HMG Advance Access published June 4, 2010 Human Molecular Genetics, 2010 doi:10.1093/hmg/ddq215

1–14

Effect of endogenous mutant and wild-type PINK1 on Parkin in fibroblasts from Parkinson disease patients Aleksandar Rakovic, Anne Gru¨newald, Philip Seibler, Alfredo Ramirez, Norman Kock, Slobodanka Orolicki, Katja Lohmann and Christine Klein ∗ Section of Clinical and Molecular Neurogenetics, Department of Neurology, University of Lu¨beck, 23562 Lu¨beck, Germany Received April 13, 2010; Revised and Accepted May 21, 2010

Mutations in the PTEN-induced putative kinase 1 (PINK1), a mitochondrial serine – threonine kinase, and Parkin, an E3 ubiquitin ligase, are associated with autosomal-recessive forms of Parkinson disease (PD). Both are involved in the maintenance of mitochondrial integrity and protection from multiple stressors. Recently, Parkin was demonstrated to be recruited to impaired mitochondria in a PINK1-dependent manner, where it triggers mitophagy. Using primary human dermal fibroblasts originating from PD patients with various PINK1 mutations, we showed at the endogenous level that (i) PINK1 regulates the stress-induced decrease of endogenous Parkin; (ii) mitochondrially localized PINK1 mediates the stress-induced mitochondrial translocation of Parkin; (iii) endogenous PINK1 is stabilized on depolarized mitochondria; and (iv) mitochondrial accumulation of full-length PINK1 is sufficient but not necessary for the stress-induced loss of Parkin signal and its mitochondrial translocation. Furthermore, we showed that different stressors, depolarizing or non-depolarizing, led to the same effect on detectable Parkin levels and its mitochondrial targeting. Although this effect on Parkin was independent of the mitochondrial membrane potential, we demonstrate a differential effect of depolarizing versus non-depolarizing stressors on endogenous levels of PINK1. Our study shows the necessity to introduce an environmental factor, i.e. stress, to visualize the differences in the interaction of PINK1 and Parkin in mutants versus controls. Establishing human fibroblasts as a suitable model for studying this interaction, we extend data from animal and other cellular models and provide experimental evidence for the generally held notion of PD as a condition with a combined genetic and environmental etiology.

INTRODUCTION Parkinson disease (PD) is the second most common neurodegenerative disorder characterized by degeneration of dopaminergic neurons in the midbrain. About 2–3% of ‘idiopathic’ PD cases can currently be linked to a single genetic factor. Among these, recessively inherited Parkin (1) and PINK1 (PTENinduced putative kinase 1) (2) mutations are known causes of a clinical syndrome closely resembling idiopathic PD with the exception of an overall earlier age of onset and slower disease progression. Interestingly, the clinical features associated with Parkin and PINK1 mutations are indistinguishable from one

another; likewise, loss of Drosophila parkin shows phenotypes similar to loss of pink1 function (3,4). In agreement with these observations, Parkin and PINK1 have been shown to function, at least in part, in the same pathway with PINK1 acting upstream of Parkin (5–8). In addition, PINK1 and Parkin may form a complex in conjunction with DJ-1 to promote degradation of un-/misfolded proteins (9). More recently, involvement of both Parkin and PINK1 has also been demonstrated in mitophagy with mitochondrial translocation of Parkin depending on PINK1 (10–12). Human PINK1 is a 581-aa polypeptide with a predicted N-terminal mitochondrial targeting signal (MTS) and a

∗ To whom correspondence should be addressed at: Section of Clinical and Molecular Neurogenetics, Department of Neurology, University of Lu¨beck, Ratzeburger Allee 160, 23538, Lu¨beck, Germany. Tel: +49 4512903353; Fax: +49 4512903355; Email: [email protected]

# The Author 2010. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected]

2

Human Molecular Genetics, 2010

serine/threonine kinase domain similar to that in the Ca2+/calmodulin kinase family (13,14). PINK1 is synthesized as a fulllength form (66 kDa) which is proteolytically processed upon entry into mitochondria to its cleaved 55 kDa form (15). According to the most recently proposed model, PINK1 is localized on the outer mitochondrial membrane, where it is cleaved in a voltage-dependent manner (12). Inhibition of cleavage and accumulation of full-length PINK1 can be induced by loss of the mitochondrial membrane potential (12). However, the physiological relevance of cleavage and the exact cleavage site as well as the localization of both forms currently remain under debate (15). Although wild-type PINK1 protects cells against mitochondrial toxins (16) and apoptosis induced by proteasomal stress (17), loss of PINK1 results in reduced complex I activity, increased levels of oxidative stress (18) and loss of the mitochondrial membrane potential (19). Parkin encodes a 465-aa protein with a modular structure containing an N-terminal ubiquitin-like domain and a C-terminal RING-IBR-RING domain. Similar to other RING finger proteins, Parkin functions as an E3 ubiquitin ligase in the ubiquitin proteasome system (UPS), where it ubiquinates a number of substrates (20). More recently, Parkin has been implicated in several other cellular functions including a key role in maintaining mitochondrial function and integrity (21), mitochondrial DNA repair (22) and involvement in mitochondrial quality control and mitophagy (10 – 12,23). These findings collectively demonstrate that both PINK1 and Parkin are intimately involved in preventing mitochondrial dysfunction. Notably, most of these studies were performed in animal and cellular models with overexpressed protein. More recently, dermal fibroblasts from PD patients with endogenously mutated PINK1 (18,24) and Parkin (21) have emerged as a reliable human model system to study PD. This prompted us to explore in more detail the connection between PINK1 and Parkin by investigating the effect of PINK1 mutations on the protein level and subcellular localization of Parkin under basal and stress conditions in fibroblasts from PINK1 mutation carriers and controls.

RESULTS We analyzed four primary human dermal fibroblast cultures carrying two different PD-causing homozygous PINK1 mutations, c.1366C.T (p.Q456X; nonsense mutation; n ¼ 3) and c.509T.G (p.V170G; missense mutation; n ¼ 1), and fibroblast cultures from four age-matched, healthy controls without PINK1 mutations (two family members and two unrelated controls). Clinical features of these mutation carriers were compatible with idiopathic PD with an average age of onset of 48.0 + 12.7 years (for details, see 25 – 27). The localization of the PINK1 mutations and their effect on the length of the protein are schematically shown in Figure 1A. We confirmed the specificity of our PINK1 antibody and validated the knockdown efficiency of two different PINK1 siRNAs by overexpressing wild-type PINK1 (PINK1-V5) in the presence of scramble siRNA or PINK1 siRNA in fibroblasts from PINK1 nonsense mutation carriers (Fig. 1B). In order to validate the sensitivity and specificity of our antibody against Parkin that is directed against the C-terminus of

