CHIP Targets Toxic α-Synuclein Oligomers for Degradation*

α-Synuclein (αSyn) can self-associate, forming oligomers, fibrils, and Lewy bodies, the pathological hallmark of Parkinson disease. Current dogma suggests that oligomeric αSyn intermediates may represent the most toxic αSyn species. Here, we studied the effect of a potent molecular chaperone, CHIP (carboxyl terminus of Hsp70-interacting protein), on αSyn oligomerization using a novel bimolecular fluorescence complementation assay. CHIP is a multidomain chaperone, utilizing both a tetratricopeptide/Hsp70 binding domain and a U-box/ubiquitin ligase domain to differentially impact the fate of misfolded proteins. In the current study, we found that co-expression of CHIP selectively reduced αSyn oligomerization and toxicity in a tetratricopeptide domain-dependent, U-box-independent manner by specifically degrading toxic αSyn oligomers. We conclude that CHIP preferentially recognizes and mediates degradation of toxic, oligomeric forms of αSyn. Further elucidation of the mechanisms of CHIP-induced degradation of oligomeric αSyn may contribute to the successful development of drug therapies that target oligomeric αSyn by mimicking or enhancing the powerful effects of CHIP.

␣-Synuclein (␣Syn) can self-associate, forming oligomers, fibrils, and Lewy bodies, the pathological hallmark of Parkinson disease. Current dogma suggests that oligomeric ␣Syn intermediates may represent the most toxic ␣Syn species. Here, we studied the effect of a potent molecular chaperone, CHIP (carboxyl terminus of Hsp70-interacting protein), on ␣Syn oligomerization using a novel bimolecular fluorescence complementation assay. CHIP is a multidomain chaperone, utilizing both a tetratricopeptide/Hsp70 binding domain and a U-box/ubiquitin ligase domain to differentially impact the fate of misfolded proteins. In the current study, we found that co-expression of CHIP selectively reduced ␣Syn oligomerization and toxicity in a tetratricopeptide domain-dependent, U-box-independent manner by specifically degrading toxic ␣Syn oligomers. We conclude that CHIP preferentially recognizes and mediates degradation of toxic, oligomeric forms of ␣Syn. Further elucidation of the mechanisms of CHIP-induced degradation of oligomeric ␣Syn may contribute to the successful development of drug therapies that target oligomeric ␣Syn by mimicking or enhancing the powerful effects of CHIP.
The clinical symptoms of Parkinson disease (PD) 2 are caused by an unknown toxic insult that selectively destroys dopamine-producing cells in the substantia nigra of affected humans (1). ␣-Synuclein (␣Syn)-containing Lewy bodies have long been associated with PD and PD-associated cellular toxicity (2). Recent evidence indicates that ␣Syn can form prefibrillar, oligomeric assemblies as well as fibrillar cellular aggregates, and it has been suggested that these oligomeric assemblies mediate ␣Syn neurotoxicity (3)(4)(5)(6)(7). Therefore, understanding the molecular mechanisms of ␣Syn oligomerization will provide insight into targeting therapeutics that can reduce ␣Syn oligomerization or clear existing ␣Syn oligomers.
Molecular chaperones such as heat shock proteins (Hsp) 70 and 90 and their associated co-chaperone CHIP (carboxyl terminus of Hsp70-interacting protein) have a well documented ability to reduce toxicity mediated by neurodegenerative disease states, stress, or injury (8 -17). CHIP functions as an important link between the molecular chaperone system and the protein degradation system. CHIP has two specialized protein domains, each conferring different yet complementary functions on CHIP activity. The amino terminus domain of CHIP contains a tetratricopeptide (TPR) domain that links it to the molecular chaperones Hsp70 and 90 (18). The carboxyl terminus domain of CHIP contains a U-box domain that confers the E3 ubiquitin ligase function of CHIP.
Our earlier studies suggest that CHIP can enhance degradation of ␣Syn and protect cells from its toxicity. We have now taken advantage of a bimolecular complementation approach to generate stable ␣Syn oligomers and demonstrate that CHIP selectively promotes degradation of this potentially more toxic, oligomeric form.
