Loss of the deubiquitinase USP36 destabilizes the RNA helicase DHX33 and causes preimplantation lethality in mice

Deubiquitinases are proteases with a wide functional diversity that profoundly impact multiple biological processes. Among them, the ubiquitin-specific protease 36 (USP36) has been implicated in the regulation of nucleolar activity. However, its functional relevance in vivo has not yet been fully described. Here, we report the generation of an Usp36-deficient mouse model to examine the function of this enzyme. We show that Usp36 depletion is lethal in preimplantation mouse embryos, where it blocks the transition from morula to blastocyst during embryonic development. USP36 reduces the ubiquitination levels and increases the stability of the DEAH-box RNA helicase DHX33, which is critically involved in ribosomal RNA synthesis and mRNA translation. In agreement with this finding, O-propargyl-puromycin incorporation experiments, Northern blot, and electron microscopy analyses demonstrated the role of USP36 in ribosomal RNA and protein synthesis. Finally, we show that USP36 down-regulation alters cell proliferation in human cancer cells by inducing both apoptosis and cell cycle arrest, and that reducing DHX33 levels through short hairpin RNA interference has the same effect. Collectively, these results support that Usp36 is essential for cell and organism viability because of its role in ribosomal RNA processing and protein synthesis, which is mediated, at least in part, by regulating DHX33 stability.

uitin chains to target proteins (1)(2)(3)(4). There are six families of DUBs grouped according to sequence and structural similarities: ubiquitin-specific proteases (USPs), ubiquitin carboxylterminal hydrolases, ovarian-tumor proteases, Machado-Joseph disease protein domain proteases (MJDs), JAMM/MPN domain-associated metallopeptidases (JAMMs), and monocyte chemotactic protein-induced protein family (5). Over the last years, DUBs have attracted especial interest due to their wide functional diversity and profound impact on the regulation of multiple biological processes (2,6,7). The cysteine proteases of the USP family represent the largest group of DUBs, with more than 50 members in humans (8 -10). Among them, USP36 has been related to nucleolar activity regulation, through the deubiquitination of nucleophosmin/B23 and fibrillarin (11,12). Additionally, USP36 activity has also been implicated in autophagy activation (13), as well as in oxidative stress regulation by the stabilization of mitochondrial superoxide dismutase SOD2 (14). Interestingly, the Drosophila gene scrawny (scny), which encodes an USP family protein homologous to human USP36, has been described to inhibit H2B ubiquitination, repressing the premature expression in stem cells of key differentiation genes (15). The role of USP36 in the stabilization of the oncogene c-Myc has also been reported (16). Finally, it has been proposed that USP36 is critical for ribosome biogenesis through the regulation of RNA polymerase I stability (17,18).
Over the last years, several studies have shown that RNA helicases comprise the largest family of enzymes that regulate RNA metabolic processes, such as pre-mRNA splicing, RNA editing, mRNA translation, and ribosome biogenesis (19). In fact, it has been widely described that most RNA helicases contribute to several steps of RNA metabolism, coordinating and orchestrating gene expression patterns. The human genome encodes 15 DEAH-box RNA helicases, which are RNA-binding proteins with the ability to modify RNA structure through the hydrolysis of ATP. One of these family members, DHX33, has been implicated in nucleolar organization, ribosome biogene-sis, and protein synthesis. Thus, DHX33 promotes mRNA translation initiation by regulating elongation-competent 80S ribosome assembly (20) and modulates RNA polymerase I-directed transcriptional activity through its interaction with the DNA chromatin-modulating protein UBF (21).
In this work, the generation and characterization of mutant mice deficient in Usp36 revealed that this DUB is an essential enzyme for embryonic development. Additionally, this study provides evidence that Usp36 regulates ribosomal RNA and protein synthesis, and proposes a regulatory function for this protease on the ubiquitination status and stability of Dhx33. Altogether, these data identify USP36 as a deubiquitinating enzyme whose function in rRNA processing and mRNA translation is critical for cell and organism viability.

