Address correspondence and reprint requests to Philipp Kahle, Laboratory of Functional Neurogenetics, Department of Neurodegeneration, Hertie Institute for Clinical Brain Research, Otfried Müller Str. 27, 72076 Tübingen, Germany. E-mail: email@example.com
DJ-1 is a ubiquitous protein regulating cellular viability. Recessive mutations in the PARK7/DJ-1 gene are linked to Parkinson's disease (PD). Although the most dramatic L166P point mutation practically eliminates DJ-1 protein and function, the effects of other PD-linked mutations are subtler. Here, we investigated two recently described PD-associated DJ-1 point mutations, the A179T substitution and the P158Δ in-frame deletion. [A179T]DJ-1 protein was as stable as wild-type [wt]DJ-1, but the P158Δ mutant protein was less stable. In accord with the notion that dimer formation is essential for DJ-1 protein stability, [P158Δ]DJ-1 was impaired in dimer formation. Similar to our previous findings for [M26I]DJ-1, [P158Δ]DJ-1 bound aberrantly to apoptosis signal-regulating kinase 1. Thus, the PD-associated P158Δ mutation destabilizes DJ-1 protein and function. As there is also evidence for an involvement of DJ-1 in multiple system atrophy, a PD-related α-synucleinopathy characterized by oligodendroglial cytoplasmic inclusions, we studied an oligodendroglial cell line stably expressing α-synuclein. α-Synuclein aggregate dependent microtubule retraction upon co-transfection with tubulin polymerization-promoting protein p25α was ameliorated by [wt]DJ-1. In contrast, DJ-1 mutants including P158Δ failed to protect in this system, where we found evidence of apoptosis signal-regulating kinase 1 (ASK1) involvement. In conclusion, the P158Δ point mutation may contribute to neurodegeneration by protein destabilization and hence loss of DJ-1 function.
DJ-1 was originally cloned as an oncogene promoting Ras-dependent transformation (Nagakubo et al. 1997), and it has been found associated with cancer. More recently, the gene mutated in the PARK7 locus (Bonifati et al. 2003) linked to autosomal-recessive Parkinson's disease (PD) was found to be DJ-1 and the protein is altered both in stroke as well as in chronic neurodegenerative diseases (Choi et al. 2006; Aleyasin et al. 2007). DJ-1 is an evolutionary conserved cytoprotective protein that is capable of modulating anti-oxidative, anti-apoptotic, and anti-inflammatory pathways (Kahle et al. 2009).
Among the PD-associated DJ-1 point mutations, it is established that the most aggressive L166P practically eliminates DJ-1 protein stability and function (Macedo et al. 2003; Miller et al. 2003; Moore et al. 2003; Görner et al. 2004; Olzmann et al. 2004). This may be because of strong structural perturbations within the C-terminal helix-kink-helix motif, which is pivotal for DJ-1 dimerization and protein stability (Görner et al. 2007). Novel PD-associated mutations in the DJ-1 C-terminal region (Fig. 1) were recently identified (Macedo et al. 2009). We find the [A179T]DJ-1 protein to be as stable as wild-type [wt]DJ-1. In contrast, the P158Δ point deletion reduced DJ-1 protein stability and dimerization, in accord with a recent report (Ramsey and Giasson 2010). Interestingly, similar to our previous findings with the comparably reduced stability mutant [M26I]DJ-1, the P158Δ mutation led to aberrant DJ-1 binding to apoptosis signal-regulated kinase 1 (ASK1). In addition, [P158Δ]DJ-1 was less effective than [wt]DJ-1 in preventing downstream nuclear export of the death-associated protein Daxx.
As we had found some evidence of DJ-1 immunoreactivity in multiple system atrophy (MSA) (oligodendro)glial cytoplasmic inclusions (Neumann et al. 2004), which are composed of α-synuclein (AS) fibrils as is the neuronal Lewy pathology that characterizes PD and also contain the oligodendroglial tubulin polymerization-promoting protein p25α (Kovács et al. 2004; Lindersson et al. 2005; Song et al. 2007), we studied for the first time the role of DJ-1 in a cell culture model for MSA. Human AS was stably expressed in rat oligodendroglial cells (OLN-AS7) and the effects of DJ-1 variants on co-transfected p25α induced degeneration (Kragh et al. 2009) were measured. This was restored by [wt]DJ-1 but not mutants, including P158Δ. In conclusion, the P158Δ mutation positioned in the turn connecting the C-terminal helix-kink-helix motif leads to reduced DJ-1 protein stability and impaired cytoprotective functions, which may be relevant for neurodegenerative diseases.
