DNA samples of MJD patients

An overview of the MJD patient cohorts from which DNA samples were obtained and their detailed specifications can be found in Supplementary Table S1.

PCR

The amplification was carried out with standard PCR conditions using a G-Storm thermal cycler (AlphaMetrix Biotech GmbH, Rödermark, Germany). The complete list of employed primers can be found in Supplementary Table S2. The reaction mixtures and thermal cycler programmes for each SNP are provided in Supplementary Table S3 and S4, respectively.

High-resolution melting analysis

Genotyping of selected SNPs in the PRKN gene was performed by high-resolution melting analysis carried out on a LightCycler 480 (Roche Diagnostics, Mannheim, Germany) using the LightCycler 480 High Resolution Melting Master Kit (Roche Diagnostics) following the manufacturer's instructions. For a reliable detection of genotype-specific melting peaks, an unlabelled probe was added. To ensure a sufficient amplification of DNA strands complementary to the probe, an asymmetric PCR with a ratio of reverse primer to forward primer of 10:1 was performed. The complete list of employed primers and probes can be found in Supplementary Table S2. Alternatively, genotyping was performed via TaqMan. The reaction conditions for each SNP can be found in the Supplementary Table S5.

Sanger sequencing

Results from the high-resolution melting analysis were validated by Sanger sequencing. After PCR amplification, samples were purified using the QIAquick PCR Purification Kit (QIAGEN, Hilden, Germany), according to the manufacturer's instructions. Sequencing reactions were carried out on the CEQ 8000 Genetic Analysis System Sequencer (Beckman/AB SCIEX, Krefeld, Germany), following instructions of the GenomeLab Dye Terminator Cycle Sequencing with Quick Start Kit Manual (Beckman Coulter). Supplementary Table S2 provides a comprehensive list of employed sequencing primers.

Expression constructs

For V5-, Xpress-, or GFP-tagged ataxin-3 overexpression, pcDNA3.1 (Thermo Fisher Scientific, Waltham, MA USA) and pEGFP-N2 (Clontech, Mountain View, CA, US) vectors encoding canonical isoform 2 of ataxin-3 (UniProt ID: P54252-2; ataxin-3c) with 15, 77, or 148 glutamines (15Q, 77Q, 148Q) were utilised. Overexpression of parkin (UniProt ID: O60260-1) was achieved using pcDNA3.1 vectors encoding wild-type parkin (parkin WT) and parkin carrying an Val380Leu amino acid exchange (parkin V380L), both N-terminally fused to a 6xMyc tag. The Val380Leu exchange was achieved by site-directed mutagenesis using forward primer GGAGTGCAGTGCCCTATTTGAAGCCTC and reverse primer GAGGCTTCAAATAGGGCACTGCACTCC. Correct integration of mutations was confirmed by Sanger sequencing using forward primer CTGCCGGGAATGTAAAGAAG. For Tet-off system-based expression of parkin WT or V380L, the respective cDNA was cloned into a pTRE responder vector, and used in combination with a pTET-RCA2 plasmid as described earlier [55]. For microscopically visualising mitochondria, a mammalian expression vector coding for the fluorescent protein DsRed2 and fused to a mitochondrial targeting signal (pDsRed2-Mito; Takara Bio USA, Inc., San Jose CA, US) was employed.

Cell culture

For cell culture experiments, HEK293T wild-type (293T WT) cells (ATCC: CRL-11268) and HEK293T ATXN3 knockout (293T ATXN3−/−) cells [55] were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% foetal bovine serum (FBS), 1% non-essential amino acids (MEM NEAA), and 1% Antibiotic–Antimycotic (A/A) (all Gibco, Thermo Fisher Scientific) in 5% CO2 at 37 °C. Transient cell transfection was conducted for 72 h using Attractene (QIAGEN) or Turbofectin 8.0 (OriGene Technologies, Inc., Rockville, MD, US) reagents according to the manufacturers’ protocols, with an approximate transfection efficiency of 50% transfected cells as assessed using a representative pEGFP-N2 reporter construct and fluorescence microscopy. For depolarization of mitochondria, cells were treated with various concentrations of carbonyl cyanide m-chlorophenylhydrazone (CCCP; Merck, Darmstadt, Germany) for 24 h. For inhibition of proteasomal degradation, cells were incubated with 2.5 µM MG132 (Merck) or 0.5 µM epoxomicin (Selleck Chemicals, Cologne, Germany) for 24 h. For inhibiting autophagy, 25 nM bafilomycin A1 (Selleck Chemicals) was administered. Dimethyl sulphoxide (DMSO) was used as vehicle control.

