PINK1-Interacting Proteins: Proteomic Analysis of Overexpressed PINK1

Recent publications suggest that the Parkinson's disease- (PD-) related PINK1/Parkin pathway promotes elimination of dysfunctional mitochondria by autophagy. We used tandem affinity purification (TAP), SDS-PAGE, and mass spectrometry as a first step towards identification of possible substrates for PINK1. The cellular abundance of selected identified interactors was investigated by Western blotting. Furthermore, one candidate gene was sequenced in 46 patients with atypical PD. In addition to two known binding partners (HSP90, CDC37), 12 proteins were identified using the TAP assay; four of which are mitochondrially localized (GRP75, HSP60, LRPPRC, and TUFM). Western blot analysis showed no differences in cellular abundance of these proteins comparing PINK1 mutant and control fibroblasts. When sequencing LRPPRC, four exonic synonymous changes and 20 polymorphisms in noncoding regions were detected. Our study provides a list of putative PINK1 binding partners, confirming previously described interactions, but also introducing novel mitochondrial proteins as potential components of the PINK1/Parkin mitophagy pathway.


Introduction
Parkinson's disease (PD) is a progressive neurodegenerative disorder with three cardinal manifestations: tremor, rigidity, and bradykinesia. In about 25% of all PD patients, at least one additional affected family member can be found, likely pointing to a direct, genetic cause of the disease. To date, eight genes have been confirmed to be associated with PD [1].
Besides the mutational analysis of PD-linked genes, the search for potential interactions between the protein products of these genes has recently gained increasing importance. Since the identification of the second monogenetic PD gene product Parkin, a common pathway leading to dopaminergic neurodegeneration has been proposed. To date, several studies reported such connections.
First, Shimura et al. hypothesized that Parkin plays a role in the coregulation of Alpha-synuclein (SNCA). The group identified a protein complex in normal human brain that included the E3 ubiquitin ligase Parkin, UBCH7 as its associated E2 ubiquitin-conjugating enzyme, and a novel form of SNCA as its substrate [2]. Later, an early-onset PD patient with a heterozygous missense mutation in both the DJ-1 and the PTEN-induced putative kinase 1 (PINK1) gene has been described. Additionally, overexpression of DJ-1 and PINK1 in SHSY-5Y cells revealed that the wildtype as well as the mutant forms of both proteins interact, and DJ-1 stabilized PINK1 [3]. Furthermore, Drosophila pink1 and parkin lossof-function mutants showed a similar mitochondrial phenotype. Since the pink1-related abnormalities could be rescued by parkin overexpression but not vice versa, it was suggested that pink1 acts upstream of parkin in a common pathway [4][5][6]. Recent studies provided evidence that the PINK1/Parkin pathway promotes mitochondrial fission as an initial step of mitophagy by ubiquitination of Mitofusins [7][8][9].
Still, only the minority of hereditary forms of PD can be explained by a mutation in one of the nine PD-associated genes [10]. Therefore, this research focuses not only on proteins known to be involved in PD but also on novel interactors.
In the present study, we employed tandem affinity purification (TAP) to isolate proteins that are directly associated with PINK1. Using this approach, we aimed at a better characterization of the PINK1/Parkin mitophagy pathway.

Material and Methods
2.1. Patients. All patients underwent a standardized neurological examination performed by a movement disorders specialist. For sequencing of one of the candidate genes, Leucine-rich PPR motif-containing (LRPPRC), 46 patients with atypical Parkinsonism including features of dementia, depression, or rapid disease progression were included in the present study. After obtaining informed consent, we collected blood from all patients for DNA extraction according to a published protocol [11].
Furthermore, fibroblasts from three PD patients with a homozygous p.Q456X PINK1 mutation and from two agematched mutation-negative healthy controls were included in the study. Clinical features of these mutation carriers are described elsewhere [12,13].

Tissue
Culture. HEK cells and fibroblasts 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 maintained at 37 • C in a saturated humidity atmosphere containing 5% CO 2 . Passage numbers <10 were used for all experiments. To inhibit the mitochondrial membrane potential, fibroblasts were treated with the potassium ionophore valinomycin (1 μM, Sigma).