Materials and methods
All pCDNA constructs harboring ASK1-HA variants and pRK5-Flag-Daxx were kind gifts from Hidenori Ichijo, University of Tokyo, Japan. Human AS and p25α in the pcDNA3.1 zeo(-) vector was described before (Kragh et al. 2009). To create the DJ-1 P158Δ point deletion, pCMV-myc-DJ-1 wt (Waak et al. 2009) was used as template for a two step site specific mutagenesis using the following primers: P158Δforward (5′-CCTGATTCTTACAAGCCGGGG GGGGACCAGCTTCG-3′) and pCMV-STOP-NotI-reverse (5′-AGCGGCCGCGTCTTTAAGAACAAGTGGAGCC-3′) or P158Δ-reverse (5′-CGAAGCTGGTCCCCCCCCGGCTTGTAAGAATCAGG-3′) and pCMV-EcoRI forward (5′-ACGAATTCGA ATGGCTTCCAAAAGAGCTCTGGT-3′). The two PCR products were used as templates for a second PCR using pCMV-EcoRI forward and pCMV-STOP-NotI-reverse primers. To create the A179T mutation, pCMV-myc-DJ-1 was used as template and pCMV-EcoRI forward and A179T-Stop-NotI (5-'AGCGGCCGCCTAGTCTTTAAGAACAAGTGGAGCCTTCACTTGAGTCGCCACCTCCTTGC-3′) reverse as primers. The PCR products were subcloned into pCMV-myc and pCMV-HA (Clontech, Saint-Germain-en-Laye, France) at EcoRI/NotI site. pCMV-myc-DJ-1 constructs were used as templates with pCDNA-NcoI forwards (5′-AACCATGGGAATGGCTTCCAAAAGAGCT CTG-3′) and pCDNA-BamHI-Stop reverse (5′-CGCGGATCCGATGTCTTTAAGAACAAGTGGAGCC-3′) as primers and these PCR products were cloned into the pCDNA3.1-V5/His6 TOPO vector (Invitrogen, Life Technologies GmbH, Darmstadt, Germany). The nucleic acid sequences of all cloned constructs were confirmed using BigDye Terminator v3.1 (Applied Biosystems, Darmstadt, Germany) according to the manufacturer's instructions. Reaction products were analyzed using an ABI 3100 Genetic Analyzer (Applied Biosystems).
Primary antibodies: anti-DJ-1 [Stressgen 3E8 (Ann Arbor, MI, USA); abcam ab18257 (Cambridge, UK); Cell Signaling Technology #5933 (Danvers, MA, USA)], rabbit anti-DJ-1 (kindly provided by Hiroyoshi Ariga, Hokkaido University, Japan), anti-p25α (Kragh et al. 2009), peridoxidase-conjugated anti-myc (9E10, Roche, Mannheim, Germany), anti-myc (9E10, Sigma Aldrich, St. Louis, MO, USA), anti-V5 (R960-25, Invitrogen), perioxidase-conjugated anti-HA (3F10, Roche), anti-HA (3F10, Roche), anti-Flag (affinity purified clone M2, Sigma Aldrich, St. Louis, MO, USA), anti-α-tubulin (SDL.3D10 or T9026, Sigma), anti-GAPDH (H86504M, Biodesign International, Memphis, TN, USA), and anti-β-actin (AC15, Sigma). Secondary peroxidase, as well as Cy2- and Cy3-conjugated antibodies, was purchased from Jackson ImmunoResearch Laboratories (West Grove, PA, USA) and secondary Alexa-Fluor conjugated antibodies were purchased from Invitrogen.
ASK-1 inhibitors AF40197, AF40405, AF39619, and AF39992 were synthesized based on patent applications WO 2011008709 and WO 2009027283, and the IC50 values for human ASK1 were confirmed using the KinaseProfiler service at Millipore according to protocol of the manufacturer.
Cell lines and cell culture
Mouse embryonic fibroblasts (MEF) lacking DJ-1 and littermate controls (Görner et al. 2007) as well as human embryonic kidney (HEK) 293E (American Tissue Culture Collection) cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Biochrom, Berlin, Germany) containing 10% fetal bovine serum in humid 37°C and 5% CO2.
OLN-AS7 cells (Kragh et al. 2009) expressing human AS were made from OLN-93 oligodendroglial cells (Richter-Landsberg and Heinrich 1996) after transfection with a pcDNA3.1 zeo(-) AS. Positive clones were initially selected by addition of 1 mg/mL zeocin. After selection, OLN-AS7 clonal cells were maintained in 100 μg/mL zeocin. These cells were maintained at 37°C under 5% CO2 and grown in DMEM (Lonza, Verviers, Belgium) supplemented with 10% fetal calf serum, 50 units/mL penicillin, and 50 μg/mL streptomycin.
FuGENE 6 or X-tremeGENE 9 (Roche Applied Science) was used according to the manufacturer's instructions for transient transfections.
Steady-state protein and mRNA determinations
MEF Dj-1+/+, MEF Dj-1−/− and HEK293E cells were transiently transfected with pCMV-myc-DJ-1 (wt, A179T or P158Δ) or pCDNA-DJ-1-V5/His6 (wt, A179T or P158Δ) and proteins were expressed 48 h. Cells were lysed in 1% Triton X-100, 150 mM NaCl, 1 mM EDTA, 10 mM Tris-HCl (pH 7.5) plus Cømplete protease inhibitor cocktail (Roche). Protein lysates were analyzed through western blotting using polyvinylidene difluoride membranes (Millipore, Schwalbach, Germany) and Amersham Hyperfilm™ ECL (GE Healthcare, Munich, Germany).
For mRNA measurements, cells that were treated as described above were lysed in RLT-RNA lysis buffer (Qiagen Inc., Hilden, Germany). RNA was isolated using the RNeasy Mini kit (Qiagen Inc.). cDNA was produced with Transcriptor High Fidelity cDNA Synthesis kit (Roche) and the cDNA was used as template for a PCR reaction with pCMV-EcoRI forward (see above) and pCMV-Stop-NotI (see above) reverse primers or, for normalization, β-actin-forward (5′-CTAAGGCCAACCGTGAA-3′) and β-actin-reverse (5′-CCGGAGTCCATCACAAT-3′) primers. The DNA fragments were separated in 1.5% agarose gels containing ethidium bromide. The DNA bands were detected with a Vilber Lourmat (Vilber, Marne La Vallee, France).
