Quantitative Analysis of α-Synuclein Solubility in Living Cells Using Split GFP Complementation

Presently incurable, Parkinson's disease (PD) is the most common neurodegenerative movement disorder and affects 1% of the population over 60 years of age. The hallmarks of PD pathogenesis are the loss of dopaminergic neurons in the substantia nigra pars compacta, and the occurrence of proteinaceous cytoplasmic inclusions (Lewy bodies) in surviving neurons. Lewy bodies are mainly composed of the pre-synaptic protein alpha-synuclein (αsyn), an intrinsically unstructured, misfolding-prone protein with high propensity to aggregate. Quantifying the pool of soluble αsyn and monitoring αsyn aggregation in living cells is fundamental to study the molecular mechanisms of αsyn-induced cytotoxicity and develop therapeutic strategies to prevent αsyn aggregation. In this study, we report the use of a split GFP complementation assay to quantify αsyn solubility. Particularly, we investigated a series of naturally occurring and rationally designed αsyn variants and showed that this method can be used to study how αsyn sequence specificity affects its solubility. Furthermore, we demonstrated the utility of this assay to explore the influence of the cellular folding network on αsyn solubility. The results presented underscore the utility of the split GFP assay to quantify αsyn solubility in living cells.


Introduction
Parkinson's disease (PD) is the most prevalent neurodegenerative movement disorder, affecting 1% of the world's population over the age of 60 years [1]. The hallmarks of PD pathogenesis are the loss of dopaminergic neurons in the substantia nigra pars compacta and the occurrence of cytoplasmic inclusions called Lewy bodies (LB) in surviving dopaminergic neurons [2]. Post mortem analyses revealed that the main component of LB is the pre-synaptic protein alpha-synuclein (asyn) and of trace amounts of ubiquitin and molecular chaperones [3], suggesting that they result from the aberrant accumulation and aggregation of misfolded, undegraded asyn. Duplications or triplications of the asyn locus [4,5], as well as mutations in asyn-encoding gene -A53T, A30P & E46K -lead to increased aggregation and have been linked to familial cases of PD [6][7][8][9][10]. Overexpression of asyn results in the formation of inclusion bodies, cytotoxicity and cell death in animal models and cell cultures [11][12][13]. Misfolding and aggregation of asyn has been associated with impairment of proteasomal degradation, another common trait of PD pathogenesis [14][15][16]. In summary, aberrant accumulation of misfolded asyn plays a key role in development of PD pathogenesis. Therefore, monitoring asyn aggregation in living cells in a quantitative fashion is important to study the molecular mechanisms associated with asyn-induced cytotoxicity and develop therapeutic strategies for the treatment of PD.
A number of asyn variants containing mutations that alter the protein's rate of aggregation have been characterized [6][7][8][9]. Among mutations linked to familial cases of PD, the A53T asyn variant was shown to aggregate at a much faster rate than wt asyn in cell cultures and in vitro [9,10,17]. C-terminal truncations have also been reported to aggregate at higher rates than wt asyn [18][19][20], demonstrating that the proline-rich C-terminal region plays a fundamental role in limiting asyn misfolding and aggregation [18][19][20][21][22]. A recent study demonstrated that a truncation variant of asyn consisting of amino acids 1-123 (asyn123) readily formed aggregates in vitro [19]. Interestingly, it was also shown that truncated asyn accumulates in LB [23], suggesting that lower molecular weight truncated asyn species may have a role in PD pathology. Recently, in an effort to decipher the determinants of asyn aggregation, a rationally designed mutant containing three proline substitutions (TP asyn, containing substitutions A30P, A56P and A76P) was also constructed and demonstrated to resist aggregation in vitro [24]. Its solubility in cell cultures, however, is not known.
