Proteolytic Cleavage of Extracellular Secreted α-Synuclein via Matrix Metalloproteinases*

Although α-synuclein is the main structural component of the insoluble filaments that form Lewy bodies in Parkinson disease (PD), its physiological function and exact role in neuronal death remain poorly understood. In the present study, we examined the possible functional relationship between α-synuclein and several forms of matrix metalloproteinases (MMPs) in the human dopaminergic neuroblastoma (SK-N-BE) cell line. When SK-N-BE cells were transiently transfected with α-synuclein, it was secreted into the extracellular culture media, concomitantly with a significant decrease in cell viability. Also the addition of nitric oxide-generating compounds to the cells caused the secreted α-synuclein to be digested, producing a small fragment whose size was similar to that of the fragment generated during the incubation of α-synuclein with various MMPs in vitro. Among several forms of MMPs, α-synuclein was cleaved most efficiently by MMP-3, and MALDI-TOF mass spectra analysis showed that α-synuclein is cleaved from its C-terminal end with at least four cleavage sites within the non-Aβ component of AD amyloid sequence. Compared with the intact form, the protein aggregation of α-synuclein was remarkably facilitated in the presence of the proteolytic fragments, and the fragment-induced aggregates showed more toxic effect on cell viability. Moreover, the levels of MMP-3 were also found to be increased significantly in the rat PD brain model produced by the cerebral injection of 6-hydroxydopamine into the substantia nigra. The present study suggests that the extracellularly secreted α-synuclein could be processed via the activation of MMP-3 in a selective manner.

Parkinson's disease (PD) 1 is pathologically characterized by the progressive degeneration and death of dopaminergic neurons in the substantia nigra and by the presence of cytoplasmic inclusions known as Lewy bodies (LBs). There is a wealth of evidence to suggest that the abnormal accumulation and aggregation of ␣-synuclein play a primary role in PD pathogenesis (1,2). LBs are composed of polymers made up of full-length ␣-synuclein proteins (3), and the formation of ␣-synuclein-positive inclusions is also seen in other synucleinopathies, such as diffuse Lewy body disease or multiple system atrophy (4). In addition, point mutations of the ␣-synuclein gene are associated with early-onset familial PD (5,6).
Considerable changes in the organization of the extracellular matrix occur in neurodegeneration. Remodeling of the extracellular matrix and its degradation are controlled by matrix metalloproteinases (MMPs), a family of extracellular soluble or membrane-bound and structurally related zinc-dependent endopeptidases (7,8). It has also been suggested that MMPs play a role in the pathogenesis of both acute and chronic neurodegenerative disorders such as stroke (9), Alzheimer's disease (AD) (10 -13), and multiple sclerosis (14).
There is increasing evidence to suggest that oxidative or nitrative injury is implicated in the pathogenesis of several neurodegenerative disorders, including PD (15,16). For example, there is ample evidence that oxidative injury is significantly enhanced in PD (17). Oxidative stress is also known to be an early feature of PD, because the oxidation-dependent aggregation of proteins has been observed in the form of advanced glycation end products in Lewy bodies at a time when no phenotype of a neurodegenerative disorder was evident. In addition, experimental models of PD show higher levels of free radical generation and the degeneration of dopaminergic neurons (18).
One consequence of increased oxidative stress may be the activation of MMPs, which are synthesized primarily by astrocytes, microglia, and neurons (19). Therefore, the excessive expression of MMPs may result in neuronal damage. There is accumulating evidence to suggest that MMPs are involved in the pathogenesis of PD (20). Inducers of MMP expression and activity, such as cytokines, nitric oxide, reactive oxygen species, and metabolites of the arachidonic pathway (21)(22)(23), are also implicated in the pathophysiology of PD. Based on the hypothesis that the generation of free radicals stimulates the expression/activity of MMPs, and that free radical-mediated damage is a key mediator of neuronal damage in PD, it is tempting to speculate that MMPs are involved in the pathophysiology of PD.
The cellular and physiological function and processing of ␣-synuclein are still poorly known. This protein was originally found to be located in presynaptic terminals and was thought to play a role in neuronal plasticity (24). However, several pieces of evidences have subsequently emerged that indicate that intact ␣-synuclein plays a functional role in non-cytoplasmic locations, such as nucleus and extracellular space, and in its processing. In the present study, we examined whether ␣-synuclein proteins could be secreted extracellularly in the human dopaminergic neuroblastoma cell (SK-N-BE) line. Moreover, we attempted to identify the cleaved ␣-synuclein fragments present in extracellular space and the possible functional relationship between several forms of MMPs and the processing mechanism of ␣-synuclein. Herein we show for the first time that the extracellularly secreted ␣-synuclein is selectively processed via the activation of MMP-3 upon the stimulation of dopaminergic neuronal cells with nitric oxide.