OLN-AS7 cells were transfected with mock plasmid and DJ-1 or myc-DJ-1 variants. After 24 h, the cells were harvested in radio-immunoprecipitation assay buffer (50 mM Tris (pH 7.4), 150 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.1% sodium dodecylsulfate, 1 mM EDTA, and Cømplete protease inhibitor cocktail). The concentration of the total protein lysates was determined using bicinchoninic acid (BCA) protein assay kit (Pierce, Bonn, Germany). Lysates were resolved on 10–16% polyacrylamide gels (30 μg/lane) and transferred to nitrocellulose membranes. The membranes were blocked in 10 mM Tris, pH 7.4, 150 mM NaCl (Tris-buffered saline) supplemented with 0.1% Tween 20 and 5% non-fat milk and then probed with the different antibodies. After washing, the membranes were incubated with horseradish peroxidase-conjugated anti-mouse IgG, and proteins were visualized with enhanced chemiluminescence using a Fuji LAS-3000 imager.
Dj-1−/− and Dj-1+/+ MEF cells (500 000 ea.) were plated into 10 cm cell culture dishes. The following day, cells were transiently transfected with pCMV-myc-DJ-1 constructs. After 24 h, the cells from the 10 cm dishes were split to six wells in six-well-plates. Another day later, 40 μg/mL cycloheximide (Sigma) was added to the cells and incubated for 0–12 h. Cells were lysed in Triton containing lysis buffer (see above) and protein lysates were analyzed through western blotting.
HEK 293E cells (2 × 106 ea.) were plated into 14 cm cell culture plates. On the following day, cells were transiently transfected with pCMV-myc-DJ-1 constructs. Eight hours after transfection, the cells were split into six 6 cm dishes. One day later, the cells were washed once in warm phosphate buffer saline and starved 1 h in cysteine and methionine free media (DMEM 21013-024, Invitrogen), 2 mM l-glutamine, 10 U/mL penicillinG, 10 μg/mL streptomycin), after which a 30 min long pulse with 80 μCi/mL EasyTag™ EXPRESS35S Protein Labeling Mix (Perkin Elmer, Mechelen, Belgium) was given. Cells were washed twice and there after chased for 0–24 h in chase media [DMEM (PAA Laboratories, Pasching, Austria), 2 mM l-glutamine, 1 mM l-cysteine, 1 mM l-methionine, 10% (v/v) fetal bovine serum, 10 U/mL Penicillin, and 10 μg/mL Streptomycin]. Cells were lysed in Triton containing lysis buffer (see above), and the protein lysates were quick frozen using dry ice and stored at −20°C until all samples were collected. Equal amounts of protein (determined by BCA protein assay, Pierce) were immunoprecipitated overnight using EZview™ Red Anti-myc affinity Gel (Sigma). Beads were washed three times in lysis buffer and boiled in 6x Laemmli buffer at 95°C for 5 min and proteins were separated on a 12.5% sodium dodecyl sulfate (SDS)-gel. The gel was Coomassie stained and incubated in Amplify™ (GE Healthcare) for 30 min, after which the gel was dried. Radioactive signals were detected on Amersham Hyperfilm™ ECL (GE Healthcare) using approximately 7 days of exposure at −80°C. Band densities were quantified using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). Band intensities of the different time points were expressed as percentage to the DJ-1 band intensity at time point zero. Half-life times were calculated by curve fitting assuming an exponential decay rate (x = e−kt).
Co-immunoprecipitation and gel filtration chromatography
HEK 293E cells were transiently co-transfected with pCMV-myc-DJ-1 and pCMV-HA-DJ-1 constructs or pCDNA3-ASK1-HA. After 48 h, expression cells were washed once in pyruvate-free DMEM (PAA) containing 5 mM d-glucose and treated or left untreated with 1 mM H2O2 for 30 min. For mapping co-immunoprecipitations and gel filtration chromatography, HEK 293E cells were transiently co-transfected with pCMV-myc-DJ-1 constructs and pCDNA3-ASK1-HA. After 24 h, medium was changed to pyruvate-free DMEM containing 5 mM d-glucose and 16 h later cells were treated or left untreated with 1 mM H2O2 for 30 min.
Cells were lysed in CoIP buffer [0.2% NP-40, 10 mM KCl, 50 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 1.5 mM MgCl2, 10% glycerol, 10 mM sodium pyrophosphate, 50 mM HEPES (pH 7.5) plus Cømplete protease inhibitor cocktail]. Equal amounts of protein were immunoprecipitated 1 h or 3 h using anti-myc or EZview™ Red Anti-HA affinity Gel beads (Sigma Aldrich). Beads were washed three times in CoIP buffer and boiled in 6× Laemmli buffer at 95°C for 15 min. Proteins were separated on 12.5% SDS-gels and western blotted.
Alternatively, after lysis in coIP buffer (see above) the lysates were applied on a Superdex 200 10/300 column (GE Healthcare) pre-equilibrated with running buffer [10 mM KCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1.5 mM MgCl2, 5% glycerol, 50 mM HEPES-KOH, (pH 7.5)]. The proteins were separated with a 0.45 mL/min flow rate and the eluate was collected in 0.5 mL fractions. Fractions between 7 and 20.5 mL elution volume were applied on 10% SDS-gels and western blotting was performed.
