was not due to irreversible changes of tubulin. Thus, the inhibitory effect of WTS seeds on Tau-promoted MT formation might result from the binding of Tau by WTS seeds. To prove this possibility, MT formation was performed using higher amounts of Tau (2.7 M instead of 2.1 M). Indeed, at a higher Tau/WTS ratio (1:3), MTs were efficiently assembled despite ␣-Syn Oligomers Alter Microtubule-based Cytoskeleton JULY 26, 2013 • VOLUME 288 • NUMBER 30 the presence of WTS seeds (Fig. 3D). In line with this, the inhibitory effect of WTS seeds on Tau-promoted MT assembly was dose-dependent (Fig. 3, H and I Fig. 3I). At the same time, no dose dependence was observed for WTS monomers on Tau-promoted MT assembly (Fig. 3, G and I; R 2 for a linear regression fit 0.2416, p value for regression 0.4). These results indicate the potency of WTS seeds to interplay with Tau-controlled MT polymerization. ␣-Syn oligomeric and fibrillar forms did not influence this process.
␣-Syn Oligomers but Not Fibrils Slow Down MT Gliding Velocity in Vitro-Because WTS, ␣-Syn variants, and different multimeric forms of ␣-Syn were able to bind to KIF5A as well as tubulin and MTs (Fig. 1), we aimed to investigate whether they are able to affect kinesin-MT interaction. To monitor the functionality of the kinesin-MT system, we took advantage of an in vitro gliding assay that mimics the anterograde axonal trans-port process (29). Monomers and multimers of WTS and ␣-Syn variants were added at a concentration of 10 M (mimicking physiological conditions) (33,34), and gliding of MTs across the KIF5A-coated surface was monitored by live imaging (Fig. 4A).  Fig. 4D). Intriguingly, among ␣-Syn multimers, only oligomers were able to influence the MT gliding character and velocity (Fig. 4, B, C, and E). The gliding MTs were straight in control and in the presence of WTS seeds, and ␣-Syn fibrils (control as representative for WTS seeds and ␣-Syn fibrils in Fig. 4, B and C; see supplemental Movie 1). In the presence of ␣-Syn oligomers, MTs were more likely bent (E57K in Fig. 4, B and C, and supplemental Movie 2) (data not shown). Moreover, ␣-Syn oligomers significantly decreased the gliding velocity of MTs (E57K (0.69 Ϯ 0.08 m/s) and A30P (0.64 Ϯ 0.06 m/s) versus control (0.85 Ϯ 0.04 m/s); p Ͻ 0.05; Fig. 4E). These data indicate impairment of kinesin-MT motility by ␣-Syn oligomers.

␣-Syn Oligomers Alter Microtubule-based Cytoskeleton
Overexpression of Oligomers of ␣-Syn Leads to Disrupted Neurite Morphology in a Dopaminergic Neuronal Cell Line-Our findings from the cell-free system show an inhibitory capacity of WTS seeds on Tau-promoted MT formation and ␣-Syn oligomer-mediated reduction of kinesin-MT gliding. The question arises of whether specific ␣-Syn species (seeds or oligomers) that critically affect MT-kinesin cytoskeleton in vitro would cause neurite pathology in neuronal cells. We therefore aimed to specifically investigate the influence of ␣-Syn seeds and oligomers on neurite morphology and the MT-kinesin system in a neuronal cell culture. Therefore, specifically WTS and E57K, as examples of seeds and oligomers, respectively, were chosen. They were overexpressed in LUHMES cells (a human dopaminergic neuronal cell line) using LV infection.
Different MOIs (2 and 4) were applied to achieve different levels of ␣-Syn overexpression. Confirming previous data, a high level of ␣-Syn overexpression (LV infection at MOI 4) resulted in increased cell death in E57K-infected neurons  JULY 26, 2013 • VOLUME 288 • NUMBER 30 JOURNAL OF BIOLOGICAL CHEMISTRY 21747 (Fig. 5, A and D) (12,13). Importantly, mild levels of ␣-Syn overexpression (WTS and E57K after LV infection at MOI 2) did not significantly increase neuronal cell toxicity (Fig. 5B).

␣-Syn Oligomers Alter Microtubule-based Cytoskeleton
The following experiments were performed in LUHMES cells overexpressing mild levels of ␣-Syn (MOI 2).
First, LUHMES cells were analyzed for their neurite network morphology following LV infection. The overexpression of mild amounts of the oligomerizing ␣-Syn variant led to significant disruptions in the neurite network morphology compared with mock-infected cells (Fig. 5, C and E). Mild WTS overexpression also changed the neurite morphology, although this effect was not as pronounced as in E57K-overexpressing cells (Fig. 5, C and E). Potentiating a possible interplay of ␣-Syn seeds and oligomers with a neuritic MT-based system and in line with the in vitro MT assembly data, colocalization of ␣-Syn with ␤3-tubulin and Tau was present endogenously in mock-infected LUHMES cells and in LUHMES cells overexpressing WTS and E57K (Fig. 5, F and G).

