α-Synuclein Aggregates Interfere with Parkin Solubility and Distribution

Parkinson disease (PD) belongs to a heterogeneous group of neurodegenerative disorders with movement alterations, cognitive impairment, and α-synuclein accumulation in cortical and subcortical regions. Jointly, these disorders are denominated Lewy body disease. Mutations in the parkin gene are the most common cause of familial parkinsonism, and a growing number of studies have shown that stress factors associated with sporadic PD promote parkin accumulation in the insoluble fraction. α-Synuclein and parkin accumulation and mutations in these genes have been associated with familial PD. To investigate whether α-synuclein accumulation might be involved in the pathogenesis of these disorders by interfering with parkin solubility, synuclein-transfected neuronal cells were transduced with lentiviral vectors expressing parkin. Challenging neurons with proteasome inhibitors or amyloid-β resulted in accumulation of insoluble parkin and, to a lesser extent, α-tubulin. Similarly to neurons in the brains of patients with Lewy body disease, in co-transduced cells α-synuclein and parkin colocalized and co-immunoprecipitated. These effects resulted in decreased parkin and α-tubulin ubiquitination, accumulation of insoluble parkin, and cytoskeletal alterations with reduced neurite outgrowth. Taken together, accumulation of α-synuclein might contribute to the pathogenesis of PD and other Lewy body diseases by promoting alterations in parkin and tubulin solubility, which in turn might compromise neural function by damaging the neuronal cytoskeleton. These studies provide a new perspective on the potential nature of pathogenic α-synuclein and parkin interactions in Parkinson disease.

Although the identification and distribution of ␣-syn-immunoreactive LBs is a useful neuropathological marker for the diagnosis of PD and LBD (7)(8)(9), recent studies suggest that abnormal neuronal accumulation of ␣-syn oligomers and protofibrils (10 -12) might be centrally involved in the pathogenesis of the neurodegenerative process in these disorders. ␣-Synuclein is an abundant synaptic protein (13) that interacts with a variety of proteins (14,15), including those involved in regulating the vesicular release of dopamine (16,17).
While the cause of sporadic PD is still unclear, familial forms of PD have been linked to mutations in various genes including ␣-syn, parkin, DJ1, PTEN-induced kinase 1 (PINK1), and leucine-rich repeat kinase-2 (LRRK2) (18 -21). Missense mutations (A30P, A53T, and E46K) and multiplications in the ␣-syn gene (22,23) that accelerate aggregation and toxic conversion of ␣-syn have been described in a few families with autosomal dominant PD (24). Mutations in parkin are the most common cause of familial parkinsonism (25)(26)(27). Several reports indicate that parkin functions as an E3 ubiquitin protein ligase and that familial-linked mutations in parkin disrupt its ligase activity (28,29) and de-stabilize its ubiquitin-like domain (30). In sporadic forms of PD and LBD, parkin accumulates in the insoluble fraction (31). In addition to incorporating ubiquitin to a number of substrates (32) such as the aminoacyl tRNA synthetase cofactor p38/JTV-1 (p38), ␣-tubulin, cell division control-related protein-1 (CDCrel-1, also known as the septin, Sept5), glycosylated ␣-syn (33), Parkin-associated endothelin receptorlike receptor (Pael-R) (34), and synphilin-1, parkin ubiquitinates itself as an early step in its proteasome-mediated degradation process (29,35,36). Of these substrates, more recent studies in parkin knock-out mice have shown that p38 is the most important parkin substrate (37); however, p38 is unlikely to be a target of parkin-mediated degradation because a previous study showed that p38 is largely mono-ubiquitinated in the presence of parkin, and poly-ubiquitinated p38 is difficult to detect (38).
Based on the genetic evidence and the known role of parkin as a ubiquitin ligase, most studies have focused at investigating the role of parkin alterations and proteasomal dysfunction on ␣-syn accumulation in the pathogenesis of PD (28,39). However, recent evidence suggests that mutations (40) and stress factors might lead to the accumulation of parkin, and that translocation of parkin into the insoluble fraction might impair neuronal cell function (41)(42)(43). Moreover, because parkin and ␣-syn associate (44) and co-localize in LBs (31), and proteasomal alterations are often found in patients with sporadic PD and LBD (45,46), then it is also possible that ␣-syn might interfere with parkin by promoting aggregation and accumulation in the insoluble fraction.
