Identification of substrates for human deubiquitinating enzymes (DUBs): An up-to-date review and a case study for neurodevelopmental disorders.

Similar to the reversal of kinase-mediated protein phosphorylation by phosphatases, deubiquitinating enzymes (DUBs) oppose the action of E3 ubiquitin ligases and reverse the ubiquitination of proteins. A total of 99 human DUBs, classified in 7 families, allow in this way for a precise control of cellular function and homeostasis. Ubiquitination regulates a myriad of cellular processes, and is altered in many pathological conditions. Thus, ubiquitination-regulating enzymes are increasingly regarded as potential candidates for therapeutic intervention. In this context, given the predicted easier pharmacological control of DUBs relative to E3 ligases, a significant effort is now being directed to better understand the processes and substrates regulated by each DUB. Classical studies have identified specific DUB substrate candidates by traditional molecular biology techniques in a case-by-case manner. Lately, single experiments can identify thousands of ubiquitinated proteins at a specific cellular context and narrow down which of those are regulated by a given DUB, thanks to the development of new strategies to isolate and enrich ubiquitinated material and to improvements in mass spectrometry detection capabilities. Here we present an overview of both types of studies, discussing the criteria that, in our view, need to be fulfilled for a protein to be considered as a high-confidence substrate of a given DUB. Applying these criteria, we have manually reviewed the relevant literature currently available in a systematic manner, and identified 650 high-confidence substrates of human DUBs. We make this information easily accessible to the research community through an updated version of the DUBase website (https://ehubio.ehu.eus/dubase/). Finally, in order to illustrate how this information can contribute to a better understanding of the physiopathological role of DUBs, we place a special emphasis on a subset of these enzymes that have been associated with neurodevelopmental disorders.

Biotin-Based Strategies to Explore the World of Ubiquitin and Ubiquitin-Like Modifiers.

A complex code of cellular signals is mediated by ubiquitin and ubiquitin-like (Ub/UbL) modifications on substrate proteins. The so-called Ubiquitin Code specifies protein fates, such as stability, subcellular localization, functional activation or suppression, and interactions. Hundreds of enzymes are involved in placing and removing Ub/UbL on thousands of substrates, while the consequences of modifications and the mechanisms of specificity are still poorly defined. Challenges include rapid and transient engagement of enzymes and Ub/UbL interactors, low stoichiometry of modified versus non-modified cellular substrates, and protease-mediated loss of Ub/UbL in lysates. To decipher this complexity and confront the challenges, many tools have been created to trap and identify substrates and interactors linked to Ub/UbL modification. This review focuses on an assortment of biotin-based tools developed for this purpose (for example BioUbLs, UbL-ID, BioE3, BioID), taking advantage of the strong affinity of biotin-streptavidin and the stringent lysis/washing approach allowed by it, paired with sensitive mass-spectrometry-based proteomic methods. Knowing how substrates change during development and disease, the consequences of substrate modification, and matching substrates to particular UbL-ligating enzymes will contribute new insights into how Ub/UbL signaling works and how it can be exploited for therapies.

HERC2 deficiency activates C-RAF/MKK3/p38 signalling pathway altering the cellular response to oxidative stress.

HERC2 gene encodes an E3 ubiquitin ligase involved in several cellular processes by regulating the ubiquitylation of different protein substrates. Biallelic pathogenic sequence variants in the HERC2 gene are associated with HERC2 Angelman-like syndrome. In pathogenic HERC2 variants, complete absence or marked reduction in HERC2 protein levels are observed. The most common pathological variant, c.1781C > T (p.Pro594Leu), encodes an unstable HERC2 protein. A better understanding of how pathologic HERC2 variants affect intracellular signalling may aid definition of potential new therapies for these disorders. For this purpose, we studied patient-derived cells with the HERC2 Pro594Leu variant. We observed alteration of mitogen-activated protein kinase signalling pathways, reflected by increased levels of C-RAF protein and p38 phosphorylation. HERC2 knockdown experiments reproduced the same effects in other human and mouse cells. Moreover, we demonstrated that HERC2 and RAF proteins form molecular complexes, pull-down and proteomic experiments showed that HERC2 regulates C-RAF ubiquitylation and we found out that the p38 activation due to HERC2 depletion occurs in a RAF/MKK3-dependent manner. The displayed cellular response was that patient-derived and other human cells with HERC2 deficiency showed higher resistance to oxidative stress with an increase in the master regulator of the antioxidant response NRF2 and its target genes. This resistance was independent of p53 and abolished by RAF or p38 inhibitors. Altogether, these findings identify the activation of C-RAF/MKK3/p38 signalling pathway in HERC2 Angelman-like syndrome and highlight the inhibition of RAF activity as a potential therapeutic option for individuals affected with these rare diseases.