Protein stability analysis

Analysis of protein stability was performed using a Tet-off system as previously described [55]. In brief, 293T ATXN3−/− cells were transfected with pTRE-parkin WT or V380L responder constructs and a pTET-RCA2 vector in a 1:1 ratio, and, if desired, additionally in combination with a pcDNA3.1 Xpress-Atx3 148Q plasmid. Expression was terminated by the addition of doxycycline (Merck; 4.5 µM) at desired time points before cell harvest.

Cell viability assay

Assessment of cell viability was performed as previously described [36] with the following specifications: 5,000 cells/well were seeded in 96-well cell culture plates (ViewPlate-96 Black, Perkin Elmer, Massachusetts, USA) and transfected 24 h later using Turbofectin 8.0 for 72 h. Cells were treated with 12.5 µM CCCP or vehicle control DMSO for the last 24 h. Afterwards, culture medium was aspirated, cells incubated in fresh medium containing the resazurin-based PrestoBlue™ Cell Viability Reagent (Thermo Fisher Scientific) in a 1:10 ratio under standard culture conditions for 60 min. Fluorescence signals were measured at 535 nm (excitation)/615 nm (emission) using a Synergy HT plate reader and the Gen5 software (both BioTek Instruments, Winooski, VT, USA).

Fluorescence-activated cell sorting (FACS) analysis

Cell death was measured using fluorescence-activated cell sorting (FACS) analysis as previously described [41]. Briefly, transfected and treated 293T WT cells were trypsinised, pelleted, washed with 1× DBPS and stained with 2.5% (v/v) of 7-aminoactinomycin D (7-AAD) (BD Biosciences, San Jose, CA, USA) in CliniMACS® PBS/EDTA buffer (Miltenyi Biotec, Bergisch Gladbach, Germany) at room temperature for 3–5 min. Afterwards, cells were analysed using a BD LSRFortessa™ X-20 cytofluorometer with the BD FACSDiva™ v9.0 software (both BD Biosciences). The flow cytometry results were further analysed using FlowJo™ v10.10 Software (FlowJo LLC, BD Life Sciences, Ashland, OR, USA).

Immunofluorescence staining and microscopy

Immunofluorescence staining of cells was performed as previously described [53]. In brief, 5000 293T cells per well were seeded on an 8-well chamber slide (80841, Ibidi, Gräfelfing, Germany), treated as desired, and subjected to fixation in 4% (w/v) paraformaldehyde in 1× Dulbecco’s phosphate-buffered saline (DPBS), followed by a 1-h permeabilization and fixation step in 10% (w/v) bovine serum albumin, 0.5% (v/v) Triton X-100, and 0.02% (w/v) NaN3 in 1× DPBS. Cells were incubated with primary antibodies mouse anti-Myc-tag (1:500; clone 9B11, #3739, Cell Signaling, Danvers, MA, US) or goat mouse anti-parkin (1:1,000; clone PRK8, MAB5512, Merck) at 4 °C overnight. The next day, cells were washed and incubated with goat anti-mouse IgG Alexa Fluor 555 (1:500; A-21424, Thermo Fisher Scientific) or goat anti-mouse IgG Alexa Fluor Plus 594 (1:500; A32742, Thermo Fisher Scientific) secondary antibodies at room temperature for 1 h. After washing, cells were mounted with VECTASHIELD® Antifade Mounting Medium with DAPI (H-1200, Vector Laboratories, Newark, CA, US) using coverslips and sealed with transparent nail polish.