In this context, we explored the pathological interactions between ␣-syn and parkin. We found that ␣-syn accumulation interfered with parkin and ␣-tubulin solubility and distribution, leading the cytoskeletal alterations and neuronal dysfunction. This study provides a new perspective about the nature of the interactions between ␣-syn and parkin.

EXPERIMENTAL PROCEDURES
Tissue Culture-Briefly, as previously described (47), B103 neuronal cell lines stably transfected with either human ␣-syn, ␤-syn, or empty vector (pCEP4; Invitrogen, Carlsbad, CA) were used. These cells were routinely maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Irvine Scientific, Irvine, CA) in the presence of 50 g/ml Hygromycine B (Calbiochem, San Diego, CA) in a 5% CO 2 atmosphere. This neuronal cell line, derived from rat neuroblastoma, was selected because ␣-syn overexpression results in the formation of discrete insoluble aggregates in the cell body and processes accompanied by reduced neurite outgrowth (47), mimicking another important aspect of LBD, namely compromised axonal plasticity (48).
Primary rat cortex (P0) was dissected, digested by trypsin, mechanically dissociated, and plated at 8 ϫ 10 5 -1 ϫ 10 6 cells per dish/well under serum-free conditions in neurobasal A media supplemented with B27. Primary cultured neurons were maintained for a minimum of 14 divisions prior to co-transfection with lentiviruses expressing parkin, ␣-syn, or GFP.
Antibodies-Mouse monoclonal anti-␣-tubulin was obtained from Sigma. Rabbit polyclonal anti-ubiquitin, mouse monoclonal anti-actin, and rabbit polyclonal anti-␤-syn were from Chemicon (Temecula, CA). Mouse monoclonal anti-␣-syn was from BD Transduction Laboratories (Lexington, KY). Mouse monoclonal anti-c-Myc was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal anti-parkin was from Cell Signaling Technology (Beverly, MA). Mouse monoclonal anti-A␤ was obtained from Signet Laboratories (Dedham, MA).
Immunoblot and Immunoprecipitation Analyses-Immunoblot analysis was performed as previously described (50). Briefly, cells were lysed in buffer A (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM NaVO 4 , 50 mM NaF, with protease inhibitors (Roche Applied Science)) containing 1% Triton X-100. After incubation at 4°C for 20 min, cell lysates were separated into detergent-soluble and -insoluble fractions by centrifugation at 15,000 ϫ g at 4°C for 20 min. Frozen tissues from the temporal cortex of six LBD and five age-matched control cases (Table 1) were analyzed for parkin accumulation in the insoluble fraction. Cases were obtained from the UCSD Alzheimer Disease Research Center. For each case, 100 g of frozen superior temporal neocortical sample were homogenized in 3ϫ volume buffer A and centrifuged at 1,000 ϫ g for 15 min. 1% Triton X-100 was added to supernatants, and samples were separated into soluble and insoluble fractions by centrifugation at 15,000 ϫ g for 20 min. Fractions were then used for immunoblotting.
The detergent-soluble and -insoluble fractions were resolved with SDS-PAGE and transferred to a polyvinylidene difluoride membrane filter (Immobilon P; Millipore, Bedford, MA) The membranes were blocked with Tris-buffered saline (25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.2% Tween-20) containing 5% skim milk, flowed by incubation with primary antibody in Trisbuffered saline containing 5% skim milk. After washing with Tris-buffered saline, proteins were visualized by enhanced chemiluminescence and analyzed with a Versadoc XL imaging apparatus (Bio-Rad). Analysis of actin levels was used as a loading control.