The ubiquitin ligase Ariadne-1 regulates neurotransmitter release via ubiquitination of NSF.

Ariadne-1 (Ari-1) is an E3 ubiquitin-ligase essential for neuronal development, but whose neuronal substrates are yet to be identified. To search for putative Ari-1 substrates, we used an in vivo ubiquitin biotinylation strategy coupled to quantitative proteomics of Drosophila heads. We identified 16 candidates that met the established criteria: a significant change of at least twofold increase on ubiquitination, with at least two unique peptides identified. Among those candidates, we identified Comatose (Comt), the homologue of the N-ethylmaleimide sensitive factor (NSF), which is involved in neurotransmitter release. Using a pull-down approach that relies on the overexpression and stringent isolation of a GFP-fused construct, we validate Comt/NSF to be an ubiquitination substrate of Ari-1 in fly neurons, resulting in the preferential monoubiquitination of Comt/NSF. We tested the possible functional relevance of this modification using Ari-1 loss-of-function mutants, which displayed a lower rate of spontaneous neurotransmitter release due to failures at the presynaptic side. By contrast, evoked release in Ari-1 mutants was enhanced compared with controls in a Ca2+-dependent manner without modifications in the number of active zones, indicating that the probability of release per synapse is increased in these mutants. This phenotype distinction between spontaneous and evoked release suggests that NSF activity may discriminate between these two types of vesicle fusion. Our results thus provide a mechanism to regulate NSF activity in the synapse through Ari-1-dependent ubiquitination.

Impaired proteostasis in rare neurological diseases.

Rare diseases are classified as such when their prevalence is 1:2000 or lower, but even if each of them is so infrequent, altogether more than 300 million people in the world suffer one of the ∼7000 diseases considered as rare. Over 1200 of these disorders are known to affect the brain or other parts of our nervous system, and their symptoms can affect cognition, motor function and/or social interaction of the patients; we refer collectively to them as rare neurological disorders or RNDs. We have focused this review on RNDs known to have compromised protein homeostasis pathways. Proteostasis can be regulated and/or altered by a chain of cellular mechanisms, from protein synthesis and folding, to aggregation and degradation. Overall, we provide a list comprised of above 215 genes responsible for causing more than 170 distinct RNDs, deepening on some representative diseases, including as well a clinical view of how those diseases are diagnosed and dealt with. Additionally, we review existing methodologies for diagnosis and treatment, discussing the potential of specific deubiquitinating enzyme inhibition as a future therapeutic avenue for RNDs.

A Proteomic Approach for Systematic Mapping of Substrates of Human Deubiquitinating Enzymes.

The human genome contains nearly 100 deubiquitinating enzymes (DUBs) responsible for removing ubiquitin moieties from a large variety of substrates. Which DUBs are responsible for targeting which substrates remain mostly unknown. Here we implement the bioUb approach to identify DUB substrates in a systematic manner, combining gene silencing and proteomics analyses. Silencing of individual DUB enzymes is used to reduce their ubiquitin deconjugating activity, leading to an increase of the ubiquitination of their substrates, which can then be isolated and identified. We report here quantitative proteomic data of the putative substrates of 5 human DUBs. Furthermore, we have built a novel interactive database of DUB substrates to provide easy access to our data and collect DUB proteome data from other groups as a reference resource in the DUB substrates research field.

Quantitative proteomics reveals neuronal ubiquitination of Rngo/Ddi1 and several proteasomal subunits by Ube3a, accounting for the complexity of Angelman syndrome.

