Parkin drives pS65‐Ub turnover independently of canonical autophagy in Drosophila

Abstract

Parkinson's disease‐related proteins, PINK1 and Parkin, act in a common pathway to maintain mitochondrial quality control. While the PINK1‐Parkin pathway can promote autophagic mitochondrial turnover (mitophagy) following mitochondrial toxification in cell culture, alternative quality control pathways are suggested. To analyse the mechanisms by which the PINK1–Parkin pathway operates in vivo, we developed methods to detect Ser65‐phosphorylated ubiquitin (pS65‐Ub) in Drosophila. Exposure to the oxidant paraquat led to robust, Pink1‐dependent pS65‐Ub production, while pS65‐Ub accumulates in unstimulated parkin‐null flies, consistent with blocked degradation. Additionally, we show that pS65‐Ub specifically accumulates on disrupted mitochondria in vivo. Depletion of the core autophagy proteins Atg1, Atg5 and Atg8a did not cause pS65‐Ub accumulation to the same extent as loss of parkin, and overexpression of parkin promoted turnover of both basal and paraquat‐induced pS65‐Ub in an Atg5‐null background. Thus, we have established that pS65‐Ub immunodetection can be used to analyse Pink1‐Parkin function in vivo as an alternative to reporter constructs. Moreover, our findings suggest that the Pink1‐Parkin pathway can promote mitochondrial turnover independently of canonical autophagy in vivo.

Pink1‐mediated mitochondrial phospho‐ubiquitination (pS65‐Ub) is activated by oxidative stress in vivo, selectively accumulates on disrupted mitochondria, and is degraded via the action of parkin independently of canonical autophagy genes.

Introduction

Parkinson's disease (PD) is the second most common neurodegenerative disease, with the global burden of disease having more than doubled between 1990 and 2016 (GBD, ²⁰¹⁸). Autosomal recessive mutations in the genes encoding the mitochondria‐targeted kinase PINK1 and the E3 ubiquitin (Ub) ligase Parkin are associated with parkinsonism (Kitada et al, ¹⁹⁹⁸; Valente et al, ²⁰⁰⁴). Loss of either homologue in Drosophila (designated as Pink1 and parkin, respectively) results in strikingly similar phenotypes of severe mitochondrial dysfunction and degeneration of the indirect flight muscles, as well as the degeneration of a subset of dopaminergic neurons, thus mimicking a key hallmark of PD (Greene et al, ²⁰⁰³; Whitworth et al, ²⁰⁰⁵; Clark et al, ²⁰⁰⁶; Park et al, ²⁰⁰⁶). Genetic interaction studies subsequently placed Pink1 and parkin in a common pathway, with parkin downstream of Pink1 (Clark et al, ²⁰⁰⁶; Park et al, ²⁰⁰⁶).

PINK1 phosphorylates both Ub and Parkin, each at their respective Ser65 residues (Kondapalli et al, ²⁰¹²; Shiba‐Fukushima et al, ²⁰¹²; Kane et al, ²⁰¹⁴; Kazlauskaite et al, ²⁰¹⁴; Koyano et al, ²⁰¹⁴). PINK1 is partially imported into healthy, polarised mitochondria via its N‐terminal mitochondrial targeting sequence where it is cleaved and subsequently degraded in the cytosol by the N‐end rule pathway (Yamano & Youle, ²⁰¹³). PINK1 is activated upon stalling on the outer mitochondrial membrane (OMM), where it phosphorylates Ub (pS65‐Ub) that is conjugated at low abundance to OMM proteins (Okatsu et al, ^{2015a, 2015b}). Parkin resides in the cytosol in an autoinhibited state and is recruited to mitochondria by binding pS65‐Ub (Okatsu et al, ^2015b). pS65‐Ub binding partially displaces Parkin's Ubl domain, which allows it to be phosphorylated by PINK1 (Gladkova et al, ²⁰¹⁸). This second phosphorylation event results in a dramatic domain rearrangement that relieves Parkin's autoinhibitory contacts and allows it to ubiquitinate proteins in close proximity (Gladkova et al, ²⁰¹⁸; Sauvé et al, ²⁰¹⁸). The Ub conjugated by Parkin provides the further substrate for phosphorylation by PINK1, which in turn promotes further Parkin recruitment, thus constituting a feed‐forward mechanism of mitochondrial pS65‐ubiquitination (Ordureau et al, ²⁰¹⁴). Both the structure of active Parkin and cell‐based studies suggest that Parkin has low substrate selectivity (Gladkova et al, ²⁰¹⁸; Koyano et al, ²⁰¹⁹), and it has been found to predominantly produce K6, K11, K48 and K63 chains in vitro (Ordureau et al, ²⁰¹⁴).

Much of our understanding of the function of PINK1 and Parkin have been derived from utilising chemical depolarisation of mitochondria in cultured cells in conjunction with Parkin overexpression (Narendra et al, ²⁰⁰⁸; Vives‐Bauza et al, ²⁰¹⁰). These experiments established a paradigm for studying the PINK1‐Parkin pathway; upon depolarisation, PINK1‐ and Parkin‐mediated ubiquitination of OMM proteins leads to the recruitment of the Ub‐binding mitophagy receptors OPTN and NDP52 (Lazarou et al, ²⁰¹⁵), which in turn promote autophagosome initiation (Boyle et al, ²⁰¹⁹; Yamano et al, ²⁰²⁰), ultimately leading to degradation of the damaged mitochondria via the autophagy system. However, studies in animal models have provided mixed results as to the contribution of PINK1 and Parkin to mitophagy. While proteomic analysis of mitochondrial protein turnover in Drosophila supports a role for Pink1/parkin in promoting the degradation of mitochondrial membrane components (Vincow et al, ²⁰¹³), imaging analysis of pH‐sensitive fluorescent mitophagy reporters has yielded conflicting results (Cornelissen et al, ²⁰¹⁸; Lee et al, ²⁰¹⁸; McWilliams et al, ²⁰¹⁸; Kim et al, ²⁰¹⁹; Liu et al, ²⁰²¹). It has also been shown in cell culture models that treatment with antimycin A or expression of an aggregation‐prone matrix protein, ΔOTC, which does not abolish mitochondrial membrane potential, led to the production of mitochondria‐derived vesicles (MDVs) in a PINK1‐ and Parkin‐dependent manner (McLelland et al, ²⁰¹⁴; Burman et al, ²⁰¹⁷). Other studies have focussed on the role of PINK1 and Parkin in mitochondrial biogenesis, protein import and in the regulation of the fission and fusion machinery (Poole et al, ²⁰⁰⁸; Stevens et al, ²⁰¹⁵; Jacoupy et al, ²⁰¹⁹). However, many questions remain about the mechanisms of PINK1‐Parkin‐mediated mitochondrial quality control in vivo.

We sought to illuminate the physiological mechanisms of the PINK1‐Parkin pathway by monitoring pS65‐Ub levels as a direct measure of PINK1 activity. We developed complementary mass spectrometry, immunoblotting and immunostaining methods to detect pS65‐Ub using Drosophila as a model system (hence, we adopt the FlyBase nomenclature “Pink1” and “parkin”). We confirm that pS65‐Ub is produced by Drosophila Pink1 in vivo and can therefore be utilised to follow activation of the Pink1‐parkin pathway and the downstream mechanisms of mitochondrial turnover. We identify exposure to the oxidant and parkinsonian toxin paraquat as a potent activator of Pink1 and establish this approach as a new paradigm to study the Pink1‐parkin pathway in vivo.

Results

Development of methods to detect pS65‐Ub in vivo

To understand the role of the Pink1‐parkin pathway in maintaining mitochondrial quality control in vivo, we developed a mass spectrometry approach to detect low‐abundance pS65‐Ub. Adapting a recently described sample preparation pipeline (Swatek et al, ²⁰¹⁹), we first determined the absolute abundance of total Ub and pS65‐Ub in mitochondrial extracts from young (2–3 days) and aged (50–60 days) flies from a wild‐type background (w ¹¹¹⁸). In young flies, Ub was present on mitochondria, but pS65‐Ub was not reliably detected with this method (Fig 1A, young). By contrast, aged flies displayed elevated total mitochondrial Ub, and pS65‐Ub was robustly detected (Fig 1A, aged). To gain a clearer insight into the basal levels of pS65‐Ub in young wild‐type animals, we adjusted our protocol in order to optimise pS65‐Ub detection (see Materials and Methods). Consequently, we were indeed able to detect pS65‐Ub in mitochondrial fractions from young flies (Fig 1B), thus confirming that pS65‐Ub is present in young flies, albeit in very low abundance.