the Parkin protein, we tested levels of endogenous Parkin protein in human dermal fibroblasts from controls and from homozygous or heterozygous Parkin mutation carriers. These mutations (del exon 7 and c.1072delT) generate a truncated protein, which cannot be detected by our antibody (Fig. 1C). For experiments under stress conditions, we chose valinomycin and carbonyl cyanide m-chlorophenylhydrazone (CCCP) as mitochondrial stressors. They trigger rapid loss of mitochondrial membrane potential (28,29), however, through different mechanisms: CCCP is a respiratory chain uncoupler via release of hydrogen, and valinomycin is a potassium ionophore. We also used hydrogen peroxide (H2O2) as a generator of free radicals. PINK1 decreases detectable Parkin levels, under stress conditions To test whether PINK1 regulates Parkin levels, we extracted proteins from both PINK1 mutants and controls under basal conditions and upon valinomycin treatment using RadioImmunoprecipitation Assay (RIPA) buffer containing 0.1% SDS. Under basal conditions, protein levels of Parkin were comparable between the group of mutants and controls as shown by western blotting (Fig. 2A, upper panel). In contrast, under stress conditions using valinomycin, we detected a loss of Parkin signal in cells from healthy individuals but not in those from PINK1 mutation carriers (Fig. 2A, middle panel). It was previously suggested that PINK1 reduces the solubility of Parkin (30,31). Therefore, the observed loss of signal in controls may be caused by a shift of Parkin into the less soluble fraction of proteins, where it would be difficult to extract and detect. To test this hypothesis, we dissolved the pellet that remained after protein extraction (for details, see Materials and Methods) in RIPA buffer containing 2% SDS (‘insoluble fraction’) followed by western blotting. For simplicity, in the remainder of the paper, we will use the term ‘soluble fraction’ for proteins soluble in RIPA buffer containing 0.1% SDS and ‘insoluble fraction’ for proteins of the pellet dissolved in RIPA buffer containing 2% SDS. Parkin signal was absent in the insoluble fraction in both controls and mutants (Fig. 2A, lower panel), indicating that reduced solubility of endogenous Parkin is not the cause of loss of detectable Parkin signal in the soluble fraction of controls. Furthermore, H2O2 (Fig. 2B) and CCCP (data not shown) had the same effect on detectable Parkin levels as observed under valinomycin-induced stress. To test whether these findings on the protein level were due to changes in mRNA expression, we measured mRNA levels of PINK1 and Parkin before and after valinomycin treatment in mutants and controls by real-time PCR using the housekeeping gene YWHAZ for normalization. As mentioned above, the PINK1 nonsense mutation leads to a premature stop codon exerting a major effect on the PINK1 mRNA level (80 – 90% reduction compared with the PINK1 mRNA levels in controls; Fig. 2D) most likely via nonsense-mediated mRNA decay (26). In contrast, PINK1 expression levels were comparable between the PINK1 missense mutant and healthy controls (Fig. 2D). Importantly, we detected no differences in Parkin mRNA levels between PINK1 mutants and controls before or after valinomycin-induced stress (Fig. 2E), thereby

Human Molecular Genetics, 2010

3

Figure 1. Location of PINK1 mutations studied, confirmation of antibody specificities and validation of knockdown efficiency of PINK1 siRNAs. (A) Schematic representation of the PINK1 structure with positions of the two PINK1 mutations studied that are both located in the kinase domain of the protein. The nonsense mutation results in protein truncation. (B) Whole-cell lysate of fibroblasts of a homozygous PINK1 nonsense mutation carrier transfected with PINK1-V5 in conjunction with either scramble siRNA or two different PINK1 siRNAs. A western blot probed with an antibody against the V5 tag (left upper panel) showed bands of the same size as a blot probed with an antibody against PINK1 (left lower panel), confirming specificity of the PINK1 antibody. Protein levels of overexpressed full-length PINK1 (FL PINK1) and cleaved 55 kDa PINK1 are abolished (siRNA1) or decreased (siRNA2) specifically by both PINK1 siRNAs. Transfection efficiency was determined using an antibody against the Neomycin cassette (left middle panel). Knockdown efficiency was estimated densitometrically (right panel). (C) Western blot using whole-cell lysates from fibroblasts from controls and individuals carrying homozygous or heterozygous Parkin mutations. Protein levels of Parkin were estimated using an antibody able to recognize only Parkin with an intact C-terminus of the protein. In fibroblasts from heterozygous mutation carriers, protein levels were 50% of those detected in controls, and no Parkin protein was detectable in fibroblasts of homozygous Parkin mutation carriers lacking the C-terminal part of Parkin.

excluding a possible effect of gene expression on protein levels. As most of the previous studies showing that PINK1 reduces the solubility of Parkin were carried out with overexpressed proteins, we wanted to test whether this is also the case in our cell model. For this, we overexpressed wild-type Parkin (FLAG-Parkin) in fibroblasts from a PINK1 nonsense mutation carrier and two sets of controls, one co-transfected with a vector expressing wild-type PINK1 (PINK1-V5) and

the other co-transfected with empty vector (Fig. 3). We analyzed both the soluble and the insoluble fractions of the nontreated cells and of cells treated with valinomycin. First, we confirmed overexpression of PINK1-V5 with an antibody against PINK1 and detected both the full-length and the major cleaved fragment. Upon treatment with valinomycin, we observed an increase in intensity of the full-length form of PINK1 (Fig. 3A, upper panel). We next examined levels of FLAG-Parkin using an antibody against the FLAG tag.

4

Human Molecular Genetics, 2010

Figure 2. Treatment with valinomycin or hydrogen peroxide promotes loss of detectable endogenous Parkin in fibroblasts from controls but not from PINK1 mutation carriers and is independent of mRNA expression. (A) Western blot analysis of whole-cell lysate from non-treated (NT) cells (upper panel) and from cells treated with valinomycin (middle and lower panels) from four controls and four mutants. The middle panel shows a western blot analysis of proteins extracted with RIPA buffer containing 0.1% SDS (‘soluble fraction’); the lower panel shows a western blot analysis of proteins from pellets remaining after 0.1% SDS RIPA buffer protein extraction, which were then dissolved in RIPA buffer containing 2.0% SDS (‘insoluble fraction’). (B) H2O2 induces decrease in the detectable levels of endogenous Parkin in cells from healthy controls but not in cells from PINK1 mutation carriers. In non-treated fibroblasts, the level of endogenous Parkin is comparable between controls and mutants. (C) Densitometric quantification of detectable Parkin levels normalized for levels of b-actin. (D) Fibroblasts from PINK1 mutation carriers and from controls were treated with valinomycin for 12 h. Total RNA was then extracted from stressed and nonstressed cells and quantified by real-time PCR. Relative PINK1 mRNA expression levels remain unchanged upon valinomycin treatment. Expression levels of the nonsense mutation carriers are reduced to 10– 20% compared with those of controls and the missense mutation carrier due to nonsense-mediated mRNA decay in the former. (E) Relative Parkin mRNA expression is not significantly different between controls and mutants before or after treatment. Values represent means + SD. C, healthy control; Mm, PINK1 missense mutation carrier (c.509T.G; p.V170G) mutation carrier; Mn, PINK1 nonsense mutation carriers (c.1366C.T; p.Q456X) mutation carriers.

In mutants, levels of Parkin remained unchanged upon stress. In controls expressing only FLAG-Parkin, we detected a decrease in Parkin levels upon valinomycin treatment. In controls co-overexpressing PINK1-V5, the level of Parkin under basal conditions was lower than in controls transfected with empty vector (i.e. having only endogenous PINK1) and the decrease of Parkin levels upon stress was even more pronounced (Fig. 3A, middle panel). In the insoluble fraction of control cells, we found the opposite situation than in the soluble fraction with increased Parkin levels upon stress, which was more pronounced in controls co-expressing PINK1. In mutants, levels of Parkin were low, suggesting that most of the overexpressed Parkin remains in the soluble fraction. The observed loss of Parkin signal in controls after mitochondrial stress indicates that wild-type PINK1 (PINK1-V5) is involved in this process. In order to confirm this hypothesis, we performed experimental knockdown of PINK1 in controls and a rescue experiment in PINK1 mutants (Fig. 4). We first tested the effect of PINK1 knockdown on levels of endogenous Parkin in control cells (Fig. 4B). Upon valinomycin

treatment, controls transfected with PINK1 siRNA had higher Parkin levels than controls transfected with scramble siRNA. The same effect was observed even under basal conditions, suggesting that PINK1 is indeed involved in the reduction of detectable Parkin levels. In mutant cells, we expressed PINK1 using a mammalian expression vector containing PINK1-V5. Under basal conditions, Parkin levels were comparable between mutants transfected with PINK1 or with empty vector. As expected, valinomycin treatment induced a decrease of Parkin signal in mutants transfected with a vector containing PINK1-V5, but not in those transfected with an empty vector. Of note, we detected an increase in the intensity of the full-length form of endogenous and of transgenically expressed PINK1, respectively (Fig. 4A). PINK1 mediates stress-induced mitochondrial translocation of Parkin Subcellular localization of Parkin was determined using immunocytochemistry. For this, we overexpressed wild-type Parkin (FLAG-Parkin) in human dermal fibroblasts. Subcellular