Usp36 deletion leads to embryonic lethality in mice
To study the biological function of the deubiquitinase USP36, the Usp36 gene was disrupted in mice by homologous recombination using a gene trap strategy (Fig. 1A). Heterozygosity for the mutation was confirmed by Southern blot analysis (Fig. 1B). Usp36 heterozygous mice were fertile and healthy with no obvious abnormalities. However, when these mice were intercrossed, no homozygous pups were detected at weaning (Fig. 1C), suggesting that the complete absence of Usp36 causes embryonic lethality. To further analyze how Usp36 deficiency affects embryogenesis, embryos were collected at different stages of development, from 2.5 to 14.5 days postcoitum (dpc).
As shown in Fig. 1C, homozygous mice were detected only at the morula stage. To characterize the role played by Usp36 during the critical transition from morula to blastocyst, embryos at 2.5 dpc were extracted and cultured in vitro. The next day, some of them had developed into blastocysts, whereas other remained at the morula stage. A high proportion of these morulae carried the homozygous mutation (Fig. 1D), whereas, by contrast, all the blastocysts were wildtype or heterozygous for Usp36.
To elucidate the mechanism by which Usp36 Ϫ/Ϫ blastocysts are unviable, TUNEL analysis was performed in these embryos. Only 1 of 4 wildtype and 2 of 5 heterozygous morulae were positive for TUNEL staining, whereas 4 of 4 Usp36-deficient arrested morulae showed apoptotic markers, suggesting that Usp36 deficiency activates apoptotic pathways that could impede the transition from morula to blastocyst. Taken together, these results demonstrate that Usp36 is essential for embryo development and that its deficiency induces an apoptotic response at the morula stage that leads to preimplantation lethality.

USP36 regulates DHX33 ubiquitination levels
To investigate the biochemical function of USP36, characterization of the proteomic profile of wildtype and Usp36-deficient cells was undertaken. With this purpose, quantitative mass spectrometry analysis using a tandem mass tag (TMT) technique was performed with protein extracts from wildtype and Usp36-heterozygous MEFs by triplicate. This proteomic Figure 1. Usp36 depletion leads to preimplantation lethality. A, schematic representation of the gene trap strategy used for the generation of the Usp36-deficient mouse model. B, Southern blot analysis performed on wildtype and Usp36-heterozygous mice. C, distribution of genotyped offspring and embryos from Usp36 ϩ/Ϫ intercrosses. Usp36 Ϫ/Ϫ embryos were not found beyond the morula stage. D, representative images of embryos obtained from heterozygous intercrosses of Usp36 ϩ/Ϫ mice obtained at 2.5 dpc (morulae) and cultured in vitro for 24 h (blastocysts, 3.5 dpc). E, TUNEL staining of 3.5 dpc embryos revealed apoptosis in Usp36-depleted morulae.

USP36 loss destabilizes DHX33 and causes embryonic lethality
approach allowed us to quantify over 7,200 proteins, revealing changes in 251 proteins ( Fig. 2A and Table S1). This analysis showed that Usp36 protein levels in Usp36 ϩ/Ϫ MEFs were about 70% of those in Usp36 ϩ/ϩ fibroblasts ( Fig. 2B and Table  S1). Similarly, the levels of Dhx33, a pivotal DEAH-box RNA helicase involved in RNA polymerase I-mediated transcription (21), were also reduced in Usp36 ϩ/Ϫ MEFs to 70% of control samples (Fig. 2C and Table S1). Dhx33 down-regulation in MEFs was validated by Western blot (Fig. 3A). Next, Usp36 ϩ/ϩ and Usp36 ϩ/Ϫ MEFs were treated with 1 M bortezomib during 18 h. As shown in Fig. 3A, proteasome inhibition induced the accumulation of Dhx33 in Usp36 ϩ/Ϫ MEFs, to the same extent as in wildtype MEFs, supporting that Dhx33 is regulated by the ubiquitin-proteasome system. To investigate if USP36 and DHX33 proteins interact each other, the human USP36 cDNA was cloned fused to the C-terminal of GFP and overexpressed into HEK-293T cells, together with pLVX-FLAG-DHX33. Co-immunoprecipitation assays confirmed that ectopically expressed GFP-USP36 could be detected in FLAG-tagged DHX33 immunoprecipitates (Fig. 3B). In agreement with these results, colocalization of USP36 and DHX33 was observed through immunofluorescence analysis in HeLa cells transfected with pLVX-FLAG-DHX33 and GFP-USP36 (Fig. 3C). Furthermore, a Duolink In Situ Proximity Ligation Assay (P-LISA) (22) showed that GFP-USP36 directly interacted with FLAG-tagged DHX33 protein (Fig. 3D), supporting the interaction between both proteins. These findings are consistent with the results from reported global proteomic studies (23, 24) (Fig. S1) and a yeast two-hybrid assay (25), which confirms the interaction between USP36 and DHX33.
To investigate whether USP36 regulates DHX33 levels through deubiquitination, HEK-293T cells were transfected with HA-ubiquitin and pLVX-FLAG-DHX33, with or without USP36 overexpression. After treatment with bortezomib, DHX33 was immunoprecipitated and its ubiquitination status was analyzed using both anti-HA and anti-DHX33 antibodies. As shown in Fig. 3E, USP36 overexpression reduced DHX33 ubiquitination and, accordingly, led to increased levels of ectopically expressed DHX33. On the contrary, HEK-293T transfected with an USP36-specific shRNA showed an increase in DHX33 ubiquitination levels (Fig. 3F). Altogether, these results support that USP36 stabilizes DHX33 through the regulation of its ubiquitination status.