MEF DJ-1−/− cells were plated onto poly-d-lysine (Sigma Aldrich) coated coverslips. After 24 h cells were transiently co-transfected with pRK5-Flag-Daxx and pCMV-HA-DJ-1 constructs or empty pCMV-HA. The next day medium was changed to pyruvate-free DMEM containing 5 mM d-glucose, after 16 h 500 μM H2O2 was added for 30 min, controls were left untreated. The cells were fixed in 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min at 20°C and permeabilized in 1% (v/v) Triton-X-100 in PBS for 5 min at 20°C. Unspecific binding of the antibodies was decreased with a blocking step with 10% (v/v) normal goat serum in PBS for 1 h at 20°C. Primary and secondary antibodies were diluted (dilutions were 1 : 500 for primary and 1 : 1000 for secondary) in 1% (w/v) bovine serum albumin (BSA) in PBS and incubated for 1 h at 20°C or overnight at 4°C in a humidified chamber. Nuclei were counter-stained with 2 μg/mL Hoechst 33342 in PBS for 10 min at 20°C. Coverslips were mounted in Mowiol/DABCO solution (Carl Roth GmbH, Karlsruhe, Germany) onto glass slides. After each step of the staining protocol, coverslips were washed three times in PBS at 20°C. The cells were analyzed with Axioimager microscope equipped with ApoTome Imaging system (Carl Zeiss, GmbH, Jena, Germany) using 25x objective. The images were processed with AxioVision software (Carl Zeiss). The cytosolic versus nuclear localization of Flag-Daxx was quantified from the images; more than 50 cells per sample were used. Following formula were used: [N(Cytosol)+H2O2:N(total)+H2O2]:[N(Cytosol)-H2O2:N(total)-H2O2], where N is the amount of scored cells.
OLN-AS7 viability assays
For the comparison of DJ-1 wt and mutants, poly-l-lysine-coated coverslips were used to culture the cells for 24 h followed by transfection with FuGENE transfection reagent. Then cells were fixed with 3.7% formaldehyde for 10 min, permeabilized with 0.1% Triton X-100 for 30 min, and blocked in 3% BSA solution for 30 min at 20°C. Cell preparations were incubated with primary antibodies for 1 h at 20°C, and proteins were visualized by Alexa-Fluor 488, 568-or-647-conjugated secondary antibodies. Nuclei were stained with 4, 6-diamidine-2-phenylindole dihydrochloride (DAPI). Signals were analyzed on a fluorescence microscope (Axiovert 200M, Zeiss, Germany) equipped with an ApoTome.
Microtubule (MT) retraction as a measure of degeneration was quantified as previously described (Kragh et al. 2009). Briefly, OLN-AS7 cells were immunostained for α-tubulin, p25α, and DJ-1 with nuclei counter-stained, and analyzed by fluorescence microscopy. Assessment of MT retraction in p25α expressing cells was performed using antibodies against DJ-1, p25α, and α-tubulin. MT retraction was defined by retraction of MT from the cellular processes to the perinuclear region resulting in an intense tubulin staining surrounding the nucleus. MT retraction was quantified by counting p25α-positive cells with a clear perinuclear localization of MT compared with the total number of p25α-positive cells. In each experiment, 120–200 transfected cells localized in five randomly chosen microscopic fields were examined at 10× magnification.
For investigation of ASK1 inhibitors, OLN AS 7 cells were seeded on pre-coated black Corning Costar 96 well plates (cat no.3904), 4500 cells/well. One day later half the plate was transfected with empty pcDNA3.1(-) or pcDNA3.1(-)-p25α (0.375 μL FuGene 6 to 0.125 μg DNA) in a total of 50 μL per well. Four hours later one of the four ASK1 inhibitors or 2-dimethylamino-4,5,6,7-tetrabromo-1H-benzimidazole (DMAT) as positive control was added in 50 μL volume. Twenty-four hours after transfection the cells were fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.1% Triton X-100 for 20 min and blocked with 1% BSA for 20 min at 20°C. The cells were stained for α-tubulin and p25α by incubation with primary antibodies for 1 h (1 : 1500) in 1% BSA followed by 1 h incubation in Cy2- or Cy3 labeled secondary antibodies (1 : 400) and Hoechst dye (1 : 1000) to visualize nuclei.
The Cellomics Array Scanner™ (Thermo Fisher Scientific Inc., Waltham, MA, USA) was used for automated quantification of MT retraction. Images were acquired automatically at 10× magnification. Transfected cells were identified by the nuclear staining and the p25α immunostaining and cell morphology was scored based on the shape and area of the cytoplasm visualized by the α-tubulin immunostaining. The Bioapplication Morphology Explorer V3 (Thermo Fisher Scientific Inc., Waltham, MA, USA) was used to analyze the morphology of the cells and the bioapplication Cell health profiling V3 was used to count the number of total cells as well as the number of transfected cells. The protocol for the Morphology explorer V3 was set to select cells based on the following criteria and thresholds: Mean average intensity of the α-tubulin immunostaining ‘MeanAvgIntensityCH1 < 140 (this value was set individually for each preparation of plates), Area of the cytoplasm ‘MeanAvgAreaCH1 < 1000, perimeter squared to 4π*area ‘P2A’ < 2.0 and length to width ratio ‘LWR’ < 2.1. For each well, 25 images was analyzed at 10× magnification, corresponding to app. 300 transfected cells. Data points were normalized to OLN-AS7 + p25 on each plate and shown as mean ± SD. Student's t-test was performed using GraphPadPrism software (GraphPad Software Inc., San Diego, CA, USA) to calculate the significant difference of the mean value relative to non-treated OLN-AS7 + p25 expressing cells.