Figure 1. Detection of pS65‐Ub in vivo .

1. Total Ub (left) and normalised pS65‐Ub abundance (right) in 500 μg Ub‐Clippase‐treated, TUBE‐enriched mitochondrial fractions from young (2–3 days) and aged (50–60 days) wild‐type flies, an mtDNA mutator model (daG4 > UAS‐mito‐APOBEC1), and wild‐type flies that had been exposed to paraquat (5 mM) for 3 days.
2. Absolute abundance of pS65‐Ub in 4 mg mitochondria from young (2–3 days) wild‐type flies following sodium carbonate extraction, protease treatment and phospho‐peptide enrichment.
3. pS65‐Ub (left) and total Ub (right) immunoblots of mitochondria‐enriched fractions from wild‐type flies treated for 3 days with either paraquat or vehicle (sucrose).
4. Immunoblots for the indicated antibodies following subcellular fractionation of flies treated with paraquat for 3 days. T, total lysate; N, nuclear‐enriched fraction; M, mitochondria‐enriched fraction; S, postmitochondrial supernatant.
5. Immunoblots for the indicated antibodies of (T) total lysates or (M) mitochondria‐enriched fractions from fly heads (with proboscises removed) and bodies (thoraces and abdomens).
6. Representative image of flight muscle from aged (50‐day‐old) wild‐type animals immunostained with the indicated antibodies. Scale bars = 20 μm (inset, 5 μm).

Data information: Charts show mean ± SEM, n = 3–4 independent biological replicates as shown.

We next sought to identify naturalistic stimuli of the Pink1‐parkin pathway in vivo. We found that mitochondrial stress induced by an mtDNA mutator model (daG4 > mito‐APOBEC1; Andreazza et al, ²⁰¹⁹) was able to modestly induce pS65‐Ub in young flies (Fig 1A, mutator). By contrast, 3‐day exposure to paraquat, an oxidant that has been epidemiologically linked to PD (Tanner et al, ²⁰¹¹) and a potent inducer of reactive oxygen species (Filograna et al, ²⁰¹⁶), led to a robust increase in both total and pS65‐Ub in mitochondrial fractions in young flies (Fig 1A, paraquat). In paraquat‐treated flies, pS65‐Ub comprised approximately 6% of the total mitochondrial Ub, which is less than the 10–30% Ub phosphorylation that has been observed using similar methods in depolarised cells (Ordureau et al, ²⁰¹⁴; Swatek et al, ²⁰¹⁹).

As an orthogonal validation of our mass spectrometry results, we evaluated immunodetection methods using an antibody recently characterised to specifically detect pS65‐Ub (Watzlawik et al, ²⁰²¹). Immunoblotting confirmed the robust induction of pS65‐Ub and total mitochondrial Ub upon paraquat treatment, while pS65‐Ub was not detected in response to amino acid starvation from a sucrose‐only diet (Fig 1C). The paraquat‐induced pS65‐Ub co‐enriched with mitochondria in crude subcellular fractions (Fig 1D). Interestingly, pS65‐Ub levels appeared to be similar between heads (with proboscises removed to eliminate muscle contamination) and bodies (thorax and abdomens) with a possible enrichment in heads (Fig 1E). We also determined that after removal of paraquat there was a progressive reduction in pS65‐Ub levels, presumably due to mitochondrial turnover (Fig EV1A).

Figure EV1. Dynamics of pS65‐Ub production.

1. Total protein and immunoblots for the indicated antibodies of mitochondria‐enriched fractions from wild‐type flies treated for the indicated number of days with paraquat. Recovery, return to normal food.
2. Immunostaining of flight muscles of Pink1 ⁻ and wild‐type flies. Signal acquisition, brightness and contrast settings for pS65‐Ub are identical to those presented in Fig 3D.
3. pS65‐Ub immunoblot of mitochondrial fractions from wild‐type and Pink1 ⁻ flies following either no treatment (−) or after treatment for 3 days with sucrose (S) or paraquat (PQ).

Flight muscles are a major tissue affected by the loss of Pink1 and parkin and so are a key site of Pink1/parkin activity. Immunofluorescence microscopy of the flight muscles of untoxified, aged flies revealed low but consistent detection of mitochondria (ATP5A‐positive) that were enveloped in pS65‐Ub (Fig 1F), while pS65‐Ub‐positive structures were rarely observed in young flies (Fig EV1B). These results suggest that the Pink1‐parkin pathway is basally active in Drosophila, but that pS65‐Ub levels are kept very low in young animals, likely due to infrequent and transient induction and/or efficient turnover.

pS65‐Ub production in response to paraquat requires Pink1 but not parkin

We next sought to confirm the requirements for Pink1 and parkin in the ubiquitination of mitochondria under basal and paraquat‐induced conditions. To this end, we determined the abundance of total Ub and pS65‐Ub in mitochondrial extracts from wild‐type, Pink1 ⁻ (Pink1 ^B9) and park ^−/− (park ²⁵) mutant flies by mass spectrometry. Under basal conditions, loss of Pink1 resulted in elevated total Ub levels that did not further increase upon exposure to paraquat (Fig 2A). Importantly, pS65‐Ub was not detectable above background in Pink1 ⁻ flies even upon exposure to paraquat (Figs 2B and EV1C), confirming the conserved and essential role of Pink1 in the phosphorylation of Ub at S65 in Drosophila. park ^−/− flies displayed modestly elevated total mitochondrial Ub that did not significantly increase further in response to paraquat (Fig 2A). By contrast, the increase in pS65‐Ub levels observed upon exposure to paraquat was largely unaffected by the loss of parkin (Fig 2B).

Figure 2. Analysis of the paraquat‐induced mitochondrial ubiquitome from Pink1 ⁻ and park ^−/− flies.

• A, B
  (A) total Ub and (B) normalised pS65‐Ub abundance in 500 μg TUBE‐enriched, protease‐treated mitochondrial fractions from wild‐type, Pink1 ⁻ and park ^−/− flies, either untreated young flies (2–3 days) or exposed to paraquat for 3 days.
• C
  pS65‐Ub immunoblot of mitochondria‐enriched fractions following the paraquat pulse‐chase assay in wild‐type and park ^−/− flies. UT, untreated; PQ, 1‐day paraquat treatment; Recovery, flies removed from paraquat and returned to normal food. Asterisk (*) denotes nonspecific band. See Fig EV1C for related control blots.
• D
  pS65‐Ub lane densitometry analysis of n = 3 independent replicates from (C), expressed relative to the most intense lane signal in each blot. Charts show mean ± SEM.
• E–H
  Relative abundance of (E) K6 chains, (F) K11 chains, (G) K48 chains and (H) K63 chains in mitochondrial fractions treated as in A, normalised to the total mitochondrial Ub in each sample. Charts show mean ± SEM from n = 3 independent biological replicates.

Data information: Statistical analysis used one‐way ANOVA with the Šidák's correction for multiple comparisons. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, ns = nonsignificant. The full list of multiplicity‐adjusted P‐values is presented in Dataset EV1.

As an intermediate product in the Pink1‐parkin pathway, pS65‐Ub levels will be affected by the kinetics of both its production and downstream turnover, both of which could be impacted by the loss of parkin. We therefore devised a paraquat pulse‐chase assay to probe the effect of loss of parkin on pS65‐Ub dynamics. Flies were exposed to paraquat for 1 day (pulse) and then kept under normal conditions without paraquat (chase) before being analysed for pS65‐Ub levels by immunoblotting. pS65‐Ub accumulated upon paraquat exposure then declined over several days post‐treatment, which we took to indicate turnover of the conjugated substrates (Fig 2C and D). Importantly, the loss of pS65‐Ub was not due to the known pS65‐Ub phosphatase PPEF2 (Wall et al, ²⁰¹⁹), since the loss of the Drosophila homologue rdgC did not affect pS65‐Ub degradation (Fig EV2). Another pS65‐Ub phosphatase, PTEN‐L (Wang et al, ²⁰¹⁸), is not conserved in flies thus precluding its assessment.