Usp36 deficiency impairs mRNA translation and rRNA synthesis
The essentiality of Usp36 at an early stage of development demonstrates its involvement in critical biological functions. Among the molecular changes occurring at the transition from morula to blastocyst, rDNA transcription and nucleoli maturation have been reported to occur beyond the morula stage. In this regard, DHX33 has a pivotal role as a promoter of mRNA translation initiation (20). These facts suggested that USP36 deubiquitinase activity could have a role in protein synthesis, at least in part, through the stabilization of DHX33. To investigate in vivo the role of Usp36 in this physiological process, morulae at 2.5 dpc from intercrosses between Usp36-heterozygous mice were extracted. At 3.5 dpc, embryos were treated with O-propargyl-puromycin (OP-Puro), an alkyne analog of puromycin that forms covalent conjugates with nascent polypeptide chains and Usp36 ϩ/Ϫ MEFs that were analyzed by mass spectrometry using TMT technique. Data are represented as the ratio Usp36 ϩ/ϩ versus Usp36 ϩ/Ϫ , which was calculated as the mean of the normalized signal obtained in Usp36 ϩ/ϩ MEFs divided by the mean of the normalized signal obtained in Usp36 ϩ/Ϫ MEFs (n ϭ 3 biological replicates). B and C, Usp36 and Dhx33 protein levels from wildtype and Usp36-heterozygous MEFs, assessed by mass spectrometry, two-tailed Student's t test (**, p Ͻ 0.01).

USP36 loss destabilizes DHX33 and causes embryonic lethality
and can be visualized by copper(I)-catalyzed azide-alkyne cycloaddition (26). As it is shown in Fig. 4A, deficiency of Usp36 at the morula stage decreased OP-Puro incorporation, suggesting an impairment of mRNA translation.
Because DHX33 has also been described as a mediator or rRNA synthesis (21), the role of USP36 in this phenomenon was next analyzed. With this purpose, 45S pre-rRNA levels of USP36-depleted and control HCT116 colorectal cancer cells were quantified by Northern blot analysis, finding that downregulation of USP36 significantly decreased 45S pre-rRNA accumulation (Fig. 4, B and C). Moreover, silencing of USP36 also caused more alterations in pre-rRNA processing, such as diminished 21S levels (Fig. 4, B and D), and consequently increased the 28S/18S ratio (Fig. 4E). Because these results support a direct role of USP36 in the regulation of rRNA pro-cessing, the hypothetical impact of Usp36 down-regulation on the nucleolar structure was investigated by analyzing MEFs processed for electron microscopy. Interestingly, Usp36-heterozygous MEFs showed nucleoli significantly larger than wildtype MEFs (Fig. 5A). Although Usp36-heterozygous MEFs also exhibited an increase in nucleus size (Fig. 5B), the ratio of nucleolar to nuclear area was consistently higher when compared with that of wildtype cells (Fig. 5, C and D). Finally, the effect of USP36 down-regulation was examined through the meta-analysis of the data derived from a genomewide shRNA screen in 216 cancer cell lines from multiple tumor types (27), finding that the antiproliferative effects of silencing USP36 correlated positively with the antiproliferative effects of silencing genes involved in translation and ribosome biogenesis (Fig. 5E).   Fluorescence baseline ϭ 20. Data are represented as mean Ϯ S.E. and statistical significance was assessed by using a non-parametric Mann-Whitney-Wilcoxon test (*, p Ͻ 0.05). B, representative image of Northern blot analysis of RNA from HCT116 cells transduced with control (pLKO.1) or two USP36-specific shRNAs (shUSP36.446 and shUSP36.447) and densitometry quantification of six technical replicates. The densitometry quantification of each band was normalized to pLKO.1 signal from the same blot to compare the data. C and D, Northern blot analysis showed that down-regulation of USP36 decreased the levels of 45S pre-rRNA (C) and 21S pre-rRNA (D), two-tailed Student's t test (*, p Ͻ 0.05; **, p Ͻ 0.01). E, bioanalyzer experiments demonstrate that downregulation of USP36 increased the 28S/18S ratio, two-tailed Student's t test (**, p Ͻ 0.01).

USP36 loss destabilizes DHX33 and causes embryonic lethality
Collectively, these results strongly support that USP36, similarly to DHX33, is essential for the regulation of rRNA synthesis and mRNA translation processes, whose optimal function is required during embryonic development. In this sense, any dysfunctional mutation within Usp36 could lead to a devastating effect, as demonstrated by the finding of preimplantation lethality in Usp36-deficient mice.