For the assessment of ASK1 effects, OLN-AS7 cells on poly-d-lysine-coated coverslips were transiently co-transfected with pCDNA p25α and pCMV-HA, pCMV HA-DJ-1 wt or pCDNA3 ASK1-HA or transfected only with pCDNA3 ASK1-HA. After 16 h of expression cells were fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.1% Triton X-100 for 30 min, and blocked in 3% BSA solution for 30 min at 20°C. Primary and secondary antibodies were diluted (dilutions were 1 : 500–1 : 1000 for primary and 1 : 1000 for secondary antibodies) in 1% (w/v) BSA in PBS and incubated for 1 h at 20°C or overnight at +4°C in a humidified chamber. Nuclei were counter-stained with 2 μg/mL Hoechst 33342 in PBS for 10 min at 20°C. Coverslips were mounted in Mowiol/DABCO solution (Carl Roth GmbH) onto glass slides. After each step in the staining protocol, coverslips were washed three times in PBS at 20°C. Images were taken with Axioimager microscope equipped with ApoTome Imaging system (Carl Zeiss) using a 25x objective. The images were processed with AxioVision software (Carl Zeiss) and double transfected cells with retracted versus normal MT were scored in more than 100 cells per condition. Cells with abnormal MT bundles detected by intense tubulin staining were scored as positive for MT retraction.
Each experiment was performed independently at least three times unless stated otherwise. Comparisons were analyzed with either Student's t-test or one-way anova as stated in the Figure legends. p-values less than 0.05 were considered statistically significant.
Proline-158 deletion reduces steady-state protein levels of DJ-1
To assess the effects of the novel DJ-1 mutations A179T and P158Δ on cellular steady-state expression, HEK293E cells were transfected with epitope-tagged DJ-1 cDNAs. Steady-state expression levels of [A179T]DJ-1 were no different from [wt]DJ-1, whereas [P158Δ]DJ-1 protein levels were significantly reduced (Fig. 2a). This effect was seen regardless of the nature and position of the epitope tag, both for the relatively high expressing N-terminal myc-tagged and the comparably lower expressing C-terminal V5/HIS-tag, and observed with antibodies against the tag and DJ-1 itself. Thus, the P158Δ mutation appears to reduce DJ-1 protein levels.
Steady-state expression levels were next analyzed in immortalized MEF cells derived from Dj-1−/− mice and Dj-1+/+ littermate controls. As in HEK293E cells, [A179T]DJ-1 was not consistently altered when compared with wt, whereas the steady-state expression levels of [P158Δ]DJ-1 proteins were significantly reduced (Fig. 2b). Semi-quantitative RT-PCR experiments confirmed equal mRNA expression (Fig. 2c), indicating that the P158Δ mutation acts at the protein level.
Proline-158 deletion reduces DJ-1 protein stability
The above experiments suggest that the P158Δ mutation reduces DJ-1 protein stability. To test this directly, DJ-1 protein decay was measured after cycloheximide translation inhibition as well as by pulse-chase experiments. [wt]DJ-1 showed the known slow decay in cycloheximide-treated MEF cells. [A179T]DJ-1 showed similar decay rates under these conditions, whereas [P158Δ] decay was clearly accelerated (Fig. 3a, Figure S1).
Consistent with the cycloheximide experiments, pulse-chases confirmed the long half-life time of [wt]DJ-1 (24.5 ± 5.4 h). By comparison, the aggressive L166P mutation led to the known very short half-life time (1.2 ± 0.3 h, p < 0.001). [A179T]DJ-1 decay was not significantly shorter than that of [wt]DJ-1 (29.6 ± 6.1 h, p > 0.84), whereas the P158Δ mutant showed a significant reduction of DJ-1 protein half-life time (2.1 ± 0.2 h, p < 0.001) (Fig. 3b, c). The p-values were calculated with one-way anova and the errors are standard mean errors. Thus, the P158Δ reduces DJ-1 protein stability, accounting for a loss of steady-state protein levels.
Proline-158 deletion impairs DJ-1 dimerization
As proline-158 is situated near the C-terminal helix-kink-helix motif that is an important interface of the strong DJ-1 dimer (Fig. 1), which is known to be essential for DJ-1 protein stability, DJ-1 dimerization experiments were performed. For this purpose, HEK293E cells were co-transfected with myc- and HA-tagged DJ-1 proteins. Cell lysates were immunoprecipitated with anti-myc and western blots probed for co-immunoprecipitation of the HA-tagged DJ-1. [wt]DJ-1 clearly forms homodimers, and the A179T mutation alters neither DJ-1 homodimerization nor heterodimerization with [wt]DJ-1. In contrast, [P158Δ]DJ-1 forms less heterodimers with [wt]DJ-1 and practically no homodimers (Fig. 4a). This is in agreement with the very recent report of Ramsey and Giasson (2010). In conclusion, the reduced stability of [P158Δ]DJ-1, particularly in the recessive state lacking [wt]DJ-1, may be because of impaired dimerization because of structural perturbations within the critical dimer interface.
Proline-158 deletion alters DJ-1 binding to ASK1
Among the cellular functions of DJ-1, recruitment into the ASK1 signalosome may contribute to cytoprotection (Junn et al. 2005; Waak et al. 2009; Mo et al. 2010). Thus, we transiently co-transfected HEK293E cells with HA-tagged ASK1 and myc-tagged DJ-1 cDNAs. [wt]DJ-1 co-immunoprecipitation with ASK1-HA was clearly enhanced when cells had been exposed to H2O2. The same was seen for the stable [A179T]DJ-1 mutant (Fig. 4b). However, [P158Δ]DJ-1 aberrantly bound to ASK1 even in the absence of H2O2, similar to [M26I]DJ-1 (Fig. 4b) as was reported before (Waak et al. 2009).