Cell Culture and Transient Transfections-Human H4 neuroglioma cells (HTB-148; ATCC) were maintained in OPTI-MEM medium supplemented with 10% fetal bovine serum (both from Invitrogen) and incubated at 37°C, 5%CO 2 . H4 cells were plated 24 h prior to transfection and grown to 80 -90% confluency. Cells were transfected with an equimolar ratio of DNA constructs using Superfect (Qiagen) according to the manufacturer's instructions. Co-transfection of pcDNA3.1ϩ vector was used as a control.
GN-␣Syn/␣SynGC Fluorescence Complementation Assay-For optimal fluorophore maturation, transiently transfected cells were incubated overnight at 30°C (21) after an initial 4-h incubation at 37°C. 24 h after transfection cells were either observed using a Zeiss LSM510 confocal microscope or harvested for cell lysate preparation.
Live Cell Imaging-Cells were plated on poly-D-lysine-coated coverslipped 35-mm dishes (MatTek Cultureware, Ashland, MA) and transfected the following day. Cells were imaged 24 h post-transfection on a Zeiss LSM510 confocal microscope system. Cells were observed at ϫ63 for quantification of pixel intensity. For each experimental group the average fluorescence intensity from 50 cells, across three independent experiments, was measured using Adobe Photoshop software. Adobe Photoshop translated the average GFP fluorescence to average pixel intensity. Therefore, the average pixel intensity was recorded for each cell. Data were normalized to empty vector (pcDNA3.1ϩ) controls.
Fluorescent-activated Cell Sorting-Cells were plated and transfected in 10-cm dishes. 24 h post transfection, cells were trypsinized, centrifuged, and the pellet reconstituted in phosphate-buffered saline. The resulting suspension was filtered with cell strainer caps into polypropylene tubes (both from BD Biosciences). Fluorescence was measured on a FACSCanto (BD Biosciences) flow cytometer.
Toxicity Assay-24 h post-transfection levels of toxicity were measured utilizing the ToxiLight TM non-destructive cytotoxicity assay kit (Cambrex, Rockland, ME). This assay quantitatively measures the release of adenylate kinase from damaged cells into the culture medium and was performed according to the manufacturer's protocol.
Detergent Solubility Fractionation and Gel Electrophoresis-After 24 h, cells were washed with cold phosphate-buffered saline, harvested by scraping in cold lysis buffer with or without detergent for SDS-PAGE and native gels, respectively, (50 mM Tris/HCl, pH 7.4, 175 mM NaCl, 5 mM EDTA, pH 8.0, and protease inhibitor mixture), and sheared once through a 28-gauge needle followed by sonication for 10 s (total cell lysates). Fractionation was performed by adding Triton X-100 (Tx) to total cell lysates (final concentration of 1%) and incubating for 30 min on ice followed by centrifugation (15,000 ϫ g, 60 min, 4°C). The supernatant was designated as the Tx-soluble fraction, and the pellet was redissolved in 2% SDS-containing lysis buffer and sonicated for 10 s (Tx-insoluble fraction). Protein concentration was determined using a BCA protein assay. 15-40 g of protein of each cell lysate was loaded onto a 10 -20% Tris/glycine gel (Invitrogen) or a Native PAGE Novex 4 -16% Bis-Tris gel (Invitrogen) for Western blot analysis. SDS-PAGE was performed with SDS-containing running and sample loading buffer (Invitrogen), whereas native electrophoresis was performed using detergent-free running and sample buffers (Invitrogen). Proteins resolved on the gels were transferred onto polyvinylidene difluoride membrane (Millipore, Billerica, MA) and blocked in 5% milk in Tris-buffered saline (TBS) containing 0.05% Tween 20 (TBS-T) for 2 h prior to the addition of primary antibody, anti-synuclein (610787, 1:1000; BD Biosciences), or anti-GFP (ab6556, 1:3000; Abcam, Cambridge, MA) at room temperature for 2 h or 4°C overnight. Horseradish peroxidase-labeled anti-mouse or anti-rabbit antibodies (1:2000; Jackson ImmunoResearch, West Grove, PA) were incubated at room temperature for 1 h, processed, and quantified. The results were visualized with ECL (GE Healthcare) or SuperSignal Femto (Pierce) chemiluminescent reagents. Denatured blots were probed with anti-glyceraldehyde-3-phosphate dehydrogenase (ab9485, 1:3000; Abcam) and anti-rabbit horseradish peroxidase (1:2000) as a loading control. A nonspecific protein band at ϳ150 -200 kDa across all samples in the series (Coomassie Blue staining of the gel) served as a loading control for the native blots. ImageJ and ImageQuant were used to quantify optical densities of total, soluble, and insoluble blots and native blots, respectively.

