Alpha-Synuclein is degraded by both autophagy and the proteasome.

Parkinson's disease (PD) is characterized by the loss of dopaminergic neurons in the substantia nigra and the formation of aggregates (Lewy bodies) in neurons. alpha-Synuclein is the major protein in Lewy bodies and rare mutations in alpha-synuclein cause early-onset PD. Consequently, alpha-synuclein is implicated in the pathogenesis of PD. Here, we have investigated the degradation pathways of alpha-synuclein, using a stable inducible PC12 cell model, where the expression of exogenous human wild-type, A30P, or A53T alpha-synuclein can be switched on and off. We have used a panel of inhibitors/stimulators of autophagy and proteasome function and followed alpha-synuclein degradation in these cells. We found that not only is alpha-synuclein degraded by the proteasome, but it is also degraded by autophagy. A role for autophagy was further supported by the presence of alpha-synuclein in organelles with the ultrastructural features of autophagic vesicles. Since rapamycin, a stimulator of autophagy, increased clearance of alpha-synuclein, it merits consideration as a potential therapeutic for Parkinsons disease, as it is designed for chronic use in humans.

Parkinson's disease (PD) 1 is caused by the degeneration of dopaminergic neurons in the substantia nigra. The pathogenic hallmark of PD is the accumulation and aggregation of ␣-synuclein in susceptible neurons. The cytoplasmic aggregates/inclusions characteristic of PD are called Lewy bodies and their major constituent is ␣-synuclein (1). Lewy pathology is also found in dementia with Lewy bodies, the Lewy body variant of Alzheimer's disease, in neurodegeneration with brain iron accumulation type I and in glial cytoplasmic inclusions of multiple system atrophy. These diseases are collectively known as "␣-synucleinopathies" (2,3). ␣-Synuclein is likely to play a role in PD, since two missense mutations in ␣-synuclein (A53T and A30P) cause autosomal dominant, early-onset PD (4,5), and animal models overexpressing ␣-synuclein develop a disease phenotype with Lewy body-like pathology and locomotor impairment (6,7). However, while ␣-synuclein knock-out mice show some disruption in synaptic neurotransmission, they do not manifest obvious disease and develop normally, with a normal life span and behavior (8,9,10).
Two other forms of inherited PD are caused by mutations in Parkin (an E3 ubiquitin ligase) (11) and ubiquitin carboxyLterminal hydrolase L1 (UCH-L1) (12). Both of these proteins are involved in the ubiquitin-dependent degradation of intracellular proteins (13). This evidence and the identification of proteasome subunits in Lewy bodies (14) has led to the speculation that an impairment of the ubiquitin-proteasome system may contribute to the progression of PD. A number of studies have investigated the effect of proteasomal inhibition on ␣-synuclein metabolism, with conflicting results. Some groups reported that proteasomal inhibition caused an accumulation of ␣-synuclein, inclusion formation and increased cell death (15,16,17), while other studies suggested ␣-synuclein is not a proteasome substrate (18,19).
Another pathway that may be relevant to ␣-synuclein clearance is autophagy, a process mediating bulk degradation of cytoplasmic proteins or organelles in the lytic compartment. Autophagy involves the formation of double-membrane structures called autophagosomes, which fuse with primary lysosomes to become an autophagolysosome where their content is degraded and then either disposed of or recycled back to the cell (20). We have recently shown that aggregate-prone proteins with polyglutamine and polyalanine expansions are degraded by autophagy (21). Since ␣-synuclein is also an aggregate-prone protein, we have tested whether it is degraded by this pathway. Our previous data suggested that aggregate-prone proteins were more likely to be cleared by autophagy than the more soluble species that did not have expanded polyglutamines or polyalanines. In the context of ␣-synuclein, this model predicts that the A53T mutation that forms aggregates most readily may be more dependent on autophagy than the wild-type protein or A30P mutation (that forms oligomers efficiently but not aggregates) (22).
In this study, we have used PC12 cells to establish stable, doxycycline-inducible lines as they are dopaminergic and can be differentiated into a neuron-like phenotype with nerve growth factor (23). We used the Tet-On system, where regulation of expression is achieved through a tetracycline-controlled transactivator. This allows us to turn on and off ␣-synuclein expression by adding and then removing doxycycline and therefore enables us to follow ␣-synuclein clearance. In our cell lines we found that ␣-synuclein was degraded by both proteasome and autophagy pathways.

