p62/Sequestosome 1 Regulates Aggresome Formation of Pathogenic Ataxin-3 with Expanded Polyglutamine

The cellular protein quality control system in association with aggresome formation contributes to protecting cells against aggregation-prone protein-induced toxicity. p62/Sequestosome 1 (p62) is a multifunctional protein which plays an important role in protein degradation and aggregation. Although poly-ubiquitination is usually required for p62-mediated protein degradation and aggresome formation, several p62 substrates are processed to form aggregate in an ubiquitination-independent manner. In this study we demonstrate that p62 directly interacts with pathogenic Machado Joseph Disease (MJD)-associated protein ataxin-3 with polyglutamine (polyQ) expansion. Moreover, p62 could regulate the aggresome formation of pathogenic ataxin-3 and protect cells against pathogenic ataxin-3-induced cell death.


Introduction
Machado Joseph Disease (MJD)/Spinocerebellar ataxia type 3 (SCA3), the most common SCA subtype, is a polyglutamine disease caused by the expansion of a CAG stretch in MJD1 gene [1,2]. The MJD1 gene in unaffected persons contains 12-40 CAG repeats, whereas mutant MJD1 gene in patients has 62-86 CAG repeats [1]. The CAG repeat expansion in the MJD1 gene results in an expanded polyglutamine (polyQ) tract in the encoded protein ataxin-3 [3].
p62/Sequestosome1 (p62), mutation of which is associated with human disorders such as ALS and Paget disease of the bone, is a common component of protein aggregates and aggresomes in neurodegenerative disorders [22][23][24]. p62 interacts with poly-ubiquitinated proteins and microtubule-associated protein 1 light chain 3 (LC3) to function in autophagy-lysosome pathway, and p62 also functions in protein aggregate and aggresome formation [23][24][25][26]. Although many misfolded proteins undergo poly-ubiquitination, which is a signal for recognition by p62 and for aggregate formation, several interesting studies showed that p62 substrates could be processed to form aggregate in an ubiquitination independent manner. For example, p62 regulates mutant SOD1 aggregation by directly interacting with mutant SOD1 [22,23,27].
In the current study we show that ataxin-3-Q80, a type of pathogenic ataxin-3 with polyQ expansion, forms aggresome under proteasome dysfunction in cultured cells. We also found that p62 can regulate aggresome formation of pathogenic ataxin-3, and p62 physically interacts with pathogenic ataxin-3, but not normal ataxin-3. Moreover, p62 regulates the aggresome formation of polyQ expanded ataxin-3 in a microtubule dependent manner, and protects cells against the polyQ expanded ataxin-3-induced cell death.

p62 Promotes the Aggresome Formation of Ataxin-3-Q80
As p62 can regulate the protein aggregation of mutant huntingtin and SOD1 [23,25] to play a protective role, we wondered whether p62 plays a role in the aggregation of ataxin-3-Q80. In cultured cell model, ataxin-3-Q80 displayed three distinct distribution forms: (1) diffusion throughout the cells; (2) multiple small aggregates; (3) large inclusions ( Figure 2A). As the large inclusions were singularity and juxtanuclear (Figure 2A,B,F), and the inclusion formation depended on microtubule integrity ( Figures 4B,C), we speculated that they are aggresomes. As predicted, results confirmed that those inclusions formed by ataxin-3-80Q and p62 colocalized with γ-tubulin, an MTOC and aggresome marker ( Figure 2D). When we used MG-132, a proteasome inhibitor, the rate of aggresome formation was strikingly increased in cells ( Figure 2B). Overexpression of p62 promoted the aggresome formation of ataxin-3-Q80 ( Figure 2B,C), whereas knockdown of p62 inhibited the aggresome formation of ataxin-3-Q80 under both normal or MG-132 treated condition ( Figure 2D-F). These data suggest that p62 can promote the aggresome formation of ataxin-3-Q80.

p62 Has no Effect on the Protein Expression of Ataxin-3
Given that p62 is a key regulator in protein degradation [23,25,28], to determine whether p62 affects ataxin-3 degradation, we overexpressed or knocked down p62 to assess the protein level of ataxin-3-Q20/Q80. Results showed neither overexpression nor knockdown of p62 affected the protein level of ataxin-3s in HeLa cells ( Figure 3A-D), indicating that p62 regulates aggresome formation of ataxin-3-Q80, but has no effect on ataxin-3 turnover.