RESULTS
CHIP Reduces ␣Syn Oligomerization-We have previously shown that CHIP modifies and prevents ␣-Syn inclusion formation by interacting (either directly or indirectly) with ␣Syn (22). To further investigate the specific ␣Syn species that is targeted by CHIP, we employed a novel BiFC assay that we have previously demonstrated to be a specific and effective method by which to visualize and monitor ␣Syn oligomerization in vitro (19). BiFC involves the reconstitution of non-fluorescent fragments of GFP when fused to proteins of interest. Upon proteinprotein interaction the two fragments are able to come together and reconstitute the fluorophore. We investigated the ability of CHIP to reduce ␣Syn oligomerization by transfecting H4 human neuroglioma cells with the ␣Syn BiFC constructs ␣SynGC and GN-␣Syn in the presence or absence of CHIP (Fig.  1) and found that CHIP dramatically reduced ␣Syn oligomerization as measured by intensity of GFP fluorescence (Fig. 2, A  and B). As a control, we confirmed that CHIP had no effect on fluorescence intensity of enhanced green fluorescent protein (EGFP) (Fig. 2, A and B). To confirm that CHIP's prevention of ␣Syn oligomerization is specific to ␣Syn, and not EGFP, we conducted co-immunoprecipitation experiments and deter- mined that CHIP did not co-immunoprecipitates with EGFP alone (data not shown).
To confirm that the effect of CHIP on ␣Syn oligomerization is specific, we overexpressed domain mutant constructs of CHIP (CHIP⌬U and CHIP⌬TPR) with ␣SynGC and GN-␣Syn (Fig. 2C) and determined that CHIP⌬U, but not CHIP⌬TPR, was able to significantly reduce the fluorescence produced by ␣Syn oligomer formation (Fig. 2D). Because the TPR domain of CHIP is essential for Hsp70 interactions, these data suggest that the CHIP-mediated reduction in ␣Syn oligomer formation occurs in an Hsp70-associated manner.
To quantify the CHIP domain mutant effects, we co-transfected H4 human neuroglioma cells with ␣SynGC, GN-␣Syn, and either empty vector, CHIP, CHIP⌬U, or CHIP⌬TPR and measured GFP fluorescence intensity with fluorescent-activated cell sorting. In agreement with our observations of GFP pixel intensity (Fig. 2), fluorescent-activated cell sorting analysis demonstrated that CHIP and CHIP⌬U mediate a reduction in ␣Syn oligomerization (Fig. 3, A and B). As a control, we verified that CHIP had no effect on ␣Syn-EGFP fluorescence, confirming that CHIP reduces ␣Syn oligomer formation in a TPR domain-dependent manner.
CHIP Reduces ␣Syn Oligomer-induced Toxicity-We have shown previously that overexpression of the BiFC pairs ␣SynGC and GN-␣Syn results in increased ␣Syn oligomerization formation and cytotoxicity (19). To determine whether CHIP can rescue cytotoxicity conferred by oligomeric ␣Syn, H4 cells were cotransfected with ␣SynGC, GN-␣Syn, and ␣SynGC and GN-␣Syn together, in the presence or absence of CHIP or its domain mutants (CHIP⌬U and CHIP⌬TPR), and toxicity was monitored. As expected, CHIP and CHIP⌬U rescued toxicity conferred by ␣SynGC and GN-␣Syn coexpression (Fig. 4). By contrast, CHIP had no effect on ␣Syn toxicity induced by overexpression of individual ␣Syn constructs, i.e. ␣SynGC expressed alone (Fig. 4). These data suggest that CHIP preferentially reduces cytotoxicity conferred by a stabilized ␣Syn oligomeric species and not that of the less toxic, transient ␣Syn constructs. This experiment also confirmed that the TPR domain is critical for the toxicity-rescuing effects of CHIP because CHIP⌬TPR was unable to rescue toxicity. These data are in agreement with CHIP preventing the formation of ␣Syn oligomers (Figs. 2 and 3).