EXPERIMENTAL PROCEDURES
Establishment of Inducible PC12 Cell Lines-Human ␣-synuclein with a HA tag at the NH 2 terminus and a His 6 tag at the COOH terminus was amplified from pHM6␣-synuclein constructs (24) and inserted into pTRE2hyg vector (Clontech) using NheI/SalI sites. Con-structs were confirmed by sequencing before use. The pTRE2hyg␣synuclein construct and pTet-tTS (Clontech) were co-transfected into PC12 Tet-On cells (Clontech), using LipofectAMINE (Invitrogen) and Plus Reagent (Invitrogen). Single colonies were isolated using cloning cylinders (Sigma), and cells were grown and tested for ␣-synuclein expression upon induction. Cell lines were subjected to a further round of purification from single colonies to ensure that lines were pure.
Immunofluorescence-Coverslips were placed in six-well dishes and coated with 0.01% poly-L-lysine (Sigma). Cells were seeded and induced with 2 g/ml doxycycline and incubated as necessary. Cells were fixed with 4% paraformaldehyde (Sigma) for 30 min and then washed with PBS and permeabilized with 0.1% Triton-X-100 (Sigma) for 15 min. Cells were blocked in 10% fetal calf serum for at least 30 min. Anti-␣synuclein monoclonal antibody (BD Biosciences) was used at 1:200 for 2-16 h, cells were washed and 1:200 Cy3-conjugated anti-mouse antibody (Jackson ImmunoResearch Laboratories) was added for 1 h. This anti-␣-synuclein antibody could detect both human and rat ␣-synuclein; therefore, uninduced controls were performed in parallel to compare endogenous ␣-synuclein levels. Staining was always considerably fainter in the uninduced controls. For double staining with LysoTracker Red (Molecular Probes) cells were incubated in Earle's balanced salts solution (Sigma) with 75 nM LysoTracker for 2 h at 37°C. Cells were fixed and stained with anti-␣-synuclein antibody and 1:200 fluorescein isothiocyanate-conjugated anti-mouse antibody (Jackson Immuno-Research Laboratories) was used as secondary antibody. Slides were mounted in Citifluor (Citifluor Ltd.) with 3 g/ml 4Ј,6-diamidino-2phenylindole (DAPI, Sigma). Cells were visualized using a Zeiss LSM510 confocal microscope.
Treatment with Autophagy/Proteasome Drugs-Cells were induced for 24 h and then washed twice with medium to remove doxycycline. Then cells were incubated in media containing either 10 mM 3-methyladenine (3-MA, Sigma), 200 nM bafilomycin A1 (Sigma), 0.2 mg/ml rapamycin (Sigma), 10 M lactacystin (Sigma), 10 M epoxomicin (Affinity Research Products Ltd.) or carrier controls (water or Me 2 SO (Sigma)). 3-MA and lactacystin were made up in water and rapamycin, epoxomicin, and bafilomycin were dissolved in Me 2 SO. After 24 h the medium was replaced with fresh medium plus drug, and after a further 24 h cell pellets were collected and stored at Ϫ80°C until required.
Immunogold Labeling for Electron Microscopy-Cells were induced with 2 g/ml doxycycline for 48 h. The cells were fixed in situ with 2% formaldehyde and 0.05% glutaraldehyde in 0.1 M PIPES buffer and harvested by scraping. The cells were incubated in 0.1 M PIPES buffer containing 5% bovine serum albumin and 20% polypropylene glycol, concentrated by centrifugation at 2000 rpm in a Hermle Z 160 M centrifuge (Hermle Labortechnik), and the supernatant was removed. Small droplets (5 l) of the cells were mounted onto aluminum foil and quench-frozen by plunging them into melting propane cooled in liquid nitrogen. After freezing, the cells were transferred into a Leica AFS freeze substitution unit in vials of frozen, dry methanol, containing 0.5% uranyl acetate. They were maintained at Ϫ90°C for 24 h followed by 24 h at Ϫ70°C and another 24 h at Ϫ50°C. They were infiltrated with Lowicryl HM20 over 3 days and polymerized by irradiation with UV light for 48 h. Thin sections were cut using a Leica Ultracut S and mounted on Formvar-coated nickel grids. The sections were incubated overnight in mouse anti-HA primary antibodies (Covance), diluted 1:5 in Tris-buffered saline at pH 7.4 containing 0.1% Tween 20, 0.1% Triton-X-100, 0.5% fetal calf serum, and 10% normal goat serum. The sections were washed six times in Tris-buffered saline and incubated with goat anti-mouse immunoglobulins conjugated to 10-nm gold particles (British Biocell), diluted 1:100 in the diluent for the primary antibody at pH 8.5 without added goat serum for 1 h (25). They were rinsed six times in Tris-buffered saline and twice in deionized water and stained with uranyl acetate and lead citrate before viewing in a Philips CM100 transmission electron microscope.