Discussion
Although previous studies showed that ataxin-3 could regulate aggresome formation of poly-ubiquitinated proteins, such as mutant CFTR∆F508 and mutant SOD1 [15,33], little is known about its own aggresome formation. Using cultured cell models, we firstly demonstrate that polyQ expanded ataxin-3 could form aggresome. Moreover, we identified p62 as a master regulator of ataxin-3 aggresome formation. p62 has been reported to play a role in the aggregate formation of neurodegenerative disease associated proteins, such as ALS-and Huntington's disease-associated proteins [23,25]. However, little is known about the role of p62 in aggresome formation. Interestingly, a very recent study showed that p62 is involved in the regulation of aggresome formation of pathogenic prion protein, which is associated with prion disease [24]. Taken together with our observations (Figure 2), we hypothesize that p62 may be broadly involved in the aggresome formation of many neurodegenerative disease associated proteins.
Although it is still not clear whether protein aggregation, the most common feature in neurodegenerative disorders, has a protective or toxic role, growing evidences suggest that the smaller aggregates/oligomers may exert toxic effects and the deposit of these aggregates/oligomers to aggresomes may help cells to maintain homeostasis [18,34,35]. Thus, aggresome formation is usually considered to be a cellular protective mechanism by which the cell handles toxic misfolded protein oligomers or aggregates [36,37]. In agreement with this hypothesis, our results showed a cytoprotective role of p62-mediated aggresome formation of pathogenic ataxin-3 ( Figure 5). A better understanding of the physiological role of ataxin-3 in MJD requires the identification of the specific binding partners of polyQ expanded ataxin-3. Many cell signaling pathway relative proteins, especially the key regulators involved in protein quality control system, have been identified to be associated with pathogenic ataxin-3, such as hHR23A and B, E4B (UFD2a), p45 and VCP/p97 [38][39][40][41][42][43].
In this study, we identified another regulator in protein quality control system, p62, as a novel partner of pathogenic ataxin-3. In our immunoprecipitation experiments, the pathogenic ataxin-3-Q80 immunoprecipitates p62 to a greater extent, compared with wild type ataxin-3-Q20 in cells, (Figure 1C,D). This is consistent with previous studies that pathogenic ataxin-3 shows enhanced binding to its binding partners [40,42]. The association between p62 and ataxin-3 varies in vitro and in cells, which shows that p62 does not associate with normal ataxin-3-Q20 in vitro but associates with ataxin-3-Q20 in cells, possibly due to the interaction with poly-ubiquitin modified ataxin-3-Q20 in those cells. Like other ataxin-3 binding partners (such as hHR23A and B, E4B (UFD2a) and VCP/p97), p62 binds poly-ubiquitinated proteins and acts as a shuttling factor in the protein degradation pathways. p62 can regulate the degradation of misfolded tau through the proteasome pathway [28], and degradation of misfolded huntingtin through autophagy pathway [25], thereby playing a protective role. In our observations, the protein levels of ataxin-3-Q20/Q80 were not changed when we overexpressed or knocked down p62 (Figure 3). PolyQ expanded ataxin-3 has a self-aggregation property in vitro [29,30]. When we knocked p62 down, the self-association of ataxin-3-Q80 was not affected ( Figure 4A). Also, we speculated that p62 could directly associate with ataxin-3-Q80 before targets ataxin-3-Q80 to the aggresomes. As predicted, p62 still co-localized with ataxin-3-Q80 on the small aggregates in the nocodazole treated cells ( Figure 4B,C), suggesting that p62 promotes the transport of ataxin-3-Q80 aggregates along microtubule to form aggresome, but is not simply recruited to the aggresomes, or affects ataxin-3 expression and self-aggregation. Furthermore, given the fact that the p62-mediated regulation of autophagy and protein aggregation is tightly coupled with mTOR pathway [44], and recent reports have shown that mTOR signaling pathway is associated with neurodegeneration and oxidative stress [45][46][47], it is possible that p62 cooperates with mTOR signaling to function a role in the regulation of aggresome formation of ataxin-3. Thus, the role of mTOR signaling in p62-mediated aggresome formation of ataxin-3 needs to be further explored.