CHIP Preferentially Degrades ␣Syn Oligomeric Species-To determine the mechanism of CHIP-mediated effects on ␣Syn, we examined the expression levels of ␣Syn in the presence or absence of CHIP under two conditions in which oligomeric species are to determine the mechanism of CHIP-mediated effects on ␣Syn, we examined the expression levels of ␣Syn in the presence or absence of CHIP under two conditions in which oligomeric species are, 1) less stabilized: ␣SynGC or GN-␣Syn expressed individually or, 2) more stabilized: ␣SynGC and GN-␣Syn expressed together. Total cell lysate and Tx-soluble and Tx-insoluble lysates were prepared. Interestingly, in the total cell lysate fraction, CHIP had only a modest effect on ␣Syn levels when ␣SynGC and GN-␣Syn were expressed individually (Fig. 5A). However, when the ␣Syn BiFC proteins were coexpressed, CHIP significantly reduced the protein levels (Fig. 5A).
To determine the effect of CHIP on ␣Syn solubility we extracted Tx-soluble and -insoluble fractions. We found that CHIP had little or no effect on levels of ␣Syn in the soluble or insoluble fraction when ␣SynGC and GN-␣Syn were expressed individually (Fig. 5, B and C). However, when coexpressed, CHIP degrades ␣Syn in both the soluble and insoluble fractions (Figs. 5, B and C).
CHIP Reduces High Molecular Weight ␣Syn Oligomeric Species-The co-expression of ␣SynGC and GN-␣Syn results in high molecular weight (HMW) oligomeric species when examined by native Western blot. To determine whether CHIP could degrade HMW ␣Syn oligomers we co-transfected H4 cells with ␣SynGC and GN-␣Syn with an empty vector (pcDNA), CHIP, CHIP⌬U, or CHIP⌬TPR. Interestingly, CHIP efficiently degrades the HMW ␣Syn population. Furthermore, consistent with our previous data, the TPR domain, but not the U-box domain, is critical for this oligomeric degradation because CHIP⌬TPR is ineffective (Fig. 6, A and B).
CHIP, but Not CHIP⌬TPR, Coimmunoprecipitates with Hsp70-CHIP and Hsp70 form a complex between the aminoterminal TPR domain of CHIP and the carboxyl-terminal EEVD motif of Hsp70. Our data indicate that the TPR domain of CHIP is critical for the CHIP-mediated reduction of ␣Syn oligomers and associated toxicity. To determine whether the association of Hsp70 with CHIP, ␣SynGC, and GN-␣Syn is critical, we co-transfected H4 cells with Hsp70, ␣SynGC, GN-␣Syn, and CHIP or CHIP⌬TPR. Both CHIP and CHIP⌬TPR coimmunoprecipitate with GN-␣Syn and ␣SynGC; however, only fulllength CHIP, and not CHIP⌬TPR, coimmunoprecipitates with Hsp70 (Fig. 7). These data imply that Hsp70 is required for the CHIP-mediated reduction of ␣Syn oligomers and their associated toxicity.

DISCUSSION
CHIP has been shown to mediate the degradation of misfolded proteins associated with Alzheimer disease, Huntington disease, Spinocerebellar ataxia, and PD (22,24,26,27). This degradation leads to an overall reduction of cellular toxicity and enhancement of cell survival (23,24,26,28). In PD, current thinking identifies ␣Syn oligomeric species as one of the primary mediators of toxicity (3)(4)(5)(6)(7). In the current study we hypothesized that CHIP preferentially targets ␣Syn oligomers for degradation.
The stabilization of ␣Syn in an oligomeric conformation using the novel BiFC assay leads to an increase in cytotoxicity, which can be rescued by CHIP in an Hsp70-dependent manner. By contrast, CHIP did not rescue the cytotoxicity associated with the more transient ␣Syn interactions, which suggests that CHIP targets a specific ␣Syn conformation for degradation. Moreover, CHIP degrades ␣Syn species in both Tx-100-soluble and Tx-100-insoluble fractions and significantly reduces the formation of HMW ␣Syn oligomers. These data are consistent with our previous study where overexpression of Hsp70 prevented ␣Syn oligomerization (19).