␣-Syn Oligomers Alter Microtubule-based Cytoskeleton
WTS and E57K ␣-Syn Reduce MT Stability and Alter KIF5dependent Cargo Distribution in LUHMES Cells-To clarify whether the observed disruption of the neurite network morphology by ␣-Syn seeds and oligomers correlates with a dysfunction of the MT-based system, we investigated the stability of MTs and KIF5-dependent cargo distribution in neuronal cells mildly overexpressing WTS or E57K ␣-Syn. We therefore determined the amount of acetylated tubulin, as a measure of stabilized MTs. We found a significant reduction of acetylated tubulin in LUHMES cells overexpressing WTS and E57K compared with controls (Fig. 6, A and D, p Ͻ 0.05). Because the loss of tubulin acetylation was shown to decrease binding and motility of KIF5 (35), we next investigated KIF5 and KIF5-dependent cargo distribution within neurites. We detected decreased amounts of KIF5 within neurites of LUHMES cells overexpressing WTS and a further reduction in E57K-overexpressing cells compared with control LUHMES (Fig. 6, B and E), suggesting diminished KIF5/MT interaction in the presence of ␣-Syn seeds and oligomers in neurites. Moreover, we found significantly fewer mitochondria (a KIF5-dependent cargo) distributed within neurites of LUHMES cells overexpressing WTS and even fewer in E57K-overexpressing cells compared with control cells (Fig. 6, C and F). These data suggest that ␣-Syn seeds and, more significantly, oligomers impair MT stability and MT-kinesin functionality in neuronal cells, and this is associated with a disruption of the neurite network morphology.

DISCUSSION
The sequence of pathological events caused by ␣-Syn aggregation and the impact of different ␣-Syn species is still unknown. In the present study, the interplay between proteins involved in axonal transport and different species of ␣-Syn,