A number of methods to study asyn aggregation in vitro have been reported and include microscopy [21], size-exclusion chromatography [25], and NMR spectroscopy [26]. These techniques rely on the use of purified proteins for analysis. Hence, they preclude the study of asyn aggregation in living cells, which is necessary to decipher the pathogenic mechanisms that lead to increased levels of misfolded and aggregated asyn and to identify gene targets for therapy.
Microscopy based techniques have been used to monitor protein aggregation in living cells [27,28]. Particularly, asyn aggregation can be detected using asyn-specific antibodies [11,29] or by overexpressing asyn variants fused to fluorescent reporters such as GFP [17,30,31]. The main limitation of using GFP fusions as aggregation reporters is that aggregation events that occur after the formation of the GFP chromophore do not alter fluorescence emission, leading to detection of GFP fluorescence irrespective of asyn aggregation state. To overcome this limitation, techniques that rely on fluorescence complementation have been developed. Particularly, asyn was fused to non-fluorescent complementary GFP fragments and the resulting fusion molecules were coexpressed in mammalian cells. asyn self-association causes close proximity of the two GFP fragments and results in bimolecular fluorescence complementation (BiFC). Hence, the intensity of the fluorescence signal is a measurement of asyn self-association [32][33][34]. Fluorescence energy resonance transfer (FRET) has also been used to quantify asyn aggregation by fusing two fluorophores to the N-and C-terminals of asyn [35]. BiFC and FRET, however, suffer from inherent limitations. Fusion of asyn to highly stable chromophores or to large protein fragments can perturb asyn folding and alter its misfolding-propensity. In addition, these techniques are not optimal to measure protein self-association because they fail to detect homotypic interactions.
In this study, we developed an expression system that allows detecting and quantifying soluble asyn in living cells. We adapted a previously reported split GFP molecule specifically engineered to study protein solubility [36]. This GFP variant is cleaved into two unequal size fragments, a 15-amino acid ''sensor'' fragment and a large ''detector'' fragment, that spontaneously complement upon chemical interaction, giving rise to a fluorescence signal [36]. asyn was fused to the sensor fragment, which has minimal effect on the folding and solubility of its fusion partners and can therefore be used as a sensor of asyn solubility. The resulting asyn fusion protein was co-expressed with the large detector fragment in cell cultures. Fluorescent complementation is directly proportional to asyn solubility as it occurs only if the sensor fragment escapes aggregation and is accessible to the detector fragment. The fluorescence of cells expressing wild type asyn was compared to that of cells expressing asyn variants with different aggregation properties: A53T asyn, a C-terminal truncation variant (asyn123), and a rationally designed triple proline mutant (A30P, A56P and A76P) with low propensity to aggregate (TP asyn). Cell fluorescence was also evaluated upon inhibition of proteasomal degradation and was observed to correlate with asyn solubility as predicted from in vitro studies. Our results indicate that this method provides a robust platform to quantify asyn solubility in living cells and can be used to study asyn sequence specificity and to monitor the influence of the cell folding network on asyn aggregation.