EXPERIMENTAL PROCEDURES
Materials-Calf serum, Dulbecco's modified Eagle's medium, penicillin, and streptomycin were purchased from Invitrogen, and MMP inhibitor II was obtained from Calbiochem. The anti-MMP-3 and ␣-synuclein antibodies were purchased from BD Transduction Laboratory. The plasmid used to express wild type ␣-synuclein was provided by R. Jakes (Medical Research Council Laboratory of Molecular Biology, Cambridge, UK). To prepare bacterially recombinant ␣-synuclein, ␣-synuclein proteins were either purchased from ATGen (Seongnam-si, Gyeonggi-do, Korea) or purified as described elsewhere in the literature (25). All other chemicals were purchased from Sigma.
Cell Culture and Transfection-Human neuroblastoma SK-N-BE cells were cultured in Dulbecco's modified Eagle's medium containing 10% calf serum and maintained at 37°C. The cells were transfected with various expression vectors using Lipofectamine plus reagent (Invitrogen), according to the manufacturer's instructions. The total amount of DNA in each individual transfection experiment was adjusted by using parental empty vector DNA. The cells were cultured for at least 24 h after transfection, and used in the subsequent immunoprecipitation, Western blot analysis, and cell viability assay experiments.
Assessment of Cell Viability-Quantitation of cell survival was performed using the tetrazolium salt extraction method (28). 62.5 l of the 5 mg/ml stock solution of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was added to each well of a 24-well plate in which 250 l of medium was previously placed. After 2-h incubation at 37°C, 250 l of extraction buffer containing 20% SDS and 50% N,Ndimethylformamide at pH 7.4 was added. After incubating the reagents overnight at 37°C, the optical density was measured at 570 nm using a SpectraMAX340 enzyme-linked immunosorbent assay reader (Molecular Devices, Sunnyvale, CA), using extraction buffer as a standard. Statistical analyses were completed with the aid of the StatView II program for Macintosh computers (Abacus Concepts, Berkeley, CA). All data were analyzed by one-way analysis of variance and preplanned comparisons with the control were performed by means of Dunnett's t statistic.
Nitrite Content Measurement-Nitrite was quantified by means of Griess reaction (29). HCl (4 M) was added to the supernatant and left to stand for 10 min, and then 2 mg/ml sulfanilic acid and 1 mg/ml N-(1naphthyl)ethylenediamine were added. After incubation for 30 min, the absorbance was measured using a spectrophotometer at a wavelength of 550 nm. The absorbance of each sample was compared with that of standard sodium nitrite solutions.
Analysis of Protein Aggregation-The protein aggregation of ␣-synuclein was monitored with both turbidity and thioflavin-T binding fluorescence. Following ultracentrifugation at 100,000 ϫ g for 10 min to remove any protein clumps that could act as a nucleation centers during the aggregation process, ␣-synuclein (138 M) was incubated at room temperature with continuous shaking. The turbidity caused by protein aggregation was examined with absorbance at 405 nm. The extent of amyloid formation, on the other hand, was evaluated with thioflavin-T binding fluorescence (30). During incubation, small aliquots (20 l) were combined with 5 M thioflavin-T in 50 mM glycine at pH 8.5 to produce a final volume of 100 l and, in each case, the fluorescence was measured at 482 nm with an excitation at 446 nm (FL500 Microplate Fluorescence Reader, Bio-Tek Instruments). The morphological appearances of the protein aggregates were also examined with a transmission electron microscope (JEM1010, JEOL). Aliquots (5 l) of the aggregates were adsorbed onto a carbon-coated copper grid (300 mesh) and air-dried for 1 min. After negative staining with 2% uranyl acetate for another 1 min, the aggregates were observed under the electron microscope.
Preparation of Pro-MMPs and MMP Catalytic Domains-The recombinant human pro-MMP-2 was expressed in Sf9 cells by infection of the 72Gel baculovirus and purified by gelatin-agarose column chromatography as described previously (31). The human full-length MMP-9 cDNA (a generous gift from Dr. G. Goldberg) was cloned into pBlueBac4 (Invitrogen), and 92Gel baculovirus was produced using Bac-N-Blue TM Baculovirus Expression System (Invitrogen). Expression and purification of pro-MMP-9 was performed by the same method as in pro-MMP-2. Pro-MMP-1 (Calbiochem) and pro-MMP-3 (Oncogene) were purchased. The catalytic and hinge domains of human MMP-14 (Tyr 112 -Ile 318 ), containing an N-terminal methionine and C-terminal hexa-histidines, were expressed as an inclusion body in Escherichia coli. A functional catalytic domain of MMP-14 (cMMP-14) was obtained by refolding of the recombinant MMP-14 polypeptide (39). A functional catalytic domain of the human MMP-3 was prepared by refolding of the catalytic domains (cMMP-3; Phe 100 -Pro 273 ) expressed as inclusion body in E. coli (40).