Immunoprecipitation assays were carried out essentially as previously described (51). Cells were lysed in buffer A containing 1% Triton X-100. The lysates were then centrifuged for 20 min at 12,000 rpm, and the protein concentrations were determined with a protein assay kit (Bio-Rad). 300 g of the supernatants were incubated with 1 g of antibody for parkin, ␣-tubulin, or ␣-syn overnight at 4°C, and then immunocomplexes were adsorbed to protein A-or G-Sepharose beads (Amersham Biosciences). After washing extensively with buffer A contain- ing 1% Triton X-100, samples were heated in SDS sample buffer for 5 min and subjected to immunoblot analysis. Immunocytochemistry and Laser-scanning Confocal Microscopy-Cells were grown on poly-L-lysine-coated glass coverslips to 70% confluence, and fixed for 20 min in 4% paraformaldehyde. The fixed cells were first incubated overnight at 4°C with primary antibody for ␣-syn, parkin, or ␣-tubulin. The next day, antibodies were detected with the Tyramide Signal Amplification-Direct (Red) system (NEN Life Sciences, Boston, MA) and sections were imaged with a Zeiss 63ϫ (N.A. 1.4) objective on an Axiovert 35 microscope (Zeiss, Germany) with an attached MRC1024 laser scanning confocal microscope (LSCM) system (Bio-Rad) (52).
Small Interfering RNA (siRNA) Studies-Both control siRNA and siRNA for ␣-syn were purchased from Santa Cruz Biotechnology. Transfection of the siRNA (final concentration, 100 nM) was carried out using OligofectAMINE (Invitrogen) for 72 h with lentivirus transfection. Cells were then lysed and used for immunoblot analysis.
In Vitro Ubiquitination Assay-For immunoprecipitation, parkin was bound to anti-Myc IgG-linked protein G beads from extracts of 293T cells transfected with pcDNA3.1 (ϩ) Mycparkin. To prepare aggregates of ␣-syn, recombinant ␣-syn was incubated in 20 l of distilled water at 65°C for 24 h essentially as previously described (53). Reactions were performed in a 50-l mixture containing 50 mM Tris-HCl, pH 7.5, 2 mM MgCl 2 , 2 mM ATP, 10 g of ubiquitin (Boston Biochem, Bos-ton, MA), 100 ng of E1 (Calbiochem, San Diego, CA), 200 ng of UbcH7 (Boston Biochem, Boston, MA), immunoprecipitated Myc-tagged parkin, and 5 g of purified ␣-tubulin (Cytoskeleton Inc, Denver, CO) with 0.25 nmol of bovine serum albumin, 0.25 nmol of recombinant soluble or aggregated ␣-syn. The reactions were carried out for 2 h at 37°C before terminating with an equal volume of 2ϫ SDS sample buffer. The reaction mixtures were then analyzed by immunoblotting for ubiquitin antibody.
Statistical Analyses-Analyses were carried out with the StatView 5.0 program (SAS Institute Inc., Cary, NC). Differences among means were assessed by one-way analysis of variance with post-hoc Dunnetts test. Comparisons between two groups were done with the unpaired two-tailed Student's t test. All values in the figures are expressed as means Ϯ S.E., and the null hypothesis was rejected at the 0.05 level.

␣-Synuclein Promotes the Abnormal Accumulation of Parkin and Tubulin into the Detergent-insoluble Fraction of Neurons-
To investigate the effects of ␣-syn on parkin accumulation and cellular distribution, stably transfected B103 neuroblastoma cells expressing ␣-syn-, ␤-syn-, or vector control were infected with a lentiviral vector expressing parkin. For this, we first analyzed the levels of ␣and ␤-syn in the presence or absence of MG132 (Fig. 1A) because previous studies have shown that proteasomal inhibition facilitates the pathological effects of ␣-syn For all panels, cells were lysed in buffer containing 1% Triton X-100 and fractionated into detergent-soluble and -insoluble fractions, followed by immunoblot analysis. A, immunoblot analysis of the soluble fraction of synuclein-transfected B103 cells that were infected with lentivirus (lenti) expressing parkin or GFP and treated with 10 M MG132 for 18 h. B, immunoblot analysis with an antibody against parkin shows that in ␣-syn-transfected cells, parkin accumulated in the insoluble fraction in the presence of MG132. C, parkin accumulation in the insoluble fraction of ␣-syn-expressing neuronal cells infected with lenti-Parkin and amyloid precursor protein (APP)sw (high level of A␤ 1-42 expression) but not in cells infected with APPsw(Ϫ) (a mutant that cannot be cleaved and therefore no A␤ is generated). A di-ubiquitinated (di-Ub) form of parkin was also detected in blots developed with long exposures. D, blocking ␣-syn expression with a specific siRNA resulted in