To gain more insight into the binding site(s) of DJ-1 variants, we performed mapping experiments using ASK1 fragments. HEK 293E cells were transiently co-transfected with myc-tagged DJ-1 variants and HA-tagged ASK1 FL, ASK1 ΔC (amino acids 1-947) or ASK1 CT (amino acids 945-1374) (Fig. 4c) and immunoprecipitation with anti-HA was performed. As reported before, [wt]DJ-1 co-immunoprecipitation with full-length ASK1 FL was strongly promoted by H2O2 treatment. DJ-1 was also co-immunoprecipitated with ASK1 ΔC, which could be enhanced by treatment with 1 mM H2O2 for 30 min (Fig. 4d, e, Figure S2). This is consistent with N-terminal binding of ASK1 effector proteins, most prominently the redox inhibitor thioredoxin (Saitoh et al. 1998). Surprisingly, under oxidative conditions [wt]DJ-1 also co-immunoprecipitated with ASK1 CT (Fig. 4d, e, Figure S2). This is different from what we previously observed for a longer C-terminal fragment, ASK1 ΔN (amino acids 648-1374), which was apparently less stable and no DJ-1 co-immunoprecipitation was detectable (Waak et al. 2009). [P158Δ]DJ-1 co-immunoprecipated with full-length, ΔC as well as the CT. Interestingly, [P158Δ]DJ-1 binding with full-length ASK1 was less dependent on H2O2 treatment, whereas the binding with both ΔC and CT was slightly enhanced upon H2O2 treatment (Fig. 4d, e, Figure S2). On H2O2 treatment [M26I]DJ-1 was co-immunoprecipitated as was the P158Δ mutant, but in absence of H2O2 the M26I mutant bound less to ASK1 ΔC and CT (Fig. 4d, e, Figure S2). These data suggest that DJ-1 might interact with multiple binding sites within the ASK1 signalosome, differentially affected by cellular redox status and DJ-1 sequence, with potentially modulating roles of ASK1 domains.
To confirm the binding behavior of DJ-1 to ASK1 FL and variants observed in the above co-immunoprecipitation experiments, we determined native protein co-fractionation using gel filtration chromatography (Fig. 5). HEK 293E cell lysates with co-over-expressed ASK1-HA and myc-DJ-1 were applied on a Superdex 200 10/300 column and the elution volumes of the over-expressed proteins were examined by western blotting. The UV-absorption curves from the runs with different samples were comparable (not shown). In the absence of H2O2, ASK1 FL eluted between 7.5 and 10.5 mL, the ΔC variant showed two peaks (one at 8–10 mL and one at 11.5–13.5 mL) and the CT variant eluted unexpected for its size from 7.5 to 10 mL, possibly reflecting some partially unfolded conformation of the ASK1 fragment (Fig. 5). The main portion of [wt]DJ-1 eluted between 13.5 and 15.5 mL. When the cells were co-transfected with ASK1 FL and treated with H2O2 prior to lysis, [wt]DJ-1 was also detected in earlier fractions, reflecting the incorporation of DJ-1 into ASK1 signalosomes or stabilization of DJ-1/ASK1 complexes under oxidative stress. This shift was also seen with ASK1ΔC, but much less with ASK1 CT, suggesting that [wt]DJ-1 is predominantly recruited to N-terminal ASK1 binds sites (Fig. 5).
The [P158Δ]DJ-1 peak eluted from 14.5 to 16 mL and also smeared into earlier fractions under basal conditions, confirming a certain oligomerisation and aggregation propensity of [P158Δ]DJ-1 (Ramsey and Giasson 2010). Beyond this basal smear, and in marked contrast to [wt]DJ-1, [P158Δ]DJ-1 co-fractionated with all tested ASK1 proteins irrespective of the presence or absence of H2O2, consistent with the co-immunoprecipitation results (Fig. 5).
In conclusion, the P158Δ (Macedo et al. 2009) and M26I (Abou-Sleiman et al. 2003) mutations appear to confer aberrancies to DJ-1/ASK1 complex, thereby possibly contributing to alterations of cytoprotective functions in the respective PARK7 patients.
More [P158Δ]DJ-1 is required for the repression of Daxx nuclear export
One of the downstream events during ASK1-mediated apoptosis is the nuclear export of the death protein Daxx, which is regulated by DJ-1 (Junn et al. 2005; Waak et al. 2009; Saeed et al. 2010). Transfection of [wt]DJ-1 was confirmed to suppress nuclear export of Daxx in Dj-1-/- MEF cells (Fig. 6). Although the A179T mutant was as effective as [wt]DJ-1, five times more [P158Δ]DJ-1 cDNA was necessary to see the same effect on Daxx nuclear translocation (Fig. 6b). Taken together, the PD-associated P158Δ mutation destabilizes the DJ-1 dimer protein and thus reduces its anti-apoptotic potential.