Angelman syndrome is a complex neurodevelopmental disorder caused by the lack of function in the brain of a single gene, UBE3A. The E3 ligase coded by this gene is known to build K48-linked ubiquitin chains, a modification historically considered to target substrates for degradation by the proteasome. However, a change in protein abundance is not proof that a candidate UBE3A substrate is indeed ubiquitinated by UBE3A. We have here used an unbiased ubiquitin proteomics approach, the bioUb strategy, to identify 79 proteins that appear more ubiquitinated in the Drosophila photoreceptor cells when Ube3a is over-expressed. We found a significantly high number of those proteins to be proteasomal subunits or proteasome-interacting proteins, suggesting a wide proteasomal perturbation in the brain of Angelman patients. We focused on validating the ubiquitination by Ube3a of Rngo, a proteasomal component conserved from yeast (Ddi1) to humans (DDI1 and DDI2), but yet scarcely characterized. Ube3a-mediated Rngo ubiquitination in fly neurons was confirmed by immunoblotting. Using human neuroblastoma SH-SY5Y cells in culture, we also observed that human DDI1 is ubiquitinated by UBE3A, without being targeted for degradation. The novel observation that DDI1 is expressed in the developing mice brain, with a significant peak at E16.5, strongly suggests that DDI1 has biological functions not yet described that could be of relevance for Angelman syndrome clinical research.

Ube3a, the E3 ubiquitin ligase causing Angelman syndrome and linked to autism, regulates protein homeostasis through the proteasomal shuttle Rpn10.

Ubiquitination, the covalent attachment of ubiquitin to a target protein, regulates most cellular processes and is involved in several neurological disorders. In particular, Angelman syndrome and one of the most common genomic forms of autism, dup15q, are caused respectively by lack of or excess of UBE3A, a ubiquitin E3 ligase. Its Drosophila orthologue, Ube3a, is also active during brain development. We have now devised a protocol to screen for substrates of this particular ubiquitin ligase. In a neuronal cell system, we find direct ubiquitination by Ube3a of three proteasome-related proteins Rpn10, Uch-L5, and CG8209, as well as of the ribosomal protein Rps10b. Only one of these, Rpn10, is targeted for degradation upon ubiquitination by Ube3a, indicating that degradation might not be the only effect of Ube3a on its substrates. Furthermore, we report the genetic interaction in vivo between Ube3a and the C-terminal part of Rpn10. Overexpression of these proteins leads to an enhanced accumulation of ubiquitinated proteins, further supporting the biochemical evidence of interaction obtained in neuronal cells.

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.

Linoleic acid improves PIEZO2 dysfunction in a mouse model of Angelman Syndrome.

Angelman syndrome (AS) is a neurogenetic disorder characterized by intellectual disability and atypical behaviors. AS results from loss of expression of the E3 ubiquitin-protein ligase UBE3A from the maternal allele in neurons. Individuals with AS display impaired coordination, poor balance, and gait ataxia. PIEZO2 is a mechanosensitive ion channel essential for coordination and balance. Here, we report that PIEZO2 activity is reduced in Ube3a deficient male and female mouse sensory neurons, a human Merkel cell carcinoma cell line and female human iPSC-derived sensory neurons with UBE3A knock-down, and de-identified stem cell-derived neurons from individuals with AS. We find that loss of UBE3A decreases actin filaments and reduces PIEZO2 expression and function. A linoleic acid (LA)-enriched diet increases PIEZO2 activity, mechano-excitability, and improves gait in male AS mice. Finally, LA supplementation increases PIEZO2 function in stem cell-derived neurons from individuals with AS. We propose a mechanism whereby loss of UBE3A expression reduces PIEZO2 function and identified a fatty acid that enhances channel activity and ameliorates AS-associated mechano-sensory deficits.

Mass Spectrometry-Based Characterization of Ub- and UbL-Modified Proteins.

Regulation by ubiquitin (Ub) and ubiquitin-like (UbL) modifiers can confer their substrate proteins a myriad of assignments, such as inducing protein-protein interactions, the internalization of membrane proteins, or their degradation via the proteasome. The underlying code regulating those diverse endpoints appears to be based on the topology of the ubiquitin chains formed.Experimental characterization of the specific regulation mediated by Ub and UbLs is not trivial. The substoichiometric levels of Ub- and UbL-modified proteins greatly limit their analytical detection in a background of more abundant proteins. Therefore, modified proteins or peptides must be enriched prior to any downstream detection analysis. For that purpose, we recently developed a GFP-tag based isolation strategy. Here we illustrate the usefulness of combining GFP-tag isolation strategy with mass spectrometry (MS) to identify Ub- and UbL-modified residues within the GFP-tagged protein, as well as to uncover the types of Ub and UbL chains formed.