USP36 depletion induces p53-independent apoptosis and p53-dependent cell cycle arrest
To investigate the mechanistic intricacies underlying the lethality caused by Usp36 depletion, HCT116 colorectal cancer cells were chosen as an experimental model. As a first approach, the proliferation rate of human HCT116 colorectal cancer cells transduced with either control or USP36-specific shRNAs was analyzed. As shown in Fig. 6A, USP36 depletion with three independent shRNAs, either separately or in combination, decreased the proliferation rate of HCT116 cells. To demonstrate the absence of noticeable off-target effects, a lentiviral vector encoding a shRNA-resistant USP36 cDNA was transduced into HCT116 cells. USP36 protein levels were analyzed by Western blot (Fig. 6B). As shown in Fig. 6C, ectopic expres-sion of this cDNA abrogated the effects of the shRNA-mediated USP36 silencing in this cell line, corroborating the specificity of our approach. Moreover, down-regulation of USP36 in HCT116 cells reduced their ability to grow in soft agar (Fig. 6D). Accordingly, USP36 depletion with three independent shRNAs induced apoptosis, as evidenced by an increase in the percentage of Annexin V-positive cells (Fig. 6E). In agreement with these results, Western blot analysis of PARP cleavage also confirmed apoptosis induction upon USP36 down-regulation (Fig.  6F), corroborating the results obtained from TUNEL analysis performed in Usp36 Ϫ/Ϫ morulae. To investigate whether USP36 depletion alters cell cycle progression, the cell cycle profile of wildtype and USP36-silenced cells was analyzed by flow cytometry. Interestingly, USP36 down-regulation induced an increased accumulation of cells in the G 1 phase compared with controls, suggesting a G 1 cell cycle arrest (Fig. 6G). Altogether, these results indicate that USP36 deficiency alters cell proliferation, inducing both apoptosis and cell cycle arrest.
To determine whether p53 plays a role in the apoptotic response induced by USP36 silencing, USP36 was next downregulated in p53-deficient HCT116 cells. Remarkably, p53 deficiency did not rescue the cell lethality caused by USP36 down-

USP36 loss destabilizes DHX33 and causes embryonic lethality
regulation (Fig. 7A). Furthermore, USP36-depleted cells underwent some degree of apoptosis even in the absence of p53 (Fig. 7, B and C), indicating that USP36 down-regulation induces a p53-independent apoptotic response. By contrast, a significant increase in G 1 population cells was not observed after USP36 silencing, although the accumulation of cells in G 2 phase of the cycle was noteworthy (Fig. 7D). Altogether, these results suggest that USP36 depletion induces both p53-dependent cell cycle arrest and p53-independent apoptosis. Next, we decided to investigate the possible relevance of these mechanisms in the embryonic lethality caused by Usp36 deficiency. With this purpose, p53 Ϫ/Ϫ mice were crossed with Usp36 ϩ/Ϫ mice, which were then inbred to obtain Usp36 ϩ/Ϫ p53 Ϫ/Ϫ mice. Finally, these mice were intercrossed and the resulting offspring genotyped and analyzed. As shown in Fig. 7E, no Usp36 Ϫ/Ϫ p53 Ϫ/Ϫ mice were detected at weaning. Next, 2.5 dpc embryos were collected and cultured in vitro for 2 days to obtain blastocysts. However, no Usp36 Ϫ/Ϫ p53 Ϫ/Ϫ blastocysts were found. By contrast, we obtained some Usp36 Ϫ/Ϫ p53 Ϫ/Ϫ arrested morulae (Fig. 7E). Therefore, these results demonstrate that Usp36 Ϫ/Ϫ embryos exhibit preimplantation lethality at the morula stage in a p53-independent manner.
Finally, to further evaluate if DHX33 down-regulation mimics the biological effect caused by USP36 depletion, an empty vector (pLKO.1) and two independent DHX33-specific shR-NAs (shDHX33.61 and shDHX33.62) were transduced into HCT116 cells. As observed upon USP36 targeting, silencing of DHX33 in HCT116 cells significantly reduced their prolifera-tion rate (Fig. 8A). Western blot analyses were performed to analyze DHX33 protein levels (Fig. 8B). Furthermore, cells transduced with DHX33-specific shRNAs showed an increase in apoptosis, as evidenced by analyzing both PARP cleavage (Fig. 8C) and the percentage of Annexin V-FITC positive cells (Fig. 8D). Altogether, these results corroborate that DHX33 down-regulation phenocopies USP36 silencing in human cancer cells, and support our hypothesis that USP36 is essential for cell viability due to its critical role in rRNA processing and protein synthesis by regulating the ubiquitination levels of DHX33.