We adapted a BiFC assay (28) to stabilize and detect the formation of ␣Syn oligomers (19). Here we show that CHIP targets an ␣Syn oligomeric species for degradation. This activity is associated with cytoprotection, consistent with the idea that oligomeric species are toxic moieties (3,29). The degradative preference of CHIP for the more toxic protein species is also evident in other models of neurodegeneration. For example, CHIP preferentially degrades the more toxic A␤42, but not the less toxic soluble APP, holoprotein (28). Furthermore, CHIP also preferentially inhibits toxicity induced by the more toxic four repeat tau compared with the less toxic three repeat tau (23). Thus, the degradative preference of CHIP for a more toxic ␣Syn species versus a less toxic ␣Syn species is consistent with other models of neurodegeneration.
The ability of CHIP to affect aggregate formation and degrade disease-causing proteins is well documented as is its modulatory effects on these proteins within soluble and insoluble protein fractions (15, 16, 24 -26). In models of tau and ataxin-1 overexpression, CHIP increases aggregate formation and accumulation of tau and ataxin-1 within the insoluble fraction (15,16,28). However, in instances of huntingtin and ␣Syn overexpression, CHIP reduces aggregate formation (22,24). This CHIP-mediated decrease in aggregation leads to a decrease in huntingtin protein from the insoluble fraction. Our data, which also show a CHIP-mediated decrease in ␣Syn from the insoluble fraction, agree with this finding (24). To our knowledge, these are the first experiments to investigate the effects of CHIP on oligomer formation in neurodegenerative disease. ␣Syn-induced toxicity is largely attributed to the formation of ␣Syn oligomers (3-7), and therefore, therapeutically, it will be important to reduce their formation and accumulation. Utilizing the BiFC assay to model ␣Syn oligomer formation, we determined that CHIP reduces the amount of higher order ␣Syn oligomers. This result has important implications for CHIP and chaperone proteins as a potential therapeutic for ␣Syn-associated toxicity and also for other models of neurodegeneration where oligomer formation is a contributing causative agent.
The mechanism of CHIP-induced protein degradation is not completely understood. It is known that CHIP interacts with Hsp70 via the amino-terminal TPR domain and the EEVD motif of Hsp70. However, it has been postulated that CHIP exhibits intrinsic molecular chaperone activities and can associate with misfolded proteins in an Hsp70-independent manner (30). Our data support this hypothesis in that CHIP⌬TPR can associate with GN-␣Syn and ␣SynGC in the absence of  . CHIP preferentially reduces the toxic ␣Syn oligomers from the total cell lysate and Tx-soluble and Tx-insoluble fractions. H4 cells were co-transfected with either GN-␣Syn or ␣SynGC alone or with GN-␣Syn and ␣SynGC together, plus CHIP (CH) or Empty Vector (EV). The optical density of ␣SynGC and GN-␣Syn was measured from the total (A), Tx-soluble (B), and Tx-insoluble (C) fractions for EV-or CH-transfected cells. Each bar represents protein levels from cells transfected with CH and are relative to EV. To generate each bar, the ratio between CH and EV was calculated from the optical densities of the Western blot. Representative GFP-probed blots are provided; the loading controls for A/B and C are glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and Coomassie blue, respectively.
Hsp70. In further support of this hypothesis, our laboratory has shown that CHIP⌬TPR can also associate with misfolded ␣Syn in the absence of Hsp70 (22). However, while CHIP associates with different ␣Syn species in an Hsp70-independent manner, we have previously shown that degradation of different ␣Syn species by CHIP can occur with or without Hsp70 (22). In the present study, CHIP degraded ␣Syn oligomers and reduced toxicity in an Hsp70-dependent manner. These data suggest that CHIP has the ability to process different species of ␣Syn via different mechanisms. Further experimentation will be required to elucidate the specific degradation mechanisms involved in the CHIP-mediated degradation of oligomeric ␣Syn.
In summary, we have shown that CHIP has the ability to preferentially promote degradation of ␣Syn oligomeric species from both the soluble and insoluble cellular compartments. We postulate that HMW forms of ␣Syn are toxic and that the CHIP-mediated reduction in ␣Syn-induced toxicity is produced by the ability of CHIP to preferentially degrade these toxic ␣Syn species. Understanding the molecular mechanisms of CHIP-induced toxicity will elucidate more effective therapeutic strategies for the treatment of neurodegenerative disease.