In Vitro Cleavage of the Purified ␣-Synuclein with MMPs-1 g of purified ␣-synuclein (ATGen) was incubated with various concentrations of MMPs and for various time intervals in an MMP-reaction buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 5 mM CaCl 2 , 0.5 mM ZnCl 2 ) at 37°C. The reactions were terminated by addition of SDS sample buffer, and the reaction products were subjected to SDS-PAGE.
Determination of Cleavage Sites of ␣-Synuclein-To sequence the N-terminal amino acids of the ␣-synuclein fragments, the cMMP-3cleaved ␣-synuclein mixture was separated in 15% Tris-Tricine SDS-PAGE followed by electroblotting onto a polyvinylidene difluoride membrane in 10 mM CAPS, pH 11, containing 10% methanol (41). Proteins were visualized by staining with Coomassie Brilliant Blue R-250, and the bands of interest were cut out and sequenced by Edman degradation using the 477A protein sequencer (Applied Biosystems, Framingham, MA) in the Tufts Core Facility (Tufts University). To determine molecular mass of the ␣-synuclein fragments, the cMMP-3-cleaved ␣-synuclein mixture was passed through a self-packed POROS 20 R2 (Applied Biosystems) cartridge to capture the protein. A solution of 0.1% trifluoroacetic acid was used to wash unbound components from the cartridge, and the proteins were eluted with 10 mg/ml ␣-cyano-4hydroxycinnamic acid dissolved in 0.1% trifluoroacetic acid, 70% acetonitrile. The elutes were dispensed onto a MALDI-TOF sample plate. The sample was allowed to air-dry at room temperature and then subjected to MALDI-TOF analysis. Mass analysis of the fragments was performed on a Voyager-DE STR MALDI-TOF mass spectrometer (Applied Biosystems) in the linear mode using a nitrogen laser (337 nM). Mass spectra were collected in the positive ion mode using an acceleration voltage of 25 kV and a delay of 800 ns. The grid voltage, low mass gate, and laser intensity were set to 96%, 1000.0 m/z, and 2,500, respectively. Each mass spectrum collected represents the sum of the data from 500 laser shots. Cytochrome c and aprotinin were used as internal calibrants. For each proteolytic fragment, the mean and standard deviation of the experimental molecular mass (m/z) were determined from six independent experiments. The theoretical molecular mass (m/z) was determined using an ExPASy-computed pI/MW tool program (us.expasy.org/tools/pi_tool).
Electron Microscopic Observation-The samples were fixed with 0.1 M cacodylate buffer (pH 7.4), including 2% glutaraldehyde, 2% paraformaldehyde, and 0.5% CaCl 2 for 6 h and washed with 0.1 M cacodylate buffer. The samples were post-fixed with 1.33% OsO 4 in cacodylate buffer for 2 h. The fixed samples were washed with 0.1% cacodylate buffer for 10 min. After washing, the samples were dehydrated with ethanol and incubated with propylene oxide for 10 min. Embedding Epon mixture (Epon812, methylnalic anhydride, dodecenylsuccinic anhydride, and 2,4,6-tri(dimethylaminomethyl)phenol) was prepared, and the samples were sectioned using an ultramicrotome followed by double staining with uranyl acetate and lead citrate. The samples were observed by transmission electron microscopy (Philips CM-10).
Surgical Procedure for the Striatal 6-Hydroxydopamine Lesion in Rats-All animal experiments were approved by the Committee for the Care and Use of Laboratory Animals at Yonsei University. All animals were cared for according to The Guide for Animal Experiments, 2000, edited by the Korean Academy of Medical Sciences, which was consistent with the NIH Guideline for the Care and Use of Laboratory Animals, 1996 revised edition. Male Sprague-Dawley rats, weighing 290 -330 g, were used. They were housed in cages containing two animals each, under a 12-h light-dark cycle with free access to food and water. The rats were deeply anesthetized with chloral hydrate (0.4 mg/g, intraperitoneal). For stereotaxic lesion surgery, 8 M 6-hydroxydopamine (6-OHDA, Sigma) was dissolved in 0.9% saline containing 0.02% ascorbic acid and was kept on ice protected from light. 6-OHDA was injected stereotaxically at three different points in the right striatum: 1) anterior-posterior ϩ1.0 mm, medial-lateral Ϫ3.0 mm, dorsal-ventral Ϫ5.0 mm; 2) anterior-posterior Ϫ0.1 mm, medial-lateral Ϫ3.7 mm, dorsal-ventral Ϫ5.0 mm; and 3) anterior-posterior Ϫ1.2 mm, mediallateral Ϫ4.5 mm, dorsal-ventral Ϫ5.0 mm from the Bregma, according to the Atlas of Paxinos and Watson (32). Triple 6-OHDA injections were prepared, each containing 10 g of 6-OHDA per 5 l, and administered using a 26-gauge cannula fitted to a Hamilton microsyringe at a rate of 1 l/min. At the completion of each injection, the needle was left in place for 5 min and then slowly withdrawn.