Fluorescent Microscopy
HEK293 Cells were washed with PBS (pH 7.4) and fixed with 4% paraformaldehyde for 10 min at room temperature, then the cells were treated with 0.25% Triton X-100 for 5 min. After blocking with 1% fetal bovine serum for 30 min, cells were incubated with anti-Flag (Sigma, St. Louis, MO, USA), anti-γ-tubulin (Sigma) or p62 (Enzo life science, Farmingdale, NY, USA) antibodies and then with rhodamine (red) or Alexa fluor 350 (Blue) (Invitrogen) conjugated secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 2 h. The nuclei were stained with DAPI (Sigma). The cells were visualized using an IX71 inverted system microscope (Olympus, Tokyo, Japan).

Immunoprecipitation
RIPA buffer was used for preparing the cell lysates. The RIPA-insoluble debris was removed after centrifugation at 12,000 rpm for 30 min at 4 °C. The supernatants were subjected to immunoprecipitation with rabbit polyclonal anti-GFP antibody and protein G Sepharose (Roche) for overnight at 4 °C. The protein G sepharose were washed with RIPA buffer six times and then eluted with SDS sample buffer for immunoblot analysis.

GST Pulldown Assay
In vitro binding experiment was performed using an aliquot containing approximately 20 µg of GST, GST-ataxin-3-Q20 or GST-ataxin-3-Q80 expressed by Escherichia coli, to be incubated with 30 µL of glutathione agarose beads (Pharmacia, Stockholm, Sweden) for 30 min at 4 °C. After washing two times with PBS, the beads were incubated with approximately 40 µg of p62 expressed by E. coli BL21 strain for 3 h at 4 °C. At last, the beads were washed five times with the ice-cold PBS. Bound proteins were eluted with SDS loading buffer for immunoblot analysis.

MTT Assay
The transfected cells were washed with DMEM (without phenol red) and incubated with 0.5 mg/mL MTT (3-(4,5)-dim-ethylthiahiazo(-z-y1)-3,5-di-phenytetrazoliumromide) (Sigma) in DMEM. Three hours after incubation, the media were removed and the formazan crystals were dissolved in dimethyl sulfoxide (DMSO). The optical density (OD) was measured by photometer at 570 nm, with background subtraction at 630 nm. The data from three transfection experiments were normalized to control and the ratios were presented.

Statistical Analysis
The western blot densitometry analyses of immunoblots from three independent experiments were performed by photoshop7.0 software (Adobe, San Jose, CA, USA). The data were analyzed using Origin 6.0 (Originlab, Northampton, MA, USA).

Author Contributions
Liang Zhou, Hongfeng Wang, Zheng Ying and Guanghui Wang designed the study. Liang Zhou, Hongfeng Wang, Dong Chen, Feng Gao and Zheng Ying performed experiments. Liang Zhou, Hongfeng Wang, Zheng Ying and Guanghui Wang analyzed the data. Liang Zhou drafted the manuscript and Zheng Ying and Guanghui Wang revised the manuscript. Zheng Ying and Guanghui Wang are the principal investigators. All authors have read and approved the final manuscript.