Figure EV2. Loss of rdgC (PPEF2) does not affect pS65‐Ub degradation in vivo .

Immunoblots for the indicated antibodies of whole‐fly lysates treated with paraquat (PQ) followed by recovery on normal food. UT, untreated flies. LacZ RNAi serves as control for comparison with knockdown of rdgC (the fly homologue of PPEF2). Chart shows the mean +/− SEM of equivalent blots from three biological replicates.

In park ^−/− mitochondria, pS65‐Ub levels were reduced at early time points compared with mitochondria from wild‐type animals (Figs 2C and D, and EV3A), consistent with a diminished feed‐forward cycle of Pink1‐parkin‐dependent pS65‐Ub production in response to mitochondrial damage as previously described (Ordureau et al, ²⁰¹⁴). By contrast, at later time points, pS65‐Ub levels were elevated in park ^−/− mitochondria compared with wild‐type flies, which likely reflects a defect in the turnover of damaged mitochondria (Figs 2C and D, and EV3A). It is therefore likely that parkin participates in the feed‐forward cycle to promote further parkin recruitment to damaged mitochondria but is not strictly required for the production of pS65‐Ub on mitochondria in response to paraquat.

Figure EV3. pS65‐Ub accumulates in park ^−/− flies.

1. Immunoblots for the indicated antibodies as controls for samples analysed in Fig 2C.
2. pS65‐Ub immunoblot following subcellular fractionation and USP2 treatment as indicated. (1) 10,000 × g pellet; (2) 21,000 × g pellet; (3) 21,000 × g supernatant. Asterisk (*) denotes nonspecific bands.

To identify the turnover signal produced by parkin on damaged mitochondria, we next determined the pattern of paraquat‐stimulated mitochondrial ubiquitination in the presence and absence of Pink1 or parkin. Analysing the four Ub chain types (linked at K6, K11, K48 and K63) that are most abundant on depolarised mitochondria and produced by Parkin in vitro (Ordureau et al, ²⁰¹⁴), the relative proportions of all four chain types were unchanged in mitochondria from Pink1 ⁻ and park ^−/− flies compared with wild‐type animals in basal conditions (Fig 2E–H). In response to paraquat, only K6 chains increased in abundance on wild‐type mitochondria, while K11 chains remained unchanged. Surprisingly, K48 and K63 chains decreased as a proportion of the total mitochondrial Ub, presumably due to a more substantial increase in monoubiquitination as previously described (Swatek et al, ²⁰¹⁹). The paraquat‐induced increase in K6 chains appeared to depend on Pink1 and parkin (Fig 2E). Our results are therefore consistent with other reports that the molecular function of parkin, rather than to amplify pS65‐Ub, may be to produce either K6 chains or another Ub signal on the OMM following recruitment to damaged mitochondria by binding to pS65‐Ub (Ordureau et al, ²⁰¹⁵; Gersch et al, ²⁰¹⁷).

pS65‐Ub accumulates on disrupted mitochondria in parkin mutants

In our initial pipeline for the detection of pS65‐Ub, we enriched mitochondria by relatively crude differential centrifugation. The pS65‐Ub levels of untreated park ^−/− flies were not substantially elevated in these fractions, as determined by mass spectrometry (Fig 2B) and immunoblotting (Fig 2C, UT). However, Shiba‐Fukushima et al (2017) had previously reported that pS65‐Ub accumulates in park ^−/− flies by immunoblotting. When we analysed whole cell lysates, untreated park ^−/− animals displayed a striking abundance of pS65‐Ub that was readily detectable by immunoblotting (Fig 3A), verifying the previous observation. We confirmed that this signal represented pS65‐Ub as it was sensitive to treatment with the deubiquitinase USP2 (Fig EV3B). The effect was also observed upon ubiquitous knockdown of parkin (daG4 > UAS‐park RNAi), confirming the specificity of the effect for loss of parkin (Fig 3B). The inducible RNAi line allowed us to assess the tissue distribution of the pS65‐Ub in these flies using tissue‐specific knockdowns. Interestingly, we found that the majority of the pS65‐Ub originated from the muscle rather than neurons (Fig 3B). However, when we enriched for neural tissues by harvesting heads with proboscis muscle removed, some pS65‐Ub was detectable in park ^−/− flies (Fig 3C). The higher relative abundance in muscle likely reflects that flight muscles have a very high mitochondrial content and are highly energetic, making them more prone to mitochondrial stress. This is consistent with a major requirement for Pink1‐parkin in this tissue in flies and the occurrence of major phenotypes here (Greene et al, ²⁰⁰³; Clark et al, ²⁰⁰⁶; Park et al, ²⁰⁰⁶).

Figure 3. Tissue level analysis of pS65‐Ub accumulation in park ^−/− flies.

• A, B
  pS65‐Ub immunoblot of whole‐animal lysates from untreated flies of the indicated genotypes. Park RNAi was induced by da‐GAL4 (ubiquitous), Mef2‐GAL4 (muscle) or nSyb‐GAL4 (neurons).
• C
  pS65‐Ub immunoblot of lysates from bodies or heads (without proboscises), as indicated, from park ^−/− animals.
• D
  Flight muscles from approximately 3‐day‐old untreated park ^−/− flies immunostained for ATP5A and pS65‐Ub. Scale bars = 10 μm.
• E
  Higher magnification image of pS65‐Ub‐positive structures in flight muscle counterstained for ATP5A. Scale bar = 5 μm. Chart shows quantification of ATP5A levels of pS65‐Ub‐positive (+ve) versus ‐negative (−ve) mitochondrial structures from three biological replicates (4–7 images per animal; average per animal shown in open circles). Line indicates mean. A.U., arbitrary units. ****P < 0.0001, Mann–Whitney test.

To better understand the subcellular localisation of pS65‐Ub in park ^−/− flies, we first confirmed by immunoblot of biochemical fractions that pS65‐Ub localises to cellular membranes as opposed to the cytosol (Fig EV1D). We then turned to an immunohistochemistry approach and saw numerous pS65‐Ub‐positive structures in the flight muscles of park ^−/− flies that were absent in wild‐type and Pink1 ⁻ flight muscles (Figs 3D and EV1B). These structures were highly heterogeneous and ranged from small punctate structures (< 1 μm³) to, more commonly, very large objects that resembled hyperfused or swollen mitochondria. Notably, mitochondria positive for pS65‐Ub showed a consistently reduced immunostaining of ATP5A, an essential mitochondrial component (Fig 3E). Importantly, pS65‐Ub structures clearly colocalised with a total Ub marker (Fig EV4A) and were absent in Pink1 ⁻ flight muscles, despite the presence of similar structures that stained for total Ub (Fig EV4B), confirming them as bone fide pS65‐Ub. Overall, these results are consistent with pS65‐Ub accumulating on the OMM of dysfunctional mitochondria in the absence of parkin.

Figure EV4. Pink1‐dependent pS65‐Ub colocalises with Ub on park ^−/− mitochondria.

• A, B
  Flight muscles from young, untreated (A) park ^−/− and (B) Pink1 ⁻ flies immunostained for conjugated Ub (FK2) and pS65‐Ub.

Data information: Scale bars = 10 μm.

Since there is only very limited data on the ultrastructural distribution of pS65‐Ub (Hou et al, ²⁰¹⁸), and none to our knowledge in vivo, we further explored this by immuno‐electron microscopy. As expected, wild‐type and Pink1 ⁻ flight muscle showed very little background pS65‐Ub staining, distributed randomly with cytosol or myofibrils (Fig 4A and B; arrowheads). By contrast, park ^−/− flies showed abundant pS65‐Ub immunostaining around mitochondrial cristae (Fig 4C). Interestingly, pS65‐Ub was only observed on electrolucent mitochondria and was absent from more intact, electron‐dense mitochondria (Fig 4Ci and ii). This neatly mirrors our observation of ATP5A depletion on pS65‐Ub‐positive structures by immunofluorescence. Together, these data indicate that pS65‐Ub accumulates selectively on disrupted mitochondria in vivo in the absence of parkin.

Figure 4. pS65‐Ub accumulates on disrupted mitochondria in vivo .