Establishment of Inducible ␣-Synuclein
Cell Lines-We established stable, inducible lines for human wild-type, A30P, and A53T ␣-synuclein in PC12 (rat phaeochromocytoma) cells using the Tet-On system, where addition of doxycycline switches on transgene expression. We selected two different clonal lines for each form of ␣-synuclein, on the basis of low background transgene expression and high inducibility and tested these independent lines in all our experiments.
We initially looked at ␣-synuclein localization in our induced cells. We found that ␣-synuclein was evenly distributed across the cells, often with a vesicular pattern of staining (Fig. 1A). The diffuse cytoplasmic and nuclear localization of ␣-synuclein was similar to previous observation with this protein in PC12 cells (18). The image shown is representative of ␣-synuclein localization in all our cell lines expressing wild-type, A30P, and A53T with either mitotic or differentiated cells, induced for 2, FIG. 1. ␣-Synuclein is evenly distributed across the cells, often with a vesicular pattern of staining. A, immunofluoresence image of stable, inducible PC12 cells expressing wild-type ␣-synuclein. The cells were differentiated with 100 ng/ml NGF in low serum medium for 5 days and then induced to express wild-type ␣-synuclein with 2 g/ml doxycycline. The cells were treated with doxycycline for 10 days before fixing and staining for ␣-synuclein. B, a selection of cells from the same experiment as in A comparing immunofluorescence signals in induced (I) and uninduced (U) cells. When uninduced cells were imaged by confocal microscopy using the same settings used for the induced cells, no signals were detected. However, staining was detected when the gain settings were increased (Uϩ).

FIG. 2. Degradation of ␣-synuclein after expression is switched off in the inducible cell lines.
Western blot analysis of total exogenous human ␣-synuclein in the inducible PC12 cell lines. Cells were left uninduced (U) or induced with 2 g/ml doxycycline for 24 h (I) or induced for 24 h and then washed to remove doxycycline and left for a further 72 h (OFF) in medium. The top panels show the ␣-synuclein band detected with an anti-HA antibody, and the bottom panels show ␣-tubulin as a loading control. Note that where the uninduced and induced lanes appear as separate gel "strips," these were derived from non-adjacent lanes from the same exposure of blots from single gels where lysates from uninduced and induced cells were loaded; intervening lanes were excised to simplify presentation. WT, wild-type.
␣-Synuclein Is Degraded by Both Autophagy and the Proteasome 6, or 10 days. At the light microscope level, it was unclear whether the vesicular structures were aggregates. We did not see very large aggregates characteristic of polyglutamine and polyalanine expansions (e.g. Ref. 26) for either wild-type, A30P, or A53T, even after expression of ␣-synuclein for 10 days. The staining for ␣-synuclein in uninduced cells was below the level of detection when analyzed with confocal microscopy using the same settings that gave clear signals for induced cells (Fig. 1B). However, ␣-synuclein immunoreactivity in the uninduced cells was observed when the gain was increased; this staining is likely to reflect predominantly endogenous ␣-synuclein in the cells, since there is minimal leakiness of transgene expression in uninduced cells (Fig. 2). We also examined cell death by fluorescence-activated cell sorter analysis and inspection of nuclear morphology after DAPI staining. We did not find a significant change in numbers of dead cells after induction of ␣-synuclein expression in any of the lines expressing wild-type, A30P, or A53T (data not shown). In both cycling and differentiated cells, cell death was Ͻ10% at all times studied up to 10 days of induction (assessment by DAPI staining for wild-type, A30P, or A53T).
In this study, we have assayed ␣-synuclein levels by Western blotting using an anti-HA antibody to enable specific detection of the transgene products. All blots in this paper are representative and were performed at least twice with each clonal cell line. We used our cell model to switch on expression of ␣-synuclein with doxycycline and then switch off expression by removing doxycycline from the medium, to follow ␣-synuclein degradation. Significantly lower levels of ␣-synuclein were observed 72 h after expression was switched off, compared with the time point at which doxycycline was removed (Fig. 2).

␣-Synuclein Is Degraded by the Proteasome in Our Cell
Model-We switched on ␣-synuclein expression for 24 h, then removed doxycycline and added the proteasome inhibitors epoxomicin or lactacystin for 48 h. We consistently saw an increase in ␣-synuclein levels with treatment with the proteasome inhibitors epoxomicin or lactacystin, in both cycling and differentiated cells in all cell lines (Fig. 3). We used both inhibitors as recent reports have shown that lactacystin also inhibits cathepsin A activity of the lysosome and thus may not be proteasome-specific (27,28). Addition of these drugs did not cause increased cell death or overt aggregate formation.