Protein Isolation by TAP and Mass Spectrometry (MS).
The InterPlay Mammalian TAP System was employed according to the manufacturer's (Stratagene) protocol. In brief, human control RNA was extracted using the RNeasy Mini Kit (Qiagen) and PINK1 cDNA was synthesized using Superscript II Reverse Transcriptase in combination with oligo dT primers (Invitrogen). Full-length PINK1 cDNA was cloned into the pCTAP expression vector, which encodes two tandem affinity tags (a streptavidin binding peptide and a calmodulin binding peptide) after its multiple cloning site. Next, 10 8 HEK cells were transiently transfected with the pCTAP-PINK1 vector by means of the Ca 2+ PO 4 method [14]. After 24 h, cells were harvested and resuspended in lysis buffer supplemented with protease inhibitors (Sigma). Cells underwent three rounds of freeze-thawing, cell debris was pelleted by centrifugation at 16,000× g for 10 min, and the supernatant collected. Next, 2 mM EDTA, 10 mM βmercaptoethanol, and the InterPlay streptavidin resin were added to the cell lysate and incubated at 4 • C while rotating for 2 h. The resin was collected by centrifugation at 1,500× g for 5 min, washed twice, and incubated in biotin-containing buffer for 30 min at 4 • C to elute the bound protein complexes. To further purify the protein complexes, calmodulin resin and a calcium-containing buffer were added to the supernatant. After 2 h of incubation at 4 • C, the resin was collected (centrifugation at 1,500× g for 5 min) and washed twice. From the calmodulin resin, protein complexes were eluted by adding EDTA-containing buffer and incubation for 30 min at 4 • C on a rotator. The resulting eluate was concentrated by means of trichloroacetic precipitation and subsequently resuspended in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl pH7.6, 150 mM NaCl, 1% DOC, and 1% NP-40) with protease and phosphatase inhibitors (Roche Diagnostics). Next, purified proteins were resolved by one-dimensional SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and visualized on the gel by silver staining as published [15]. In a control experiment, the pulldown approach was performed with beads only. Bands which appeared on the gel from the PINK1 overexpression experiment but not the control experiment were excised and sent for MS analysis to the Taplin Biological Mass Spectrometry Facility, Harvard Medical School, Boston, USA. According to the quality requirements of the facility, a protein can only be considered as interactor if two or more peptides match the respective protein in the Swiss-Prot database.

Knock-down Approach.
For PINK1 and LRPPRC knockdown, Hs PINK1 4 HP (Qiagen) and LRPPRC UTR (complementary to the segment of the 3 UTR region of LRP-PRC [16]) siRNAs were used with a final concentration of 50 nM. Scrambled siRNA (Silencer negative control 1 siRNA [Ambion]) with no known mammalian homology served as negative control (final concentration 50 nM). For transfection, the Nucleofector Device (Lonza) was used.

Mitochondrial Preparation.
Mitochondria were isolated from fibroblasts as previously described [17]. In brief, cells were harvested and homogenized in buffer containing 250 mM sucrose, 10 mM Tris, and 1 mM EDTA, pH7.4. After that, nuclei and unbroken cells were removed by centrifugation at 1,500× g for 20 min. The supernatant containing intact mitochondria was transferred into a new tube and centrifuged at 12,000× g for 10 min. The resulting supernatant (cytosolic fraction) was transferred into another new tube and the mitochondria-enriched pellet (mitochondrial fraction) was dissolved in RIPA buffer containing protease and phosphatase inhibitors (Roche Diagnostics). Cytoplasmic fractions were concentrated by using Centricon YM-10 devices (Millipore) according to the manufacturer's instructions.

Protein Extraction.
Proteins were extracted using RIPA buffer containing 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 16,000× g for 20 min at 4 • C. The supernatant was transferred into a new tube for further processing.

Results and Discussion
3.1. New PINK1 Interactors. Using a TAP approach with overexpressed PINK1 (PINK1+), or empty beads (PINK1−) we identified interactors of the target protein. The resulting TAP eluates were resolved by SDS-PAGE. Silver staining was used to visualize protein bands on the PAGE gels. Seven bands were excised from the PINK1+ gel (Figure 1(a), a-g) which were not detectable on the PINK1− gel (Figure 1(b)) and analyzed by MS (Table 1). A total of 14 proteins were identified which met the requirement of having two or more peptides matched to it by the database-searching program. Out of those, seven are predominantly found in the cytoplasm (Heat shock 90 kDa proteins alpha and beta, Heat shock 70 kDa proteins 1, 2, 8, and 1-like, and Hsp90 cochaperone Cdc37), two are components of microtubuli (Tubulin alpha-1C chain and Tubulin alpha-3C/D chain), one is associated with the endoplasmic reticulum (78 kDa Glucose-regulated protein) and four are mitochondrially localized (GRP75, HSP60, LRPPRC, and TUFM) (for details on subcellular localization see: http://expasy.org/sprot/). Of note, also PINK1 itself was identified by MS in most of the analyzed bands on the PAGE gel (Figure 1(a), b-g).
To our knowledge, only the interaction between the HSP90/CDC37 chaperone system and PINK1 has been described so far [19,20]. CDC37 is a molecular cochaperone that functions with HSP90 to promote folding of kinases [21]. With respect to PINK1, the chaperone system was found to influence the protein's subcellular distribution. The authors of this study proposed that the HSP90/CDC37/PINK1 complex is destined for a translocation that leads to PINK1 processing, whereas in the absence of HSP90 in the complex, PINK1 might be attached to mitochondria as full-length precursor [19].
Given that PINK1 was reported to be associated with mitochondria [17,22], we focused primarily on the mitochondrial proteins GRP75, HSP60, LRPPRC, and TUFM in the ensuing experiments.