Epi-fluorescence images were taken at a 200×, 400×, and 630× magnification on an Axioplan 2 imaging microscope equipped with an ApoTome, Plan-Neofluar 20 × /0.50, Plan-Neofluar 40×/0.75, and Plan-Neofluar 63×/1.4 Oil objectives, and an AxioCam MRm camera, using the AxioVision 4.3 imaging software (all Zeiss, Oberkochen, Germany).

Confocal fluorescence images were acquired at a 1000× magnification on an ECLIPSE Ti2 microscope equipped with an CSU-W1 SoRa confocal system and an CFI SR HP Plan Apochromat Lambda S 100 × C Sil objective, using NIS-Elements AR 5.42 (all Nikon, Tokio, Japan).

Morphological assessment of GFP-Atx3 148Q aggregates was performed using the Fiji software [40], and an in-house established macro for semi-automated analysis to measure cross-sectional area and roundness of detected GFP-positive particles. Based on an equivalent circular area (ECA) to the assessed aggregate area (Aag), a corresponding aggregate diameter was extrapolated (dECA) [21], using the following equation: \(_} =\sqrt_}}}.\)

Protein extraction

For protein extraction, 293T cells were dissociated by gentle pipetting and transferred to 2.0 mL tubes. Cell pellets were obtained by centrifugation at 500×g for 5 min followed by aspiration of the supernatant. Pellets were washed once with cold 1× DPBS. Homogenization was conducted by resuspending the cell pellet in RIPA buffer (50 mM Tris pH 7.5, 150 mM NaCl, 0.1% SDS, 0.5% sodium deoxycholate and 1% Triton X-100) containing cOmplete™ protease inhibitor cocktail (Roche Diagnostics) and ultrasonication using a Sonopuls ultrasonic homogenizer (Bandelin electronic, Berlin, Germany) for 3 s and 10% pulse duration at 10% power. Homogenates were mixed with glycerol to a final concentration of 10%. For cell lysate preparation, protein extracts were incubated for 15 min on ice followed by a 15-min centrifugation at 4 °C and 16,100×g. Supernatants were transferred to a fresh pre-cooled tube with addition of 10% glycerol. Protein concentrations were measured spectrophotometrically in a microtiter plate using Bradford reagent (Bio-Rad Laboratories, Basel, Switzerland). Samples were stored at -80 °C until further analysis.

Immunoprecipitation

GFP-tagged proteins were immunoprecipitated using GFP-Trap agarose according to the manufacturer’s protocol (ChromoTek and Proteintech Germany, Planegg-Martinsried, Germany) with the following specifications: For immunoprecipitation (IP), 1000 µg of total protein was incubated with 20 µL of agarose bead slurry in a 1.5 mL reaction tube, rotating end-over-end at 4 °C and 10 rpm for 1 h. After IP, samples were eluted in 80 µL of 4× LDS sample buffer (1 M Tris pH 8.5, 50% (v/v) glycerol, 8% (w/v) LDS, 2 mM EDTA, 0.1% (w/v) Orange G) mixed with Trap dilution buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 0.5 mM EDTA) in a ratio 1:1, supplemented with 100 mM dithiothreitol (DTT) and heat-denatured at 70 °C and 600 rpm for 10 min. Samples were subsequently analysed by western blotting.

Western blotting

Western blotting was performed according to standard procedures. Briefly, protein extracts were mixed with a 4× LDS sample buffer in a ratio 3:1 and supplemented with 100 mM DTT. After heat-denaturing for 10 min at 70 °C, protein samples were electrophoretically separated using custom-made Bis–Tris gels and MOPS electrophoresis buffer (50 mM MOPS, 50 mM Tris pH 7.7, 0.1% SDS, 1 mM EDTA). Proteins were transferred on Amersham™ Protran™ Premium 0.2 µm nitrocellulose membranes (Cytiva, Freiburg, Germany) using Bicine/Bis–Tris transfer buffer (25 mM Bicine, 25 mM Bis–Tris pH 7.2, 1 mM EDTA, 15% methanol) and a TE22 Transfer Tank (Hoefer, Inc., Holliston, MA, US) at 80 V and a maximum of 250 mA for 2 h.