Quantification of asyn Solubility using the asyn-split GFP Assay
To study asyn solubility in living cells we adapted a previously reported assay based on split GFP complementation [36]. In this assay, GFP is split into two moieties, GFP 1-10 , the bulk of the bbarrel (detector fragment), and GFP 11 , a 15-amino acid b-sheet (sensor fragment). GFP fragment complementation was shown to be inversely proportional to aggregation by comparing sequential expression and co-expression of GFP 11 -tagged proteins and GFP 1-10 [36]. The small GFP 11 tag was previously shown not to affect the folding of the fusion protein [36,37] and was therefore fused to the C-terminal of asyn in this study. The large GFP 1-10 fragment was co-expressed with asyn-GFP 11 in the cytoplasm of mammalian cells. We hypothesized that if asyn is maintained in a soluble state, the GFP 11 tag is exposed to the solvent and can complement with GFP 1-10 , giving rise to a fluorescence signal. On the other hand, asyn aggregation would preclude accessibility of GFP 11 to GFP 1-10 , thus preventing fluorescence complementation. Hence, GFP fluorescence is expected to be proportional to asyn solubility.
HeLa cells were transfected for the expression of asyn-GFP 11 and GFP 1-10 and GFP fluorescence was evaluated by flow cytometry and fluorescence microscopy ( Figure 1). As expected, cells expressing only GFP 1-10 did not display detectable fluorescent signal ( Figure 1A), whereas cells co-expressing asyn-GFP 11 and GFP 1-10 exhibited GFP fluorescence when tested 18 hrs post transfection ( Figure 1B). Fluorescence microscopy validated these results (Figure 1, inset), confirming that cell fluorescence is due to GFP fragment complementation. To ensure that the intensity of the fluorescence signal is not limited by the amount of GFP 1-10 available for complementation with GFP 11 and is therefore an accurate measurement of the concentration of soluble asyn, a series of experiments were conducted in which increasing concentrations of plasmid encoding for GFP 1-10 were used in the transfection procedure. A GFP 1-10 to GFP 11 ratio of 2:1 was sufficient to ensure that fluorescent complementation is not limited by the concentration of GFP 1-10 but rather depends on the amount of soluble asyn-GFP 11 (data not shown), in agreement with previously published work [36,37]. This ratio of plasmid concentrations was used for all subsequent experiments.
Next, we compared wild type asyn to three asyn variants -A53T asyn, TP asyn and a C-terminal truncation mutant consisting of amino acids 1-123 (asyn123). A53T asyn was shown to aggregate faster than wild type asyn in cells and in vitro [9,10,17]. The truncated asyn123 has a shortened proline rich C terminal region, making it prone to aggregation in in vitro studies [18][19][20]. TP asyn contains three proline substitutions (A30P, A56P and A76P) that disrupt the protein's ability to form aggregates in vitro, therefore preventing the formation of fibrils even after two weeks of incubation [24]. The solubility and aggregation propensity of TP asyn in living cells, however, is not known. The mutations were introduced in the asyn-GFP 11 encoding gene. HeLa cells were transfected with three plasmid encoding the asynGFP 11 variants, GFP 1-10 , and mCherry, a highly photostable red fluorescent protein mutant [38] (a gracious gift from Dr. Jonathan Silberg, Rice University), used here as a transfection control. Cells were cultured for 18 hrs and GFP fluorescence measured by flow cytometry. Cells expressing TP asyn exhibited 50% higher fluorescence than cells expressing wild type asyn, whereas, GFP fluorescence was 25% lower in cells expressing A53T and asyn123 ( Figure 2A). Results obtained using HeLa cells suggest that the asyn-split GFP assay can be used to quantify asyn solubility in living cells. To validate the use of the asyn-split GFP assay in a cell type more relevant to study the phenotype associated with PD cellular pathogenesis, these experiments were repeated using neuroglioma (H4) cells. As shown in Figure 2B, H4 cells expressing A53T asyn and asyn123 exhibited significantly lower fluorescence than wild type asyn, while GFP fluorescence was significantly higher in H4 cells transfected with TP asyn, confirming the results obtained in HeLa cells.
Fluorescence microscopy images of HeLa cells expressing the asyn-split GFP system are reported in Figure 2C-D and include detection of GFP fluorescence (left column) and detection of mCherry fluorescence (right column). GFP fluorescence was observed to decrease in cells expressing A53T asyn and asyn123 and to increase in cells expressing TP asyn, confirming results obtained with flow cytometry. The intensity of mCherry fluorescence, however, did not change in cells expressing different asyn variants, demonstrating that the differences in GFP fluorescence complementation detected are not due to differences in transfection or expression efficiency, but are rather due to GFP fluorescence complementation. These differences in GFP fluores-cence complementation were equally observed upon visualization of the whole cell population ( Figure 2C) as well as in individual cells ( Figure 2D), confirming the results obtained from flow cytometry. In summary, these results demonstrate that A53T asyn and asyn123 have a higher propensity to form aggregates and, therefore, lead to lower fluorescence complementation than wt and TP asyn.