decreased parkin accumulation in the insoluble fraction. Synuclein-transfected B103 cells were infected with lenti-parkin or GFP and treated with siRNA to block ␣-syn expression. Cells were subsequently treated with MG132. Cell lysates were probed by Western blot for ␣-syn and parkin. E, immunoblot analysis with an antibody against ␣-tubulin showed that in cells expressing ␣-syn, ␣-tubulin accumulates in the insoluble fraction in the presence of MG132. For each blot, actin levels were used as a loading control. (54). Immunoblot analysis showed that both ␣and ␤-syn levels were increased in the presence of MG132 compared with controls (Fig. 1A). To assess the effects of ␣-syn on parkin accumulation under conditions of cellular stress, parkin levels were determined. Western blot analysis showed that treatment with MG132 increased the levels of parkin (Fig. 1B). Remarkably, this effect was enhanced by ␣-syn but not by ␤-syn. Next, we examined whether the excess parkin might be translocated into detergent-insoluble fractions. For this, cells were lysed with Triton X-100 and separated into detergent-soluble and -insoluble fractions. Immunoblotting analysis showed that in the presence of MG132, levels of parkin in the detergent-insoluble fraction were increased in ␣-syn-transfected cells, compared with vector-or ␤-syn-transfected cells (Fig. 1B). Similarly, insoluble parkin accumulated in ␣-syn-transfected cells when cells were treated with lactacystin, another proteasome inhibitor (data not shown), or infected with a lentiviral vector expressing the AD-related mutant amyloid precursor protein (APPsw) (Fig. 1C). Actin levels were not affected under either condition (Fig. 1C). To confirm the effects of ␣-syn on parkin, siRNA experiments were performed. When cells were transfected with siRNA targeting ␣-syn mRNA, expression levels of ␣-syn significantly decreased in ␣-syn-transfected cells both in the presence and absence of MG132 (Fig. 1D). Under these experimental conditions, immunoblotting analysis revealed that the amount of insoluble parkin was dramatically decreased by ␣-syn siRNA compared with control siRNA in ␣-syn-transfected cells, whereas insoluble parkin was not changed by ␣-syn siRNA in vector-transfected cells (Fig. 1D).
Recent studies have shown that parkin-mediated ubiquitination promotes degradation of target proteins such as ␣-tubulin (55). Therefore, we reasoned that, similarly to parkin, ␣-syn might interfere with ␣-tubulin clearance. Immunoblot analysis showed that in the presence of MG132, the levels of insoluble ␣-tubulin were considerably increased in ␣-syn-transfected cells even when parkin was overexpressed (Fig. 1E).
We next examined whether ␣-syn accumulation may promote similar effects on parkin and ␣-tubulin in primary neurons. For this purpose, rat cortical primary neurons were coinfected with lentiviral vectors expressing parkin and ␣-syn. Immunoblot analysis showed that in the presence of MG132, ␣-syn expression promoted accumulation of insoluble parkin and, to a lesser extent, ␣-tubulin in the detergent-insoluble fractions compared with lenti-GFP control (Fig. 2). These effects were more pronounced in primary neurons infected with a lentiviral vector expressing familial-linked A53T mutant ␣-syn (Fig. 2). ␤-Synuclein had no effect on insoluble parkin or ␣-tubulin accumulation in primary neurons (not shown). Together, these results support the possibility that ␣-syn might promote abnormal accumulation of parkin and ␣-tubulin in the detergent-insoluble fraction that might affect the neuronal cytoskeleton.
Parkin and ␣-Tubulin Are Increased in the Insoluble Fraction in the Brains of LBD-To investigate whether ␣-syn accumulation in patients with LBD results in similar alterations in parkin as observed in neuronal cell lines, immunoblot analysis with detergent-insoluble fractions was performed. This analysis showed that in patients with LBD, levels of parkin and ␣-tubu-lin in the detergent-insoluble fraction were significantly increased compared with unimpaired controls (Fig. 3, A-C). In the detergent-soluble fraction, no differences were detected between the two groups (Fig. 3, A-C).
Double immunocytochemical analysis confirmed that compared with controls (Fig. 3, D-F), in LBD ␣-syn and parkin were co-localized in the neocortical LBs and neurites (but not in synapses) (Fig. 3, G-I). These results support the notion that aggregated ␣-syn might lead to an increase of insoluble parkin and ␣-tubulin in vivo.