Tissue Preparation for Histological Studies-At 3 weeks after lesioning, the animals were deeply anesthetized with chloral hydrate, which was perfused through the ascending aorta with saline, followed by 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brain then was removed carefully and post-fixed for 1 day in the same solution. Each brain was dehydrated in 30% sucrose in 0.1 M phosphate buffer at 4°C for at least 3 days before it was sectioned. Six series of 40 m-thick coronal sections were cut on a freezing microtome (MICROM).

The Overexpression of ␣-Synuclein Caused the Formation of Its Cytoplasmic Inclusion and Subsequent Neuronal Cell
Death-At first we examined whether ␣-synuclein is expressed in the human dopaminergic neuroblastoma cell line (SK-N-BE). After the cell lysates were prepared from the proliferating cells, the presence of the ␣-synuclein protein was determined by Western blot analysis using anti-␣-synuclein antibodies. As shown in Fig. 1A, the endogenous protein level of ␣-synuclein in the SK-N-BE cells was too low to be detected. When the cells were transiently transfected with a plasmid encoding ␣-synuclein, the protein was properly expressed inside the cells, as compared with the mock-transfected control cells (Fig.  1A). In addition to the ␣-synuclein monomer, the formation of its oligomers was also observed (Fig. 1A). Immunocytochemical visualization of the cells clearly showed that the distribution of the transiently transfected ␣-synuclein was uneven and took the form of granular aggregates in the cytoplasm (Fig. 1A). Following the transient transfection of a plasmid encoding ␣-synuclein into the cells, the change in cell viability was measured after 24 h by means of the MTT extraction method. As shown in Fig. 1B, the overexpression of ␣-synuclein significantly decreased the cell viability in the SK-N-BE cells. We have previously shown that the addition of recombinant ␣-synuclein or the transient transfection of ␣-synuclein into hippocampal neuroprogenitor cells (H19-7) leads to neuronal cell death, and this result appears to be closely associated with the formation of intracytoplasmic ␣-synuclein-positive inclusions, having similar components to the LBs found in PD patients (28,33).

Extracellular Secretion of Intact ␣-Synuclein Transiently Transfected into Dopaminergic Neuroblastoma SK-N-BE
Cells-To investigate whether intracellular ␣-synuclein could be secreted into extracellular space, SK-N-BE cells were transiently transfected with the plasmid of ␣-synuclein, and the presence of ␣-synuclein in the extracellular location was examined by Western blot analysis. As shown in Fig. 2A, as well as being found in the cytoplasm, ␣-synuclein was also detected in the extracellular culture media after 24 h of its transfection, suggesting that extracellular secretion did indeed occur. However, the generation of a proteolytic ␣-synuclein fragment, which is similar to the partial fragment of ␣-synuclein observed as the non-A␤ component of amyloid plaque in AD, was not observed in the proliferating SK-N-BE cells in the absence of any stimulation ( Fig. 2A).
Addition of NO Donor Resulted in the Fragmentation of Secreted ␣-Synuclein via MMPs in SK-N-BE Cells-There is increasing evidence to suggest that oxidative injury is implicated in the pathogenesis of several neurodegenerative disorders, including PD, and the nitration of tyrosine residues in ␣-synuclein induced by oxidative injury is closely associated with the formation of the inclusions characteristic of these synucleinopathies (15). Based on these previous findings, we next examined whether the intracellular generation of nitrating agents affects the secretion of ␣-synuclein in SK-N-BE cells and the mechanism by which the cleavage of ␣-synuclein takes place under these conditions, if indeed this was to occur. To assess the generation of reactive NO molecules, the formation of nitrite (NO 2 Ϫ ), a stable breakdown product of NO, was assessed quantitatively (Table I). After transfection with a plas- mid encoding ␣-synuclein in a transient manner, SK-N-BE cells grown in regular 10% fetal bovine serum generated very little nitrite, with the quantity measured being close to the background level. In contrast, further treatment of the cells with an NO donor, 3-morpholinosydnonimine (SIN-1), brought about a ϳ5.5-fold increase of nitrite formation. When other NO donors, such as sodium nitroprusside and S-nitroso-Nacetylpenicillamine, were added to the media and left for 24 h, the resulting nitrite formation was increased by ϳ4.7and ϳ4.3-fold, respectively, as expected (Table I).
Following the transient transfection of SK-N-BE cells with the plasmid encoding ␣-synuclein, the addition of SIN-1 resulted in an increase in the extracellular-secreted ␣-synuclein levels, as compared with the ␣-synuclein-transfected and vehicle-treated control cells. Interestingly, the treatment with SIN-1 at a concentration of 100 M generated a partially digested fragment of ␣-synuclein having a size of 6.5 kDa, suggesting that the processing of ␣-synuclein could be brought about via the intracellular generation of nitric oxide (Fig. 2B). When other NO donors, such as sodium nitroprusside or Snitroso-N-acetylpenicillamine, were present during the pretreatment, a similar cleavage pattern of intact ␣-synuclein was observed in the SK-N-BE cells (Fig. 2C, S-nitroso-N-acetylpenicillamine data is not shown).