After transfer, membranes were blocked for 45 min with 5% skim milk powder (Merck) in 1× TBS (10 mM Tris, pH 7.5, 150 mM NaCl) at room temperature, and probed overnight at 4 °C with primary antibodies diluted in TBS-T (TBS with 0.1% Tween 20). A detailed listing of applied primary antibodies can be found in Supplementary Table S6. Afterwards, membranes were washed with 1 × TBS-T and incubated at room temperature for 1 h with the respective secondary IRDye® antibodies goat anti-mouse 680LT (P/N 926-68020), goat anti-mouse 800CW (P/N 926-32210), and goat anti-rabbit 800CW (P/N 926-32211) (all 1:10,000; LI-COR Biosciences, Lincoln, NE, US). After final washing with 1× TBS-T, fluorescence signals were detected using the LI-COR ODYSSEY® FC and quantified with Image Studio 4.0 software (both LI-COR Biosciences).

Filter retardation assay

Detection of SDS-insoluble ataxin-3 was performed as previously described [54]. Briefly, 1 µg of cell homogenates was diluted in 1× DPBS containing 2% SDS and 50 mM DTT. Afterwards, samples were heat-denatured at 95 °C for 5 min and filtered through an Amersham™ Protran™ 0.45 µm nitrocellulose membrane (Cytiva) using a Minifold® II Slot Blot System (Schleicher & Schuell, Dassel, Germany). After blocking in 5% skim milk powder (Merck) in 1× TBS, membranes were incubated with respective primary and secondary antibodies. Fluorescence signals were detected using the LI-COR ODYSSEY® FC and quantified with Image Studio 4.0 software (both LI-COR Biosciences).

Statistical analysis

The statistical analysis to test the influence of rs1801582 in PRKN on the AAO of MJD patients was performed using the JMP software (JMP, Cary, NC, US; version 16.2.0). A P-value of ≤ 0.05 was considered statistically significant. The Hardy–Weinberg distribution of the polymorphism’s allele frequencies was verified using a chi-square (χ2) test. Non-evaluable samples were excluded from the initial cohort of MJD patients after the following criteria: no reliable genotyping result or missing values for CAG repeats or AAO. To exclude potential differences in the geographical distribution of rs1801582 genotypes in the three analysed MJD patient sub-cohorts, a Fisher–Freeman–Halton exact test was applied. To account for related samples, a family factor was applied. In the statistical analysis, the weighting of each family member was divided by the total number of family members. Thereby, each whole family was considered as one independent sample only. Testing for statistical significance between the distribution of CAG repeat length and AAO for each genotype was performed using a two-tailed Kruskal–Wallis test, followed by a Steel–Dwass all pairs analysis. To analyse the effect of the polymorphism in consideration with the already known effect of the CAG repeat length on the AAO, a multivariate linear regression analysis model was used with AAO as dependent and polymorphism and expanded CAG repeats as independent variables. The combined effect of CAG repeat length and polymorphism in additive and interactive models were compared with the effect of the CAG repeat length alone, using the coefficient of determination (r2) and the effect test function. The mean AAO for each genotype was adjusted to the mean CAG repeat length in the expanded allele and compared between the groups using the least square means adjustment tool of JMP. Linear equations for the AAO in dependence on the CAG repeat length were established for each genotype in prediction models, allowing a more accurate prediction of the AAO. Violin plots of respective datasets were generated with GraphPad Prism 10.1.1 (GraphPad Software, Dotmatics, Boston, MA, USA).

Data from the functional analysis were statistically analysed using GraphPad Prism 10.1.1 (GraphPad Software). The results are presented as bar charts with bars representing mean + standard error of the mean (s.e.m.) or violin plots featuring medians as well as the 25% and 75% percentiles, respectively. Statistical outliers were determined using the ROUT method with a default Q = 1%. One-sample t-test, Student's t-test, or one-way ANOVA with the respective post hoc analysis was applied. Significance was assumed with a P-value ≤ 0.05. For further details, see the respective figure legends.