Studies on the activity of ataxin-3 and miyochondrial dysfunction in models of Machado-Joseph disease

Abstract

Machado-Joseph disease (MJD), also known as spinocerebellar ataxia type 3 (SCA3) is an autosomal, dominant, hereditary neurodegenerative disorder and the most common cause of autosomal inherited ataxia in the world. The first symptoms of MJD usually become apparent in adulthood, around the third or fourth decade of life, although in severe cases, the onset may occur even in childhood. MJD patients exhibit a variety of phenotypes, in a wide combination of motor and non-motor symptoms. Ataxia and gait impairment are the main and primary manifestations of this neurological disorder, associated to a variable degree of oculomotor deficits, dysarthria, pyramidal signs, progressive sensory loss and parkinsonism. Symptoms progressively worsen through the course of the disease until patient’s death, 20 or 30 years after the first symptoms. MJD belongs to the polyglutamine (polyQ) expansion disorders, a group of nine neurological pathologies caused by a dynamic mutation, an expansion in the number of CAG trinucleotide repeats. In MJD, the CAG expansion is present on exon 10 of the MJD1 gene, which codes for the deubiquitinating enzyme, ataxin-3. The onset of MJD symptoms occurs when the CAG expansion extends beyond 52 repeats. The CAG expansion is translated into an expanded polyQ domain within the Cterminus of ataxin-3. A pathological polyQ expansion destabilizes ataxin-3 structure, promotes its misfolding and aggregation and bestows the protein with new toxic properties, which lead to neuronal dysfunction, and consequent neurodegeneration. Although both wild-type and pathological expanded ataxin-3 are expressed throughout all tissues of the body, degeneration occurs only in particular regions of the central nervous system (CNS). Subcortical regions as the pontine nuclei, the dentate nucleus, subthalamic nucleus and the spinal cord are among the most affected. The reasons for this pattern of neurodegeneration and the specific susceptibility of these neurons to expanded ataxin-3 are still not understood. Ataxin-3 is a 42kDa intracellular protein with ability to bind and cleave polyubiquitin chains, and currently, it is classified as deubiquitinating enzyme of the Josephin domain proteases class. Ataxin-3 has higher affinity for long ubiquitin chains, with more than four ubiquitin moieties, and a catalytic preference for lysine 63 (K63) mediated linkages between two ubiquitin molecules. Ataxin-3 is regulated by monoubiquitination and phosphorylation events, which determine its activity and intracellular localization. The biological role of ataxin-3 is not entirely clarified. It is suggested that ataxin-3 acts as an ubiquitin chain editing enzyme and it has been implicated in ubiquitin-proteasome pathways, endoplasmic reticulum associated degradation, the cytoprotective response to heat shock stress and in the determination of Caenorhabditis elegans longevity. The implication in a multitude of pathways is coupled to a growing number of proteins found to interact with ataxin-3. Recent ataxin-3 interactors have been added to the previously known p97/valosin-containing protein (VCP/p97) and hHR23A/hHR23B (human homologs of yeast Rad23 protein), namely tubulin, microtubule-associated protein 2 (MAP2), dynein, histone deacetylase (HDAC) 6, several transcriptional activators/repressors and parkin. Some interactions are specific of wild-type and are disrupted by the expanded polyQ domain, while others only occur for expanded ataxin-3. In vitro studies have not found significant differences between the activity of human wild-type and expanded ataxin-3. Therefore, it was a logical step to investigate the role of protein interactions in ataxin-3 activity and how the deubiquitinating activity influences ataxin-3 properties in the cell. Moreover, it was crucial to clarify the contribution of these features and the involvement of mitochondrial dysfunction in the pathogenesis of MJD. To achieve these goals, we performed in vitro deubiquitinating studies with wild-type and expanded ataxin-3 in the absence or presence of wellestablished protein interactors and used several MJD cell and animal models to evaluate expanded ataxin-3 toxicity, particularly focusing on mitochondrial dysfunction, a pathological mechanism shared by several polyQ diseases In Chapter 2, the effect of two well-known ataxin-3-interacting proteins, hHR23A and VCP/p97, was investigated over the deubiquitinating activity of ataxin- 3. Endogenous hHR23A and endogenous ataxin-3 were predominantly