Discussion
In this work, we have generated an Usp36-deficient mouse model that shows a lethal phenotype during the preimplantation stage, demonstrating the essentiality of this deubiquitinase for embryonic development. Moreover, we have evaluated the functional role of USP36 by analyzing the proteomic profile of wildtype and heterozygous mice for this deubiquitinase. Our studies have shown that USP36 directly interacts with DHX33, a DEAH-box RNA helicase involved in RNA polymerase I-mediated transcription and mRNA translation initiation (20,21). Furthermore, we have described for the first time the regulatory role of USP36 on the ubiquitination levels and stability of DHX33.
So far, many nucleolar proteins associated with RNA processing have been described as essential for proliferation and survival of human cells (28,29). Furthermore, several mouse

USP36 loss destabilizes DHX33 and causes embryonic lethality
models deficient for ribosome biogenesis regulators also show embryonic lethality at the preimplantation stage. Thus, Pescadillo-deficient mouse embryos arrest at the morula stage of development (30). Similarly, mice carrying a truncated mutation in the gene encoding polypeptide B of RNA polymerase I display nucleolus disruption and apoptotic cell death at the morula stage (31). Furthermore, fibrillarin-deficient mouse embryos also die beyond the morula stage (32) and Surf6 knockdown by RNAi results in development arrest at the morula stage (33). Finally, UBF and nucleostemin are also essential for embryo development beyond morula (34,35). The results derived from our study have demonstrated in vivo that Usp36 is a nucleolar protein related to ribosome biogenesis and RNA processing regulation. Thus, OP-Puro treatment has demonstrated that Usp36-deficient morulae exhibit impaired protein synthesis, highlighting the importance of Usp36 in this cellular process. Moreover, Northern blot analysis has shown that USP36 deficiency induce aberrations in rRNA synthesis. According to USP36 nucleolar activity, Usp36-heterozygous MEFs exhibit alterations in their nucleolar structure, a phenomenon that has been tightly correlated with ribosome biogenesis aberrations (36). In agreement with these results, metaanalysis of a genome-scale shRNA screen of 216 cancer cell lines of multiple histological origins has revealed that the antiproliferative effects of silencing USP36 and genes involved in translation and ribosome biogenesis positively correlate, supporting the direct role of this deubiquitinating enzyme in both physiological processes. Altogether, these results strongly support the critical role of USP36 in rDNA transcription and mRNA translation, at least in part, by regulating the stability of DHX33. However, further studies on other putative USP36 substrates will be required to completely understand its functional role in these biological processes.
It is widely accepted that inappropriate ribosome biogenesis, perturbedrRNAprocessing,andribosomeassemblyinduceapoptosis and block cell proliferation (36,37). Accordingly, the results obtained in TUNEL experiments have revealed that Usp36 deficiency induces apoptosis at the preimplantation stage of development. Furthermore, we have demonstrated that USP36 down-regulation with specific shRNAs reduces the proliferation and the anchorage-independent growth of human cancer cells, through both p53-dependent and -independent mechanisms. Thus, even in the absence of p53, USP36 ablation leads to cell death, both in human colorectal cancer cells and in preimplantation mouse embryos. Additionally, we have demonstrated that reducing DHX33 protein levels by an USP36independent mechanism also abrogates the proliferation of HCT116 cells and increases their apoptotic rate, supporting our hypothesis that the USP36 role in regulating DHX33 stability is critically required for cell viability.
In summary, the generation of Usp36-deficient mice has allowed us to identify USP36 as an essential gene for cell and organism viability. We describe for the first time the regulatory role of USP36 on the ubiquitination levels and stability of DEAH-box RNA helicase DHX33, which constitutes a new mechanism of USP36 functional implication in rRNA processing and protein synthesis.

Generation of Usp36 ؊/؊ mice and genotyping
Usp36-mutant embryonic stem cells were purchased from EUKOMP (EUCE00141e03, from the 129P2/OlaHsd mouse

USP36 loss destabilizes DHX33 and causes embryonic lethality
strain). Two different clones of targeted embryonic stem cells were microinjected into C57BL/6 mouse blastocysts to produce chimeric mice that were then subsequently crossed with C57BL/6 mice to generate Usp36-heterozygous mice. Usp36knockout first allele was identified by Southern blotting and PCR. For Southern blot experiments, the recombinant allele (3,640 bp) was detected by a specific probe in BclI-digested genomic DNA. For PCR genotyping, genomic DNA from tail samples was used under the following conditions: denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s, 30 cycles. The following primers were used for genotyping: wildtype-specific forward 5Ј-ccaggtctctcaaggactagga-3Ј, mutation-specific forward 5Ј-gaatgaaagaccccacctgtag-3Ј, and common reverse 5Ј-aggatgccaaagagcatgtagt-3Ј. p53-deficient mice were kindly provided by Dr. M. Serrano.