Parkin Self-ubiquitination and Ubiquitination of ␣-Tubulin Are Altered by ␣-Syn Overexpression-It has been previously shown that parkin self-ubiquitinates and that ubiquitinated parkin is rapidly degraded by the proteasome (29,35,36,56). Therefore we reasoned that ␣-syn might interfere with parkin ubiquitination, resulting in reduced solubility. To examine this possibility, parkin immunoprecipitates were analyzed with an antibody against ubiquitin. This study showed that levels of ubiquitinated parkin were decreased in ␣-syn-transfected cells treated with MG132 compared with controls (Fig. 4, A and B). Consistent with previous reports (57-59), high molecular weight ubiquitinated parkin was only detected in the presence of a proteasome inhibitor.
Because insoluble ␣-tubulin also accumulates in cells co-expressing ␣-syn and parkin, it is possible that ubiquitination of ␣-tubulin might also be altered. To test this possibility, cell extracts were immunoprecipitated with an ␣-tubulin monoclonal antibody, followed by analysis with an antibody against ubiquitin (Fig. 4C). Immunoblot analysis showed that in the presence of MG132, ubiquitinated ␣-tubulin is reduced in ␣-syn-transfected cells when compared with vector-and ␤-syn-transfected cells co-transfected with parkin (Fig. 4, C and  D). Thus, ␣-syn overexpression reduces ubiquitinated ␣-tubulin and this might lead to accumulation of insoluble ␣-tubulin. To confirm that total cellular levels of ubiquitinated proteins were not affected by ␣-syn overexpression, total cell extracts were subjected to immunoblot analysis for ubiquitin. These results demonstrated that synuclein expression did not alter the total levels of ubiquitinated proteins (Fig. 4, E and F). Taken together, these results suggest that under stress conditions ␣-syn might reduce ubiquitinated parkin and ␣-tubulin by interfering with parkin ubiquitin ligase activity.

␣-Synuclein and Parkin Co-immunoprecipitate in Neuronal
Cell Lines-The alterations in parkin and ␣-tubulin solubility in ␣-syn-overexpressing cells might be related to direct interactions between ␣-syn and parkin. Because ␣-syn and parkin colocalize in LBs, and we have shown that ␣-syn promotes parkin and ␣-tubulin accumulation in the detergent-insoluble fraction of brain homogenates, we hypothesized that ␣-syn might interact with parkin. First, to investigate the cellular localization of ␣-syn and parkin, double immunolabeling experiments were performed (Fig. 5, A-L). These studies showed that in the absence of MG132, ␣-syn and parkin were colocalized to the cytoplasm and neuritic processes with a diffuse pattern (Fig. 5, G-I). In contrast, in the presence of MG132, increased parkin immunoreactivity was detected in ␣-syn-transfected cells. Furthermore, both of these molecules were colocalized to inclusion-like structures (Fig. 5, J-L).
To confirm the association between ␣-syn and parkin in ␣-syn-transfected neuronal cells, co-immunoprecipitation experiments were performed. For this purpose, cell extracts were subjected to immunoprecipitation with an anti-␣-syn antibody, followed by immunoblotting with an antibody against parkin. Compared with vector and ␤-syn-transfected cells, ␣-syn was efficiently immunoprecipitated in ␣-syn-transfected cells, and parkin was found to consistently co-immunoprecipitate with ␣-syn in the presence of MG132 (Fig. 5M). Because of the inherent challenge in demonstrating the specificity of immunoprecipitation of insoluble proteins, co-immunoprecipitation of parkin and ␣-syn was confirmed by immunoprecipitation with an anti-parkin antibody, followed by immunoblotting with an antibody against ␣-syn. This confirmed that parkin and ␣-syn co-immunoprecipitate in the presence of MG132 (Fig. 5M). Parkin and ␣-syn also coimmunoprecipitated in the absence of MG132, albeit to a much lesser extent (Fig. 5M). Taken together, these results support the possibility that interactions between ␣-syn and parkin lead to the formation of insoluble aggregates that might directly damage the neuronal cytoskeleton rather than impairing parkin function.