Proline-158 deletion abrogates DJ-1 cytoprotection against AS toxicity
AS fibril formation is a hallmark of PD and related brain proteopathies. Thus, we investigated the effects of DJ-1 on AS toxicity. As there is some evidence for an involvement of DJ-1 in the (oligodendro)glial cytoplasmic inclusions of AS that characterize multiple system atrophy (Neumann et al. 2004), we employed the rat oligodendrocyte OLN-AS7 cell culture model that expresses human AS. Similar to the other cell types (see above), [P158Δ]DJ-1 steady-state protein levels were as reduced as those of [M26I]DJ-1 when compared with [wt]DJ-1 transfections (Fig. 7a). Surprisingly, [L166P]DJ-1 levels were not as much reduced in this particular cell model. We then performed an assay were p25α induces MT retraction in the OLN-AS7 cells. The MT retraction has been shown to be dependent on the double over-expression of p25α and AS, whereas non-transfected cells and AS single transfected OLN clones do not show microtubule retraction (Kragh et al. 2009). The MT degeneration induced in the model by co-expression of the AS aggregate promoting protein p25α could be rescued by coexpression with [wt]DJ-1 (Fig. 7b, c). In contrast, the L166P mutation and mutation of the critical redox residue cysteine-106 (Canet-Avilés et al. 2004; Kinumi et al. 2004; Meulener et al. 2006; Waak et al. 2009; Wilson 2011) abolished DJ-1 cytoprotective activity in this model. Interestingly, [P158Δ]DJ-1 also failed to protect OLN-93 cells against AS/p25α toxicity (Fig. 7b, c), similar to [M26I]DJ-1 that appears to confer comparable dysfunctions to the DJ-1 protein.
Finally, we investigated if the MT retraction upon AS plus p25α expression involved ASK1 (Fig. 8). For this OLN-AS7 cells were transiently transfected with p25α and treated with increasing amounts of four different ASK1 inhibitors (see Fig. 8b). The inhibitors did not affect the AS and p25α protein levels (Fig. 8c). The cells were scored for MT retraction (Fig. 8a). Interestingly, these inhibitors could decrease the MT retraction upon p25α expression in the OLN-AS7 cell line in a concentration dependent manner (Fig. 8a). The effects of ASK1 inhibitors were comparable to those of a previously validated inhibitor, DMAT (Kragh et al. 2009). Next, we tested the effects of ASK1 transfection in the MT retraction assay (Fig. 8d). Expression time was decreased to 16 h, since after 24 h of co-expression of the apoptosis inducing protein ASK1 and p25α very few transfected cells could be detected. Significantly, more cells with retracted MT were found when the cells were transfected with ASK1 together with p25α than when transfected with empty vector and p25α (Fig. 8d). ASK1-HA alone caused significantly less MT retraction in the absence of p25α. As before cells expressing DJ-1wt and p25α had significantly less MT retraction compared with the vector control. To directly demonstrate ASK1 activation in the p25α/OLN-AS7 model, we performed an in vitro immunocomplex kinase assay, where the activity of ASK1-HA is measured by 32PO4 incorporation into the pseudo-substrate myelin basic protein (MBP). When ASK1 was co-transfected together with p25α, a stronger MBP phosphorylation signal could be detected than when ASK1 was transfected alone (Figure S3). This indicates that over-expression of p25α induces an activation of ASK1 in the OLN-AS7 cells. However, the result needs to be taken with caution because of fluctuations of experimental conditions. Unfortunately, a triple transfection with ASK1, DJ-1 variants, and p25α turned out to give highly variable results.
Collectively, these data suggest at least an involvement of ASK1 in the OLN-AS7 MT retraction model. ASK1 might be one of the effectors of DJ-1 cytoprotective activity, impaired by disease-associated mutations.
Our investigations of the novel PARK7 P158Δ and A179T mutations (Macedo et al. 2009) show that [P158Δ]DJ-1 is unstable and leads to loss of function, whereas [A179T]DJ-1 does not show any obvious alterations in our assays. The A179 is not conserved (Fig. 1a) and the A179T substitution does not seem to alter the overall DJ-1 structure. The reported [A179T]DJ-1 PARK7 patient (Macedo et al. 2009) carried the mutation in the heterozygous state and did have a family history of PD (affected mother). If this genetic linkage is not coincidental, the introduction of a potentially phosphorylatable threonine residue into the most C-terminal α-helix might confer some dominant pathogenic effects that were not addressed. In contrast, the effects of the P158Δ mutation are more obvious. Point deletion of P158 that participates in the highly conserved β-turn (Fig. 1) possibly connecting and orienting the C-terminal helix-kink-helix domain (Görner et al. 2007) within the DJ-1 dimer causes strong loss of dimerization and consequently reduced protein stability. Our findings for [P158Δ]DJ-1 are in perfect accord with a recent study (Ramsey and Giasson 2010). Such structural impairments because of proline-158 deletion caused loss of DJ-1 functions, evidently contributing to PD pathogenesis in the affected homozygous patient (Macedo et al. 2009).
DJ-1 is an evolutionary ancient cytoprotective protein that appears to act via several anti-oxidative, anti-apoptotic, and anti-aggregative pathways (Kahle et al. 2009). Our data suggest that the P158Δ mutation alters binding of DJ-1 to ASK1 (Fig. 4b), similar to the M26I mutation (Waak et al. 2009). More [P158Δ]DJ-1 cDNA needs to be transfected to suppress the downstream nuclear export of Daxx. Thus, one of the functional impairments of [P158Δ]DJ-1 could be reduced suppression of the apoptosis signal-regulating ASK1-Daxx pathway (Junn et al. 2005; Waak et al. 2009; Im et al. 2010; Mo et al. 2010).