MMPs are implicated in the pathogenesis of neurodegenerative diseases and stroke (34,35). Moreover, the recent finding that S-nitrosylation activates MMP-9, subsequently leading to neuronal apoptosis, strongly implicated the involvement of a potential extracellular proteolysis pathway in the observed neuronal cell death, in which S-nitrosylation activates the MMPs. To confirm the involvement of MMPs in the NOinduced extracellular degradation of ␣-synuclein, the cells were pretreated with a potent inhibitor of matrix metalloproteinase, N-hydroxy-1,3-di-(4-methoxybenzenesulfonyl)-5,5-dimethyl-(1,3)-piperazine-2-carboxamide (also known as MMP inhibitor II), treated with SIN-1, and then the processing of ␣-synuclein was assessed. MMP inhibitor II is reported to inhibit MMP-1, -3, -7, and -9 with an IC 50 range of 2ϳ30 nM. As shown in Fig.  2B, pretreatment with this MMP inhibitor completely blocked the cleavage of ␣-synuclein, indicating that some MMPs play an active role during the NO-induced processing of ␣-synuclein in SK-N-BE cells.
Cleavage of ␣-Synuclein with Various MMPs in Vitro-As shown in Fig. 3A, the purified ␣-synuclein was cleaved by the tested MMPs to some extents. Referring the amount of the intact form, the purified ␣-synuclein was cleaved by MMP-3 most efficiently, and then in the order of MMP-14, MMP-2, MMP-1, and MMP-9. To test effect of hemopexin-like domain of MMP-3 in the cleavage of ␣-synuclein, the susceptibility of ␣-synuclein to the activated MMP-3 and the cMMP-3 that lacks its hemopexin-like domain was compared (Fig. 3B). Pattern of the purified ␣-synuclein cleaved by the cMMP-3 was the same as those by MMP-3. Thus, the hemopexin-like domain of MMP-3 has no effect on the cleavage of ␣-synuclein.
To estimate cleavage efficiency by MMP-3, we compared MMP-3-dependent degradation of ␣-synuclein with that of apolipoprotein A-I (apoA-I). ApoA-I, a major protein component of high density lipoprotein, is known to be cleaved very efficiently by MMP-3, moderately by MMP-7 and MMP-12, and slightly by MMP-1, but not by MMP-9 (53). When ␣-synuclein and apoA-I were digested by cMMP-3 at the same condition, their intact forms were disappeared at similar levels (Fig. 3, C and D). Thus, we confirmed that ␣-synuclein is a substrate for MMP-3 as efficiently as apoA-I.
Determination of Cleavage Sites of ␣-Synuclein by MMP-3-To examine the patterns of ␣-synuclein degradation, the purified ␣-synuclein was incubated for 0 -180 min with the catalytic domain of cMMP-3 at a 1:0.02 substrate/enzyme molar ratio. Upon incubation of ␣-synuclein with cMMP-3, we detected three light fragments (A, B, and C bands) and dark doublet fragments (D and E bands) at earlier time points, and two dark fragments (F and G bands) later (Fig. 4A). The N-terminal sequences of the four major fragments of ␣-synuclein (D, E, F, and G bands) were determined to be Met 1 -Asp-Val-Phe-Met. The intact form of ␣-synuclein was completely degraded at 180-min incubation with the catalytic domain of MMP-3 (Fig. 4A).    Fig. 4B). Four major peaks, m/z 7929.4 Ϯ 3.5, 7804.2 Ϯ 2.0, 5796.9 Ϯ 2.5, and 5496.7 Ϯ 1.2, which were detected in reproducible positions, were assigned as the residues Met 1 -Gln 79 , Met 1 -Ala 78 , Met 1 -Glu 57 , and Met 1 -Thr 54 (Fig. 4, B and C), respectively. Comparing intensities and relative locations of bands (Fig. 4A), along with their N-terminal sequences, with intensities and molecular masses of the peaks (Fig. 4B), we noticed that the four major peaks correspond to the D, E, F, and G fragments on SDS-PAGE (Fig.  4A), respectively. In addition, three minor peaks, m/z 10193.6 Ϯ 3.6, 9143.0 Ϯ 2.3, and 8987.2 Ϯ 5.7, had possible assignments of the residues Met 1 -Lys 102 , Met 1 -Gly 93 , and Met 1 -Ala 91 (Fig. 4, B and C), corresponding to the A, B, and C fragments on SDS-PAGE (Fig. 4A), respectively. Moreover, two peaks of m/z 6677.6 Ϯ 2.7 and 6552.2 Ϯ 0.6 were assigned as the residues Lys 80 -Ala 140 and Gln 79 -Ala 140 , which correspond to the m/z values of counterpart fragments for D and E fragments (DЈ and EЈ), respectively (Fig. 4, B and C). Therefore, ␣-synuclein was gradually broken down by MMP-3 from its C-terminal end. It is interesting to note that MMP-3 has at least four different cleavage sites, Ala 78 2Gln 79 , Gln 79 2Lys 80 , Ala 91 2Thr 92 , Gly 93 2Phe 94 , within the non-amyloid plaque sequence (Glu 61 -Val 95 ) of ␣-synuclein. In this regard, it is likely that MMP-3 would have an effect on the aggregation of ␣-synuclein.