Human Molecular Genetics, 2010

5

Figure 4. PINK1 is necessary for valinomycin-induced reduction of endogenous Parkin. Control cells (C) transfected with PINK1 siRNA and nonsense mutant cells (M) transfected with a vector expressing wild-type PINK1 (WT PINK1) tagged with V5 at the C-terminus (PINK1-V5) were stressed with valinomycin for 12 h. After valinomycin treatment, proteins were extracted and analyzed by western blotting using antibodies against (A) PINK1 and (B) Parkin. b-Actin was used as loading control. In control cells, knockdown of PINK1 prevented the reduction of Parkin signal compared with controls transfected with scramble siRNA both under basal conditions and under valinomycin treatment. In mutant cells, expression of WT PINK1 promotes loss of Parkin signal upon valinomycin-induced stress. FL PINK1, full-length form of PINK1. ∗ Endogenous full-length form of PINK1.

basal conditions, Parkin was localized in the cytosol in both controls and PINK1 mutants (Fig. 5, left lane). Treatment with valinomycin (Fig. 5, middle lane) or H2O2 (Fig. 5, right lane) induced loss of the mitochondrial network and its fragmentation into multiple, smaller organelles in both controls and mutants. In addition, both stressors resulted in mitochondrial translocation of Parkin in controls (Fig. 5A) but not in PINK1 mutants (Fig. 5B). Stress-induced loss of Parkin can be prevented by the proteasome inhibitors MG132 and epoxomicin Figure 3. PINK1 decreases the solubility of overexpressed Parkin upon stress in a dose-dependent manner. (A) Western blot analysis of PINK1 nonsense mutant and control fibroblasts overexpressing FLAG-Parkin under basal conditions (NT) and under valinomycin-induced stress. Mutant cells were co-transfected only with empty vector, whereas control cells were co-transfected either with a vector containing full-length PINK1 (PINK1-V5) or with empty vector. The upper panel confirms overexpression of PINK1 in controls co-transfected with V5-tagged PINK1 using an antibody against PINK1. The intensity of the band representing full-length PINK1 increases upon treatment with valinomycin. To detect FLAG-Parkin, the same blot was washed and reprobed with an antibody against the FLAG tag. The middle panel shows levels of overexpressed FLAGParkin soluble in 0.1% SDS RIPA buffer (‘soluble fraction’). A marked valinomycin-induced drop of the level of FLAG-Parkin was observed in controls co-transfected with PINK-V5 but not in controls containing only endogenous levels of PINK1. The lower panel represents proteins from the insoluble pellet remaining after removal of the ‘soluble fraction’ and dissolved in 2% SDS RIPA buffer (‘insoluble fraction’). In mutant cells, FLAG-Parkin was barely detectable, whereas FLAG-Parkin was present at high levels in both controls (with and without PINK1-V5 co-transfection) and accumulated upon valinomycin treatment. Equal loading and transfection efficiency were confirmed using antibodies against b-actin and neomycin. (B) Densitometric quantification of FLAG-Parkin levels in the soluble fraction from three independent experiments.

localization of FLAG-Parkin in control cells and in PINK1 nonsense mutant cells was determined by immunocytochemistry using an anti-FLAG antibody (Fig. 5A and B). Under

As Parkin can promote self-ubiquitination and degradation (32 – 34) through the proteasomal system, we sought to explore whether the PINK1-mediated stress-induced loss of Parkin can be prevented by inhibition of the proteasome with MG132. For this, we tested the effect of MG132 on Parkin levels over time in cells treated with valinomycin (Fig. 6). When treated with valinomycin only, a timedependent decrease in Parkin signal intensity was observed only in control cells. In mutant cells, Parkin levels were not affected by this treatment (Fig. 6A). In contrast, in an experiment combining valinomycin and MG132 treatment, there was no change in intensity of Parkin signal in either controls or mutants (Fig. 6B). As MG132 has a more general function as a protease inhibitor, i.e. MG132 is also able to inhibit the activity of mitochondrial proteases (35), we decided to use epoxomicin as a more selective inhibitor of the proteasome (36). For this, control cells were treated with epoxomicin in the presence of valinomycin for 12 h followed by western blotting (Fig. 6C). Our results showed that epoxomycin prevented valinomycin-induced degradation of Parkin, as did MG132, suggesting that the UPS is indeed involved in stress-induced reduction of Parkin levels.

6

Human Molecular Genetics, 2010

Figure 5. Valinomycin- and hydrogen peroxide-induced stress result in mitochondrial localization of Parkin in cells from healthy controls but not in cells from PINK1 nonsense mutation carriers. Fibroblasts transfected with FLAG-Parkin were either untreated (left lane) or treated with valinomycin (middle lane) or with H2O2 (right lane) for 12 h. After treatment, cells were fixed and immunostained with antibodies against the FLAG-tag (red) and a mitochondrial marker GRP75 (green). Upon stress, FLAG-Parkin colocalizes with mitochondria in fibroblasts originating from (A) controls but not in fibroblasts from (B) PINK1 mutation carriers. Scale bar, 50 mm, NT, non-treated.

Human Molecular Genetics, 2010

7

we loaded the same amount of proteins as in Fig. 1A, in this case, proteins from concentrated cytosolic and mitochondrial fractions were used. Surprisingly, however, when combining valinomycin with MG132 treatment, the reduction in Parkin signal could be prevented not only in the cytosolic, but also in the mitochondrial fractions (Fig. 7A). In mutant cells, Parkin was localized exclusively in the cytosolic fraction, independent of the treatment conditions (Fig. 7B). Treatment with MG132 alone had no effect on protein levels or subcellular localization of Parkin in either controls or mutants when compared with non-treated cells (Fig. 7C and D). These data collectively demonstrate that PINK1 is involved both in the stress-induced decrease and in the mitochondrial translocation of Parkin. Overexpression of Parkin in PINK1 mutants increases the levels of mitochondrial Parkin

Figure 6. Inhibitors of the UPS protect valinomycin-induced degradation of Parkin. Fibroblasts from a healthy control and from a PINK1 nonsense mutation carrier were treated with (A) valinomycin or (B) with valinomycin plus MG132. (C) In addition, control cells were treated with valinomycin together with epoxomicin or with MG132 for 12 h. Proteins were extracted at different time points and analyzed by western blotting. Only in the control cells but not in mutants, there was valinomycin-induced decrease in Parkin signal (A). This reduction could be prevented by MG132 treatment (B). The same effect was observed when using epoxomicin, a potent and selective inhibitor of the UPS (C).

In light of our immunocytochemistry findings showing stress-induced mitochondrial translocation of overexpressed Parkin, we now wanted to clarify in which cellular compartment (mitochondrial or cytosolic) endogenous Parkin accumulates after treatment with MG132. For this, we treated both controls and mutants with MG132 alone, valinomycin alone or MG132 plus valinomycin in combination. After 12 h of treatment, we harvested the cells and extracted mitochondrial and cytosolic fractions by differential centrifugation. Purity of both fractions was verified by western blotting using antibodies against b-tubulin and b-actin (cytosolically localized proteins) and antibodies against Hsp60 and VDAC1 (mitochondrially localized proteins) (Fig. 7). As expected, in control cells treated with valinomycin alone, we detected the presence of Parkin in both the mitochondrial and the cytosolic fractions. Compared with previous experiments, we were now able to detect Parkin signal even after 12 h of valinomycin treatment. This can be explained by the fact that, although