␣-Synuclein Induces Disruption of Neurite Outgrowth of Neuronal Cells Expressing Parkin-Because abnormal accumulation of insoluble ␣-tubulin was observed in neuronal cells overexpressing ␣-syn and parkin, and previous studies have shown that ␣-syn promotes cytoskeletal alterations (60, 61), we then hypothesized that ␣-syn, in combination with parkin, might affect the neuronal microtubule structure. To examine the pattern of the cellular microtubule network in synuclein-and parkin-transfected cells, we performed immunohistochemical studies for ␣-tubulin. A well organized ␣-tubulin-immunoreactive microtubule network was detected under baseline conditions (Fig. 6, A-C and M); in contrast, treatment with MG132 resulted in neurite retraction in ␣-syn-transfected cells, but not in vector or ␤-syn-transfected cells (Fig. 6, D-F, and M). In combination with parkin and MG132, the cytoskeletal alterations were exacerbated in the ␣-syn-transfected cells compared with vector control or ␤-syn (Fig. 6, G-L, and M).

DISCUSSION
Although both ␣-syn and parkin are linked to the molecular pathogenesis of familial PD (39), whether ␣-syn aggregation may affect parkin in PD and LBD has not been previously investigated. For the present study, we found that in the presence of stress conditions (namely proteasomal inhibition with MG132  's t test). In the LBD cases, parkin accumulated in the insoluble fraction. D-I, double immunocytochemical and laser scanning confocal microscopy analysis of sections from the temporal cortex labeled with antibodies against parkin (red) and ␣-synuclein (␣-syn, green). Co-localization of both signals is displayed in a merged image (yellow). D-F, in control cases, parkin and ␣-syn immunoreactivity was primarily associated with the neuropil. G-I, in LBD cases, parkin and ␣-syn immunolabeling co-localized in the Lewy bodies (arrows). Scale bar, 20 m. or lactacystin, or A␤ treatment), ␣-syn promoted parkin and, to a lesser extent, ␣-tubulin accumulation in the insoluble fraction of neurons and resulted in cytoskeletal alterations.
These findings are consistent with recent studies showing that pro-oxidants (nitric oxide, iron, hydrogen peroxide), neurotoxins (1-methyl-4-phenylpyridinium, rotenone, paraquat), and dopamine analogs promote similar alterations in parkin solubility (42). Moreover, alterations in parkin solubility were associated with an increased tendency of parkin to aggregate and precluded parkin from exercising its protective functions (42). Similarly, mutations associated with familial PD interfere with parkin solubility, localization, and ubiquitination properties (54,62). The mechanisms through which ␣-syn aggregates in the presence of stress factors might promote parkin accumulation are not completely understood. However, because ␣-syn aggregates trigger oxidative stress (63,64) that in turn could further stimulate ␣-syn aggregation (and ␣-syn oligomers are more active than monomers at promoting parkin aggregation) then it is possible that ␣-syn-mediated oxidative stress might be in part responsible for the parkin alterations. In addition, direct interactions between ␣-syn and parkin might also promote parkin accumulation. Supporting this possibility, co-immunoprecipitation studies have shown that ␣-syn and parkin interact (44). Moreover, mutant ␣-syn (that is more prone to aggregate) is more effective at stimulating parkin accumulation, while silencing the expression of ␣-syn results in a considerable reduction in the accumulation of insoluble parkin. In addition, recent studies have shown that co-expression of mutant ␣-syn and parkin in Caenorhabditis elegans enhances parkin aggregation and mislocalization (65). Together, these findings suggest that ␣-syn, in conjunction with PD-associated stress factors, promote accumulation of parkin and ␣-tubulin that might lead to neurodegeneration.
The mechanisms through which the effects of ␣-syn on parkin and ␣-tubulin might result in neuronal damage are under investigation. One possibility is that ␣-syn might impair parkin ubiquitin ligase function. This is an important aspect that will be focused on in the future. Alternatively, and in view of recent studies showing that PD-related stress factors promote parkin insolubility (41)(42)(43), the parkin aggregates might be toxic and disrupt the neuronal cytoskeleton by compromising ␣-tubulin solubility.
