Search

DJ-1 has been identified as a gene responsible for recessive familial Parkinson's disease (familial Parkinsonism), which is caused by a mutation in the PARK7 locus. Consistent with the inferred correlation between Parkinson's disease and mitochondrial impairment, mitochondrial localization of DJ-1 and its implied role in mitochondrial quality control have been reported. However, the mechanism by which DJ-1 affects mitochondrial function remains poorly defined, and the mitochondrial localization of DJ-1 is still controversial. Here, we show the mitochondrial matrix localization of various pathogenic and artificial DJ-1 mutants by multiple independent experimental approaches including cellular fractionation, proteinase K protection assays, and specific immunocytochemistry. Localization of various DJ-1 mutants to the matrix is dependent on the membrane potential and translocase activity in both the outer and the inner membranes. Nevertheless, DJ-1 possesses neither an amino-terminal alpha-helix nor a predictable matrix-targeting signal, and a post-translocation processing-derived molecular weight change is not observed. In fact, wild-type DJ-1 does not show any evidence of mitochondrial localization at all. Such a mode of matrix localization of DJ-1 is difficult to explain by conventional mechanisms and implies a unique matrix import mechanism for DJ-1 mutants.

VCP/p97 is a hexameric ring-shaped AAA(+) ATPase that participates in various ubiquitin-associated cellular functions. Mis-sense mutations in VCP gene are associated with the pathogenesis of two inherited diseases: inclusion body myopathy associated with Paget's disease of the bone and front-temporal dementia (IBMPFD) and familial amyotrophic lateral sclerosis (ALS). These pathogenic VCPs have higher affinities for several cofactors, including Npl4, Ufd1 and p47. In Parkin-dependent mitochondrial quality control systems, VCP migrates to damaged mitochondria (e.g., those treated with uncouplers) to aid in the degradation of mitochondrial outer membrane proteins and to eliminate mitochondria. We showed that endogenous Npl4 and p47 also migrate to mitochondria after uncoupler treatment, and Npl4, Ufd1 or p47 silencing causes defective mitochondria clearance after uncoupler treatment. Moreover, pathogenic VCPs show impaired migration to mitochondria, and the exogenous pathogenic VCP expression partially inhibits Npl4 and p47 localization to mitochondria. These results suggest that the increased affinities of pathogenic VCPs for these cofactors cause the impaired movement of pathogenic VCPs. In adult flies, exogenous expression of wild-type VCP, but not pathogenic VCPs, reduces the number of abnormal mitochondria in muscles. Failure of pathogenic VCPs to function on damaged mitochondria may be related to the pathogenesis of IBMPFD and ALS.

PINK1 and PARKIN are causal genes for hereditary Parkinsonism. Recent studies have shown that PINK1 and Parkin play a pivotal role in the quality control of mitochondria, and dysfunction of either protein likely results in the accumulation of low-quality mitochondria that triggers early-onset familial Parkinsonism. As neurons are destined to degenerate in PINK1/Parkin-associated Parkinsonism, it is imperative to investigate the function of PINK1 and Parkin in neurons. However, most studies investigating PINK1/Parkin have used non-neuronal cell lines. Here we show that the principal PINK1 and Parkin cellular events that have been documented in non-neuronal lines in response to mitochondrial damage also occur in primary neurons. We found that dissipation of the mitochondrial membrane potential triggers phosphorylation of both PINK1 and Parkin and that, in response, Parkin translocates to depolarized mitochondria. Furthermore, Parkin's E3 activity is re-established concomitant with ubiquitin-ester formation at Cys431 of Parkin. As a result, mitochondrial substrates in neurons become ubiquitylated. These results underscore the relevance of the PINK1/Parkin-mediated mitochondrial quality control pathway in primary neurons and shed further light on the underlying mechanisms of the PINK1 and Parkin pathogenic mutations that predispose Parkinsonism in vivo.

Ubiquitin (Ub) ligation is implicated in active protein metabolism and subcellular trafficking and its impairment is involved in various neurologic diseases. In rat brain, we identified two novel Ub ligases, Momo and Sakura, carrying double zinc finger motif and RING finger domain. Momo expression is enriched in the brain gray matter and testis, and Sakura expression is more widely detected in the brain white matter as well as in many peripheral organs. Both proteins associate with the cell membranes of neuronal and/or glial cells. We examined their Ub ligase activity in vivo and in vitro using viral expression vectors carrying myc-tagged Momo and Sakura. Overexpression of either Momo or Sakura in mixed cortical cultures increased total polyubiquitination levels. In vitro ubiquitination assay revealed that the combination of Momo and UbcH4 and H5c, or of Sakura and UbcH4, H5c and H6 is required for the reaction. Deletion mutagenesis suggested that the E3 Ub ligase activity of Momo and Sakura depended on their C-terminal domains containing RING finger structure, while their N-terminal domains influenced their membrane association. In agreement, Sakura associating with the membrane was specifically palmitoylated. Although the molecular targets of their Ub ligation remain to be identified, these findings imply a novel function of the palmitoylated E3 Ub ligase(s).

PINK1 and Parkin were first identified as the causal genes responsible for familial forms of early-onset Parkinson’s disease (PD), a prevalent neurodegenerative disorder. PINK1 encodes a mitochondrial serine/threonine protein kinase, whereas Parkin encodes an ubiquitin-protein ligase. PINK1 and Parkin cooperate to maintain mitochondrial integrity; however, the detailed molecular mechanism of how Parkin-catalyzed ubiquitylation results in mitochondrial integrity remains an enigma. In this study, we show that Parkin-catalyzed K63-linked polyubiquitylation of depolarized mitochondria resulted in ubiquitylated mitochondria being transported along microtubules to cluster in the perinuclear region, which was interfered by pathogenic mutations of Parkin. In addition, p62/SQSTM1 (hereafter referred to as p62) was recruited to depolarized mitochondria after Parkin-directed ubiquitylation. Intriguingly, deletion of p62 in mouse embryonic fibroblasts resulted in a gross loss of mitochondrial perinuclear clustering but did not hinder mitochondrial degradation. Thus, p62 is required for ubiquitylation-dependent clustering of damaged mitochondria, which resembles p62-mediated ‘aggresome’ formation of misfolded/unfolded proteins after ubiquitylation.