␣-Synuclein Is Degraded by Autophagy-We then tested whether ␣-synuclein is also degraded by autophagy using a panel of inhibitors/activators, a strategy we have used previously in the context of polyglutamine/polyalanine expansions. 3-MA is a specific inhibitor of autophagy (20) inhibiting autophagy at the sequestration stage, where a double membrane structure forms around a portion of the cytosol. Bafilomycin A1 (baf A1) is a vacuolar ATPase inhibitor that interferes with the FIG. 3. The proteasome inhibitors epoxomicin and lactacystin disrupt ␣-synuclein degradation. Cells were induced with 2 g/ml doxycycline for 24 h and then washed twice to remove doxycycline and switch off ␣-synuclein expression. The cells were then treated with 10 M epoxomicin (epoxo) or 10 M lactacystin (lact) or appropriate carrier control (c) water for lactacystin, Me 2 SO for epoxomicin, in tissue culture medium for 48 h. Cell pellets were collected, cells were lysed, and total protein was run on denaturing acrylamide gels. ␣-Synuclein was detected using an antibody to the HA tag (top panels), and loading was checked by probing for ␣-tubulin (lower panels). A, mitotic cells. B, cells induced to differentiate with 100 ng/ml NGF for 5 days before adding doxycyline to induce ␣-synuclein expression. Note that where the control and treated lanes appear as separate gel strips, these were derived from non-adjacent lanes from the same exposure of blots from single gels where lysates from control and treated cells were loaded; intervening lanes were excised to simplify presentation. WT, wild-type. Total cell lysate was analyzed for ␣-synuclein content, and loading was checked using ␣-tubulin. A, mitotic cells; B, differentiated cells. Note that where the control and treated lanes appear as separate gel strips, these were derived from non-adjacent lanes from the same exposure of blots from single gels where lysates from control and treated cells were loaded; intervening lanes were excised to simplify presentation. WT, wild-type.

␣-Synuclein Is Degraded by Both Autophagy and the Proteasome
autophagosome-lysosome fusion step (29), and rapamycin is an antifungal macrolide antibiotic that stimulates autophagy.
To test the role of autophagy on ␣-synuclein clearance in our cell lines, they were induced for 24 h, then doxycycline was removed and the drugs were added for 48 h. The trends we observed were similar in cells treated with drugs for 24 and 72 h (data not shown). Inhibiting autophagy with 3-MA and baf A1 led to an obvious increase of ␣-synuclein levels in the A53T lines (Fig. 4). These inhibitors only induced slight changes, at best, in ␣-synuclein levels in mitotic wild-type and A30P lines (e.g. 3-MA with wild-type). Rapamycin, which stimulates autophagy, increased clearance of wild-type, A30P, and A53T ␣-synuclein in mitotic cells (Fig. 4A).
␣-Synuclein Is Seen in Vesicles with Autophagic Morphology-We examined the subcellular localization of ␣-synuclein in our cell lines using immunogold electron microscopy. We used an antibody to the HA-tag on the ␣-synuclein transgene products to detect ␣-synuclein (Fig. 5). Gold labeling for ␣-synuclein was apparent over vesicles with autophagic morphology. A degree of caution is required in interpreting these data, since there are no available antibodies to allow specific labeling of mammalian autophagosomes or autolysosomes in general or in PC12 cells. Nevertheless, the vesicles we observed had morphologies consistent with those described in previous studies of autophagy (30), Huntington's disease (31), and ␣-synuclein (32). (In these studies, vesicles were identified as autophagic on similar morphologic grounds in the absence of specific antibodies.) ␣-Synuclein was observed either "free" or associated with electron dense bodies of around 100 nm in diameter. There were often two or more gold particles over these bodies, which could indicate ␣-synuclein microaggregates. Such bodies were also seen in the cytoplasm, sometimes associated with ␣-synuclein. It is possible that these electron dense bodies may be dense core granules transporting neurotransmitters.