• A–C
  Immuno‐electron microscopy of flight muscle from (A) wild‐type, (B) Pink1 ⁻ or (C) park ^−/− flies stained for pS65‐Ub. Immuno‐gold particles are highlighted (arrowheads). Scale bar = 500 nm. The numerous pS65‐Ub immuno‐gold particles are not labelled in (C), to avoid obscuring the image, but dashed boxes are magnified in the insets (Ci and Cii). Scale bar (insets) = 100 nm.

Loss of core autophagy genes minimally affects pS65‐Ub accumulation

The striking increase in pS65‐Ub levels in park ^−/− flies indicated that the turnover of pS65‐Ub was disrupted, which presents a paradigm to investigate the turnover mechanisms downstream of Pink1 and parkin as a more physiological alternative to genetic reporter constructs. Given the abundant evidence in cell culture models that PINK1‐Parkin‐mediated turnover occurs via the canonical autophagy machinery (Lazarou et al, ²⁰⁰⁷; Nguyen et al, ²⁰¹⁶), we analysed pS65‐Ub levels in mutants of core autophagy genes, Atg1 (homologue of ULK1), Atg5 and Atg8a (homologue of LC3/GABARAP). We saw a modest age‐related increase in pS65‐Ub levels in Atg5 ⁻ (Atg5 ^5cc5) flies compared with wild‐type animals (Fig EV5A), but surprisingly, this was very low compared with the increase in pS65‐Ub levels observed in park ^−/− flies (Fig 5A). Consistent with this, neither loss of Atg1 (daG4 > Atg1 RNAi) nor Atg8a ⁻ (Atg8a ^KG07569) led to the same dramatic increase in pS65‐Ub levels as loss of park (Figs 5B and EV5B).

Figure EV5. Dynamics of pS65‐Ub production/turnover in autophagy mutants.

1. pS65‐Ub immunoblot of mitochondrial fractions from wild‐type and Atg5 ⁻ flies harvested at the indicated ages. Asterisk (*) denotes nonspecific band.
2. pS65‐Ub immunoblotting in whole‐animal lysates from wandering L3 larvae of the indicated genotypes.
3. pS65‐Ub immunoblot of mitochondrial fractions from wild‐type and Atg5 ⁻ flies following a paraquat (PQ) pulse‐chase assay. UT, untreated; Recovery, return to normal food.
4. Quantification of pS65‐Ub lane densitometry from n = 3 independent replicates of (B), expressed relative to the most intense band in each blot. Charts show mean ± SEM.

Figure 5. Loss of Atgs does not lead to the same degree of pS65‐Ub accumulation as loss of parkin.

1. pS65‐Ub immunoblot of whole‐animal lysates from wild‐type, Atg5 ⁻ and park ^−/− animals harvested at the indicated ages.
2. pS65‐Ub immunoblot of whole‐animal lysates from young animals of the following genotypes: wild‐type, Pink1 ⁻, park ^−/−, park RNAi (daG4 > UAS‐park RNAi), Atg1 RNAi (daG4 > UAS‐Atg1 RNAi) and Atg8a ⁻.

Loss of parkin led to pS65‐Ub production that was readily detectable by immunoblotting as early as the larval stage of development (Fig EV5B). To further probe whether the canonical autophagy machinery affects pS65‐Ub production, we quantified the size and number of pS65‐Ub‐positive puncta in larval muscle. Wild‐type and Pink1 ⁻ larvae displayed no pS65‐Ub puncta (Fig 6A, B and H), consistent with the absence of pS65‐Ub observed by immunoblotting (Fig EV4B). By contrast, park ^−/− tissues displayed abundant pS65‐Ub puncta (Fig 6C and H). Atg5 ⁻ and Atg8a ⁻ larvae also displayed pS65‐Ub puncta, although they were markedly fewer and generally smaller than those present in park ^−/− mutants (Fig 6D, E, H and I), while Atg5 ⁻; park ^−/− and Atg8a ⁻; park ^−/− double mutants displayed puncta similar in number and size to park ^−/− alone (Fig 6F–I). These results suggest that canonical autophagy is not the major route for the turnover of pS65‐Ub‐positive structures under basal conditions. Notably, while park ^−/−, Atg5 ⁻ and Atg8a ⁻ animals are viable to the adult stage, Atg5 ⁻; park ^−/− and Atg8a ⁻; park ^−/− double mutants are generally nonviable past the pupal stage, with only a few rare escapers, indicating synthetic lethality from the combined effect of independent pathways.

Figure 6. pS65‐Ub immunostaining of larval muscle.

• A–G
  Representative images of pS65‐Ub immunostaining false‐coloured for intensity (Fire LUT) in muscle segments 6/7 from wandering L3 larvae of the following genotypes: (A) wild‐type, (B) Pink1 ⁻, (C) park ^−/−, (D) Atg5 ⁻, (E) Atg8a ⁻, (F) Atg5 ⁻; park ^−/− and (G) Atg8a ⁻; park ^−/−. Scale bar = 20 μm.
• H
  Quantification of the number of pS65‐Ub puncta from (A–G) displaying mean ± SEM from the indicated number of animals (n = 3–5 as indicated).
• I
  Puncta volume from (C–G) expressed as mean ± SEM (n = 4–5 animals).

Data information: Statistical analysis used one‐way ANOVA with the Dunnett's correction for multiple comparisons. Comparisons are against wild‐type unless indicated. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001, ns = nonsignificant.

Parkin overexpression reduces pS65‐Ub levels in the absence of Atg5

Although the loss of the core autophagy components Atg1, Atg5 and Atg8a did not result in the same extent of pS65‐Ub accumulation as loss of parkin, loss of Atg5 or Atg8a did lead to modestly increased pS65‐Ub levels compared with wild‐type animals (Figs EV5A and 6H). In addition, in the paraquat pulse‐chase assay Atg5 ⁻ flies had elevated pS65‐Ub at later time points relative to wild‐type flies, suggestive of a block in turnover (Fig EV5C and D). These results are consistent with the autophagy machinery contributing at some level to basal turnover of damaged mitochondria and more substantially upon induction of mitochondrial damage using paraquat.

In order to further dissect whether the parkin‐mediated pS65‐Ub turnover is autophagy‐dependent, we investigated the effect of parkin overexpression in an Atg5‐null background (Atg5 ^5cc5; daG4 > UAS‐park). We hypothesised that if the Pink1‐parkin pathway proceeds primarily via autophagy, then parkin overexpression should either not affect or perhaps even further increase pS65‐Ub levels in an Atg5 ⁻ background. By contrast, if parkin drives autophagy‐independent turnover, its overexpression should reduce pS65‐Ub levels even in the absence of Atg5. We found in the paraquat pulse‐chase assay that parkin overexpression substantially reduced pS65‐Ub levels in an Atg5 ⁻ background relative to an Atg5 ⁻ mutant control (Atg5 ^5cc5; daG4 > UAS‐mito‐HA‐GFP; Fig 7A and B). We further confirmed by mass spectrometry that while mitochondria from Atg5 ⁻ flies displayed modestly elevated pS65‐Ub levels, this could be reduced upon overexpression of parkin (Fig 7C). Taken together, these results indicate that parkin is able to drive pS65‐Ub turnover independently of the canonical autophagy machinery in Drosophila.

Figure 7. Parkin overexpression reduces pS65‐Ub levels in an Atg5 ⁻ background.

1. pS65‐Ub immunoblot of whole‐animal lysates following paraquat pulse‐chase assay of Atg5 ⁻ flies with parkin overexpression (OE) (Atg5 ^5cc5 daG4 > UAS‐park) compared with Atg5 ⁻ flies overexpressing mitochondrially targeted GFP (Atg5 ^5cc5 daG4 > UAS‐mito‐HA‐GFP). UT, untreated flies; PQ, 1‐day paraquat treatment; Recovery, flies removed from paraquat and returned to normal food.
2. pS65‐Ub lane densitometry, normalised to total protein (Revert™ stain), of n = 3 independent replicates of (A), expressed as pS65‐Ub intensity relative to the most intense lane in each blot. Charts show mean ± SEM.
3. Mass spectrometry analysis of normalised pS65‐Ub levels in Ub‐Clippase‐treated, TUBE‐enriched mitochondrial fractions from flies overexpressing parkin (daG4 > UAS‐park), Atg5 ⁻ flies and Atg5 ⁻ flies overexpressing parkin (Atg5 ^5cc5 daG4 > UAS‐park). Charts show mean ± SEM (n = 3). Statistical analysis used one‐way ANOVA with the Dunnett's correction for multiple comparisons. **P < 0.01; ***P < 0.001.