Inhibition of Proteasomal Degradation Lowers asyn Solubility and Prevents GFP Fluorescence Complementation
Our results demonstrate that the asyn-split GFP assay is a viable tool to study asyn aggregation. This assay can be used to study the aggregation of naturally occurring asyn variants and to predict the aggregation of rationally designed mutants such as TP asyn. To further characterize the TP asyn mutant, we tested its solubility in HeLa cells under cell culturing conditions that are expected to alter the solubility of misfolded, aggregation-prone proteins. To this end, we induced chemical inhibition of proteasomal degradation and investigated its effect on fluorescence complementation. Inhibition of the proteasome causes aberrant accumulation of misfolded proteins and formation of insoluble aggregates [39,40]. Lactacystin is a highly selective proteasome inhibitor [41] that can easily penetrate the cell membrane and irreversibly block multiple hydrolytic activities in the proteasome [42]. HeLa cells expressing GFP 1-10 and either asyn-GFP 11 or TP asyn-GFP 11 were treated with a range of concentrations of lactacystin for 24 hrs and GFP fluorescence was measured by flow cytometry. As shown in Figure 3A, cells expressing TP asyn exhibited 10% and 21% higher fluorescence than cells expressing wild type asyn after 12 and 24 hrs of incubation, respectively. Upon treatment with lactacystin, we observed a decrease in GFP fluorescence in a concentration dependent manner. Specifically, cells expressing asyn wild type displayed 68% fluorescence of untreated cells upon treatment with 5 mM lactacystin and cells expressing TP asyn displayed 70% fluorescence under the same conditions ( Figure 3B). These results suggest that proteasomal inhibition, by causing an increase in asyn aggregation, results in lowered GFP fluorescence complementation. Thus, this assay can be used to monitor the influence of the folding and degradation machinery on asyn solubility. It should be noted that even though lactacystin treatment caused similar changes in fluorescence in cells expressing wild type asyn and TP asyn relative to untreated cells, the absolute fluorescence of cells expressing TP asyn was significantly higher than that of cells expressing asyn wild type ( Figure S1), as reported before (Figure 2 & 3A). In order to confirm that the loss of GFP fluorescence observed upon lactacystin treatment is due to increase in asyn aggregation, asyn solubility was investigated by Western blot. HeLa cells expressing asyn-GFP 11 were incubated with lactacystin (5 mM) for 24 hrs. The soluble protein fraction was collected and analyzed using an asyn-specific antibody. Lactacystin-induced proteasome inhibition was observed to result in approximately 25% decrease in soluble asyn ( Figure 3C and 3D). This data indicates that the

Fluorescence Complementation Inversely Correlates with the Formation of Cellular Aggregates
To examine the correlation between fluorescence complementation and asyn aggregation, we evaluated the formation of aggregates using immunofluorescence microscopy. Cells were cultured under conditions that gave rise to maximal change in GFP complementation and analyzed by immunofluorescence microscopy. Specifically, HeLa cells were transfected for the expression of GFP 1-10 and either asyn-GFP 11 or TP asyn-GFP 11 and treated with lactacystin (5 mM) for 24 hrs. asyn accumulation into cellular aggregates was detected using an antibody specific for asyn ( Figure 4, column 1, blue) and the ProteoStatH dye ( Figure 4, column 2, red), a 488-nm excitable red fluorescent molecule that specifically interacts with denatured proteins within aggresomes [43]. Images showing co-localization of asyn and the aggregatespecific dye were analyzed with NIH ImageJ software to obtain heatmaps ( Figure 4, column 3). To quantify the aggregation of asyn, co-localization events were counted and averaged over three independent experiments. The extent of co-localization was evaluated by analyzing the image heatmaps based on the color scale reported in Table 1 as described in the Materials and Methods. Our analysis revealed that the degree of asyn aggregation induced by lactacystin treatment depends on asyn sequence. Specifically, cells expressing asyn-GFP 11 display a 3-fold increase in asyn aggregation upon treatment with lactacystin, while cells expressing TP asyn-GFP 11 exhibit only a 1.5-fold increase ( Table 1, high aggregation). The extent of aggregation detected from fluorescence microscopy studies (Figure 4) inversely correlates with measurements of cell fluorescence obtained by flow cytometry (Figure 3). We therefore concluded that the decrease in fluorescence complementation observed in cells treated with lactacystin can be attributed to the increase in asyn aggregation caused by inhibition of proteasomal degradation. Furthermore, the higher fluorescence complementation observed in cells expressing TP asyn compared to wild type asyn is a direct result of its lower rate of aggregation.