␣-Syn Oligomers Alter Microtubule-based Cytoskeleton
such as oligomers (formed by A30P and E57K ␣-Syns), seeds (WTS), and fibrils (formed by aSyn (30 -110)) was investigated. Our data show that WTS monomers and seeds bind to proteins required for MT-based anterograde axonal transport, such as KIF5A, tubulin, MTs, MAP2, and Tau. Interestingly, WTS seeds impaired MT polymerization in vitro promoted by axonspecific Tau protein in a dose-dependent manner. Moreover, in a model system of kinesin-driven transport, oligomers of ␣-Syn decreased kinesin-MT motility in vitro. In a human dopaminergic neuronal cell line, neurite network morphology was severely disrupted by mild overexpression of ␣-Syn oligomers. Neurite morphology disruption in these cells correlated with a significant reduction of MT stability and impaired amounts of KIF5 and KIF5-dependent cargo in neurites. Thus, our study The identification of proteins that interact with ␣-Syn may clarify the cellular processes accompanying ␣-Syn aggregation. Here, we observed a direct interaction of WTS with tubulin heterodimers, polymerized MTs, MAP2, Tau, and with the MT-based neuron-specific motor protein KIF5A, which confirms previous findings for tubulin (16, 22, 36 -38) and expands data for Tau (39,40), MAP2 (41), and KIF5A (42). In line with this, we observed a colocalization of ␣-Syn (WTS and E57K) with both Tau and tubulin in untreated and ␣-Syn-overexpressing LUHMES cells. The question raised is whether the binding capacity of ␣-Syn to cytoskeletal proteins is different between WTS and ␣-Syn variants. WTS and WTS seeds showed stronger binding to all tested cytoskeleton proteins compared with monomer and multimer fractions of the oligomer-promoting variants (E57K and A30P) and fibril-form-ing variant (aSyn(30 -110)) of ␣-Syn. These results might suggest that alterations in axonal transport and/or MT networking could occur in the presence of ␣-Syn mutants. For example, familial PD mutants of ␣-Syn (A30P and A53T) were reported to favor formation of amorphous tubulin polymers lacking tubular structure in contrast to WTS in vitro (36).
Our study showed that WTS seeds, but not monomers or oligomers or fibrils of ␣-Syn, were able to reduce Tau-promoted MT polymerization. This is in line with a higher interacting capacity of WTS seeds compared with other forms of ␣-Syn proteins tested with tubulin, MTs, and Tau. Only Taupromoted and not MAP2-promoted MT assembly was reduced by WTS seeds. The differences between MAP2-and Tau-directed effects of WTS seeds might highlight a preferential activity of ␣-Syn toward axon-specific proteins, such as Tau (40), compared with dendrite-specific proteins (MAP2). Because taxol was able to rescue MT polymerization to the control level, we hypothesized that WTS seeds did not cause irreversible tubulin changes. Instead, this effect could be due to the interaction of WTS seeds and/or WTS-tubulin multimers with Tau molecules that might temporally and spatially separate Tau from MTs. This idea might be potentiated by the fact that adding more Tau to the assay could rescue an inhibitory effect of WTS seeds. Moreover, Tau and ␣-Syn can promote aggregation of each other (43,44) and are colocalized in Lewy bodies in PD brains (45). Furthermore, ␣-Syn was reported to bind to the MT-binding domain of Tau and to promote Tau phosphorylation at Ser-262/356 and Ser-396/404 (40,46), both of which would enhance the dissociation of Tau from MTs. Notably, a recent large scale replication study has confirmed a strong association of the Tau genetic locus (MAPT, rs2942168) with sporadic PD, further potentiating Tau involvement in the pathogenesis of PD (47).
Disruption of MT structures after overexpression of ␣-Syn was reported previously in different cellular systems (19,22). Importantly, in our study, WTS overexpression in a neuronal cell line led to significant reduction of the acetylated tubulin level, indicating a loss of MT stability. Thus, our data provide evidence that ␣-Syn seeds might be the species that evoke MT dysfunction and destabilize MT networks under ␣-Syn-aggregating conditions. Interestingly, tubulin has been shown to promote ␣-Syn fibril formation, to form insoluble complexes with ␣-Syn in primary mouse neurons, and to colocalize with ␣-Syn in Lewy bodies (16,20,21). In turn, ␣-Syn seeds induce depolymerization of MTs and thereby increase the level of free tubulin. This might indicate that ␣-Syn aggregation-induced MT depolymerization and tubulin-induced/enhanced ␣-Syn seeding are interconnected.
MTs are important cytoskeletal filaments that control cellular processes in neurons, including neurite structure, intracellular transport of different cargoes, and neuronal cell maintenance (48). ␣-Syn aggregation has been shown to increase the amount of "vacant synapses" (49) and neuritic degeneration (19,50) in ␣-Syn-overexpressing neurons. The reason for these events might be MT aberrations and/or MT-dependent transport disruption. Our study was designed to clarify whether and which ␣-Syn species could alter MT-based functionality. We show that, whereas ␣-Syn seeds led to critical disruption of MT assembly, ␣-Syn oligomers significantly slowed down MT gliding across the KIF5A-coated surface in vitro. Fibrils of ␣-Syn did not influence MT assembly and kinesin-dependant MT gliding. These data confirm previous findings of destructive effects of ␣-Syn oligomers and seeds on different cellular processes in contrast to "inert" fibrils (10,11,13,51). In neurites of a neuronal cell line mildly overexpressing E57K, we found KIF5 significantly reduced. Previously, decreased KIF5A levels were reported at very early stages of PD (15) and in a rat PD model overexpressing A30P ␣-Syn mutant (52). In line with this, lower numbers of a KIF5-dependent cargo (mitochondria) in neurites of E57K-overexpressing cells might be a further indication for an ␣-Syn oligomer-induced transport disruption. An additional way of analyzing the effect of WTS and ␣-Syn variants on KIF5/MT interaction would be co-immunoprecipitation in neuronal cells. However, due to the ATP-dependant dynamic nature of the KIF5 interaction with MTs (35), we were not able to co-precipitate KIF5 with anti-tubulin antibody and vice versa (data not shown).
Of interest, significantly reduced levels of acetylated tubulin were caused by mild overexpression of E57K in a neuronal cell line, showing disrupted MT stability. These data are in line with previous reports showing that ␣-Syn oligomers, when applied extracellularly, caused significant disruption of tubulin polymerization in a dopaminergic cell line (MES23.5), whereas no direct effect of ␣-Syn oligomers on MT assembly in vitro was seen in this study (23). Furthermore, overexpression of E57K led to more pronounced neurite degeneration and changes of neuronal network morphology than WTS. In both conditions, neurite degeneration occurred at lower expression levels than neuronal cell death. This might indicate a more direct critical influence of ␣-Syn oligomers on kinesin-MT interplay, whereas seeds of ␣-Syn would mostly act on MT network, destabilization of which would cause further alterations of MT-based transport.
In conclusion, we propose a sequence of pathologic events in neurites induced by ␣-Syn aggregation that involve impairment of MT-kinesin interaction and functionality due to ␣-Syn oligomers and MT network disruption by ␣-Syn oligomers and seeds. Oligomers and seeds then act together with other pathologic effects of ␣-Syn oligomers, such as membrane thinning and leakage (53) (Fig. 7), and together represent critical early mechanisms of synucleinopathies. Together with other pathologic effects induced by ␣-Syn oligomers, such as membrane leakage, cytoskeleton changes might have a large spectrum of critical consequences that would together contribute to neurite degeneration, collapsed neuronal network, and finally to neuronal death. Fibrillar aggregates of ␣-Syn do not critically interfere with the axonal transport machinery and might represent "inert" ␣-Syn aggregate species.