Ugo Mayor

A Novel Strategy to Isolate Ubiquitin Conjugates Reveals Wide Role for Ubiquitination during Neural Development

Ubiquitination has essential roles in neuronal development and function. Ubiquitin proteomics studies on yeast and HeLa cells have proven very informative, but there still is a gap regarding neuronal tissue-specific ubiquitination. In an organism context, direct evidence for the ubiquitination of neuronal proteins is even scarcer. Here, we report a novel proteomics strategy based on the in vivo biotinylation of ubiquitin to isolate ubiquitin conjugates from the neurons of Drosophila melanogaster embryos. We confidently identified 48 neuronal ubiquitin substrates, none of which was yet known to be ubiquitinated. Earlier proteomics and biochemical studies in non-neuronal cell types had identified orthologs to some of those but not to others. The identification here of novel ubiquitin substrates, those with no known ubiquitinated ortholog, suggests that proteomics studies must be performed on neuronal cells to identify ubiquitination pathways not shared by other cell types. Importantly, several of those newly found neuronal ubiquitin substrates are key players in synaptogenesis. Mass spectrometry results were validated by Western blotting to confirm that those proteins are indeed ubiquitinated in the Drosophila embryonic nervous system and to elucidate whether they are mono- or polyubiquitinated. In addition to the ubiquitin substrates, we also identified the ubiquitin carriers that are active during synaptogenesis. Identifying endogenously ubiquitinated proteins in specific cell types, at specific developmental stages, and within the context of a living organism will allow understanding how the tissue-specific function of those proteins is regulated by the ubiquitin system.

USP30 deubiquitylates mitochondrial Parkin substrates and restricts apoptotic cell death

Mitochondria play a pivotal role in the orchestration of cell death pathways. Here, we show that the control of ubiquitin dynamics at mitochondria contributes to the regulation of apoptotic cell death. The unique mitochondrial deubiquitylase, USP30, opposes Parkin‐dependent ubiquitylation of TOM20, and its depletion enhances depolarization‐induced cell death in Parkin‐overexpressing cells. Importantly, USP30 also regulates BAX/BAK‐dependent apoptosis, and its depletion sensitizes cancer cells to BH3‐mimetics. These results provide the first evidence for a fundamental role of USP30 in determining the threshold for mitochondrial cell death and suggest USP30 as a potential target for combinatorial anti‐cancer therapy.

Basal mitophagy is widespread in Drosophila but minimally affected by loss of Pink1 or parkin

The Parkinson’s disease factors PINK1 and parkin are strongly implicated in stress-induced mitophagy in vitro, but little is known about their impact on basal mitophagy in vivo. We generated transgenic Drosophila melanogaster expressing fluorescent mitophagy reporters to evaluate the impact of Pink1/parkin mutations on basal mitophagy under physiological conditions. We find that mitophagy is readily detectable and abundant in many tissues, including Parkinson’s disease–relevant dopaminergic neurons. However, we did not detect mitolysosomes in flight muscle. Surprisingly, in Pink1 or parkin null flies, we did not observe any substantial impact on basal mitophagy. Because these flies exhibit locomotor defects and dopaminergic neuron loss, our findings raise questions about current assumptions of the pathogenic mechanism associated with the PINK1/parkin pathway. Our findings provide evidence that Pink1 and parkin are not essential for bulk basal mitophagy in Drosophila. They also emphasize that mechanisms underpinning basal mitophagy remain largely obscure.

Ubiquitin profiling in liver using a transgenic mouse with biotinylated ubiquitin.

Ubiquitination is behind most cellular processes, with ubiquitin substrates being regulated variously according to the number of covalently conjugated ubiquitin molecules and type of chain formed. Here we report the first mammalian system for ubiquitin proteomics allowing direct validation of the MS-identified proteins. We created a transgenic mouse expressing biotinylated ubiquitin and demonstrate its use for the isolation of ubiquitinated proteins from liver and other tissues. The specificity and strength of the biotin-avidin interaction allow very stringent washes, so only proteins conjugated to ubiquitin are isolated. In contrast with recently available antibody-based approaches, our strategy allows direct validation by immunoblotting, therefore revealing the type of ubiquitin chains (mono or poly) formed in vivo. We also identify the conjugating E2 enzymes that are ubiquitin-loaded in the mouse tissue. Furthermore, our strategy allows the identification of candidate cysteine-ubiquitinated proteins, providing a strategy to identify those on a proteomic scale. The novel in vivo system described here allows broad access to tissue-specific ubiquitomes and can be combined with established mouse disease models to investigate ubiquitin-dependent therapeutical approaches.