Previous findings demonstrated that overexpression of Parkin is able to restore normal function in PINK1-deficient animal and cellular models (5,7,8). In PINK1 mutants, we detected no Parkin signal in the mitochondrial fraction (Fig. 7B and D). Possible explanations of this result may be that, in the absence of PINK1, stress-induced mitochondrial translocation of Parkin is either completely blocked or reduced to very low levels. As shown in Fig. 7A, an increase in the level of cytosolic Parkin is followed by an increase in the level of mitochondrially localized Parkin in controls upon combined valinomycin and MG132 treatment. Therefore, we hypothesized that overexpression of Parkin can also increase the amount of its mitochondrially localized fraction in PINK1 mutants. To this end, we used a fibroblast from a nonsense mutation carrier, where PINK1 mRNA expression is reduced to 10% of controls and translates into a truncated protein lacking a large portion of its kinase domain. In these cells, we overexpressed wild-type Parkin (FLAG-Parkin), followed by subcellular fractionation (Fig. 8A). In the mitochondrial fraction, we detected a certain amount of Parkin (Fig. 8B). These data demonstrate that loss of PINK1 leads to a decrease in mitochondrial translocation of Parkin, which can be overcome by increasing the amount of cytosolic Parkin above its endogenous level. Mitochondrial localization of PINK1 is necessary for the valinomycin-induced reduction in detectable Parkin and its mitochondrial translocation We next aimed to test the hypothesis that the stress-induced reduction in Parkin levels and mitochondrial translocation of Parkin are dependent upon the mitochondrial localization of PINK1. To prevent mitochondrial targeting of PINK1, we designed vectors expressing N-terminally truncated PINK1. The MTS has unambiguously been localized within the Nterminal part of human PINK1, however its exact length and composition are still under debate (37). Although the first 93 amino acids of PINK1 are required to target green fluorescent protein to mitochondria, recent findings showed that only upon deletion within the first 110 amino acids (12,38,39), PINK1 cannot be targeted to mitochondria. On the basis of these findings, we designed truncated forms of PINK1 lacking

8

Human Molecular Genetics, 2010

Figure 7. MG132 prevents valinomycin-induced decrease of Parkin in the cytosolic fraction and promotes accumulation of Parkin in the mitochondrial fraction in cells from healthy controls. Fibroblasts from controls and from PINK1 nonsense mutation carriers (mutants) were treated with valinomycin alone, with valinomycin plus MG132 or with MG132 alone for 12 h. Non-treated cells were also analyzed. The cells were harvested to prepare mitochondrial (Mi) and cytosolic (Cy) fractions and analyzed by western blotting. (A) In control cells, upon valinomycin treatment, we detected the presence of Parkin in mitochondrial and cytosolic fractions. Furthermore, combining valinomycin with MG132, we observed an increase in Parkin levels in both the mitochondrial and the cytosolic fractions compared with control cells treated with valinomycin only. (B) In mutants, neither treatment with valinomycin alone or in combination with MG132 induced mitochondrial translocation or loss of Parkin signal. (C, D) Treatment with MG132 alone had no effect on Parkin levels and subcellular localization compared with non-treated cells. Purity of the cellular fractionation was verified using antibodies against b-tubulin and b-actin (cytosolically localized proteins) and antibodies against Hsp60 and VDAC1 (mitochondrially localized proteins). Mi, mitochondrial fraction; Cy, cytosolic fraction.

its first 93 (PINK1d93) or 110 (PINK1d110) amino acids. To determine their subcellular localization, we transiently transfected HEK293 cells with vectors containing full-length PINK1 (FL PINK1), PINK1d93 or PINK1d110. We harvested the cells and extracted mitochondrial and cytosolic fractions. As shown in Fig. 9B, full-length PINK1 and PINK1d93 were mainly localized in the mitochondrial fraction. PINK1d110, on the other hand, was predominately present in the cytosolic fraction (Fig. 9B). Notably, we detected the presence of the fulllength (66 kDa) form of PINK1 in the cytosolic fraction, which can be explained by overloading of the mitochondrial translocase system (15). We detected the major cleaved (55 kDa) form of PINK1 not only in the mitochondrial, but also in the cytosolic fractions (Fig. 9B). There are two possible explanations for this phenomenon: (i) after processing, cleaved PINK1 is translocated from mitochondria to the cytosol, as recently proposed (12); (ii) cleaved PINK1 is generated in the cytosol by a process that is alternative to mitochondrial processing (15,39). In accordance with previous publications (15,37), our data suggest that processing of full-length PINK1 into its cleaved 55 kDa form occurs around residue 110 (Fig. 9B). These findings confirm that the N-terminal end of PINK1 contains the MTS but is not restricted to the first 93 amino acids but more likely to the first 110 amino acids. Importantly,

removal of the first 110 amino acids should not affect the kinase activity of PINK1, as shown by several studies, where amino acids 112– 581 alone were enough to perform autophosphorylation (40,41), as well as phosphorylation of several other proteins (42) in vitro. We next transfected mutant cells with vectors containing either full-length PINK1, PINK1d93 or PINK1d110. In mutant cells transfected with full-length PINK1 or with PINK1d93, we detected a valinomycin-induced decrease in Parkin signal by western blotting. In contrast, the cytosolically localized form of PINK1 (PINK1d110) failed to induce a decrease in signal of endogenous and overexpressed Parkin upon valinomycin treatment (Fig. 9C and D). Finally, we explored the effect of these PINK1 forms on the subcellular localization of Parkin. For this, we co-transfected wild-type Parkin (FLAG-Parkin) together with full-length PINK1, PINK1d93 or PINK1d110 in cells from a PINK1 nonsense mutation carrier. Co-transfection of FLAG-Parkin with an empty vector served as a negative control. After transfection, the cells were divided into two groups. One group of cells was treated with valinomycin for 12 h and the other one remained non-treated. In non-treated cells, Parkin was diffusely localized in the cytosol. Valinomycin treatment nonselectively induced loss of the mitochondrial network and its

Human Molecular Genetics, 2010

Figure 8. Overexpression of Parkin in PINK1 mutant cells increases the amount of mitochondrially localized Parkin. Fibroblasts from a PINK1 nonsense mutation carrier (Mn) were transfected with a vector expressing wildtype Parkin (FLAG-tagged). Twenty-four hours after transfection, the cells were treated with valinomycin for an additional 12 h. After treatment, mitochondrial and cytosolic fractions were analyzed by western blotting. (A) Quality of the cellular fractionation is confirmed using antibodies against b-actin (cytosolic) and Hsp60 (mitochondrial). The membrane was then washed and (B) reprobed with an antibody against Parkin, which was also found in the mitochondrial fraction. Mi, mitochondrial fraction; Cy, cytosolic fraction. ∗ Degraded Parkin.

fragmentation into multiple, smaller organelles independent of the vector used for transfection. In accordance with the data observed on western blotting, only the cytosolically localized form of PINK1 (PINK1d110) failed to induce mitochondrial translocation of Parkin upon valinomycin-induced stress (Fig. 9E and F). Mitochondrial accumulation of full-length PINK1 is not necessary for the stress-induced reduction in the level of Parkin and its mitochondrial translocation and is independent of the mitochondrial membrane potential As we found a stress-induced reduction in detectable Parkin levels (Fig. 2A and B) and its mitochondrial translocation (Fig. 5) upon both valinomycin and H2O2 treatment, we sought to test whether both of these stressors had the same effect on the mitochondrial membrane potential. In this context, it was recently published that Parkin is recruited to depolarized mitochondria (12). Therefore, we analyzed the mitochondrial membrane potential of fibroblasts from both PINK1 mutation carriers and controls upon H2O2 or valinomycin treatment using the same concentrations and within the same time frame like in the remainder of the experiments in this study. Although treatment with valinomycin caused an immediate loss of the mitochondrial membrane potential, H2O2 did not cause a drop in mitochondrial membrane potential (Fig. 10A). To additionally confirm a differential effect of H2O2 and valinomycin on the mitochondrial membrane potential, we analyzed protein levels and processing of the glucose-regulated protein 75 (GRP75). GRP75 is a mitochondrial matrix chaperone,