ASK1 is regulated in a redox-dependent manner under physiological and pathological stress conditions by an extremely complex process (Takeda et al. 2008). In resting cells, the N-terminal portion of ASK1 tightly binds its major inhibitory protein, thioredoxin (Saitoh et al. 1998). Upon oxidative activation, thioredoxin dissociates from the inhibitory ASK1 binding site, and stimulatory effector proteins are incorporated into the ASK1 complex, such as tumor necrosis factor receptor-associated factors 2 and 6 (Noguchi et al. 2005; Fujino et al. 2007) and Daxx (Chang et al. 1998). In untreated cells, the low-activity ASK1 signalosome exists as a 1500–2000 kDa complex containing thioredoxin and ASK1 homo-oligomers that are probably held together via the C-terminal coiled-coil domains of ASK1 (Noguchi et al. 2005). Full activation of the large (> 3000 kDa) ASK1 complex involves in addition to the recruitment of effector proteins (Noguchi et al. 2005) further homophilic interactions via the N-terminal coiled-coil domain (Fujino et al. 2007) and/or the formation of homomeric ASK1 disulfide bonds via cysteine residues both in the C- and N-termini (Nadeau et al. 2007). Activated ASK1 signalosomes undergo autophosphorylations and eventually stimulate stress-activated protein kinase cascades (Takeda et al. 2008).
How does DJ-1 fit into this process? In addition to the originally described nuclear sequestration of Daxx, by which DJ-1 could limit the availability of a effector protein for the cytosolic ASK1 signalosome (Junn et al. 2005), wt DJ-1 suppresses ASK1 in oxidatively stressed cells also by direct binding when thioredoxin dissociates from the ASK1 signalosome (Görner et al. 2007). Thereby DJ-1 could substitute (Waak et al. 2009) and/or facilitate the inhibitory function of thioredoxin (Im et al. 2010) and prevent oligomerization of ASK1 (Mo et al. 2010). The most straightforward model would be that DJ-1 binds to the regulatory domains within the ASK1 N-terminus, as DJ-1 failed to co-immunoprecipitate with ASK1 ΔN (Waak et al. 2009). Here, we provide evidence (Figs 4c–e and 5) that DJ-1 indeed co-immunoprecipitated in an oxidation-dependent manner with the N-terminal fragment, ASK1 ΔC, but also with a C-terminal fragment, ASK1 CT, which is shorter (lacking the kinase domain) than the ASK1 ΔN used by Waak et al. (2009). Thus, DJ-1 might bind to multiple sites along the ASK1 molecule. This has been recently observed even for thioredoxin (Nadeau et al. 2009), which apparently shifts binding within the N-terminus and in addition may also interact with C-terminal residues to regulate ASK1 oligomerization. However, DJ-1 interactions with the ASK1 fragments tested here were weaker than those with the full-length ASK1. We cannot rule out the possibility that the deletions of fairly big parts of the ASK1 protein interrupt the structure of the remaining domains and thereby influence the DJ-1 binding. This may be particularly prominent in the case of the P158Δ and M26I mutants that co-immunoprecipitated with full-length ASK1 regardless of the presence or absence of H2O2, but surprisingly showed increased interactions with ASK1 fragments ΔC and CT upon H2O2 treatment (Fig. 4d–e). It is of note that both N-and C-termini of ASK1 contain disulfide bonding cysteine residues as well potential protein–protein interacting coiled-coil domains (the latter differing considerably in size and sequence though), which could be potential interaction domains for wt and mutant DJ-1. The exact interaction modes of DJ-1 molecules within the highly complex ASK1 signalosome and their functional consequences remain to be further elucidated.
As DJ-1 was also implicated with the MSA characteristic (oligodendro)glial cytoplasmatic inclusions (Neumann et al. 2004), we studied the effects of DJ-1 in OLN-AS7 cells stably transfected with AS and co-transfected with p25α. This cell culture model shows morphological impairments because of AS aggregation as evidenced by MT retraction (Kragh et al. 2009). Interestingly, [wt]DJ-1 but not mutants, including [P158Δ]DJ-1, protected in this assay (Fig. 7). This DJ-1 function has not been previously shown and it may be of importance in the regulation of the development of glial cytoplasmatic inclusion in MSA patients. Unfortunately, the genetic studies in MSA patient cohorts have not yet included PARK7 as a candidate gene. In addition, the MT retraction could also be decreased with ASK1 inhibitors and increased by over-expression of ASK1 (Fig. 8). This indicates a role of ASK1 in this cell culture model. DJ-1 has been shown to regulate AS aggregation through its chaperone activity as well as by up-regulating heat-shock protein 70 (Shendelman et al. 2004; Zhou and Freed 2005; Batelli et al. 2008; Liu et al. 2008). Therefore, the question remains open whether the protective DJ-1 effect in p25α-transfected OLN-AS7 cells is mediated by AS chaperoning, oxidative stress scavenging, and/or direct inhibition of ASK1 signaling.
In conclusion, destabilizing DJ-1 mutations, including the novel P158Δ, deprive cells of a multifaceted cytoprotective protein, thus contributing to neurodegenerative pathogenesis.
We thank Atsushi Matsuzawa and Hidenori Ichijo (University of Tokyo, Japan) for the donation of ASK1 and Daxx constructs, Kerstin Reiss and Thilo Stehle for technical advice, and Jens Waak (Nexigen, Köln, Germany) for critically reading the manuscript. This study was supported by the German Genome Network NGFNplus, the Hertie Foundation, and the German Center for Neurodegenerative Diseases. None of the authors declares a conflict of interest.