Cell culture
Cancer cell lines HEK-293T, HeLa, and HCT116 were purchased from the American Type Culture Collection. Cells were routinely maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 units/ml of penicillin, and 100 g/ml of streptomycin (Life Technologies). HCT116 p53 Ϫ/Ϫ cells were kindly provided by Dr. B. Vogelstein. MEFs were extracted and cultured as previously described (38).

Cell cycle analysis
HCT116 cells transduced with control and USP36-specific shRNAs were washed twice with PBS and permeabilized with 1 ml of 70% ethanol at Ϫ20°C added drop by drop. After 30 min on ice, cells were washed twice in PBS and resuspended in 1 ml of 20 g/ml of propidium iodide (PI), 200 g/ml of RNase in PBS for 15 min in the dark at room temperature. Finally, the cell cycle was analyzed by flow cytometry (10,000 events per sample, Cytomics FC500, Beckman Coulter).

shRNA lentiviral infection
USP36-specific shRNAs, DHX33-specific shRNAs, and empty vector (pLKO.1) (Open Biosystems, Thermo Scientific) were packaged in HEK-293T cells using a vesicular stomatitis virus glycoprotein-based package system. Viral supernatant was collected at 24 h and added in a 1:3 dilution to previously seeded HCT116 cells, supplemented with 5 mg/ml of Polybrene (Millipore). HCT116 stably transduced cells were selected with puromycin at a final concentration of 1 g/ml.

Cell-proliferation assay
To quantify cell proliferation, 5,000 cells per well were seeded (n ϭ 6) into 96-well plates and a Cell Titer 96 NonRadioactive cell proliferation kit was used following the manufacturer's instructions (Promega Corp.).

shRNA-resistant USP36 expression
Silent mutations were introduced into the shUSP36.447 target sequence of USP36 cDNA cloned in pCDH-puro, using the Q5 Site-directed Mutagenesis Kit (New England Biolabs) and the primers: 5Ј-ggcttacagccgacatcttgccatcccaggt-3Ј and 5Ј-aagacgccatcgaagacagcagacaggcccgga-3Ј. Lentiviral particles were packaged as described above. HCT116 cells were simultaneously transduced with the USP36-specific shRNAs shUSP36.447 and the shRNA-resistant USP36 cDNA or the empty vector and their proliferation was assessed without prior antibiotic selection. It is noteworthy that HCT116 cells were simultaneously transduced with two different plasmids (USP36-specific shRNAs and a shRNA-resistant USP36 cDNA clone), which shared the same antibiotic resistance, leading to a smaller fraction of silenced cells. Therefore, the anti-proliferative effect of this experiment was smaller and observed later when compared with that of Fig. 6A.

Soft agar colony formation
1,000 cells per well were seeded, maintained during 8 days at 37°C and 5% CO 2 and anchorage-independent growth was analyzed with CytoSelect TM 96-Well In Vitro Tumor Sensitivity Assay kit (Cell Biolabs), following the manufacturer's instructions.

Northern blot
A total of 10 g of denatured RNA was separated by electrophoresis on 1.2% agarose gels and transferred to Hybond Nϩ (Amersham Biosciences). Blots were prehybridized at 42°C for 3 h in 50% formamide, 5ϫ SSPE (1ϫ ϭ 150 mM NaCl, 10 mM NaH 2 PO 4 , 1 mM EDTA, pH 7.4), 10ϫ Denhardt's solution, 1.4% SDS, and 100 g/ml of denatured herring sperm DNA. Then, the blot was hybridized with a genomic DNA probe synthetized by PCR using the following primers: 5Ј-gagaactcgggagggagac-3Ј and 5Ј-gaccctcgagacgcccta-3Ј for 24 h under the same conditions. Blots were washed twice with 2ϫ SSC, 0.05% SDS for 20 min at room temperature. Finally, blots were washed three times with 0.1ϫ SSC, 0.1% SDS for 20 min at 50°C and exposed to autoradiography.

Western blot analysis
Western blot analyses were performed as previously described (39)

Annexin V analysis and PI staining
HCT116 cells transduced with either control, USP36-specific shRNAs or DHX33-specific shRNAs, were labeled with Annexin V-FITC (Inmunostep) and PI for 10 min in the dark. The percentage of positive Annexin V cells was analyzed by Flow Cytometry (15,000 events per sample).