9

synthesized as a 679-amino acid pre-protein, which contains a 51-residue N-terminal MTS sequence. After the membrane potential-dependent import into mitochondria, it is cleaved into the mature protein which is 5.5 kDa shorter than the preprotein (43 – 45). For this, we treated cells from a healthy control, from a PINK1 nonsense and from a PINK1 missense mutation carrier with H2O2 and valinomycin followed by subcellular fractionation. We detected accumulation of the fulllength form of GRP75 only upon valinomycin, but not upon H2O2 treatment. This finding was independent of the type of PINK1 mutation (nonsense versus missense), confirming our previous findings (Fig. 10A). Additionally, we tested endogenous PINK1 levels in the same experiment. Although both stressors induced the previously shown loss of detectable Parkin in controls (Fig. 10B, left upper panel) but not in mutants (Fig. 10B, left middle and lower panels), we found accumulation of full-length PINK1 in the mitochondrial fraction only in cells from controls treated with valinomycin. In contrast, no PINK1 was detected in non-treated controls or in controls treated with H2O2 (Fig. 10B, right upper panel). As expected, we did not detect the presence of any PINK1 in either fraction of the cells of the nonsense mutation carrier (Fig. 10B, right middle panel) due to nonsense-mediated decay. Likewise, in the missense mutation carrier, PINK1 was absent in both the mitochondrial and the cytosolic fraction, although the PINK1 expression level was the same in the missense mutation carrier as in controls (Fig. 10B, right lower panel).

DISCUSSION In this study, we used human fibroblasts from PD patients carrying mutations in the PINK1 gene to further elucidate the interaction between PINK1 and Parkin. The significance of this study is 2-fold: First, we confirmed at the endogenous level that (i) PINK1 regulates the stress-induced decrease in detectable Parkin levels (30,31), (ii) mediates the stress-induced mitochondrial translocation of Parkin (10 – 12) and (iii) endogenous PINK1 accumulates upon depolarization of the mitochondria (10,12). Second, our findings provide novel evidence that mitochondrial accumulation of full-length PINK1 is not necessary for the stress-induced loss of Parkin signal and its mitochondrial translocation is independent of the mitochondrial membrane potential. Additional novelty of our study is that we showed that different stressors, either depolarizing or not affecting the mitochondrial membrane potential, led to the same effect on detectable Parkin levels and its subcellular localization. In contrast, recent studies on PINK1/Parkin interaction used stressors with a depolarizing effect on the mitochondrial membrane potential. We studied two different types of human pathogenic PINK1 mutations, i.e. a nonsense mutation and a missense mutation, and observed the same phenotype. The nonsense mutation is known to cause nonsense-mediated decay (26), whereas little is known about the consequences of the homozygous missense mutation. Although mRNA expression of the missense mutation was comparable to that of controls, we did not detect accumulation of the full-length mutated protein after valinomycin treatment, suggesting that protein stability is

10

Human Molecular Genetics, 2010

Figure 9. Mitochondrial localization of PINK1 is necessary for valinomycin-induced decrease in detectable Parkin and its mitochondrial translocation. (A) Schematic representation of PINK1 constructs with the deletions used in the present study. (B) HEK-293 cells were transfected with vectors expressing FL PINK1, PINK1d93 or PINK1d110. Mitochondrial (Mi) and cytosolic (Cy) fractions were analyzed by western blotting. Although FL PINK1 and PINK1d93 were predominately localized in mitochondria, PINK1d110 was mainly localized in the cytosol. (C) Fibroblasts from a PINK1 nonsense mutation carrier were transfected with empty vector and vectors expressing FL PINK1, PINK1d93 or PINK1d110 (upper panel) or vectors expressing FL PINK1, PINK1d93 or PINK1d110 in combination with FLAG-Parkin (middle panel). After transfection, cells were treated with valinomycin for an additional 12 h. Whole-cell lysate was analyzed by western blotting. In cells expressing the mitochondrially localized forms of PINK1 (FL PINK1 and PINK1d93), treatment with valinomycin induced a decrease in Parkin signal (upper and middle panels in B). (D) Densitometric quantification of Parkin levels from three independent experiments. (E, F) Fibroblasts from a PINK1 nonsense mutation carrier co-expressing Parkin (FLAG-tagged) together with empty vector, full-length PINK1, PINK1d93 or PINK1d110. (E) In nontreated cells, co-transfection of Parkin with any of the PINK1 constructs had no effect on the subcellular localization of Parkin. In contrast, (F) in valinomycintreated cells, expression of FL PINK1 and PINK1d93 induced mitochondrial translocation of Parkin. Only the cytosolically localized form of PINK1 (PINK1d110) failed to induce mitochondrial translocation of Parkin upon valinomycin-induced stress. Scale bar, 50 mm. FL PINK1, full-length PINK1; PINK1d93, PINK1 construct lacking the first 93 amino acids; PINK1d110, PINK1 construct lacking the first 110 amino acids; NT, non-treated.

Human Molecular Genetics, 2010

11

Figure 10. Mitochondrial accumulation of full-length PINK1 is not necessary for the stress-induced reduction in the level of Parkin and its mitochondrial translocation and is independent of the mitochondrial membrane potential. (A) Fibroblasts of controls and mutants were treated with H2O2 or valinomycin for 12 h. The level of the mitochondrial membrane potential was measured at different time points using JC-1 and normalized for number of cells. Valinomycin but not H2O2 caused an immediate loss of mitochondrial potential, which remained at the same level throughout the entire experiment. (B) Fibroblasts from a control (upper panels), a nonsense mutation carrier (middle panels) and a missense mutation carrier (lower panels) were treated with H2O2 or valinomycin for 12 h. Both the mitochondrial and the cytosolic fractions were analyzed by western blotting. In controls, we detected loss of Parkin signal in cells treated with either H2O2 or valinomycin and accumulation of the full-length form of GRP75 in the cytosolic fraction only in cells treated with valinomycin (upper left panel). In the mitochondrial fraction of control cells, we found accumulation of full-length PINK1 only in cells treated with valinomycin (upper right panel). In the cells from both the nonsense and the missense mutation carriers, neither of the treatments resulted in loss of Parkin signal, whereas accumulation of full-length GRP75 was present in cells treated with valinomycin but not in those treated with H2O2 (left middle and left lower panels). In contrast to controls, there was no accumulation of PINK1 in the mitochondrial or cytosolic fraction of cells from both the nonsense and the missense mutation carriers (right middle and right lower panels). FL PINK1, full-length PINK1.

impaired. This idea is further supported by the fact that this mutation affects the kinase domain and not the MTS sequence. Our data demonstrated that wild-type but not mutant PINK1 mediates the stress-induced loss of detectable endogenous Parkin in both the soluble and the insoluble fractions. This loss could be prevented by inhibitors of the UPS, which is in keeping with previous findings showing that PINK1 promotes proteasomal degradation of Parkin (9). An alternative explanation of the loss of detectable Parkin could be a PINK1mediated decrease in Parkin solubility, as suggested in a study using overexpressed proteins (30). On the endogenous level, we did not detect any Parkin in the insoluble fraction. In contrast, when overexpressing wild-type Parkin in human dermal fibroblasts, we also observed a PINK1-mediated reduction in Parkin solubility. We detected Parkin in the insoluble fraction using two different antibodies (anti-FLAG tag and anti-Parkin) confirming the sensitivity of the Parkin antibody. Thus, these seemingly conflicting results can be explained by different experimental conditions (endogenous versus overexpressed protein). Our data obtained on the endogenous level support the hypothesis of PINK1-mediated degradation of Parkin and underline the necessity to interpret results generated in artificial systems with caution. In accordance with previous studies (10 – 12), we showed by immunostaining that valinomycin- and H2O2-induced mitochondrial translocation of Parkin is dependent upon PINK1 in human dermal fibroblasts and only takes place in healthy controls but not in affected PINK1 mutation carriers. Furthermore, on the endogenous level, we demonstrated accumulation of Parkin in the mitochondrial fraction upon valinomycin-induced stress. Importantly, this accumulation