located in the nucleus, whereas endogenous VCP/p97 was more abundant in the cytoplasm. The interactions of VCP/p97 and hHR23A with ataxin-3 were corroborated, despite the differences in intracellular localization and the assembly of a trimolecular complex containing the three endogenous proteins was observed. VCP/p97 and hHR23A interactions with ataxin-3 were also established in in vitro conditions. Human recombinant hHR23A, VCP/p97, wild-type and expanded ataxin-3 were purified and subsequently used in in vitro deubiquitinating assays to assess the influence of hHR23A and VCP/p97 on ataxin-3 activity. Ataxin-3 exhibited a slow rate of enzymatic activity and a catalytic preference for lysine-63 (K63)-linked polyubiquitin chains. Addition of hHR23A to the in vitro assays did not alter ataxin-3 kinetics or substrate preference. On the contrary, the presence of VCP/p97 was able to increase specifically wild-type ataxin-3 performance in vitro. Interestingly, this in vitro VCP/p97-mediated activation was not detected for expanded ataxin-3. In Chapter 3, fully active and catalytically inactive ataxin-3 were expressed in cells and used in in vitro assays to continue the study of ataxin-3 activity, focused on the influence of catalytic activity over ataxin-3 cellular behavior. Inactive ataxin-3 displayed higher protein levels in comparison to active ataxin-3 due partially to a slower degradation rate of catalytically inactive ataxin-3. Intriguingly, inactive ataxin-3 was more heavily ubiquitinated in the cell than its active counterpart. Despite the higher level of ubiquitination, a slower proteasomal degradation rate for inactive ataxin-3 was determined through in vitro studies. The lower proteasomal degradation was correlated with a reduced interaction of catalytically inactive ataxin-3 with VCP/p97, a protein known for shuttling substrates to the proteasome. Ataxin-3 did not interact directly with the proteasome or any of its subunits; however, ataxin-3 and the proteasome were shown to colocalize in discrete subnuclear foci. Interestingly, inactive ataxin-3 exhibited less nuclear colocalization with the proteasome and a diminished localization in the nuclear compartment. Overall, the enzymatic activity of ataxin-3 was shown to modulate its cellular turnover, intracellular localization and the level of ubiquitination. In Chapter 4, several cell and animal models of MJD were analyzed to evaluate the contribution of mitochondrial dysfunction in MJD pathology. HEK cells transiently transfected with expanded (Q84) enhanced green fluorescent protein (EGFP)-ataxin-3 fusion protein were more susceptible to inhibition of mitochondrial complex II with 3-nitropropionic (3-NP) than HEK cells expressing wild-type (Q28) EGFP-ataxin-3. Similarly, stably transfected rat pheochromocytoma (PC6-3) cell line expressing expanded (Q108) human ataxin-3, in a tetracycline-regulated manner, also displayed a higher susceptibility to 3-NP inhibition than PC6-3 cells expressing the wild-type (Q28) protein. Primary cerebellar granule cells isolated from MJD transgenic mice exhibited the same trend. A higher vulnerability to 3-NP was observed for the transgenic cerebellar granule cells when compared to wild-type cerebellar neurons. Moreover, mitochondria from MJD transgenic mouse brains and lymphoblast cell lines derived from MJD patients showed a tendency towards a reduction in mitochondrial complex II activity. The reduction in complex II activity was even more evident, and statistically significant, in mitochondria of PC6-3 (Q108) cells, stimulated into a neuronal-like phenotype with nerve growth factor (NGF). In conclusion, a functional ubiquitin protease capability determines more cellular properties of ataxin-3 than exclusively its DUB activity, influencing the turnover and intracellular localization of ataxin-3. Interestingly, catalytically compromised ataxin-3 displays a reduced ability to interact with VCP/p97. On the other way around, VCP/p97 interaction stimulates wild-type ataxin-3 enzymatic activity, suggesting a direct association between VCP/p97 binding and ataxin-3 performance as a DUB. Therefore, this interaction might become a good target not only to modulate ataxin-3 activity but other properties of this protein. More so, as the DUB activity of expanded ataxin-3 was shown to be insensitive to VCP/p97 stimulation, it what might constitute a pathological loss-of-function mechanism. Moreover, the expression of expanded ataxin-3 also exerts mitochondrial dysfunction, through reduction of mitochondrial complex II activity, adding a new pathway in MJD pathogenesis.