Cellular
Abundance of GRP75, HSP60, LRPPRC, and TUFM. First, we determined the quality of the noncommercially available LRPPRC antibody by means of a knock-down  approach. This experiment showed a drop in LRPPRC levels when siRNA against LRPPRC was employed but not when scrambled siRNA was used, confirming the specificity of the antibody. LRPPRC protein levels were also not affected by a PINK1 knockdown ( Figure 2).
Next, the abundance of GRP75, HSP60, LRPPRC, and TUFM was investigated in mitochondrial fractions from control and PINK1-mutant fibroblasts. This experiment revealed comparable levels of GRP75, HSP60, and LRPPRC in both groups (Figures 3(a) and 3(b)-left half). The abundance of TUFM was variable in all investigated samples showing no clear trend when comparing mutants and controls (Figure 3(a)).
Furthermore, we tested the quality of the GRP75 antibody by investigating GRP75 in the cytosolic fraction under valinomycin stress conditions. GRP75 is a mitochondrial matrix chaperone, synthesized as a 679-amino acid preprotein, which contains a 51-residue N-terminal mitochondrial targeting sequence (MTS). After the membrane potentialdependent import into mitochondria, it is cleaved into the mature protein which is ∼5.5 kDa shorter than the preprotein [23][24][25]. When treating cells with the mitochondrial membrane potential inhibitor valinomycin, the full-length form of GRP75 (MTS-GRP75) was detected in the cytosolic fractions from mutants and controls. Consequently, we considered the anti-GRP75 antibody as specific. In both groups, the levels of MTS-GRP75 were comparable (Figure 3(b)right half). The abundance of the processed form of GRP75 in the cytosolic fraction was likely due to contamination with the mitochondrial fraction, although previous studies have shown that GRP75 can also be cytosolically localized [26].
Though our Western blot results are not supporting a direct link between PINK1 and any of the detected mitochondrial proteins, it should be noted that GRP75, HSP60, and LRPPRC have been identified by proteomic analysis as potential interactors of Parkin earlier [27]. Furthermore, their molecular functions render them interesting targets in the context of PD.
GRP75 serves as a major mitochondrial molecular chaperone and plays a key role in the import and partitioning of nuclear-encoded proteins within the two mitochondrial membranes and the matrix [28][29][30]. Furthermore, GRP75 seems to function in the management of oxidative stress via the PD-associated protein DJ-1. Mutations in DJ-1 were found to weaken the protein's interaction with GRP75 [31,32]. GRP75 has also been described as an antiapoptotic agent. By binding of the transcription regulator p53, GRP75 prevents the formation of the proapoptotic p53/Bcl-xL/Bcl-2 complex [33,34]. Furthermore, putative mutations in GRP75 were suggested to contribute to the risk of developing PD [35] and a decrease in GRP75 expression was detected in PD patient brains compared to controls [36].
HSP60 is a mitochondrial chaperone responsible for the transport of nuclear-encoded proteins via the mitochondrial membranes and their refolding in the matrix [37,38] and has been linked to the pathogenesis of Alzheimer's disease. Apparently, HSP60 provides protection against intracellular β-amyloid stress through maintenance of mitochondrial respiratory complex IV activity [39]. Complex IV deficiency in turn, has been implicated in PD [40,41] opening the possibility for a role of HSP60 in the pathogenesis of the disease.
LRPPRC has been linked to cytochrome C oxidase deficiency. Mutations in the gene lower MT-CO1 and MT-CO3 mRNA levels and, in turn, impair complex IV assembly [42,43]. Recent functional studies further strengthen the link between LRPPRC and mitochondrial RNA metabolism [44,45]. However, when we compared MT-CO1 protein levels in fibroblasts from PINK1 mutants and controls, no differences were observed (Figure 3(a)). Furthermore, LRPPRC was identified as a component of the PGC-1α complex which itself is also linked to energy homeostasis in the cell [46]. Like in the case of HSP60, LRPPRC's impact on the respiratory chain offers a potential connection with PD.
TUFM is part of the translational apparatus of mitochondria. During protein biosynthesis, it mediates the GTPdependent binding of aminoacyl-tRNA to the A-site of ribosomes [47]. Additional described functions of TUFM comprise recognition and translocation of cotranslationally damaged proteins to the proteasome [48], rearrangement of cytoskeletal components [49,50], and regulation of cell survival [51]. Mutations in the TUFM gene cause combined oxidative phosphorylation deficiency type 4 due to decreased mitochondrial protein synthesis [52]. Interestingly, however, there is also a report connecting TUFM and PD, where TUFM was found to co-immunoprecipitate with Leucin-rich repeat kinase 2 which is encoded by the PARK8 gene LRRK2. Coincubation with recombinant TUFM reduced the kinase

Conclusions
In the current study, TAP technology was employed for the first time to identify PINK1-associated proteins. These experiments resulted in a list of 14 putative PINK1 binding partners, confirming two reported interactions (HSP90 and CDC37), but also introducing four novel mitochondrially localized proteins (GRP75, HSP60, LRPPRC, or TUFM) as potential components of the PINK1/Parkin mitophagy pathway. Although preliminary results from protein expression and DNA sequencing analyses do not strengthen a link between the PINK1/Parkin pathway and any of these interactors, it cannot be excluded that their connection with the pathway may be more complex. Additional protein function studies, for instance under mitochondrial stress conditions, will be needed to fully characterize this potential link. In addition, future perspectives include association studies with SNPs in all identified genes in a larger PD patient sample.