DJ-1 and ATP13A2: Two proteins involved in Parkinson's disease

Josephat M Asiago, Purdue University

Abstract

Parkinson's disease (PD) is the second most common neurodegenerative disorder after Alzheimer's disease, affecting approximately 0.3% of the total U.S. population, and its prevalence increases with age. Two neuropathological hallmarks of PD are the loss of dopaminergic neurons in the substantia nigra pars compacta, a region in the midbrain involved in initiating and sustaining movement, and the presence of cytosolic inclusions called Lewy bodies (LBs) in various brain regions. LBs are enriched with fibrillar forms of the presynaptic protein &agr;-synuclein (aSyn). Two autosomal recessive genes implicated in familial PD are PARK9, encoding the P-type ATPase ATP13A2, a lysosomal ATPase; and PARK7, encoding DJ-1, a protein with proposed antioxidant and chaperone activities. Understanding the biochemical mechanisms underlying the neuroprotective functions of DJ-1 and mechanistic details accounting for functional overlap between ATP13A2 and DJ-1 can provide insight into the cellular pathways relevant to PD pathogenesis. DJ-1 belongs to the DJ-1/ThiJ/PfpI superfamily, consisting of proteins that typically function as proteases and chaperones. Most members of this superfamily have shared characteristics including a conserved cysteine residue, a catalytic triad, and the ability to form oligomers. Whereas PH1704, a bacterial protease and a member of the DJ-1/ThiJ/PfpI superfamily, adopts a hexameric structure that is necessary for the formation of a catalytic triad, DJ-1 exists as a homodimer and apparently has a catalytic diad rather than a catalytic triad. A recent study has shown that DJ-1Δ15, a truncated form of DJ-1 (15 amino acids cleaved at the C terminus, resulting in the removal of helix H8, has a much greater protease activity compared to full length DJ-1. The structure of DJ-1Δ15 is similar to that of PH1704, which lacks a C-terminal helix corresponding to helix H8 of full-length DJ-1. We chose to focus on DJ-1Δ15 because it was not known whether this variant has the ability to form a hexamer, or whether such a hexameric structure is responsible for the observed protease activity (similar to PH1704). We analyzed DJ-1Δ15 via computer modelling using PyMOL to determine whether it can form a hexamer with favorable intersubunit interactions. In addition, we prepared recombinant DJ-1Δ15 and analyzed the protein via analytical ultra-centrifugation (AUC), native PAGE electrophoresis, and size exclusion chromatography coupled with multi-angle light scattering. Our PyMOL results showed that (i) DJ-1Δ15 can from a hexamer which may be stabilized by two inter-subunit interfaces, referred to as patch 1' and patch 2 and, (ii) hexamer formation may lead to the formation of a catalytic triad involving residues C106 and H126 from one subunit and E84 from another subunit. The full length DJ-1 was unable to form the hexamer because helix H8 caused steric hindrances. Data from AUC and size exclusion analysis of DJ-1Δ15 showed peaks representing species with different assembly states. There was evidence of a species with a molecular weight greater than that of the DJ-1Δ15 dimer. In addition the native gel data of DJ-1Δ15 showed a higher molecular weight species. Taken together, our results suggest that the DJ-1Δ15 protease activity may be due to the formation of an oligomer that is larger than the homodimer—potentially a hexamer. Analysis of the DJ-1 crystal structure revealed a previously undescribed potential ATP binding site that included two arginine residues, Arg 28 and Arg 48, near the oxidizable Cys106 residue. We therefore focused on elucidating whether DJ-1 binds and hydrolyzes ATP. We found that our preparations of recombinant, human WT DJ-1 had ATPase activity that was lost upon filtration or ultracentrifugation. Analysis of our protein samples via sucrose gradient sedimentation coupled with immunoblotting revealed the presence of high molecular weight (high-MW) species immunoreactive with a DJ-1 antibody in our DJ-1 preparations. Additional studies revealed that the high-MW protein fraction of our DJ-1 preparations had ATPase activity and consisted of ring-like structures that could be visualized by electron microscopy. Furthermore, a DJ-1-positive high-MW complex isolated from these preparations by sucrose gradient sedimentation was shown via mass spectrometry analysis to contain F1 ATPase subunits, which are also known to assemble into ring-like structures, suggesting that the ATPase activity in our high molecular weight fraction might be associated with assembled F1 ATPase. Consistent with this idea, the ATPase activity in our high molecular weight protein fraction was abolished in the presence of sodium azide or piceatannol, classical F1 ATPase inhibitors. Furthermore, we obtained evidence that dimeric DJ-1 may interact with purified E-coli F1 ATP synthase, and this interaction was apparently dependent on the presence of a reduced cysteine residue at position 106. These results imply that (i) F1 ATP synthase (in mitochondria) may be a target of the DJ-1 chaperone activity, and (ii) this interaction may be modulated by DJ-1 oxidation. Loss-of-function mutations in the ATP13A2 or DJ-1 gene have been shown to disrupt lysosomal autophagy and interfere with mitochondrial function and quality control. We hypothesized that dysfunction of either protein elicits neurotoxicity by triggering defects in autophagy coupled with an accumulation of dysfunctional mitochondria. To address this hypothesis, we investigated the functional interplay between ATP13A2 and DJ-1 in terms of their ability to protect neuronal cells against a PD related stress, methamphetamine (METH), an abused drug that disrupts autophagy in N27 dopaminergic neuronal cells. Our results showed that knocking down ATP13A2 or DJ-1 results in a buildup of LC3 II, an autophagic marker, whereas ATP13A2 over-expression reduces the accumulation of autophagic vesicles, termed autophagosomes. In addition, ATP13A2 or DJ-1 KD N27 cells showed a decreased rate of O2 consumption. Strikingly, the level of ATP13A2 mRNA was increased in DJ-1 knockdown in primary midbrain cultures; in contrast the level of DJ-1 mRNA was decreased in ATP13A2 knockdown in the same cultures. These results suggest DJ-1 and ATP13A2 interact functionally in regulating lysosomal degradation and mitochondrial function. Overall, these studies have yielded insights into biochemical mechanisms of DJ-1-mediated neuroprotection and of the functional interplay between DJ-1 and ATP13A2. Not only do these findings advance our understanding of neuroprotective mechanisms relevant to PD, but they also suggest new strategies to treat this devastating syndrome.

Degree

Ph.D.

Advisors

Rochet, Purdue University.

Subject Area

Molecular biology|Cellular biology|Pharmacology

COinS