Finally, in an effort to determine the major degradation route of pS65‐Ub, whether by lysosome or proteasome mechanisms, we treated wild‐type flies with classic inhibitors of lysosome (chloroquine) or proteasome (MG132) during the recovery phase of the paraquat pulse‐chase assay. Surprisingly, chemical inhibition of either lysosome or proteasome alone was not sufficient to substantially affect pS65‐Ub degradation (Fig EV6A–C), while the combination of both inhibitors substantially blocked pS65‐Ub degradation (Fig EV6D). These results suggest that both proteasome and lysosome mechanisms partially contribute to pS65‐Ub turnover; however, the mechanism remains to be elucidated.

Figure EV6. Impact of lysosome and proteasome inhibitors on pS65‐Ub degradation.

• A–D
  pS65‐Ub immunoblots of whole‐fly lysates treated with paraquat (PQ) followed by recovery on filter papers with sucrose solution only (A) or dosed with (B) chloroquine (2 mM), (C) MG132 (50 μM) or both (D). Blots are representative of duplicate experiments.

Discussion

We have developed mass spectrometry and immunodetection methods to monitor physiological levels of pS65‐Ub as a direct and specific readout of Pink1 activity in Drosophila, a preeminent model for dissecting the conserved functions of Pink1 and parkin. We have found that pS65‐Ub is produced by Pink1 under basal conditions, albeit at very low levels. Our methods revealed that ~0.5% of total Ub on mitochondria from aged flies is Ser65‐phosphorylated, and we further observed individual mitochondria that were enveloped in pS65‐Ub. Although we could not reliably detect pS65‐Ub in young flies without additional enrichment, we surmise that it is likely less than 0.1% of mitochondrial Ub. This suggests that under normal healthy conditions, in the absence of exogenous or accumulated endogenous stresses, Pink1 activation is either an extremely infrequent event or the pS65‐Ub is very quickly degraded.

Importantly, in an effort to understand the physiological stimulus of PINK1, we have found that pS65‐Ub deposition is strongly induced by transient exposure to the oxidant and PD‐linked toxin paraquat. Elevated oxidative stress is a long‐standing culprit in the pathogenesis of PD and an intuitive trigger for mitochondrial quality control systems. Interestingly, recent work has elucidated a likely mechanism for PINK1 regulation by Cys oxidation (Gan et al, ²⁰²²; Rasool et al, ²⁰²²). While these in vitro and oxidant‐induced observations are compelling, it will be interesting to determine the endogenous origin of the Cys oxidation, and particularly, how this signal is relayed from the matrix to PINK1 on the outer mitochondrial surface. Nevertheless, our data support that oxidative stress is a key trigger for Pink1/parkin activity in vivo.

We observed that pS65‐Ub accumulates to high levels in Drosophila parkin mutants in the absence of exogenous stimulation of the pathway, confirming a previous observation (Shiba‐Fukushima et al, ²⁰¹⁷). Further, we have explored the spatiotemporal dynamics of pS65‐Ub deposition, which occurs specifically on disrupted, electrolucent mitochondrial structures depleted for the OXPHOS subunit ATP5A. In response to paraquat, park ^−/− flies displayed low pS65‐Ub levels relative to wild‐type flies at early time points and elevated pS65‐Ub at later time points, suggestive of a defect in mitochondrial turnover. These observations indicate that pS65‐Ub alone is insufficient to elicit mitochondrial turnover and therefore that parkin's molecular function is not simply to amplify the pS65‐Ub signal produced by Pink1. Several earlier studies also support this conclusion: increased stoichiometry of mitochondrial Ub phosphorylation was found to be inhibitory to mitophagy receptor recruitment in cell culture studies (Ordureau et al, ²⁰¹⁸), and in vitro binding studies using Ser65‐phosphorylated Ub chains have found that Ub phosphorylation does not promote autophagy receptor binding (Heo et al, ²⁰¹⁵; Ordureau et al, ²⁰¹⁵). Indeed, to our knowledge, Parkin itself is the only protein that has been shown to bind preferentially to pS65‐Ub (Wauer et al, ²⁰¹⁵). Our findings therefore suggest that the function of pS65‐Ub is primarily to recruit Parkin to damaged mitochondria, rather than to promote downstream organelle turnover.

Our investigation of Pink1‐ and parkin‐dependent Ub chain types showed that K6 chains (and not K11, K48, K63 or bulk ubiquitination) were upregulated in response to paraquat in wild‐type but not Pink1‐ or parkin‐deficient animals. These results corroborate previous in vitro data showing that K6 chains are produced on damaged mitochondria by these enzymes and suggest that this signal could be important for parkin‐dependent turnover of damaged mitochondria in vivo. We did not investigate the specific substrates that are ubiquitinated by parkin, but previous studies have found a wealth of OMM substrates ubiquitinated in vivo and in cell culture studies in a Parkin‐dependent manner, some of which could be important signals for mitochondrial turnover (Martinez et al, ²⁰¹⁷; Ordureau et al, ^{2018, 2020}). Regardless, the increase in K6 chains that we observed is particularly interesting given that the only mitochondria‐resident deubiquitinase, USP30, preferentially hydrolyses K6 chains (Cunningham et al, ²⁰¹⁵; Wauer et al, ²⁰¹⁵; Gersch et al, ²⁰¹⁷; Sato et al, ²⁰¹⁷). The functions of K6 chains are not well understood, especially in vivo (Swatek & Komander, ²⁰¹⁶), and further work is warranted to elucidate the precise contribution of this atypical Ub chain type to mitochondrial quality control.

The dramatic increase in pS65‐Ub levels upon loss of parkin allowed us to assess the machinery responsible for the downstream turnover of pS65‐ubiquitinated mitochondria. Analysing the autophagy machinery, we observed that upon loss of the core autophagy components Atg1, Atg5 or Atg8a, pS65‐Ub levels were not affected nearly to the extent observed in park ^−/− flies, which suggests that pS65‐Ub is not primarily turned over via canonical autophagy under basal conditions. Moreover, parkin overexpression was able to reduce both basal and paraquat‐induced pS65‐Ub levels in an Atg5 ⁻ background. These results add to the growing evidence that the Pink1‐parkin pathway may, under more physiological conditions, promote the turnover of damaged mitochondrial components in an autophagy‐independent manner. For instance, Vincow et al (2013) analysed turnover rates of mitochondrial proteins in Drosophila and found that Pink1 and parkin were required for the turnover of a subset of IMM proteins that were distinct from those turned over by the core autophagy protein Atg7. Vincow et al also found some overlap between the proteins turned over by parkin and Atg7; similarly, we observed a slight accumulation of pS65‐Ub in autophagy‐deficient flies, consistent with some mitochondrial degradation occurring via a classic nonselective autophagy route. However, the Pink1‐parkin pathway clearly has roles that are divergent from canonical autophagy, as Atg5 ⁻; park ^−/− and Atg8a ⁻; park ^−/− double mutants showed synthetic lethality, and parkin overexpression was able to reduce both basal and paraquat‐induced pS65‐Ub levels in an Atg5 ⁻ background. Consistent with this, the ability of parkin overexpression to rescue Pink1 phenotypes has been shown to be unaffected by the loss of Atg1 or Atg18 (Liu & Lu, ²⁰¹⁰).

