The Machado-Joseph disease deubiquitylase ataxin-3 interacts with LC3C/GABARAP and promotes autophagy.

The pathology of spinocerebellar ataxia type 3, also known as Machado-Joseph disease, is triggered by aggregation of toxic ataxin-3 (ATXN3) variants containing expanded polyglutamine repeats. The physiological role of this deubiquitylase, however, remains largely unclear. Our recent work showed that ATX-3, the nematode orthologue of ATXN3, together with the ubiquitin-directed segregase CDC-48, regulates longevity in Caenorhabditis elegans. Here, we demonstrate that the long-lived cdc-48.1; atx-3 double mutant displays reduced viability under prolonged starvation conditions that can be attributed to the loss of catalytically active ATX-3. Reducing the levels of the autophagy protein BEC-1 sensitized worms to the effect of ATX-3 deficiency, suggesting a role of ATX-3 in autophagy. In support of this conclusion, the depletion of ATXN3 in human cells caused a reduction in autophagosomal degradation of proteins. Surprisingly, reduced degradation in ATXN3-depleted cells coincided with an increase in the number of autophagosomes while levels of lipidated LC3 remained unaffected. We identified two conserved LIR domains in the catalytic Josephin domain of ATXN3 that directly interacted with the autophagy adaptors LC3C and GABARAP in vitro. While ATXN3 localized to early autophagosomes, it was not subject to lysosomal degradation, suggesting a transient regulatory interaction early in the autophagic pathway. We propose that the deubiquitylase ATX-3/ATXN3 stimulates autophagic degradation by preventing superfluous initiation of autophagosomes, thereby promoting an efficient autophagic flux important to survive starvation.

Supporting Information Table S1: Statistical analysis for Figure 1A Average -Average survival for the designated time point (%) SEM -Standard error of the mean n = 5 * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 Supporting Information Table S3: Statistical analysis for Figure 1C Average -Average survival for the designated time point (%) SEM -Standard error of the mean n = 8 * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 Supporting Information Table S4: Statistical analysis for Figure 1D Average -Average survival for the designated time point (%) SEM -Standard error of the mean n = 8 * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 Supporting Information Table S5: Statistical analysis for Figure S1A Average -Average survival for the designated time point (%) SEM -Standard error of the mean n = 8 * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001 Supporting Information Table S6: List of primers used for cloning the indicated constructs.

Supporting Information Experimental Procedures
Caenorhabditis elegans maintenance and transgenic lines The following strains were used in this study: Bristol (N2) strain as wild-type ( Center (CGC). Mutant strains were outcrossed at least six times against the wild-type strain to provide isogenic conditions. Nematodes were grown at 20°C on nematode growth medium (NGM) agar plates seeded with Escherichia coli (E. coli) OP50 as food source (Brenner, 1974).

RNA Interference in worms
RNAi in C. elegans was performed using the RNAi feeding method (Timmons & Fire, 1998).
In L1 survival assays, the maternal worm population was placed on HT115 E. coli bacteria expressing bec-1-or unc-51-specific double-stranded RNA (dsRNA) at L4 stage to obtain partial reduction of bec-1 and unc-51 expression in the progeny. After 24 hours of incubation worms were bleached with alkaline bleach solution, eggs were hatched at 20ºC overnight and L1 synchronized larvae starved as described for the starvation assay. As a control, bacteria transformed with the empty pPD129.36 vector (Fire Lab Vector Kit) were used as food source.

SDS-soluble aggregates
Worms expressing the aggregation-prone Q40::YFP protein were grown at 20ºC on NGM plates seeded with E. coli (OP50) until they reached adulthood (72h after L1 stage). Animals were washed off the plates with M9 buffer, settled on ice and washed in PBS containing protease inhibitor (Roche). Worms were lysed on ice using a Precellys homogenizer (2 x 20s 5,500 rpm) and sonicated (4 x 10s on ice; 60% power; Sonopuls UW 2200, Bandelin). Lysates were centrifuged at 4ºC for 15min at 15500 x g. Supernatant was kept as soluble (S) fraction.
The pellet was washed once in PBS-lysis buffer and dissolved in SDS-lysis buffer (187.5 mM TrisHCl, 6% SDS, 30% sucrose) by sonication (4 x 10s (60%)). Samples were boiled in the presence of 50 mM DTT, followed by centrifugation at RT for 30 to 60min at 14,000 rpm.

C. elegans immunoblotting and antibodies
Synchronized L1 larvae were grown on NGM-agar plates seeded with E. coli (OP50) until young adult (60h after L1 stage) day 1 adult stage (72h after L1 stage). Alternatively, chloroquine treated worms have been grown on NGM-agar plates seeded with E. coli (OP50) until L4 stage (48h after L1 stage), when they were moved on chloroquine-containing (concentration indicated in figure) NGM plates seeded with OP50 with for 24h prior to sample collection. Worm lysates used for SDS-PAGE were prepared by washing worms from NGMagar plates followed by multiple washing step with M9 buffer, until bacteria were removed.
Worm samples were taken up in SDS sample buffer followed by heating to 95°C for 5min and subsequently subjected to sonication (2x 30s, on ice; 60% power; Sonopuls UW 2200, Bandelin). Lysates were cleared by centrifugation at 8000 × g for 5min, and the supernatant was run on a 4-12% electrophoresis polyacrylamide gel containing sodium dodecyl sulfate

RNA interference
The following siRNA oligonucleotides were used at a final concentration of 30 nM in this study: LDS sample buffer containing 2 x NuPAGE reducing agent, and analysed by SDS-PAGE and immunoblotting with GST and HIS antibody using a Licor Odyssey imager.

Cleavage of preGABARAP-GFP
Recombinant preGABARAP-GFP (250 ng) was incubated with 0.1 μM or 1 μM recombinant ataxin-3 or 0.1 μM recombinant Atg4B or in the absence of protease in DUB buffer (50 mM HEPES pH 8, 0.5 mM EDTA, 1 mM DTT, 0.1 μg/ml BSA) for 16h. The reaction was stopped by addition of 2 x LDS sample buffer containing 2 x NuPAGE reducing agent, samples were boiled at 95˚C and analysed by SDS-PAGE and immunoblotting.

Statistical analysis
Statistical analysis was performed using GraphPad Prism 6. Gaussian distribution of data was tested using the Shapiro-Wilk normality test. Data with normal distribution are presented as mean ± standard deviation (SD), unless stated differently, and were analysed using unpaired Student's t-test (to compare two groups) or one-way ANOVA (to compare multiple groups) followed by Bonferroni to either compare all pairs or selected pairs. Non-normally distributed data are shown as box plots with median and 5-95 percentile and were analysed using the nonparametric Kruskal-Wallis test followed by Dunn's to compare multiple groups or selected pairs. Data shown are generated from at least three independent experiments (if not stated differently in figure legends). For data obtained with worms, mean values ± SEM are shown, unless otherwise indicated. Statistical comparisons between groups were done using two-tailed Student's t test. Quantification of immunoblots was done with ImageJ 1.48v or Image Studio 4.0 software. Data are presented as fold change compared to the control. The following p-values were considered significant: * p ≤ 0.05, ** p ≤ 0.01, *** p ≤ 0.001. Statistical significance was assessed by two-tailed paired student's t-test using Excel software (Microsoft) or GraphPad Prism 6.