TUNEL and OP-Puro treatment
Embryos derived from Usp36-heterozygous mice intercrosses were flushed at day 2.5 dpc and cultured overnight. For OP-Puro experiments, embryos at 3.5 dpc were incubated with 50 M OP-Puro for 1 h at 37°C. Then, they were washed with USP36 loss destabilizes DHX33 and causes embryonic lethality PBS and fixed in 4% paraformaldehyde in PBS for 5 min at room temperature. They were subsequently washed with PBS and permeabilized in 0.1% Triton X-100, 0.1% sodium citrate for 2 min. Embryos were washed with 2% BSA in PBS and the azidealkyne cycloaddition was performed using the Click-iT Cell Reaction Buffer Kit (Life Technologies) and azide conjugated to Alexa Fluor 488 (Life Technologies) at a 5 M final concentration. After 30 min of incubation at room temperature, morulae were washed with 2% BSA in PBS and images were taken using a Leica TCS-SP8X laser confocal microscope (HCX PL FLU-OTAR L ϫ20/0.40 DRY, zoom 2.70). For TUNEL staining, embryos were fixed and permeabilized as for OP-Puro experiments. Embryos were then washed with PBS and incubated with terminal deoxynucleotidyltransferase and fluorescein-labeled dUTP for terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) (In Situ Cell Death Detection kit; Roche Applied Science) at 37°C for 1 h. Embryos were washed with PBS before they were analyzed with a Leica DMI6000B inverted microscope and photographed with a Leica DFC480 camera (Leica Image Manager software). After analysis, embryos were individually transferred into PCR tubes and processed for genotyping.

Electron microscopy
MEFs were collected and fixed in 3% glutaraldehyde in cacodylate buffer. Then, they were transferred into agar, and postfixed with 1% osmium tetroxide for 1 h. We then rinsed the samples 3 times in Sorensen's buffer and agar blocks were dehydrated in graded ethanol and infiltrated in 50% propylene oxide, 50% resin. After 1 h, samples were embedded in a resin-based mold for polymerization overnight at 37°C. Finally, ultra-thin sections (85 nm) were taken from each experiment and analyzed on a Jeol (JEM-1011) transmission electron microscope at 60 kV.

Sample preparation for mass spectrometry
Protein samples were prepared by triplicate from wildtype and Usp36-heterozygous MEFs in 150-mm dishes. Cells were washed twice with ice-cold PBS, scraped in PBS, spun down, and lysed in SDS buffer (2% SDS, 150 mM NaCl, 50 mM HEPES pH 8.0, with added complete protease inhibitors (Roche Applied Science) and PhosSTOP phosphatase inhibitors (Roche Applied Science). 25-Guage needles were used to homogenize the cell pellet and lyse it. Cell debris was spun down for 15 min at 13,000 ϫ g at 4°C. Protein concentration was determined using the BCA assay (Thermo Fisher Scientific). To reduce and alkylate the disulfide bonds, 300 g of lysate was incubated with 5 mM tris(2-carboxyethyl)phosphine for 30 min and 14 mM iodoacetamide for 30 min at room temperature in the dark. Proteins were precipitated using methanol/chloroform precipitation and resuspended in 8 M urea buffer (8 M urea, 50 mM HEPES, pH 8.0, 1 mM CaCl 2 ). Protein extracts were diluted to 4 M urea using 50 mM HEPES, pH 8.0, and digested for 2 h at 37°C with LysC (Wako) at a 1:250 LysC/ protein ratio. Then, diluted to 2 M urea, and incubated overnight at 37°C with LysC. The next day, urea was further diluted to 1 M and trypsin (Promega) was added at a 1:50 trypsin/protein ratio for 6 h at 37°C. The proteolytic peptides were then acidified with formic acid (FA) to a pH of Ͻ2 and desalted using Sep-Pak C 18 solid-phase extraction cartridges (Waters), following the manufacturer's protocol. Dried peptides were resuspended in 200 mM HEPES, pH 8.0, the concentration was determined using the micro-BCA protein assay (Thermo Fisher Scientific) and the isobaric labeling of the peptides was performed using 6-plex TMT reagents (Thermo Fisher Scientific). Thus, 100 g of peptides at 1 g/l were labeled with 0.2 mg of each TMT reagent (previously dissolved in acetonitrile (ACN)) for 2 h at room temperature. Finally, the reactions were quenched with 0.3% hydroxylamine (Sigma) at room temperature for 15 min. The resultant labeled peptides were combined equally and ACN diluted to 3% with 3% FA (final pH should be Ͻ2). The peptides were desalted again with a Sep-Pak C 18 column and dried down in the SpeedVac. The peptides were then subjected to basic-pH reverse-phase HPLC fractionation as described (40) and fractionated into 24 fractions. Half of them were dried down in the SpeedVac, dissolved in 3% FA, 3% ACN, desalted via StageTip, dried again in the SpeedVac, and resuspended in 10 l of 3% FA, 3% ACN just before mass spectrometry analysis.