Alternative mechanisms of parkin‐mediated pS65‐Ub turnover could be via direct proteasomal degradation of pS65‐ubiquitinated OMM proteins (Tanaka et al, ²⁰¹⁰; McLelland et al, ²⁰¹⁸), or via MDV delivery directly to the lysosome (McLelland et al, ²⁰¹⁴; Sugiura et al, ²⁰¹⁴; Ryan et al, ²⁰²⁰). However, in vivo evidence for the existence of MDVs is limited, due in part to technical constraints in observing such small structures in complex tissues. Our preliminary data using chemical inhibitors of proteasome and lysosome suggest some contribution from both of these mechanisms, suggesting that there is plasticity and redundancy in the mechanisms of mitochondrial turnover in vivo as has been previously discussed in other contexts (Kocaturk & Gozuacik, ²⁰¹⁸). Although the mechanistic details remain to be elucidated, we anticipate that the methods described herein could aid in analysing the presence of MDVs and their dependence on Pink1 and parkin in vivo, and to dissect the relative contribution of proteasome‐ and lysosome‐dependent mechanisms of parkin‐dependent mitochondrial turnover.

pS65‐Ub as a biomarker for neurodegeneration

pS65‐Ub has been proposed as a potential biomarker for neurodegenerative disease, with a recent study finding elevated pS65‐Ub levels in blood samples from a cohort of Alzheimer's Disease patients compared with age‐matched controls (Watzlawik et al, ²⁰²¹). The cellular pathology underlying PD precedes classical symptom onset and therefore clinical diagnosis by many years, which is likely to hamper the success of clinical trials of potentially disease‐modifying drugs as they may be given too late to halt disease progression (Stern et al, ²⁰¹²). We found that pS65‐Ub was readily detectable in park ^−/− animals in developmental stages prior to overt neurodegeneration, thereby suggesting that pS65‐Ub accumulation is an early event that, if replicated in mammals, could have potential as an early‐stage diagnostic biomarker. We also note the 100% sequence identity of Ub and conservation of the S65 phosphorylation site in Drosophila compared with mammals, meaning that the tools described herein can be applied to various species. However, pS65‐Ub levels increased with healthy ageing in flies and were absent upon loss of Pink1, and we show a complex time‐dependent effect on pS65‐Ub levels in paraquat‐treated park ^−/− flies. We therefore posit that a meaningful healthy range of pS65‐Ub abundance would need to be established in order to use pS65‐Ub levels as a clinical or preclinical biomarker.

In conclusion, we have developed methods to detect pS65‐Ub at physiological levels in Drosophila and delivered the unanticipated finding that loss of parkin results in a striking elevation in pS65‐Ub levels that is not recapitulated upon loss of the canonical autophagy genes Atg5, Atg8a or Atg1. We expect that the tools described herein will greatly aid in future studies dissecting the downstream mechanisms of mitochondrial turnover that are promoted by Pink1 and parkin in vivo.

Materials and Methods

Drosophila stocks and husbandry

Flies were raised under standard conditions in a temperature‐controlled incubator with a 12 h:12 h light:dark cycle at 25°C and 65% relative humidity, on food consisting of agar, cornmeal, molasses, propionic acid and yeast. The following strains were obtained from the Bloomington Drosophila Stock Centre (RRID:SCR_006457): w ¹¹¹⁸ (RRID: BDSC_6326), da‐GAL4 (RRID: BDSC_55850), Mef2‐GAL4 (RRID: BDSC_27390), nSyb‐GAL4 (RRID: BDSC_51941), Atg8a ^KG07569 (RRID: BDSC_14639), UAS‐mito‐HA‐GFP (RRID: BDSC_8443), UAS‐rdgC ^HMC05070 RNAi (RRID: BDSC_60076) and UAS‐park ^HMS01800 RNAi (RRID: BDSC_38333). The UAS‐Atg1 ^GD7149 RNAi (VDRC_16133) and UAS‐rdgC ^GD11966 RNAi (VDRC_35105) lines were obtained from the Vienna Drosophila Resource Center. Pink1 ^B9 flies were a kind gift from J. Chung (Park et al, ²⁰⁰⁶), and the Atg5 ^5cc5 stock was a kind gift from G. Juhasz (Kim et al, ²⁰¹⁶). UAS‐park _C2, park ²⁵ and UAS‐mito‐APOBEC1 lines have been described previously (Greene et al, ²⁰⁰³; Lee et al, ²⁰¹⁸; Andreazza et al, ²⁰¹⁹). Male flies only were used for experiments with adults, while experiments with larvae used both males and females except for animals with X chromosome balancers (Pink1 ^B9, Atg5 ^5cc5), for which only male animals were used. For ageing experiments, flies were maintained in bottles (MS experiments) or tubes (immunostaining experiments), transferred to fresh food 2–3 times per week and harvested after 50–60 days. Unless otherwise specified, young flies (2–3 days old) were used for all experiments.

Paraquat exposure assays

For MS experiments, flies were maintained in bottles (100–200 flies per replicate) containing 9 semi‐circular pieces of filter paper (90 mm diameter, Cat#1001‐090) saturated with 5% (w/v) sucrose solution containing 5 mM paraquat. Sucrose‐only starvation experiments were performed as above, with the omission of paraquat. After 3 days, the flies were anaesthetised with mild CO₂ and live flies only were harvested. For pulse‐chase experiments, 5–25 flies were harvested per replicate. The day 0 control was taken prior to paraquat treatment, and the remaining flies were incubated overnight in bottles containing paraquat as above. The next day, flies were anaesthetised with CO₂, dead flies were removed, a day 1 timepoint was taken and the remaining flies were divided among tubes of food with no more than 20 flies per tube. On day 2, all flies were flipped onto fresh food and then flipped every 2–3 days thereafter until harvest. For the lysosome and proteasome inhibitor pulse‐chase experiments, w ¹¹¹⁸ flies were collected and treated with 10 mM paraquat as described above. After 1 day, flies were placed in tubes containing filter papers saturated with 5% (w/v) sucrose solution dosed with chloroquine (2 mM, Merck Life Science Limited, Cat#C6628) and/or MG132 (50 μM, Sigma Aldrich, Cat#C2211). Fresh solutions were applied every 2–3 days thereafter until harvest.

Mitochondrial enrichment by differential centrifugation

All steps were performed on ice or at 4°C. For mass spectrometry analysis, mitochondria were harvested from fresh (i.e. not frozen) flies according to (Lazarou et al, ²⁰⁰⁷), with modifications. Whole flies (60–200 per replicate) were placed in a Dounce homogeniser. Solution A (70 mM sucrose, 20 mM HEPES pH 7.6, 220 mM mannitol, 1 mM EDTA) containing cOmplete protease inhibitors (Roche) and PhosSTOP phosphatase inhibitors (Roche) was added (approximately 10 μl per fly), and the flies were homogenised with 35 strokes of a drill‐fitted pestle. The homogenate was transferred to a 50 ml tube and incubated for 30 min, then centrifuged for 5 min at 1,000 × g. The supernatant was transferred to microcentrifuge tubes and centrifuged for 15 min at 10,000 × g to pellet mitochondria. The postmitochondrial supernatant was removed and the pellet was resuspended in Solution A. The homogenate was then clarified by centrifugation for 5 min at 800 × g, and the supernatant was transferred to a fresh tube. This clarification step was repeated once more to ensure all cuticle was removed from the sample. The supernatant was then centrifuged 10 min at 10,000 × g, and the postmitochondrial supernatant was discarded. The pellet was resuspended in Solution A and centrifuged for 10 min at 10,000 × g, and this wash step was repeated for a total of three times. The washed pellet was resuspended in Sucrose Storage Buffer (500 mM sucrose, 10 mM HEPES pH 7.6) and stored at −80°C until needed.

For immunoblotting analysis and biochemical fractionation from small numbers of flies (10–30), a modified mitochondrial enrichment procedure was performed. Flies were prepared either fresh or after flash‐freezing in liquid nitrogen, with all direct comparisons performed with flies that were prepared in the same manner. Flies were transferred into a Dounce homogeniser containing 700 μl Solution A containing protease and phosphatase inhibitors as above and manually homogenised with 50 strokes of a pestle. The homogenate was transferred to an Eppendorf tube, a further 500 μl of Solution A was added to the homogeniser and the sample was homogenised with a further 10 strokes. The homogenates were pooled and incubated for 30 min, and then centrifuged for 5 min at 800 × g. The supernatant (containing mitochondria) was transferred to a new tube and clarified twice by centrifugation for 5 min at 1,000 × g. The clarified supernatant was then centrifuged for 10 min at 10,000 × g and the postmitochondrial supernatant was discarded or, in the case of biochemical fractionation experiments, further centrifuged for 30 min at 21,000 × g, and the pellet and supernatant retained for analysis. The mitochondrial pellet was washed once in Solution A containing only protease inhibitors, and then once in Solution A without inhibitors. The washed mitochondrial pellet was resuspended in 50–200 μl sucrose storage buffer, the protein content determined by BCA assay (Thermo Pierce) and stored at −80°C until needed.