Overexpression of ␣-Synuclein Results in the Up-regulation of MMP-3 Levels in SK-N-BE Cells-We investigated whether the overexpression of ␣-synuclein affects the endogenous levels of MMP-3 inside SK-N-BE cells. As shown in Fig. 5A, RT-PCR analysis showed that the mRNA levels of MMP-3 were significantly affected by the transient transfection of ␣-synuclein, as compared with those in mock-transfected control cells. Furthermore, Western blot analysis showed that the active and cleaved 45-kDa form of MMP-3 was also significantly produced by ␣-synuclein (Fig. 5B). Based on the report that nitritegenerating NO stimuli might influence the expression of MMPs (23), we investigated whether the addition of an NO donor affects the levels of expression of MMP-3 induced by ␣-synuclein transfection. As shown in Fig. 5C, the addition of an NO donor (SIN-1) alone resulted in the induction of MMP-3 expression, as determined by RT-PCR and Western blot analysis, respectively. However, no synergistic increase in the mRNA or protein levels of MMP-3 was observed in those cells treated with ␣-synuclein plus SIN-1. In a similar way, the intracellular production of nitric oxide resulted in the production of MMP-3 proteins. Taken together, these findings suggest that the overexpression of ␣-synuclein resulted in the activation of MMP-3, as well as changes in its mRNA and protein levels, and its effect appeared to be mediated by the generation of intracellular nitrating compounds in neuronal SK-N-BE cells. Consistent with this observation, we previously reported that the expression of ␣-synuclein causes the formation of reactive oxygen species in hippocampal neuronal cells (28).

The Kinetics and Neurotoxicity Induced by Intact ␣-Synuclein Inclusions Are Different From Those Observed in the Presence of MMP-3-cleaved Fragments-Protein aggregation of ␣-synuclein
in the presence and absence of the MMP-digested ␣-synuclein fragments was evaluated by measuring the turbidity. The ␣-synuclein fragments, which were produced by MMP-3 (Fig.  6A, inset), exhibited most dramatic stimulation, whereas the aggregation observed only with the intact protein was slowly progressed to lesser extents (Fig. 6A). When the thioflavin-T binding fluorescence was simultaneously evaluated, the difference on the amyloid formations of ␣-synuclein in the presence and absence of the proteolyzed fragments by MMP-3 was narrowed, as compared with the turbidity (Fig. 6B). When the cytotoxicity of the protein aggregates by intact ␣-synuclein was compared with those prepared with MMP-3-mediated digestion fragments, the aggregates following MMP-3 treatment had a more toxic effect on the cell viability (Fig. 6C). The formation of fibrous and granular structures during the ␣-synuclein aggregation in the presence and absence of the digested fragments generated by MMP-3 was morphologically analyzed with a transmission electron microscope. As shown in Fig. 6D, the structures of the final protein aggregates are clearly distinctive from each other. The intact ␣-synuclein aggregates contained only the fibrillar forms, whereas the inclusions obtained with the MMP-3-digested ␣-synuclein showed the spherical granular forms in addition to the fibrillar aggregates. Lansbury et al. demonstrated that in vitro fibril formation by ␣-synuclein from its soluble monomeric form does not follow a simple one-step transition, but rather a complex process that involves one or more discrete intermediates, termed protofibrils (26,27). During the process, a protofibril intermediate, rather than the fibril itself, may be thought to be more pathogenic. They have also characterized several ␣-synuclein oligomers, which are much smaller than fibrils and appear early in the fibrillization process as granular forms (26,27). In consistent with this view, the morphological difference between the two protein aggregations has been reflected by the distinct patterns of aggregation kinetics and the neurotoxicities of the two final inclusions. ␣-Synuclein (1 g) was digested by cMMP-3 in a 1:0.02 substrate/enzyme molar ratio at 37°C. At the indicated time intervals, aliquots were taken and analyzed on 15% Tris-Tricine gel. B, MALDI-TOF mass spectra of the cMMP-3-cleaved ␣-synuclein. The mass spectra between m/z 5,000 and 20,000 were collected. The molecular mass (m⁄z) of each peak was labeled as a value of average mass. C, schematic representation of the cMMP-3 cleavage sites of ␣-synuclein. The observed mass (m⁄z) of each fragment was labeled as mean Ϯ S.D., which was calculated from six independent experiments. duces an almost complete and selective lesion, is recommended for studies of different functional states in the nigrostriatal dopamine system (36).