In the present study, the effects of ␣-syn on parkin and ␣-tubulin solubility were observed primarily in the presence of proteasomal inhibitors or A␤. ␣-synuclein exerted more prominent effects on parkin than on tubulin solubility. Because inhibition of proteasomal activity by itself promotes accumulation of insoluble parkin and interferes with parkin activity (42), this suggests that stress conditions may be necessary for the pathological interactions between ␣-syn and parkin to occur. This is consistent with recent studies showing that proteasomal inhibitors interfere with parkin solubility (42). Furthermore, during the aging process and in the presence of A␤ (66), proteasomal activity is reduced (67) and insoluble ␣-syn and parkin levels are increased (42,68). Therefore, the combination of oxidative stress, neurotoxins, and mutations associated with familial parkinsonism might provide the conditions to facilitate the pathological interactions between ␣-syn and parkin.
There are several possible ways through which cellular stress and ␣-syn aggregates might promote parkin accumulation. For example, recent studies have shown that nitrosative stress leads to S-nitrosylation of parkin, which in the long term results in a dramatic decrease in E3 ligase ubiquitin proteasome degradative activity (69,70). In agreement with this possibility, previous studies have shown that insults associated with a PD-like phenotype induced similar alterations in microtubules resulting in neurodegeneration (71,72), and the neurons in patients with PD and LBD exhibit significant cytoskeletal alterations (61,73,74). Moreover, a recent study showed that ␣-syn aggregates impair microtubule-dependent trafficking (60).
It was previously shown that wild-type parkin reduces the toxicity induced by dopamine in SH-SY5Y cells (75), by hydrogen peroxide, hydroxynonenal, and 1-methyl-4-phenylpyridinium in NT-2 cells (76), and is capable of rescuing the phenotype associated with ␣-syn overexpression in the rat (77, 78) FIGURE 5. Analysis of ␣-syn and parkin co-localization in synuclein-transfected neuronal cell lines. A-L, double immunocytochemical and laser scanning confocal microscopy analysis with antibodies against ␣-syn (green) and parkin (red) in synuclein-transfected B103 neuronal cells. Yellow pixels in the merged images indicate co-localization of both signals. A-C, parkin expression in control cells that were infected with lentiviral (lenti)-parkin. D-F, ␣-synuclein expression in stably transfected ␣-syn cells that were infected with an empty lentiviral vector. G-I, co-localization of ␣-syn and parkin in ␣-syn stably transfected cells that were infected with lenti-parkin. J-L, colocalization between ␣-syn and parkin was enhanced by treatment with MG132, and occurred in inclusion-like structures (arrows). M, co-immunoprecipitation of ␣-syn with parkin in synuclein-transfected B103 cells that were infected with lenti-parkin and subsequently treated with MG132. Cell extracts were immunoprecipitated (IP) with anti-␣-syn antibody and probed by Western blot for ␣-syn (upper panel) and parkin (middle panel), or IP with antiparkin antibody and probed by Western blot for ␣-syn (lower panel). Scale bar, 10 m. and in Drosophila (79). In contrast, in our model system, ␣-syn overexpression altered parkin solubility and resulted in cytoskeletal alterations. The differences in models used and the relative levels and efficiency in the expression of ␣-syn and parkin with the viral vectors might explain the apparent discrepancy between our results and those of others. In addition, the effects of ␣-syn aggregates on parkin were detected primarily in the presence of proteasomal inhibitors, and compared with wild-type ␣-syn, mutant A53T ␣-syn had a more prominent effect on parkin. Moreover, although previous studies have shown that parkin protects against a variety of neurotoxins (76,80), it has also been shown that parkin has no effects in delaying cell death associated with proteasome inhibition (76), suggesting that increased levels of cellular stress are necessary for the pathological interactions between parkin and ␣-syn to take place.
These findings have important implications for understanding the pathogenesis of PD and LBD as they provide a potential mechanistic link through which two molecules associated with familial PD-␣-syn and parkin might interact to promote neurodegeneration. However, it remains controversial as to what extent parkin alterations precede or follow ␣-syn pathology in sporadic PD and LBD, and whether in familial cases with mutations in the parkin gene, the PD phenotype might emerge independently of ␣-syn and vice versa.
In conclusion, our results indicate that ␣-syn aggregates induce abnormal accumulation of parkin and, to a lesser extent, ␣-tubulin, resulting in cytoskeletal pathology. Moreover, these studies provide a new model for the pathological interactions between ␣-syn and parkin that might help better understand the pathogenesis of PD and LBD.