is even more pronounced when inhibiting the UPS, whereby stress-induced degradation of Parkin was prevented and accumulation of Parkin occurred in both the cytosolic and the mitochondrial fractions. However, these effects were only present in the cells from control individuals but not in those derived from carriers of pathogenic PINK1 mutations. On the other hand, overexpressing wild-type Parkin in cells from PINK1 mutation carriers does lead to the presence of detectable Parkin also in the mitochondrial fraction. This observation from experiments with overexpressed protein is in agreement with the finding that normal development of PINK1 knockout Drosophila flies can be rescued by overexpressing WT Parkin but not vice versa (5,7,8). Thus, our data collectively demonstrate that wild-type PINK1 is necessary for two different processes, i.e. mitochondrial translocation of endogenous Parkin and regulation of Parkin levels via the UPS. At this point, the possible interplay of these two processes and the underlying mechanism(s) remain to be elucidated. Another important finding of our study is that it is specifically the mitochondrial localization of PINK1 that is needed for the stress-induced changes of Parkin, as demonstrated in a series of experiments using different truncated forms of PINK1. Importantly, PINK1d93 is sufficient to restore the PINK1 wild-type phenotype, i.e. the stress-induced mitochondrial translocation and loss of detectable Parkin. In contrast, PINK1d110 did not have this effect. This data is consistent with previous findings, where PINK1 lacking the first 110 amino acids, suggesting that the amino acid residues between position 93 and 110 are necessary for mitochondrial targeting of PINK1 and its normal function.

12

Human Molecular Genetics, 2010

In our experiments with the potassium ionophore valinomycin, we showed that endogenous full-length PINK1 accumulates in the mitochondrial fraction upon collapse of the mitochondrial membrane potential, followed by mitochondrial translocation and loss of endogenous Parkin. These results are in accordance with previous findings that PINK1 stabilization on impaired mitochondria is necessary for Parkin recruitment (12). In contrast, in our study, treatment with H2O2, which had the same effect on the level of endogenous Parkin and its subcellular localization as valinomycin, induced no accumulation of PINK1. Therefore, we conclude that accumulation of PINK1 is sufficient but not necessary for the stress-induced recruitment of Parkin to mitochondria or the PINK1-mediated loss of Parkin. In addition to accumulation of PINK1 upon loss of the mitochondrial membrane potential, we detected accumulation of the premature, non-processed form of the mitochondrial matrix protein GRP75 only upon valinomycin treatment but not upon H2O2 treatment. Although PINK1 and GRP75 accumulated in different cellular compartments, our data suggest that these two proteins share, at least partially, the same pathway of mitochondrial import. This is in keeping with the notion that PINK1, as well as GRP75, are synthesized as a precursor protein which is cleaved into its mature form after import into mitochondria in a mitochondrial membrane potential-dependent manner (41,46). More recently, a novel mechanism was proposed, where PINK1 is constitutively synthesized and imported into all mitochondria, but cleaved by voltage-sensitive proteolysis. On damaged mitochondria that have lost their membrane potential, PINK1 cleavage is inhibited, leading to PINK1 accumulation (12). However, this model cannot explain the H2O2-induced PINK1-mediated mitochondrial recruitment of Parkin, where no PINK1 accumulation was observed. In addition to confirming and expanding on previous findings on the PINK1/Parkin pathway, our study revealed certain discrepancies between the behavior of endogenous versus overexpressed protein. This highlights the need for more experiments to be performed at the endogenous level and for employing different stressors to explore their differential effects on the functional relationship of PINK1 and Parkin. More generally speaking, our study shows the necessity to introduce an environmental factor, i.e. stress, to visualize the differences in the interaction of PINK1 and Parkin in mutants versus controls and thus provide experimental evidence for the generally held notion of PD as a condition with a combined genetic and environmental etiology. Although our data support the proposition that additional stressors may be important in PD pathogenesis, such stressors do not necessarily have to be exogenous in terms of environment. In fact, endogenous stressors, such as free radicals produced over an individual’s lifetime, may be equally important as contributing factors in the development of PD.

MATERIALS AND METHODS

maintained at 378C in a saturated humidity atmosphere containing 5% CO2. All fibroblasts had the same passage number. Passage numbers ,10 were used for all experiments. RNA extraction and real-time PCR analysis Total RNA from fibroblasts was prepared by using the RNA easy protect kit (Qiagen) according to the manufacturer’s instructions and then reverse-transcribed into cDNA with the Super Script First-Strand Synthesis System (Invitrogen). The resulting cDNAs were quantified by real-time PCR using LightCycler DNA Master SYBR Green I on the Light Cycler 2.0 Real-Time PCR system (Roche Diagnostics). Transient transfection Fibroblasts were transiently transfected with pcDNA3.1 V5/His (Invitrogen) containing full-length wild-type PINK1 cDNA (FL PINK1). Truncated forms PINK1d93 and PINK1d110 were generated by PCR and cloned in pcDNA3.1 V5/His (Invitrogen). For overexpression of Parkin, N-terminally FLAG-tagged full-length Parkin cDNA was cloned in pcDNA3.1 (Invitrogen’s modified vector lacking V5/His tags). All constructs were confirmed by sequencing. For PINK1 knockdown, Hs_PINK1_4_HP validated siRNA (Qiagen) (final concentration 50 nM) was used and scramble siRNA (Silencer negative control 1 siRNA [Ambion]) (final concentration 50 nM) with no known mammalian homology was used as negative control. All transfections of fibroblasts were performed using the Nucleofector Device (Lonza). HEK cells were transfected with the Ca2+PO4 method (47). Stress induction Fibroblasts were treated with the potassium ionophore valinomycin (1 mM, Sigma) and with H2O2 (70 mM, Sigma). For inhibition of the proteasomal system, MG132 (10 –20 mM, Sigma) and epoxomicin (10 mM, Sigma) were used. Immunocytochemistry Fibroblasts or HEK cells were transiently transfected with a vector containing FLAG-Parkin alone or together with FL PINK1, PINK1d93 or PINK1d110. After transfection, cells were plated on glass coverslips. After 24 h, cells were incubated with valinomycin for an additional 12 h, formaldehydefixed and stained with anti-FLAG M2 (Sigma) and anti-GRP75 (AbCam) antibodies in combination with Zenon labeling technology (Invitrogen). Fixed and stained cells were imaged using an inverted microscope (Axiovert 200M) upgraded with an Apotome with a 63x/1.25 oil EC PlanNEOFLUAR objective (Carl Zeiss, Inc.). All immunostainings were analyzed by the first author and at least one additional expert blinded to experimental set-up and mutational status.

Cell culture Primary human dermal fibroblasts and human embryonic kidney (HEK) cells were cultured in Dulbecco’s modified Eagle’s medium (PAA Laboratories) and supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin. All cells were

Mitochondrial preparation Mitochondria were isolated from fibroblasts as previously described (48). In brief, cells were harvested and homogenized in buffer containing 250 mM sucrose, 10 mM Tris and 1 mM

Human Molecular Genetics, 2010

EDTA, pH 7.4. After that, nuclei and unbroken cells were removed by centrifugation at 1500 g for 20 min. The supernatant containing intact mitochondria was transferred into a new tube and centrifuged at 12 000 g for 10 min. Supernatant (‘cytosolic fraction’) was transferred into another new tube and the mitochondria-enriched pellet (‘mitochondrial fraction’) was dissolved in RIPA buffer containing a cocktail of protease and phosphatase inhibitors (Roche Diagnostics). Cytoplasmic fractions were concentrated by using centricon YM-10 devices (Millipore) according to the manufacturer’s instructions. Proteins of the mitochondrial and cytoplasmic fractions were separated by SDS – polyacrylamide gel electrophoresis (SDS – PAGE) and detected by western blot analysis using appropriate antibodies. Protein extraction Proteins were extracted using RIPA buffer containing 0.1% SDS (50 mM Tris – HCl pH 7.6, 150 mM NaCl, 1% DOC, 1% NP-40 and 0.1% SDS). Cells or mitochondria-enriched pellets were dissolved in the appropriate amount of buffer and incubated on ice for 30 min. After that, the lysates were centrifuged at 16 000g for 20 min at 48C. The supernatant (‘soluble fraction’) was transferred into a new tube. The remaining pellet was washed with 1 ml of 1× PBS and dissolved in RIPA buffer containing 2% SDS (‘insoluble fraction’). Western blot analysis SDS–PAGE was performed using NuPAGE 4–12% Bis-Tris gels (Invitrogen). After electrophoresis, proteins were transferred to the nitrocellulose membrane (Protran) and probed with antibodies raised against Parkin (Cell Signalling), b-actin (Sigma), Hsp60 (Cell Signalling), b-tubulin (Sigma), VDAC1 (AbCam), anti-FLAG M2 (Sigma), PINK1 (clone E35) and Neomycin (AbCam). We tested several anti-PINK1 antibodies but none of them was generally sensitive enough (exceptions: Figs 4 and 10) to detect endogenous PINK1, as shown previously (37). All western blot analyses were performed at least twice, and representative blots are shown in the figures.

