Role of phosphorylation of ataxin-3 and oxidative stress in the pathogenesis of spinocerebellar ataxia type (SCA3)

Spinocerebellar ataxia type 3 (SCA3) is a dominantly inherited, neurodegenerative disease caused by the presence of an expanded polyglutamine repeat inside ataxin-3, the protein encoded by MJD1 gene (located on chromosome 14). Bioinformatic analysis, using NetPhos 2.0, PHOSIDIA and ScanProsite, showed the presence of eight casein kinase 2 (CK2) phosphorylation sites and three glycogen synthase kinase 3 (GSK3) phosphorylation sites in ataxin-3 sequence. CK2 is a pleiotropic kinase involved in many cellular processes and GSK3 has turned out to have a key role in the regulation of many cell functions, both kinases are involved in neurodegenerative diseases. In order to investigate the possible role of ataxin-3 phosphorylation by CK2 and GSK3 on its sub-cellular localization, in vitro phosphorylated AT-3Q6 was analyzed by mass spectrometry phosphomapping. The analysis identified 7 phosphorylation sites for CK2: the serines 29, 272, 277, 329, 341, 344 and the threonine 271 and 3 phosphorylation sites for GSK3: the serines 29, 268 and 273. We obtained mutants lacking phosphorylation sites for CK2 and GSK3 by PCR, substituting serines and threonine with alanines: S29A, S272A, S277A, DM (S272A and S277A), TM (T271A, S272A and S277A) and QM (S268A, T271A, S272A, S273A and S277A). Sub-cellular localization of these mutants was investigated in COS7 transfected cells, through sub-cellular fractionation, mitochondria extraction and confocal microscopy. Only S29A mutant showed a phenotype, in fact it had a lower localization in the nucleus than the wild-type, suggesting that this phosphorylation site might be involved in nuclear ataxin-3 import. In order to confirm the role of S29 phosphorylation on ataxin-3 nuclear uptake, this residue was substituted by PCR with an aspartic acid, yielding S29D mutant, mimicking the presence of the phosphate negative charge. Confocal microscopy and sub-cellular fractionation showed that this mutant has the same sub-cellular localization of wild-type, demonstrating that the identified phenotype is due to lack of phosphorylation and not to amino acidic substitution. Sub-cellular localization was also investigated in SHSY-5Y cells, a human neuronal cancer cell line, and also in these cells the S29A mutant showed a lower nuclear localization than the wild-type. COS-7 cells expressing either AT-3Q6 or AT-3Q6S29D were grown in the presence of CK2 inhibitor TBB (tetrabromobenzotriazole) and GSK3 inhibitor SB216763. Inhibitors were added to growth media either separately or in combination. When the culture medium was supplemented with TBB alone, both wild-type AT-3Q6 and S29D mutant were found evenly distributed between the cytosol and the nucleus, suggesting that TBB could not prevent S29 phosphorylation. The same happened when only GSK3 inhibitor SB216763 was administered. However, when both inhibitors were added to culture medium, nuclear translocation of wild-type AT-3Q6 appeared to be reduced, a behaviour closely mirroring that of S29A mutant. These data clearly show that CK2 and GSK3 can substitute each other in the phosphorylation of S29, promoting AT-3 nuclear uptake. Mueller and coworkers (Mueller t. et al., 2009) demonstrated that serine 340 and 352 within the third ubiquitin-interacting motif of AT- 3 were particularly important for nuclear localization of normal and expanded AT-3. To demonstrate if the combination of S29A with Mueller’s mutants could suppress AT-3 nuclear uptake, we obtained AT-3Q6M substituting S29, S329 and S341 with alanines, in fact S329 and S341 correspond to S340 and S352 respectively in human AT-3. Our data demonstrated that removal of S329 and S341 does not further reduce S29A nuclear uptake. To demonstrate if S29 was also involved in nuclear uptake of AT-3 pathological form we mutated S29 of AT-3Q72 to alanine. Our results demonstrated that the mutation does not change AT-3Q72 nuclear localization. VCP/p97 is a key protein essential for the extraction of substrates from the endoplasmic reticulum (ER) to the cytosol in ER-associated degradation (ERAD). Since VCP/p97 interacts with ataxin-3 in a region between aminoacids 257 and 291 (Boeddrich A. et al.,2006), we studied the interaction between the mutants lacking phosphorylation sites in this region, and VCP to understand if phosphorylation was involved in this interaction. In this region of interaction with VCP there are 3 phosphorylation sites for CK2 (T271, S272 and S277) and 2 phosphorylation sites for GSK3 (S268 and S273). We studied the interaction between VCP/p97 and AT-3Q6 or the mutants TM (T271A, S272A and S277A) and QM (S268A, T271A, S272A, S273A and S277A) by immunoprecipitation of AT-3 from COS7 cells co-transfected with VCP and AT-3. The mutants showed the same interaction with VCP as the wild-type, suggesting that these phosphorylation sites are not involved in interaction with VCP/p97. To assess whether the phosphorylation sites had an effect on the interaction between VCP/p97 and its substrates, we treated transfected cells with DTT to obtain an ER stress. No difference between AT-3Q6 and its mutants was detected in these conditions. Nevertheless, in the presence of ERAD stress, AT-3 showed a weaker interaction with VCP than in normal conditions. This suggests that when ER is under stress, VCP/p97 must increase its interaction with its normal partner Ufd1 and with polyubiquitin chains to degrade ERAD substrates, so VCP/p97 must decrease its interaction with AT-3. We also studied the interaction between VCP/p97 and AT-3Q72 with or without inducing ER stress conditions and we showed that there was no change in the interaction between AT-3Q72 and VCP/p97 under stress condition. Nevertheless AT-3Q72 has a stronger interaction with VCP than AT-3Q6, as demonstrated by Zhong and coworkers (Zhong X. et al., 2006). Therefore the pathological form of AT-3 interacts aberrantly with VCP/p97 also in ER stress condition, leading to a deregulation of ERAD. Overproduction of reactive nitrogen species (RNS) and reactive oxygen species (ROS), which lead to neuronal cell injury and death, is a potential mediator of neurodegenerative disorders. To understand the role of nitrosative stress in Spinocerebellar Ataxia Type 3 we studied the expression of NO synthase in Neuro2a cells transfected with AT-3Q6 or AT-3Q72. The results demonstrated that in presence of AT-3 pathological form there was an increase in NOS expression. Thus we studied nitration pattern through mono and bidimensional analysis and we observed that in presence of pathological form of AT- 3 there was a remarkable increase in protein nitration, both as nitration level and as number of nitrated proteins. Therefore the presence of AT-3 pathological form causes activation of nitrosative stress.