By using immunohistochemical analysis, we examined whether the levels of MMP-3 were increased or decreased significantly in the CNS region in the rat PD brain model. We utilized triple intrastriatal injections of a convulsive dosage of 6-OHDA, to induce the Parkinsonism phenotype in the rat brain. The distribution of the MMP-3 protein levels in the 6-OHDA-treated rat brain was investigated using immunohistochemistry. The region-specific activation of MMP-3 was observed in the CNS 3 weeks after the 6-OHDA injection. As shown in Fig. 7, the intrastriatal systemic administration of a 30 g/kg dose of 6-OHDA induced significant levels of MMP-3 gene transcripts in the substantia nigra region of the CNS midbrain. However, the corresponding brain section of the control rat did not show any significant in vivo MMP-3 signals (Fig. 7). These data suggest that significant levels of MMP-3 expression are induced. Taken together, our findings suggest that the extracellularly secreted ␣-synuclein could be processed via the activation of MMP-3 in a selective manner and that this kind of specific protein digestion could play a certain role in the pathogenesis of PD. DISCUSSION Previous microscopic and ultrastructural analyses revealed that anti-␣-synuclein immunostaining was co-localized with synaptophysin-immunoreactive presynaptic terminals as well as synaptic vesicles in the rat brain, suggesting that ␣-synuclein is a presynaptic protein (24,37). In the present study, we observed that, when the ␣-synuclein is transiently expressed in dopaminergic neuroblastoma SK-N-BE cells, it is secreted into extracellular space. When the bacterial recombinant ␣-synuclein protein was added to the culture medium, it was found to be transported into the intracellular cytoplasm, resulting in necrotic cell death in H19 -7 cells (28). Based on these two findings, we speculate that ␣-synuclein has the capability to pass through the membrane as well as the binding to the membrane, and, as a result of this novel property, ␣-synuclein is localized in the synaptic cleft and postsynaptic densities, as well as in the presynaptic terminals. Although the protein was originally found to be located in the presynaptic terminals, and was thought to play a role in neuronal plasticity (24), a certain amount of evidence has emerged, which suggests that the full-length ␣-synuclein as well as its digested fragment plays a functional role in non-cytoplasmic locations, such as the nucleus and extracellular space.
Firstly, whereas classic brainstem LBs appear as intracytoplasmic circular inclusions, LBs can extend into nerve cell processes or lie free in the neuropil, referred to as "extracellular Lewy bodies" (38). Secondly, the fragment (residues 61-95) of ␣-synuclein defined as the "non-amyloid component (NAC) of senile plaques" was found to be associated with extracellular A␤ deposits in the AD patient's brain (24,42). Although NAC is tightly associated with A␤ fibrils in the AD amyloid, ␣-synuclein has also been shown to bind A␤ (43) as well as amyloid plaques in the AD brain (44). Thirdly, Borghi et al. (45) reported that the entire ␣-synuclein protein is released by neurons into the cerebrospinal fluid. Moreover, the detection of extracellular ␣-synuclein and/or its modified forms in body fluids, particularly in human plasma (46), strongly suggests that cells normally secrete ␣-synuclein into their surrounding media, both in vitro and in vivo. Finally, ␥-synuclein, a member of the synuclein superfamily, which shares 73% amino acid identity across the ␣-synuclein amphipathic domain, was found to be overexpressed in infiltrating breast carcinoma (47). Interestingly, it has an abnormal pattern of expression in some neurodegenerative diseases, and modulates the expression of certain MMP genes, suggesting its involvement in the action mechanism of neuropathologies mediated through MMPs (48,49). Notwithstanding the many previous reports on this topic, the cellular mechanism by which ␣-synuclein is secreted into the cerebrospinal fluid and the question of which proteases are involved in the processing required generating the smaller fragments of ␣-synuclein, such as NAC, have not yet been clearly elucidated.
To assess the possibility of the extracellular localization of ␣-synuclein in vivo, we analyzed the localization of ␣-synuclein in the mouse brain using electron microscopy. 2 We found that ␣-synuclein is primarily localized in the presynaptic nerve terminals, whereas a moderate intensity of ␣-synuclein immunoreactivity is present in the cytoplasm and nucleoplasm of the cerebral cortical neurons, as well as the plasma membrane, ribosomes, rough endoplasmic reticulum, mitochondria, small vesicles and nuclear outer membrane. 2 Interestingly, however, ␣-synuclein immunoreactivity is also associated with the postsynaptic areas and synaptic clefts, 2 which supports the current observation of ␣-synuclein immunoreactivity in SK-N-BE cells, tending to implicate the extracellular localization of ␣-synuclein.