USP2 treatment

For the validation of the pS65‐Ub signal in park ^−/− samples, 30 μg protein per subcellular fraction was treated with the pan‐specific deubiquitinase USP2 (BostonBiochem, E‐506). The USP2 enzyme was diluted in buffer (50 mM Tris–pH 7.5, 50 mM NaCl, 10 mM DTT; Hospenthal et al, ²⁰¹⁵) and then added to the subcellular fractions to a final USP2 concentration of 1 μM. The mixture was incubated for 45 min at 37°C prior to analysis by immunoblotting.

Mass spectrometry sample preparation and analysis

Absolute quantification (AQUA) analysis of Ub modifications was performed as previously described (Swatek et al, ²⁰¹⁹), with modifications. Five hundred microgram mitochondria (from approximately 100 flies) prepared as above were resuspended in 250 μl TUBE lysis buffer (PBS containing 1% (v/v) NP‐40, 2 mM EDTA, 10 mM chloroacetamide, cOmplete EDTA‐free protease inhibitor cocktail (Roche)) supplemented with 8 μg/ml GST‐Ubiquilin‐UBA (Hjerpe et al, ²⁰⁰⁹; Fiil et al, ²⁰¹³; Hrdinka et al, ²⁰¹⁶). The lysate was incubated on ice for 20 min and then centrifuged for 15 min at 21,000 × g, 4°C. The clarified lysate was added to 20 μl Glutathione Sepharose 4B resin (GE Healthcare) that had been washed three times in TUBE lysis buffer and was incubated for 2 h at 4°C with gentle rotation. The lysate was removed and the beads were washed twice with PBS containing 0.1% (v/v) Tween 20, then twice with PBS. Eighty microliter Lb^pro reaction buffer (50 mM NaCl, 50 mM Tris–pH 7.4, 10 mM DTT) containing 20 μM Ub‐clippase Lb^pro construct containing residues 29–195 with L102W mutation, purified as described previously (Guarne et al, ²⁰⁰⁰; Swatek et al, ²⁰¹⁹) was added and Ub was cleaved from the beads for 16 h at 37°C. The supernatant was removed, the beads were washed with Lb^pro reaction buffer, and the supernatants pooled and acidified to pH < 4 using formic acid (FA), prior to fractionation using StageTips (Rappsilber et al, ²⁰⁰⁷). StageTips were assembled using 4 plugs that were cut using a gauge 16 needle (Hamilton) from C₄ substrate (SPE‐Disks‐Bio‐C4‐300.47.20, AffiniSEP) and assembled into a P200 pipette tip using a plunger (Hamilton). The matrix was activated by the addition of 30 μl methanol and the tip was centrifuged inside a 2 ml Eppendorf tube at 800 × g for 30 s at room temperature to allow the liquid to pass through. The tip was then equilibrated by passing through 30 μl 80% (v/v) acetonitrile (ACN), 0.1% (v/v) FA, followed by 30 μl 0.1% (v/v) FA. The acidified sample was loaded and centrifuged as above until almost all the liquid had passed through. The tip was then desalted by passing through 40 μl 0.1% (v/v) FA, twice. The StageTip was then washed twice with 30 μl 20% (v/v) ACN, 0.1% (v/v) FA, then the ubiquitin was eluted into a clean tube with two elutions of 30 μl 45% (v/v) ACN, 0.1% (v/v) FA. The eluate was lyophilised and resuspended in Trypsin Resuspension Buffer (Promega) supplemented with Tris–pH 8.0 to ensure a final pH above 6. Sequencing grade modified Trypsin (Promega) was added at a concentration of 1 μg per 250 μg initial mitochondrial protein, and the samples were incubated for 16 h at 37°C. For StageTip purification after trypsin treatment, the sample was acidified to pH < 4 using FA. AQUA peptides, supplied by Cambridge Research Biochemicals (pS65‐Ub peptide) and Cell Signalling Technologies (all other peptides), were spiked at concentrations as indicated in Dataset EV1 and the sample was loaded into a StageTip containing 4 plugs of C₁₈ substrate (SPE‐Disks‐Bio‐C18‐100.47.20, AffiniSEP) that had been assembled, activated and pre‐equilibrated as above. The tip was washed 3 times in 0.1% (v/v) FA, then elution was performed twice with 30 μl 80% (v/v) ACN, 0.1% (v/v) FA. Samples were lyophilised and resuspended in 5% (v/v) ACN, 0.1% (v/v) FA, and 10 μl was injected onto a Dionex Ultimate 3000 HPLC system (Thermo Fisher Scientific) and trapped on a C18Acclaim PepMap100 (5 μm, 100 μm × 20 mm nanoViper; Thermo Scientific). Peptides were eluted with a 60‐min acetonitrile gradient (2–40%) at a flow rate of 0.3 μl/min. The analytical column outlet was directly interfaced via an EASY‐Spray electrospray ionisation source to a Q Exactive mass spectrometer (Thermo Fisher Scientific). The following settings were used: resolution, 140,000; AGC target, 3E6; maximum injection time, 200 ms; scan range, 150–2,000 m/z. Absolute abundances of Ub peptides were calculated by peak integration using Xcalibur Qual Browser (Version 2.2, Thermo Fisher Scientific). Layouts were applied according to Dataset EV1, and abundances were calculated relative to the known amount of added AQUA reference peptide using Microsoft Excel.

For the detection of pS65‐Ub in young flies (Fig 1B), the following modifications to the method were performed. Instead of TUBE‐mediated Ub pulldown, mitochondrial fractions were sodium carbonate‐extracted to enrich ubiquitinated integral membrane proteins as previously described (Swatek et al, ²⁰¹⁹). In brief, 4 mg mitochondria were resuspended in 4 ml 100 mM Na₂CO₃. The mixture was incubated for 30 min on ice with occasional vortexing and then centrifuged for 30 min, 21,000 × g, 4°C. The supernatant, containing soluble and peripheral membrane proteins, was discarded, and the pellet, containing integral membrane proteins, was then resuspended in Lb^pro reaction buffer (1 μl per 10 μg mitochondria). An equal volume of 20 μM Lb^pro was added (10 μM final concentration) and the mixture was incubated overnight at 37°C. The samples were centrifuged for 30 min, 21,000 × g, 4°C, and the supernatant was acidified and purified using StageTips as above (1 StageTip per 1 mg starting material). Trypsin treatment was performed as above and AQUA peptides were spiked in according to Dataset EV1. Phospho‐peptide enrichment was then performed using the High‐Select™ TiO₂ Phospho‐peptide Enrichment kit (Thermo Fisher Scientific). Each replicate was divided between two TiO₂ columns and prepared according to the manufacturer's instructions. The eluates were pooled, lyophilised and analysed by LC–MS as above.

Antibodies and dyes

The following mouse antibodies were used for immunoblotting (WB) and/or immunofluorescence (IF) in this study: ATP5A (Abcam, ab14748, 1:10,000 (WB), 1:300 (IF, adult muscle)), Ubiquitin (clone FK2, MBL, D058‐3, 1:2,000 (WB), 1:250 (IF, adult muscle)), Actin (Millipore, MAB1501, 1:1,000 (WB)), GAPDH (GeneTex, GTX627408, 1:1,000 (WB)). The following rabbit antibodies were used in this study: pS65‐Ub (Cell Signalling Technologies, 62802S, 1:750 (WB), 1:200 (IF, larval muscle), 1:250 (IF, adult muscle)), COXIV and SDHA (both kind gifts from Edward Owusu‐Ansah (Murari et al, ²⁰²⁰), both 1:2,000 (WB)), Porin (Calbiochem, PC548, 1:5,000 (WB)). The following secondary antibodies were used: sheep anti‐mouse (HRP‐conjugated, GE Healthcare, NXA931V, 1:10,000 (WB)), donkey anti‐rabbit (HRP‐conjugated, GE Healthcare, NA934V, 1:10,000 (WB)), goat anti‐mouse (AlexaFluor 488, Invitrogen, A11001, 1:200 (IF)), goat anti‐rabbit (AlexaFluor 594, Invitrogen, A11012, 1:200 (IF)).