ACKNOWLEDGEMENT We thank Dr E. Hartmann for providing a PINK1 antibody. Conflict of Interest statement. None declared.

FUNDING This work was supported by the Fritz Thyssen Foundation, the Volkswagen Foundation, the Hermann and Lilly Schilling Foundation, the DFG and the BMBF (NGFN plus), PNP01GS08135-3.

REFERENCES 1. Kitada, T., Asakawa, S., Hattori, N., Matsumine, H., Yamamura, Y., Minoshima, S., Yokochi, M., Mizuno, Y. and Shimizu, N. (1998) Mutations in the parkin gene cause autosomal recessive juvenile parkinsonism. Nature, 392, 605–608.

13

2. Valente, E.M., Abou-Sleiman, P.M., Caputo, V., Muqit, M.M., Harvey, K., Gispert, S., Ali, Z., Del Turco, D., Bentivoglio, A.R., Healy, D.G. et al. (2004) Hereditary early-onset Parkinson’s disease caused by mutations in PINK1. Science, 304, 1158– 1160. 3. Greene, J.C., Whitworth, A.J., Kuo, I., Andrews, L.A., Feany, M.B. and Pallanck, L.J. (2003) Mitochondrial pathology and apoptotic muscle degeneration in Drosophila parkin mutants. Proc. Natl Acad. Sci. USA, 100, 4078–4083. 4. Pesah, Y., Pham, T., Burgess, H., Middlebrooks, B., Verstreken, P., Zhou, Y., Harding, M., Bellen, H. and Mardon, G. (2004) Drosophila parkin mutants have decreased mass and cell size and increased sensitivity to oxygen radical stress. Development, 131, 2183–2194. 5. Clark, I.E., Dodson, M.W., Jiang, C., Cao, J.H., Huh, J.R., Seol, J.H., Yoo, S.J., Hay, B.A. and Guo, M. (2006) Drosophila pink1 is required for mitochondrial function and interacts genetically with parkin. Nature, 441, 1162–1166. 6. Exner, N., Treske, B., Paquet, D., Holmstrom, K., Schiesling, C., Gispert, S., Carballo-Carbajal, I., Berg, D., Hoepken, H.H., Gasser, T. et al. (2007) Loss-of-function of human PINK1 results in mitochondrial pathology and can be rescued by parkin. J. Neurosci., 27, 12413–12418. 7. Park, J., Lee, S.B., Lee, S., Kim, Y., Song, S., Kim, S., Bae, E., Kim, J., Shong, M., Kim, J.M. et al. (2006) Mitochondrial dysfunction in Drosophila PINK1 mutants is complemented by parkin. Nature, 441, 1157–1161. 8. Yang, Y., Gehrke, S., Imai, Y., Huang, Z., Ouyang, Y., Wang, J.W., Yang, L., Beal, M.F., Vogel, H. and Lu, B. (2006) Mitochondrial pathology and muscle and dopaminergic neuron degeneration caused by inactivation of Drosophila Pink1 is rescued by Parkin. Proc. Natl Acad. Sci. USA, 103, 10793– 10798. 9. Xiong, H., Wang, D., Chen, L., Choo, Y.S., Ma, H., Tang, C., Xia, K., Jiang, W., Ronai, Z., Zhuang, X. et al. (2009) Parkin, PINK1 and DJ-1 form a ubiquitin E3 ligase complex promoting unfolded protein degradation. J. Clin. Invest., 119, 650 –660. 10. Geisler, S., Holmstro¨m, K.M., Skujat, D., Fiesel, F.C., Rothfuss, O.C., Kahle, P.J. and Springer, W. (2010) PINK1/Parkin-mediated mitophagy is dependent on VDAC1 and p62/SQSTM1. Nat. Cell Biol., 12, 119–131. 11. Vives-Bauza, C., Zhou, C., Huang, Y., Cui, M., de Vries, R.L., Kim, J., May, J., Tocilescu, M.A., Liu, W., Ko, H.S. et al. (2009) PINK1-dependent recruitment of Parkin to mitochondria in mitophagy. Proc. Natl Acad. Sci. USA, 107, 378 –383. 12. Narendra, D.P., Jin, S.M., Tanaka, A., Suen, D.F., Gautier, C.A., Shen, J., Cookson, M.R. and Youle, R.J. (2010) PINK1 is selectively stabilized on impaired mitochondria to activate Parkin. PLoS Biology, 8, 1 –21. 13. Unoki, M. and Nakamura, Y. (2001) Growth-suppressive effects of BPOZ and EGR2, two genes involved in the PTEN signaling pathway. Oncogene, 20, 4457– 4465. 14. Deas, E., Plun-Favreau, H. and Wood, N.W. (2009) PINK1 function in health and disease. EMBO Mol. Med., 1, 152 –165. 15. Weihofen, A., Ostaszewski, B., Minami, Y. and Selkoe, D.J. (2008) Pink1 Parkinson mutations, the Cdc37/Hsp90 chaperones and Parkin all influence the maturation or subcellular distribution of Pink1. Hum. Mol. Genet., 17, 602–616. 16. Wood-Kaczmar, A., Gandhi, S., Yao, Z., Abramov, A.Y., Miljan, E.A., Keen, G., Stanyer, L., Hargreaves, I., Klupsch, K., Deas, E. et al. (2008) PINK1 is necessary for long term survival and mitochondrial function in human dopaminergic neurons. PLoS One, 3, e2455. 17. Klinkenberg, M., Thurow, N., Gispert, S., Ricciardi, F., Eich, F., Prehn, J.H., Auburger, G. and Kogel, D. Enhanced vulnerability of PARK6 patient skin fibroblasts to apoptosis induced by proteasomal stress. Neuroscience, 166, 422–434. 18. Hoepken, H.H., Gispert, S., Morales, B., Wingerter, O., Del Turco, D., Mulsch, A., Nussbaum, R.L., Muller, K., Drose, S., Brandt, U. et al. (2007) Mitochondrial dysfunction, peroxidation damage and changes in glutathione metabolism in PARK6. Neurobiol. Dis., 25, 401–411. 19. Gegg, M.E., Cooper, J.M., Schapira, A.H. and Taanman, J.W. (2009) Silencing of PINK1 expression affects mitochondrial DNA and oxidative phosphorylation in dopaminergic cells. PLoS One, 4, e4756. 20. Moore, D.J. (2006) Parkin: a multifaceted ubiquitin ligase. Biochem. Soc. Trans., 34, 749– 753. 21. Abou-Sleiman, P.M., Muqit, M.M. and Wood, N.W. (2006) Expanding insights of mitochondrial dysfunction in Parkinson’s disease. Nat. Rev. Neurosci., 7, 207–219.