The present data suggest that the overexpression of ␣-synuclein triggers its extracellular secretion, as well as a remarkable up-regulation of MMP-3 expression. ␣-Synuclein is cleaved most effectively in vitro by MMP-3, generating an ϳ6.5-kDa major fragment, whereas it could be digested in some extents by MMP-14, MMP-2, MMP-1, and MMP-9. Furthermore, upon its stimulation by the intracellular formation of nitric oxide, ␣-synuclein is secreted as an NAC-like smaller fragment, resulting in the increased formation of extracellular protein inclusions. The existence of such a role for MMPs during PD is further supported by the finding that significant levels of MMP-3 expression are observed arising from the substantia nigra area into the nigrostriatal dopaminergic pathways in 6-OHDA-injected PD rat brain model. With the purpose of assessing whether ␣-synuclein fragment is the same peptide as the NAC co-aggregated with A␤ in AD, MALDI-TOF mass spectral peptide fingerprinting was performed to clarify the exact amino acids cleaved by MMP-3 within the ␣-synuclein. The analysis revealed that, upon incubation of ␣-synuclein with MMP-3, ␣-synuclein was gradually and completely broken down from its C-terminal end within 3 h, which led to the generation of total four major and three minor fragments. In addition, MMP-3 was shown to have at least four different cleavage sites within the non-amyloid plaque sequence (Glu 61 -Val 95 ) of ␣-synuclein, strongly indicating that MMP-3 affects the aggregation pattern of ␣-synuclein in some way. This was confirmed by the finding that MMP-3-digested ␣-synuclein fragments have caused different kinetics of the protein aggregation and subsequent neurotoxic effect.
There are numerous reports to suggest that a variety of structural modifications in ␣-synuclein affect its inclusion morphology pattern and aggregation kinetics. For example, the A53T and A30P mutations of ␣-synuclein increase its iron-dependent aggregation and toxicity, compared with wild type ␣-synuclein (50). Furthermore, the Ser-129 residue of ␣-synuclein is selectively and extensively phosphorylated, and the phosphorylated ␣-synuclein is subsequently targeted to mono-and di-ubiquitination in synucleinopathy brains (51,52). Although the structural modification of ␣-synuclein, due to its phosphorylation, leads to a change in the aggregation kinetics and its morphology, the current data show that the proteolytic cleavage of ␣-synuclein could also affect the pattern and biochemical properties of its abnormal intracytoplasmic accumulation in a similar way.
Using zymography, Lorenzl et al. (20) observed a reduction in the MMP-2 levels in the substantia nigra obtained from postmortem brain tissue of a PD patient, as compared with agematched control cases, suggesting that the alteration of the expression of MMPs in the substantia nigra of PD may contribute to the disease pathogenesis. The present study provided additional supporting evidence to suggest that the increased levels of MMP-3 are associated with the pathogenesis of PD, as shown in the 6-OHDA-injected rat PD model.
Although MMPs are implicated in the pathogenesis of neurodegenerative diseases and stroke, the mechanism of MMP  7. Increase of MMP-3 levels in the substantia nigra region of intrastriatally 6-OHDA-injected rat brain. Triple intrastriatal injections of 8 M 6-OHDA into the rat brain were carried out. Three weeks after injection, the rat brain was fixed and sliced, and the midbrain slice sections containing the substantia nigra were prepared for histological studies. Immunohistochemical analysis was performed with anti-MMP-3 antibodies, as indicated.
thioflavin-T binding fluorescence, respectively, as described under "Experimental Procedures." C, cell viability was measured by MTT extract assay after treatment with either the intact ␣-synuclein aggregate (Syn) or the ␣-synuclein aggregate after its MMP-3-mediated cleavage (SϩMMP3) (**, p Ͻ 0.01 versus control; *, p Ͻ 0.05). D, morphological differences of the ␣-synuclein aggregates. Protein aggregates of ␣-synuclein prepared in the absence (Control) and presence (MMP-3) of the MMP-3-derived fragments of ␣-synuclein were examined with electron microscope under magnifications of ϫ80,000. The scale bars represent 100 nm. activation remains unclear. Reactive oxygen species have also been shown to be important messenger molecules in the induction of several genes. Brenneisen et al. (54) provided evidence that H 2 O 2 is an important intermediate in the downstream signaling pathway, ultimately leading to the induction of increased steady state MMP-1 mRNA levels. Gu et al. (55) have recently reported that MMP activation also involves S-nitrosylation. Consistent with these reports, the present study demonstrated that the activation of MMP-3 is induced upon the stimulation provided by a nitric oxide donor in neuronal cells, triggering the digestion of extracellular ␣-synuclein, and subsequently promoting its amyloidogenic properties, as compared with those of the intact ␣-synuclein. However, we could not rule out the possibility that more than one MMP (e.g. MMP-3) are actually involved in the digestion the ␣-synuclein in vivo.
Taken together, our current findings suggest that the impairment of cellular anti-oxidative mechanisms or the overproduction of reactive species may be a primary event leading to the onset and progression of neurodegenerative synucleopathies. Accordingly, efforts to elucidate the precise mechanism for this hazardous state could lead to the development of novel therapeutic strategies for PD and other synucleinopathies.