Discussion
Aggregation of asyn into proteinaceous inclusions [2] has been repeatedly associated with the development of PD pathogenesis [11,44]. Therefore, there is an urgent need to understand the molecular mechanisms underlying asyn misfolding and aggregation in living cells. Currently available methods to study aggregation in cell cultures, including the use of GFP fusions, BiFC and FRET, present a number of limitations mainly associated with the use of reporter molecules that alter asyn misfolding and aggregation pathway [32], preclude rapid and high-throughput quantification and, most importantly, do not afford reliable distinction between soluble and insoluble pools of asyn [45]. In this study, we report the use of a split GFP assay based on the detection of fluorescent complementation [36], previously reported for quantification of protein solubility in vitro [36], and in bacterial and mammalian cells [37]. The GFP variant used in this assay is split into a small ''sensor'' fragment, which was fused to asyn in this study, and a large ''detector'' fragment. asyn aggregation precludes accessibility of the sensor fragment to the detector fragment for fluorescence complementation. We demonstrated here that the asyn-split GFP expression system provides a reliable tool to quantify asyn solubility in living cells.
We investigated the utility of the asyn-split GFP assay to study the relationship between asyn sequence and its rate of aggregation in living cells. Mutations in the asyn-encoding gene have been associated with the development of early onset familial cases of PD [6][7][8]. asyn C-terminal truncations were observed to accumulate in LB [20,46,47]. The aggregation properties of naturally occurring and rationally designed asyn mutants have been extensively characterized in vitro [24,48,49]. To evaluate the use of the asyn-split GFP assay to study how asyn sequence specificity affects protein aggregation, we tested a rationally designed variant (TP asyn) known to resist aggregation in vitro [24]. We compared the fluorescence of cells expressing TP asyn to that of cells expressing wild type asyn, A53T asyn and a truncated asyn variant (asyn123). We observed a significant increase in fluorescence in cells expressing TP asyn compared to cells expressing wild type asyn, demonstrating higher solubility of this asyn variant in cell cultures. On the other hand, cells expressing the A53T mutant and the truncation mutant asyn123 exhibited significantly lower fluorescence than cells expressing wild type asyn, suggesting that these variants aggregate at higher rate and that aggregation lowers GFP fragment complementation and fluorescence. These results indicate that the asyn-split GFP assay can be used to quantify the effect of mutations in asyn-encoding gene on the protein aggregation propensity in living cells.
We also investigated whether the asyn-split GFP assay can be used to study the impact of environmental factors that alter the efficiency of the folding quality control system on asyn solubility. Although the causes of PD are far from understood, studies have shown that changes in the cellular environment such as oxidative stress and inflammation are involved in the progression of the disease [50]. Inducing oxidative stress or inflammation was shown to increase asyn aggregation and asyn-induced cytotoxicity [51,52]. Furthermore, the accumulation of asyn has also been associated with impairment of the proteasome [14][15][16]. In this study, proteasomal inhibition was chemically induced in cells expressing the asyn-split GFP system and observed to lower fluorescence complementation, demonstrating that proteasome dysfunction lowers asyn solubility.
Finally, we showed that the intensity of fluorescence of cells expressing the asyn-split GFP system is inversely proportional to the extent of asyn aggregation. Analysis of co-localization between asyn and an aggregate-specific dye revealed that the increase in fluorescent signal measured correlates with the decrease in aggregate formation. These results demonstrate that the asynsplit GFP assay can be used to investigate cell treatments that affect protein aggregation and that it will potentially enable molecular screenings for the discovery of compounds that modulate asyn aggregation.  In summary, our results show that the asyn-split GFP assay allows to quantitatively measure the solubility of asyn in living cells. Furthermore, we demonstrated that this assay can be used to study the aggregation properties of asyn mutants in cell cultures and elucidate the effects that modifiers of cellular protein folding have on asyn aggregation.