14

Human Molecular Genetics, 2010

22. Rothfuss, O., Fischer, H., Hasegawa, T., Maisel, M., Leitner, P., Miesel, F., Sharma, M., Bornemann, A., Berg, D., Gasser, T. et al. (2009) Parkin protects mitochondrial genome integrity and supports mitochondrial DNA repair. Hum. Mol. Genet., 18, 3832–3850. 23. Narendra, D., Tanaka, A., Suen, D.F. and Youle, R.J. (2009) Parkin-induced mitophagy in the pathogenesis of Parkinson disease. Autophagy, 5, 706–708. 24. Gru¨newald, A., Gegg, M.E., Taanman, J.W., King, R.H., Kock, N., Klein, C. and Schapira, A.H. (2009) Differential effects of PINK1 nonsense and missense mutations on mitochondrial function and morphology. Exp. Neurol., 219, 266–273. 25. Hedrich, K., Hagenah, J., Djarmati, A., Hiller, A., Lohnau, T., Lasek, K., Gru¨newald, A., Hilker, R., Steinlechner, S., Boston, H. et al. (2006) Clinical spectrum of homozygous and heterozygous PINK1 mutations in a large German family with Parkinson disease: role of a single hit? Arch. Neurol., 63, 833 –838. 26. Gru¨newald, A., Breedveld, G.J., Lohmann-Hedrich, K., Rohe, C.F., Ko¨nig, I.R., Hagenah, J., Vanacore, N., Meco, G., Antonini, A., Goldwurm, S. et al. (2007) Biological effects of the PINK1 c.1366C.T mutation: implications in Parkinson disease pathogenesis. Neurogenetics, 8, 103–109. 27. Moro, E., Volkmann, J., Konig, I.R., Winkler, S., Hiller, A., Hassin-Baer, S., Herzog, J., Schnitzler, A., Lohmann, K., Pinsker, M.O. et al. (2008) Bilateral subthalamic stimulation in Parkin and PINK1 parkinsonism. Neurology, 70, 1186– 1191. 28. Furlong, I.J., Lopez Mediavilla, C., Ascaso, R., Lopez Rivas, A. and Collins, M.K. (1998) Induction of apoptosis by valinomycin: mitochondrial permeability transition causes intracellular acidification. Cell Death Differ., 5, 214– 221. 29. Narendra, D., Tanaka, A., Suen, D.F. and Youle, R.J. (2008) Parkin is recruited selectively to impaired mitochondria and promotes their autophagy. J. Cell Biol., 183, 795– 803. 30. Shiba, K., Arai, T., Sato, S., Kubo, S., Ohba, Y., Mizuno, Y. and Hattori, N. (2009) Parkin stabilizes PINK1 through direct interaction. Biochem. Biophys. Res. Commun., 383, 331–335. 31. Um, J.W., Stichel-Gunkel, C., Lubbert, H., Lee, G. and Chung, K.C. (2009) Molecular interaction between parkin and PINK1 in mammalian neuronal cells. Mol. Cell Neurosci., 40, 421– 432. 32. Imai, Y., Soda, M. and Takahashi, R. (2000) Parkin suppresses unfolded protein stress-induced cell death through its E3 ubiquitin–protein ligase activity. J. Biol. Chem., 275, 35661–35664. 33. Shimura, H., Hattori, N., Kubo, S., Mizuno, Y., Asakawa, S., Minoshima, S., Shimizu, N., Iwai, K., Chiba, T., Tanaka, K. et al. (2000) Familial Parkinson disease gene product, parkin, is a ubiquitin–protein ligase. Nat. Genet., 25, 302– 305. 34. Zhang, Y., Gao, J., Chung, K.K., Huang, H., Dawson, V.L. and Dawson, T.M. (2000) Parkin functions as an E2-dependent ubiquitin–protein ligase and promotes the degradation of the synaptic vesicle-associated protein, CDCrel-1. Proc. Natl Acad. Sci. USA, 97, 13354– 13359. 35. Granot, Z., Melamed-Book, N., Bahat, A. and Orly, J. (2007) Turnover of StAR protein: roles for the proteasome and mitochondrial proteases. Mol. Cell Endocrinol., 265–266, 51– 58.

36. Meng, L., Mohan, R., Kwok, B.H., Elofsson, M., Sin, N. and Crews, C.M. (1999) Epoxomicin, a potent and selective proteasome inhibitor, exhibits in vivo antiinflammatory activity. Proc. Natl Acad. Sci. USA, 96, 10403– 10408. 37. Zhou, C., Huang, Y., Shao, Y., May, J., Prou, D., Perier, C., Dauer, W., Schon, E.A. and Przedborski, S. (2008) The kinase domain of mitochondrial PINK1 faces the cytoplasm. Proc. Natl Acad. Sci. USA, 105, 12022–12027. 38. Haque, M.E., Thomas, K.J., D’Souza, C., Callaghan, S., Kitada, T., Slack, R.S., Fraser, P., Cookson, M.R., Tandon, A. and Park, D.S. (2008) Cytoplasmic Pink1 activity protects neurons from dopaminergic neurotoxin MPTP. Proc. Natl Acad. Sci. USA, 105, 1716–1721. 39. Takatori, S., Ito, G. and Iwatsubo, T. (2008) Cytoplasmic localization and proteasomal degradation of N-terminally cleaved form of PINK1. Neurosci. Lett., 430, 13– 17. 40. Beilina, A., Van Der Brug, M., Ahmad, R., Kesavapany, S., Miller, D.W., Petsko, G.A. and Cookson, M.R. (2005) Mutations in PTEN-induced putative kinase 1 associated with recessive parkinsonism have differential effects on protein stability. Proc. Natl Acad. Sci. USA, 102, 5703–5708. 41. Silvestri, L., Caputo, V., Bellacchio, E., Atorino, L., Dallapiccola, B., Valente, E.M. and Casari, G. (2005) Mitochondrial import and enzymatic activity of PINK1 mutants associated to recessive parkinsonism. Hum. Mol. Genet., 14, 3477–3492. 42. Sim, C.H., Lio, D.S., Mok, S.S., Masters, C.L., Hill, A.F., Culvenor, J.G. and Cheng, H.C. (2006) C-terminal truncation and Parkinson’s disease-associated mutations down-regulate the protein serine/threonine kinase activity of PTEN-induced kinase-1. Hum. Mol. Genet., 15, 3251– 3262. 43. Mizzen, L.A., Chang, C., Garrels, J.I. and Welch, W.J. (1989) Identification, characterization, and purification of two mammalian stress proteins present in mitochondria, grp 75, a member of the hsp 70 family and hsp 58, a homolog of the bacterial groEL protein. J. Biol. Chem., 264, 20664– 20675. 44. Dahlseid, J.N., Lill, R., Green, J.M., Xu, X., Qiu, Y. and Pierce, S.K. (1994) PBP74, a new member of the mammalian 70-kDa heat shock protein family, is a mitochondrial protein. Mol. Biol. Cell, 5, 1265– 1275. 45. Szabadkai, G., Bianchi, K., Varnai, P., De Stefani, D., Wieckowski, M.R., Cavagna, D., Nagy, A.I., Balla, T. and Rizzuto, R. (2006) Chaperone-mediated coupling of endoplasmic reticulum and mitochondrial Ca2+ channels. J. Cell Biol., 175, 901– 911. 46. Rohe, C.F., Montagna, P., Breedveld, G., Cortelli, P., Oostra, B.A. and Bonifati, V. (2004) Homozygous PINK1 C-terminus mutation causing early-onset parkinsonism. Ann. Neurol., 56, 427–431. 47. Sena-Esteves, M., Tebbets, J.C., Steffens, S., Crombleholme, T. and Flake, A.W. (2004) Optimized large-scale production of high titer lentivirus vector pseudotypes. J. Virol. Methods, 122, 131– 139. 48. Almeida, A. and Medina, J.M. (1997) Isolation and characterization of tightly coupled mitochondria from neurons and astrocytes in primary culture. Brain Res., 764, 167–172.