Whole‐animal lysis and immunoblotting

For the analysis of pS65‐Ub levels in whole cell lysates by immunoblot, 180 μl cold RIPA buffer (150 mM NaCl, 1% (v/v) NP‐40, 0.5% (w/v) sodium deoxycholate, 0.1% (w/v) SDS, 50 mM Tris–pH 7.4), supplemented with cOmplete and PhosSTOP inhibitors, was added to 2 ml tubes containing 1.4 mm ceramic beads (Fisherbrand 15555799). Animals (5 to 20 per replicate) were harvested and stored on ice or flash‐frozen in liquid N₂, with all direct comparisons performed with flies that were harvested in the same manner. The flies were added to the tubes containing RIPA buffer and lysed using a Minilys homogeniser (Bertin Instruments) with the following settings: maximum speed, 10 s on, 10 s on ice, for a total of three cycles. After lysis, samples were returned to ice for 10 min and then centrifuged for 5 min at 21,000 × g, 4°C. Ninety microliter supernatant was transferred to a fresh Eppendorf tube and centrifuged a further 10 min at 21,000 × g. Fifty microliter supernatant was then transferred to a fresh Eppendorf tube and the protein content determined by BCA assay as above. Thirty microgram total protein was then diluted in NuPAGE LDS loading dye (Invitrogen) and analysed by SDS–PAGE using Novex 4–12% Bis–Tris NuPAGE gels (Invitrogen). For the analysis of mitochondria‐enriched fractions, 30–50 μg mitochondrial protein was aliquoted into a tube, centrifuged for 10 min at 16,000 × g, the supernatant removed and the pellet resuspended in LDS loading dye prior to SDS–PAGE analysis as above. Gels were transferred onto precut and ‐soaked PVDF membranes (1704157, BioRad) using the BioRad Transblot Turbo transfer system, and blots were immediately stained with Revert total protein stain (926‐11011, LiCOR) or No‐Stain total protein (A44449, Invitrogen) where indicated, according to the manufacturer's instructions. Fluorescence intensity was measured using a BioRad Chemidoc MP using the IR680 setting. Blots were then washed by gentle shaking 3 times for 5 min in PBS containing 0.1% (v/v) Tween‐20 (PBST) and blocked by incubation with PBST containing 3% (w/v) BSA for 1 h. Blots were washed a further 3 times as above and then incubated at 4°C overnight with primary antibodies in PBST containing 3% (w/v) BSA. Further three washes were performed, then the blots were incubated for 1 h in secondary antibodies made up of PBST containing 3% (w/v) BSA. Blots were then washed twice in PBST and once in PBS (twice in the case of pS65‐Ub blots) prior to incubation with the ECL reagent. For pS65‐Ub blots, SuperSignal Femto reagent (Thermo Scientific) was used, while other blots used Clarity ECL reagent (BioRad). Blots were imaged using the BioRad Chemidoc MP using exposure settings to minimise overexposure, except where high exposure is indicated. Image analysis was performed using Image Lab (Version 5.2.1 build 11, BioRad) and images were exported as TIFF files for presentation.

Immunostaining of Drosophila tissues

Larval fillet and adult thoraces dissections were performed in PBS and fixed in 4% formaldehyde, pH 7.0, for 20 (larval fillet) or 30 (adult flight muscles) minutes, respectively. Permeabilisation was performed for 30 min in PBS containing 0.3% (v/v) Triton X‐100 (PBS‐TX), then tissues were blocked for 1 h in PBS‐TX containing 1% (w/v) BSA (larval fillet) or 4% Horse Serum (adult flight muscles). Primary antibody incubation was performed overnight at 4°C in PBS‐TX containing 1% (w/v) BSA. The tissues were washed 3 times for 10 min each in PBS‐TX prior to incubation with secondary antibodies for 2 h at room temperature (larval fillet) or overnight at 4°C (adult thoraces). The tissues were washed three times for 10 min in PBS‐TX, then once for 10 min in PBS, and rinsed once in water prior to mounting in Prolong Diamond Antifade mounting media with DAPI (Thermo Fisher Scientific).

Microscopy and image analysis

Fluorescence microscopy imaging was performed using a Zeiss LSM 880 confocal microscope equipped with a 20× Plan Apochromat (air, NA 0.8), 63× Plan Apochromat (oil immersion, NA 1.4) and 100× Plan Apochromat (oil immersion, NA 1.4), objective lenses. Laser power and gain settings were adjusted depending on the fluorophore but were maintained across samples for the purpose of comparing pS65‐Ub levels among genotypes.

For the quantification of pS65‐Ub puncta in adult indirect flight muscles (Fig 3), images were processed using FIJI (ImageJ, Version: 2.3.0/1.53q): A ROI was generated for each of the large pS65‐Ub puncta using the threshold tool (Huang method) to generate a mask of the pS65‐Ub signal. Puncta smaller than 0.5 μm² were removed. The pS65‐Ub mask was then applied to the ATP5A staining. The ATP5 immunostaining intensity was extracted and represented for inside and outside pS65‐Ub structures.

For the quantification of pS65‐Ub puncta in larval muscle 6–7 (Fig 6), images were processed using FIJI (ImageJ, Version 2.1.0/1.53c): z‐stacks (5 per image, 0.28 μm step size) were cropped to remove extraneous tissue, axons and neuromuscular junctions, which we observed contained high background signal with the anti‐pS65‐Ub antibody. The retained area was measured, local background subtraction was performed, and the number of puncta was quantified using the 3D Object Counter v2.0 programme on FIJI using the same threshold value for all samples. Puncta with a size smaller than 0.1 μm³ were excluded, and the remaining puncta were considered to be true pS65‐Ub puncta. The data were imported into Prism, and a frequency distribution analysis was performed to obtain the number and mean volume of puncta.

Immuno‐electron microscopy

Thoraces were fixed in 0.1% glutaraldehyde and 4% paraformaldehyde in 0.1 M cacodylate buffer overnight at 4°C. After washing with 0.1 M cacodylate buffer, the samples were stained with 1% uranyl acetate for 1 h at 4°C. After washing with ddH₂O, the samples were dehydrated in a gradient series using 30, 50, 70, 95 and 100% ethanol solutions at −20°C. The samples were then infiltrated with an LR‐gold series and mounted in pure LR‐gold (Agar Scientific, AGR1284). After polymerisation by UV crosslinking, 90 nm sections were prepared and transformed into a block with 5% normal goat serum (NGS) for 1 h, before incubation with primary antibodies overnight at 4°C. After washing with 1% NGS, samples were incubated with 12 nm colloidal gold anti‐rabbit secondary antibodies for 30 min at room temperature (Jackson ImmunoResearch, 111‐205‐144). Samples were again washed with 1% NGS, and the sections were stained with 4% uranyl acetate and lead citrate before examination by TEM (Tecnai G2 Spirit TWIN transmission electron microscope, Thermo, USA).

Statistical analysis

Statistical analyses were performed using Prism (Version 9.1.0 (216)). For the analysis of mass spectrometry data presented in Fig 2, each Ub modification was analysed by ordinary one‐way ANOVA with the Šidák's correction for multiple comparisons (wild‐type untreated compared with Pink1 ⁻ and park ^−/− untreated and +/− paraquat comparison within each genotype, five comparisons total). The pS65‐Ub abundance presented in Fig 7C was analysed by ordinary one‐way ANOVA with the Dunnett's correction for multiple comparisons (each genotype compared with Atg5 ⁻, two comparisons total). For the analysis of the number of pS65‐Ub puncta in Fig 6H, data were first log‐transformed (Y = Ln(y + 1)) to account for heteroscedasticity in the raw data. The transformed data, as well as the raw data in Fig 6I, were analysed by one‐way ANOVA with the Dunnett's multiple comparisons.

Author contributions

Joanne L Usher: Conceptualization; data curation; formal analysis; validation; methodology; writing – original draft; writing – review and editing. Alvaro Sanchez‐Martinez: Formal analysis; visualization; methodology; writing – review and editing. Ana Terriente‐Felix: Formal analysis; methodology; writing – review and editing. Po‐Lin Chen: Formal analysis; visualization; methodology. Juliette J Lee: Formal analysis; methodology; writing – review and editing. Chun‐Hong Chen: Supervision; funding acquisition; investigation; writing – review and editing. Alexander J Whitworth: Conceptualization; formal analysis; supervision; funding acquisition; investigation; writing – original draft; project administration; writing – review and editing.

Disclosure and competing interests statement

The authors declare that they have no conflict of interest.

Supporting information

Expanded View Figures PDF

Dataset EV1

PDF+

EMBO reports (2022) 23: e53552

Data availability

This study includes no data deposited in external repositories. All data generated or analysed during this study are included in the manuscript and supporting files.