Reagents, Cell Lines, and Media
Lactacystin was purchased from Cayman Chemicals. Cell culture media were purchased from Gibco and Invitrogen. Fetal bovine serum (FBS) was purchased from Atlanta Biologicals. JetPrime TM transfection kit was purchased from Polyplus Transfection. ProteostatH Aggresome Detection Kit was purchased from Enzo Life Sciences.

Plasmids and Transient Transfections
pCMV-mGFP Cterm S11 Neo Kan and pCMV-mGFP 1-10 Hyg Amp vectors were obtained from Theranostech, Inc. The sequence encoding for GFP 1-10 was amplified from pCMV-mGFP 1-10 Hyg Amp by PCR using the primers listed in Table S1 and subcloned into pcDNA4/TO (Invitrogen) using the KpnI and XhoI restriction sites, giving rise to pcDNA4/TO/GFP 1-10 . The cDNA encoding for a-syn was amplified from pcDNA6.2+asyn-emGFP plasmid (lab collection) using the primers listed in Table  S1 and cloned into the plasmid pCMV-mGFP Cterm S11 Neo Kan using XhoI and AgeI restriction sites, giving rise to pCMV-mGFP/asyn-GFP 11 . The A53T substitution carrying mutant asyn was constructed using the QuikChangeH Site-Directed mutagenesis kit (Strategene) and KAPA HiFi HotStart PCR kit (Kapa Biosystems) following manufacturers' protocols and using primers listed in Table S1, giving rise to pCMV-mGFP/A53Tasyn-GFP 11 .The gene encoding for the asyn mutant containing A30P, A56P, and A76P substitutions was constructed using the same procedure, using primers listed in Table S1, giving rise to pCMV-mGFP/TP asyn-GFP 11 . The sequence of the truncated asyn gene was amplified using pCMV-mGFP/asyn-GFP 11 as template and primers listed in table S1. The PCR product was cloned into the empty pCMV-mGFP Cterm S11 Neo Kan plasmid using XhoI and AgeI restriction sites, giving rise to pCMV-mGFP/asyn123-GFP 11 .
Transfections were conducted in 6-well plates. 10 4 cells were seeded in each well of a 6-well plate and plates were incubated for 24 hrs at 37uC. Transient transfections were performed using the JetPrime TM DNA transfection kit (Polyplus Transfection) according to the manufacturer's procedures.

GFP Complementation Analyses
HeLa or H4 cells were plated in 6-well plates and incubated for 24 hrs at 37uC. The media was removed and replaced with fresh media containing 0.33 mg of vectors encoding for wild type asyn, A53T asyn, TP asyn or asyn123 and 0.67 mg of pcDNA4/TO/ GFP 1-10 per well and transfected as described above. Transfection reactions were incubated for 16 hrs, at which point the media was replaced again. Cells were then collected and fluorescence was measured using a flow cytometer (FACSCanto TM II, BD Biosciences).