Quantitative proteomic analysis of Parkin substrates in Drosophila neurons

BackgroundParkin (PARK2) is an E3 ubiquitin ligase that is commonly mutated in Familial Parkinson’s Disease (PD). In cell culture models, Parkin is recruited to acutely depolarised mitochondria by PINK1. PINK1 activates Parkin activity leading to ubiquitination of multiple proteins, which in turn promotes clearance of mitochondria by mitophagy. Many substrates have been identified using cell culture models in combination with depolarising drugs or proteasome inhibitors, but not in more physiological settings.MethodsHere we utilized the recently introduced BioUb strategy to isolate ubiquitinated proteins in flies. Following Parkin Wild-Type (WT) and Parkin Ligase dead (LD) expression we analysed by mass spectrometry and stringent bioinformatics analysis those proteins differentially ubiquitinated to provide the first survey of steady state Parkin substrates using an in vivo model. We further used an in vivo ubiquitination assay to validate one of those substrates in SH-SY5Y cells.ResultsWe identified 35 proteins that are more prominently ubiquitinated following Parkin over-expression. These include several mitochondrial proteins and a number of endosomal trafficking regulators such as v-ATPase sub-units, Syx5/STX5, ALiX/PDCD6IP and Vps4. We also identified the retromer component, Vps35, another PD-associated gene that has recently been shown to interact genetically with parkin. Importantly, we validated Parkin-dependent ubiquitination of VPS35 in human neuroblastoma cells.ConclusionsCollectively our results provide new leads to the possible physiological functions of Parkin activity that are not overtly biased by acute mitochondrial depolarisation.

Proteomic Analysis of the Ubiquitin Landscape in the Drosophila Embryonic Nervous System and the Adult Photoreceptor Cells.

Background Ubiquitination is known to regulate physiological neuronal functions as well as to be involved in a number of neuronal diseases. Several ubiquitin proteomic approaches have been developed during the last decade but, as they have been mostly applied to non-neuronal cell culture, very little is yet known about neuronal ubiquitination pathways in vivo. Methodology/Principal Findings Using an in vivo biotinylation strategy we have isolated and identified the ubiquitinated proteome in neurons both for the developing embryonic brain and for the adult eye of Drosophila melanogaster. Bioinformatic comparison of both datasets indicates a significant difference on the ubiquitin substrates, which logically correlates with the processes that are most active at each of the developmental stages. Detection within the isolated material of two ubiquitin E3 ligases, Parkin and Ube3a, indicates their ubiquitinating activity on the studied tissues. Further identification of the proteins that do accumulate upon interference with the proteasomal degradative pathway provides an indication of the proteins that are targeted for clearance in neurons. Last, we report the proof-of-principle validation of two lysine residues required for nSyb ubiquitination. Conclusions/Significance These data cast light on the differential and common ubiquitination pathways between the embryonic and adult neurons, and hence will contribute to the understanding of the mechanisms by which neuronal function is regulated. The in vivo biotinylation methodology described here complements other approaches for ubiquitome study and offers unique advantages, and is poised to provide further insight into disease mechanisms related to the ubiquitin proteasome system.

A comprehensive platform for the analysis of ubiquitin-like protein modifications using in vivo biotinylation

Post-translational modification by ubiquitin and ubiquitin-like proteins (UbLs) is fundamental for maintaining protein homeostasis. Efficient isolation of UbL conjugates is hampered by multiple factors, including cost and specificity of reagents, removal of UbLs by proteases, distinguishing UbL conjugates from interactors, and low quantities of modified substrates. Here we describe bioUbLs, a comprehensive set of tools for studying modifications in Drosophila and mammals, based on multicistronic expression and in vivo biotinylation using the E. coli biotin protein ligase BirA. While the bioUbLs allow rapid validation of UbL conjugation for exogenous or endogenous proteins, the single vector approach can facilitate biotinylation of most proteins of interest. Purification under denaturing conditions inactivates deconjugating enzymes and stringent washes remove UbL interactors and non-specific background. We demonstrate the utility of the method in Drosophila cells and transgenic flies, identifying an extensive set of putative SUMOylated proteins in both cases. For mammalian cells, we show conjugation and localization for many different UbLs, with the identification of novel potential substrates for UFM1. Ease of use and the flexibility to modify existing vectors will make the bioUbL system a powerful complement to existing strategies for studying this important mode of protein regulation.

Isolation of Ubiquitinated Proteins to High Purity from In Vivo Samples.