How to Inactivate Human Ubiquitin E3 Ligases by Mutation.

E3 ubiquitin ligases are the ultimate enzymes involved in the transfer of ubiquitin to substrate proteins, a process that determines the fate of the modified protein. Numerous diseases are caused by defects in the ubiquitin-proteasome machinery, including when the activity of a given E3 ligase is hampered. Thus, inactivation of E3 ligases and the resulting effects at molecular or cellular level have been the focus of many studies during the last few years. For this purpose, site-specific mutation of key residues involved in either protein interaction, substrate recognition or ubiquitin transfer have been reported to successfully inactivate E3 ligases. Nevertheless, it is not always trivial to predict which mutation(s) will block the catalytic activity of a ligase. Here we review over 250 site-specific inactivating mutations that have been carried out in 120 human E3 ubiquitin ligases. We foresee that the information gathered here will be helpful for the design of future experimental strategies.

Detailed Dissection of UBE3A-Mediated DDI1 Ubiquitination.

The ubiquitin E3 ligase UBE3A has been widely reported to interact with the proteasome, but it is still unclear how this enzyme regulates by ubiquitination the different proteasomal subunits. The proteasome receptor DDI1 has been identified both in Drosophila photoreceptor neurons and in human neuroblastoma cells in culture as a direct substrate of UBE3A. Here, we further characterize this regulation, by identifying the UBE3A-dependent ubiquitination sites and ubiquitin chains formed on DDI1. Additionally, we found one deubiquitinating enzyme that is capable of reversing the action of UBE3A on DDI1. The complete characterization of the ubiquitination pathway of an UBE3A substrate is important due to the role of this E3 ligase in rare neurological disorders as Angelman syndrome.

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.

Significant transcriptional changes in 15q duplication but not Angelman syndrome deletion stem cell-derived neurons.

Background: The inability to analyze gene expression in living neurons from Angelman (AS) and Duplication 15q (Dup15q) syndrome subjects has limited our understanding of these disorders at the molecular level. Method: Here, we use dental pulp stem cells (DPSC) from AS deletion, 15q Duplication, and neurotypical control subjects for whole transcriptome analysis. We identified 20 genes unique to AS neurons, 120 genes unique to 15q duplication, and 3 shared transcripts that were differentially expressed in DPSC neurons vs controls. Results: Copy number correlated with gene expression for most genes across the 15q11.2-q13.1 critical region. Two thirds of the genes differentially expressed in 15q duplication neurons were downregulated compared to controls including several transcription factors, while in AS differential expression was restricted primarily to the 15q region. Here, we show significant downregulation of the transcription factors FOXO1 and HAND2 in neurons from 15q duplication, but not AS deletion subjects suggesting that disruptions in transcriptional regulation may be a driving factor in the autism phenotype in Dup15q syndrome. Downstream analysis revealed downregulation of the ASD associated genes EHPB2 and RORA, both genes with FOXO1 binding sites. Genes upregulated in either Dup15q cortex or idiopathic ASD cortex both overlapped significantly with the most upregulated genes in Dup15q DPSC-derived neurons. Conclusions: Finding a significant increase in both HERC2 and UBE3A in Dup15q neurons and significant decrease in these two genes in AS deletion neurons may explain differences between AS deletion class and UBE3A specific classes of AS mutation where HERC2 is expressed at normal levels. Also, we identified an enrichment for FOXO1-regulated transcripts in Dup15q neurons including ASD-associated genes EHPB2 and RORA indicating a possible connection between this syndromic form of ASD and idiopathic cases.

Quantitative proteomic analysis of Parkin substrates in Drosophila neurons.

BACKGROUND: Parkin (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. METHODS: Here 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. RESULTS: We 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. CONCLUSIONS: Collectively our results provide new leads to the possible physiological functions of Parkin activity that are not overtly biased by acute mitochondrial depolarisation.

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.

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.