Confocal immunofluorescence studies showed that ␣-synuclein (wild-type, A30P, and A53T) was localized inside acidic vacuoles labeled using LysoTracker Red (Fig. 6). LysoTracker Red labels acid compartments including degradative autophagic vacuoles and lysosomes. Note that ␣-synuclein was not excluded from these compartments, and speckles of ␣-synuclein immunoreactivity are clearly visible within these vacuoles. These data support our in- FIG. 5. ␣-Synuclein is seen inside vesicles with autophagic morphology. Immunogold electron microscopy analysis of ␣-synuclein subcellular localization in our stable, inducible PC12 cell lines. Cells were induced to express ␣-synuclein for 48 h and then fixed and processed for electron microscopy. ␣-Synuclein was detected using an anti-HA antibody directed to the epitope tag on the transgene. Arrows indicate ␣-synuclein in organelles with autophagic morphology. Representative images from each form of ␣-synuclein are shown (label at top left-hand corner of photograph). The upper wild-type (WT) image shows an example of ␣-synuclein associated with an electron dense vesicle of around 100 nm in diameter, inside the autophagic vesicle.
FIG. 6. ␣-Synuclein is detected inside acidic organelles. Cells were induced for 48 h with 2 g/ml doxycycline and then incubated in Earle's balanced salts solution for 2 h at 37°C to induce starvationmediated autophagy. 75 nM LysoTracker Red was added to the Earle's balanced salts solution medium at the start of the 2 h incubation to label acidic organelles. The cells were then fixed and stained with anti-␣-synuclein antibody. Immunostaining was visualized by confocal microscopy. wt, wild-type.

␣-Synuclein Is Degraded by Both Autophagy and the Proteasome
hibitor/rapamycin studies that implicate a role for the autophagylysosome pathway for ␣-synuclein degradation. DISCUSSION We have shown that ␣-synuclein is degraded by both the proteasome and autophagy in our inducible PC12 cell lines. Our Western blotting data implicating autophagy as a clearance route for this protein are supported by observations of wild-type, A53T, and A30P species in autophagic vesicles by electron microscopy. This is the first report that implicates autophagy as well as the proteasome as a route for ␣-synuclein degradation. Perturbations of autophagy had a more marked effect on the degradation of A53T ␣-synuclein, which may be due to its greater propensity to aggregate, compared with wildtype and A30P (22). This would be consistent with the recent report of Verhoef et al. (33), suggesting that soluble, but not aggregated, polyglutamine expansions could be degraded by the proteasome. This is probably because aggregated proteins cannot enter the entrance to the proteasome barrel. If this is the case, then perhaps all aggregated proteins will be preferentially routed to the autophagy-lysosome pathway for degradation.
A model where soluble ␣-synuclein is cleared by the proteasome but aggregated ␣-synuclein is preferentially cleared by autophagy may partially account for the conflicting previous data regarding its metabolism by the proteasome (15)(16)(17)(18)(19). ␣-Synuclein expression and aggregation is likely to vary between different cell types, under different transfection conditions and with different promoters etc., and this may explain the apparent discrepancies (15)(16)(17)(18)(19). Indeed, recent studies do support a role for the clearance of forms of ␣-synuclein by the proteasome (34,35).
Our data suggest that the dependence of wild-type and A30P ␣-synuclein on autophagy for degradation is less in cycling versus differentiated cells. One could speculate that this may be because the lack of cell division in the differentiated cells may enhance microaggregate formation.
An interesting feature of our data is that proteasome inhibitors were more efficient than autophagy inhibitors in stabilizing ␣-synuclein. For instance, in Fig. 4A, bafilomycin A1 has almost no effect, yet rapamycin induces very efficient clearance of ␣-synuclein. This may be compatible with increased partitioning of the ␣-synuclein pool into an autophagy pathway that may be up-regulated by the overexpression of this protein (32). Neither rapamycin nor autophagy inhibitors, like 3-MA, modulate the clearance of prototypical proteasomal substrates, like transcriptional activators (36) or misfolded intracellular cystic fibrosis transmembrane conductance regulator (37). However, alternative explanations for the lower efficacy of autophagy inhibitors compared with proteasome inhibitors on ␣-synuclein clearance may be that the autophagy inhibitors are less effective or that the half-life of total intracellular ␣-synuclein clearance through the proteasome is shorter than through the autophagy route. Indeed, this latter possibility would be compatible with the idea that aggregated ␣-synuclein is cleared mainly by autophagy, while soluble forms may be efficiently degraded by the proteasome.
We have shown that the autophagy inducer rapamycin increases clearance of all forms of ␣-synuclein. This may have therapeutic potential for treatment of Parkinson's disease as rapamycin is intended for long term use in patients. A strategy aimed at down-regulating steady-state levels of ␣-synuclein in humans may be feasible as a means of preventing PD, espe-cially if humans tolerate reductions/absence of this protein as well as mice.