Ubiquitination pathways are widely used within eukaryotic cells. The complexity of ubiquitin signaling gives rise to a number of problems in the study of specific pathways. One problem is that not all processes regulated by ubiquitin are shared among the different cells of an organism (e.g., neurotransmitter release is only carried out in neuronal cells). Moreover, these processes are often highly temporally dynamic. It is essential therefore to use the right system for each biological question, so that we can characterize pathways specifically in the tissue or cells of interest. However, low stoichiometry, and the unstable nature of many ubiquitin conjugates, presents a technical barrier to studying this modification in vivo. Here, we describe two approaches to isolate ubiquitinated proteins to high purity. The first one favors isolation of the whole mixture of ubiquitinated material from a given tissue or cell type, generating a survey of the ubiquitome landscape for a specific condition. The second one favors the isolation of just one specific protein, in order to facilitate the characterization of its ubiquitinated fraction. In both cases, highly stringent denaturing buffers are used to minimize the presence of contaminating material in the sample.

Neuronal Proteomic Analysis of the Ubiquitinated Substrates of the Disease-Linked E3 Ligases Parkin and Ube3a

Both Parkin and UBE3A are E3 ubiquitin ligases whose mutations result in severe brain dysfunction. Several of their substrates have been identified using cell culture models in combination with proteasome inhibitors, but not in more physiological settings. We recently developed the bioUb strategy to isolate ubiquitinated proteins in flies and have now identified by mass spectrometry analysis the neuronal proteins differentially ubiquitinated by those ligases. This is an example of how flies can be used to provide biological material in order to reveal steady state substrates of disease causing genes. Collectively our results provide new leads to the possible physiological functions of the activity of those two disease causing E3 ligases. Particularly, in the case of Parkin the novelty of our data originates from the experimental setup, which is not overtly biased by acute mitochondrial depolarisation. In the case of UBE3A, it is the first time that a nonbiased screen for its neuronal substrates has been reported.

Multi-story Parkin

Rapid progress has been made in obtaining an inventory of Parkinson’s Disease (PD) associated genes. The challenge now is to understand their interactions and on which physiological pathways they converge. Two major themes have emerged; influence upon mitochondrial dynamics and on trafficking within the endocytic pathway [1]. Parkin (PARK2) is a ubiquitin E3-ligase, for which loss of function mutations lead to PD. Under basal conditions Parkin is largely held in an inactive state. It is recruited to mitochondria by another PD linked protein, the kinase PINK1, which specifically accumulates on depolarised mitochondria. Parkin then promotes mitophagy by ubiquitylating multiple mitochondrial proteins [2]. The rapid and dazzling progress in understanding this coupled recruitment and phosphorylation-dependent activation process has understandably taken centre stage. However, there remains no formal proof that defects in mitophagy lead to PD. Most cellular studies of mitophagy rely upon Parkin over-expression. Irrespective of this matter, it has become clear that Parkin may play more widespread roles that are associated with alternative pathophysiology. Broader links between Parkin and mitochondrial health have been made in both cardiac maturation and following myocardial infarction [3]. Parkin knockout mice placed on a high fat diet resist weight gain, steatohepatitis and insulin resistance. They also show increased susceptibility to Tuberculosis infection through a defect in xenophagy. Parkin is also an important tumour suppressor. Decreased PARK2 mRNA or focal deletions are found in more than a third of human cancers. This has been attributed to diminished ATP levels and increased ROS production leading indirectly to inactivation of PTEN [4]. Mutations in the same amino acid are found in cancer and PD. Discovery-based proteomic approaches offer the opportunity to identify E3-ligase substrates in an unbiased manner. Quantitation of protein levels may reflect differences in turnover rate but can also report indirect effects on transcription/translation. Alternatively ubiquitylated peptides can be directly detected by a signature diGly modification following trypsinisation. We have recently used an alternative method (BioUb) whereby a tagged form of ubiquitin is stably expressed and biotinylated in vivo. Ubiquitylated proteins from mice, flies or human cells can then be isolated by Neutravidin beads under stringent washing conditions. This provides greater peptide coverage of substrates although the actual site of ubiquitylation is not always identified [5]. Acute mitochondrial depolarisation provides the only known trigger for Parkin activation and therefore offers a means to facilitate the identification of substrates. Using this procedure in tissue cell culture models has revealed a broad swathe of mitochondrial substrates, with some bias towards modification with ubiquitin chains linked through Lys6 [6]. The overall picture is that a distributed ubiquitin signal across the mitochondrial surface comprised of many substrate proteins is sufficient for Parkin-dependent mitophagy. There remains a need to identify Parkin substrates under more physiological conditions. Comparative proteomic studies of Parkin -/vs wild type mice and fly brains have reported some differentially expressed proteins. However, these reports largely pre-date a step change in proteomics technology, that increases protein coverage by an order of magnitude. We have recently used the BioUb method to search for Parkin substrates in a fly model, that compares neuronal cells expressing either wild type Parkin or a ligase dead mutant [7]. Our results come with the caveat of over-expression but do not rely upon any acute drug treatments. Of >1000 proteins that were quantitated, 35 were identified as potential Parkin substrates. Some of these were mitochondrial proteins that had been identified Editorial