Mass spectrometry analysis
All spectra were acquired on an Orbitrap Fusion (Thermo Fisher Scientific) coupled to a Proxeon EASY-nLC II LC pump (Thermo Fisher Scientific) as described previously (40). Briefly, peptides were separated on an in-house packed column with ϳ35 cm of GP-18 resin (1.8 m, 200 Å; Sepax) using a gradient of 135 min from 6 to 26% ACN in 0.125% FA at a flow rate of ϳ350 nl/min. The scan sequence in the mass spectrometer began with FTMS1 spectra (resolution of 120 k; mass range 400 -1400 m/z; automatic gain control (AGC) target 2 ϫ 10 5 , maximum injection time of 100 ms) followed by the selection of the 10 most intense ions for MS/MS with a dynamic exclusion of 90 s. ITMS2 spectra were collected in the ion trap with a maximum injection time of 150 ms with an AGC target of 4 ϫ 10 3 and CID collision energy of 35%. FTMS3 spectra were collected using the multi-notch method previously described (41). Briefly, to create TMT reporter ions, a synchronous-precursorselection-MS3 scan was collected on the top 10 most intense ions in the MS2 spectrum. Synchronous-precursor-selection-MS3 precursors were fragmented by high energy collision-induced dissociation and analyzed using the Orbitrap (NCE ϭ 55%, AGC ϭ 5 ϫ 10 4 , maximum injection time ϭ 150 ms, and resolution ϭ 15 K). The data analysis were performed with a SEQUEST-based in-house software pipeline (42). In short, mass spectra were searched against the mouse Uniprot database (February 2014) and a reverse decoy database. Precursor ion tolerance was set at 20 ppm and product ion tolerance at 0.9 Da. Addition of a TMT tag (ϩ229.1629 Da) on lysine residues and peptide N termini, and cysteine carbamidomethylation (ϩ57.0215 Da) were added as static modifications, and methionine oxidation (ϩ15.9949 Da) was set as a variable modification. Peptide-spectrum matches were adjusted to a 1% false discovery rate using a linear discriminant analysis (42), quantified by extracting the signal to noise ratio and proteins were further collapsed to a final protein level false discovery rate of USP36 loss destabilizes DHX33 and causes embryonic lethality 1% (40, 41). Statistical analysis was performed using two-tailed Student's t test.
Immunoprecipitation HEK-293T cells were transfected with pLVX-FLAG-DHX33 and pEGFP-USP36 or pEGFP empty vector and immunoprecipitation was performed as previously described (43). For ubiquitination experiments, HEK-293T cells were transfected with HA-ubiquitin and pLVX-FLAG-DHX33, together with pEGFP/pEGFP-USP36 or pLKO.1/USP36-specific shRNA plasmids, treated with 500 nM bortezomib during 16 h and lysed with RIPA buffer (100 mM Tris-HCl, pH 7.4, 150 mM NaCl, 10 mM EDTA, 1% sodium deoxycholate, 1% Triton X-100, 0.1% SDS and protease and phosphatase inhibitors). Then, protein extracts were precleared for 4 h at 4°C with Dynabeads TM Protein G (Life Technologies), released from beads and incubated in rotation with anti-FLAG antibody overnight at 4°C. Beads were next washed three times with co-IP buffer and incubated in rotation with antibody-conjugated protein extracts 4 h at room temperature. Immunoprecipitates and input samples were resolved by SDS-PAGE and immunostained with anti-HA and anti-DHX33 antibodies.

P-LISA and imunofluorescent staining
The interaction between USP36 and DHX33 was detected using P-LISA (Sigma). HeLa cells were seeded on coverslips and co-transfected with pEGFP-USP36 and pLVX-FLAG-DHX33 plasmids. After 24 h, coverslips were fixed with 4% paraformaldehyde for 10 min, permeabilized with 0.5% Triton X-100 for 5 min, and blocked with Duolink blocking buffer for 30 min at 37°C in a humidity chamber. Incubation with anti-FLAG and anti-USP36 antibodies was carried out overnight at 4°C in a humidity chamber. Incubation with secondary probes, ligation, and amplification steps were all performed according to the manufacturer's instructions. For immunofluorescent staining, HeLa cells transfected with pEGFP-USP36 and pLVX-FLAG-DHX33 plasmids were fixed, permeabilized as indicated above, and blocked with 15% goat serum for 1 h at room temperature. Incubation with anti-FLAG and anti-GFP antibodies was performed overnight at 4°C in a humidity chamber. Secondary antibody incubation was carried out 1 h at room temperature. Then, cells were washed twice with PBS for 5 min and coverslips were mounted with ProLong TM Gold Antifade Mountant with DAPI (Invitrogen, Thermo Fisher Scientific). Positive cells were visualized in an Axioplan-2 fluorescence microscope (Zeiss) under a ϫ100 oil immersion objective.