Fluorescence Microscopy Analysis
HeLa cells were seeded on glass coverslips in 6-well plate and incubated for 24 hrs at 37uC. The media was removed and replaced with fresh media containing 0.33 mg/well of vectors encoding for wild type asyn, A53T, TP asyn or asyn123, 0.67 mg/ well of pcDNA4/TO/GFP 1-10 and 0.2 mg/well of plasmid encoding for mCherry. The transfection reactions were incubated for 16 hrs, at which point they were washed with 0.1% Tween-20/ PBS and fixed with 4% paraformaldehyde for 30 min. The coverslips were mounted on glass slides for fluorescence microscopy. The slides were imaged using an Olympus IX81 confocal microscope and analyzed using proprietary Fluoview software.

Western Blot Analysis
HeLa cells were plated in 6-well plates and incubated for 24 hrs at 37uC. The media was removed and replaced with fresh MEM media containing 0.5 mg of pCMV-mGFP/asyn-GFP 11 per well and transfected as described above. Cells were lysed with complete lysis-M buffer (Roche) for 30min on ice with gentle rocking. The protein concentration was determined by Bradford assay (Pierce), and each sample was diluted to the same protein concentration. Proteins were separated by 12% SDS-polyacrylamide gels and transferred to a nitrocellulose membrane. Membranes were incubated with primary antibodies (mouse anti-a-syn (Sigma-Aldrich) and rabbit anti-GAPDH (Santa Cruz Biotechnology)) and appropriate secondary antibodies (HRP conjugated goat antirabbit and goat anti-mouse antibodies (Santa Cruz)). Blots were visualized using Millipore Luminata Forte HRP chemiluminescent substrate (Fisher) and quantified using NIH ImageJ software.

Immunofluorescence and Co-localization Analyses
Cells were seeded on glass coverslips in 6-well plate, transfected, incubated in the presence of small molecules for 24 hrs and fixed with 4% paraformaldehyde for 30 min. Cells were permeabilized with a 0.5% Triton X-100, 0.6% 0.5 M EDTA solution in Assay buffer (ProteostatH Aggresome Detection Kit, Enzo) for 30 min on ice, followed by incubation in 8% BSA (blocking buffer) for 1 hr at room temperature. Cells were then incubated for 1 hr with primary antibody (mouse anti-asyn, Sigma-Aldrich), washed with 0.1% Tween-20/PBS, and incubated with secondary antibody (Dylight 649 Goat anti-mouse from KPL). Cells were washed again and incubated with ProteoStatH dye (ProteostatH Aggresome Detection Kit, Enzo) for 30 min in the dark. The coverslips were mounted on glass slides for fluorescent microscopy. The slides were imaged using an Olympus IX81 confocal microscope and analyzed using proprietary Fluoview software.
Co-localization of asyn with the ProteoStatH dye was evaluated using the ImageJ plugin Co-localization Colormap [53]. Results are reported in the form of co-localization heatmaps where hot colors represent positive co-localization, and cold colors represent negative co-localization. The co-localization heatmaps were analyzed using the ImageJ plugin Threshold Colour, which allows RGB images to be filtered based on the hue, saturation, and brightness of the pixels. Images were filtered to display RGB color hues as follows: high co-localization (RGB hue: 0-35, red pixels) and low co-localization (RGB hue: 36-60, yellow pixels). Pixels falling in the RGB hue range 60-255 were considered negative correlation and not evaluated in this study. For each sample, 85-120 cells were analyzed to count co-localization events.

Statistical Analysis
All data are presented as mean 6 S.E.M. Statistical significance was calculated using a two-tailed Student's t test. Values were considered significantly different when p was ,0.05. Figure S1 Effect of inhibition of proteasomal degradation on asyn solubility and split GFP fluorescence complementation. Representative plot of absolute GFP fluorescence in cells expressing asyn-GFP 11 and GFP 1-10 (blue) and TP asyn-GFP 11 and GFP 1-10 (red). Cells were incubated for 24 hrs with increasing concentrations of lactacystin (0-5 mM).

Author Contributions
Conceived and designed the experiments: AK LS. Performed the experiments: AK KK JAN. Analyzed the data: AK KK. Contributed reagents/materials/analysis tools: AK KK JAN. Wrote the paper: AK KK LS.