The E3 ubiquitin ligase UBR5 interacts with the H/ACA ribonucleoprotein complex and regulates ribosomal RNA biogenesis in embryonic stem cells

UBR5 is an E3 ubiquitin ligase involved in distinct processes such as transcriptional regulation and development. UBR5 is highly upregulated in embryonic stem cells (ESCs), whereas its expression decreases with differentiation, suggesting a role for UBR5 in ESC function. However, little is known about how UBR5 regulates ESC identity. Here, we define the protein interactome of UBR5 in ESCs and find interactions with distinct components of the H/ACA ribonucleoprotein complex, which is required for proper maturation of ribosomal RNA (rRNA). Notably, loss of UBR5 induces an abnormal accumulation of rRNA processing intermediates, resulting in diminished ribosomal levels. Consequently, lack of UBR5 triggers an increase in p53 levels and a concomitant decrease in cellular proliferation rates. Thus, our results indicate a link between UBR5 and rRNA maturation.

Edited by Michael Ibba UBR5 is an E3 ubiquitin ligase involved in distinct processes such as transcriptional regulation and development. UBR5 is highly upregulated in embryonic stem cells (ESCs), whereas its expression decreases with differentiation, suggesting a role for UBR5 in ESC function. However, little is known about how UBR5 regulates ESC identity. Here, we define the protein interactome of UBR5 in ESCs and find interactions with distinct components of the H/ACA ribonucleoprotein complex, which is required for proper maturation of ribosomal RNA (rRNA). Notably, loss of UBR5 induces an abnormal accumulation of rRNA processing intermediates, resulting in diminished ribosomal levels. Consequently, lack of UBR5 triggers an increase in p53 levels and a concomitant decrease in cellular proliferation rates. Thus, our results indicate a link between UBR5 and rRNA maturation.
Keywords: embryonic stem cells; protein-protein interactions; ribosomal RNA; ribosome; ubiquitin ligases UBR5 is an E3 ubiquitin ligase which belongs to the HECT (homologous to the E6-AP carboxy terminus) ligase family [1]. This 300 kDa-enzyme is mostly localized in the nucleus [2] and contains several functional domains. Besides its HECT-ubiquitin ligase domain, UBR5 also has a UBR box finger-like domain, which targets proteins containing the destabilizing N-degron signal toward degradation [3]. Moreover, it contains a PABC [polyadenylate-binding protein (PABP) C terminus] domain, also known as MLLE domain, which confers UBR5 the ability to bind mRNA and modulate eukaryotic translation initiation [4]. Notably, UBR5 exhibits E3-independent activity as a transcriptional cofactor for the progesterone receptor and acts as a binding partner for distinct proteins such as GW182, CHK2, and DUBA [5][6][7]. UBR5 was first discovered as a tumor suppressor in a progestin-modifying screening using breast cancer cells [1]. In these lines, several studies evidenced a role of UBR5 in cell cycle progression and tumorigenesis [8,9]. UBR5 is also an important mediator of the DNA damage response, since it is a substrate of DNA damage response kinases such as ATM [10,11]. Furthermore, UBR5 has a role in transcriptional regulation, as it binds and promotes the transactivation of the progesterone receptor [2], ubiquitinates the CDK9 subunit of the positive transcription elongation factor b [12], and regulates the miRNA pathway [13]. In addition, Ubr5 À/À mice embryos are not viable, indicating that this protein is essential for development [14].
Recently, we have shown that UBR5 is highly expressed in human embryonic stem cells (hESCs) and its expression decreases when hESC differentiate, suggesting a role of UBR5 in ESC identity [15]. In these lines, a study reported that Ubr5 knockdown results in significant loss of pluripotency markers in mouse ESCs (mESCs) [16]. Here we examine the role of UBR5 in ESCs by defining its interactome using immunoprecipitation assays followed by quantitative proteomics. We find that UBR5 interacts with the H/ACA ribonucleoprotein (RNP) complex, which participates in the proper maturation and processing of ribosomal RNA (rRNA). We further characterize the role of UBR5 in this system by analyzing pre-rRNA processing and find that the levels of several pre-rRNA intermediates are altered in Ubr5 À/À ESCs. Finally, we show a decreased translational profile in UBR5-defective cells accompanied by increased p53 levels and diminished proliferation rates.

Methods hESC and mESC lines and culture
The H9 (WA09) hESC line was obtained from the WiCell Research Institute and was maintained on Geltrex (ThermoFisher Scientific, Waltham, MA, USA) using mTeSR1 (Stem Cell Technologies, Cologne, Germany). Undifferentiated hESC colonies were passaged using a solution of dispase (2 mgÁmL À1 ), and scraping the colonies with a glass pipette. The mESC AN-13 mESC line was obtained from the Haplobank at the Institute of Molecular Biotechnology GmbH. The mESC was cultured in noncoated plates using DMEM supplemented with 7.25% FBS, 1% Pen/ Strep, 1% nonessential amino acids, 1% Glutamax, 1% Sodium pyruvate, 50 lM b-mercaptoethanol and human LIF (12 ng/mL). The cell lines used in this study were tested for mycoplasma contamination at least once every three weeks. No mycoplasma contamination was detected. Research involving hESCs was performed with approval of the German Federal competent authority (Robert Koch Institute).

Lentiviral infection of hESCs
Lentivirus (LV)-nontargeting shRNA control, LV-UBR5 shRNA #1 (TRCN0000003411), and LV-UBR5 shRNA #2 (TRCN0000226458) in pLKO.1-puro vector were obtained from Mission shRNA (Sigma, Hamburg, Germany). Transient infection experiments were performed as follows. H9 colonies growing on Geltrex were individualized using Accutase. Hundred thousand cells were plated on Geltrex plates and incubated with mTesR1 medium containing 10 lM ROCK inhibitor for 1 day. Then, cells were infected with 5 lL of concentrated lentivirus. Cells were fed with fresh media the day after to remove the virus. After 1 day, cells were selected for lentiviral integration using 2 lgÁmL À1 puromycin (ThermoFisher Scientific).
CRISPR/Cas9-mediated KO of mESCs UBR5 gene knockout was done in mouse embryonic stem cells carrying a haploid chromosome set [ [17]]. These cells become diploid over time and mutations thus remain homozygous. For genome engineering CRISPR/Cas9 was used as previously described [18]. Sequences for small guide RNAs were designed online (http://crispor.org) and purchased from Sigma (F Guide 1 : CACCGTAAATGATTTAC CATACGGT, R Guide 1 : AAACACCGTATGGTAAATCA TTTAC, F Guide 2 : CACCGATTGCTTTAAAACTCCAC TT, R Guide 2 : AAACAAGTGGAGTTTTAAAGCAATC). The primers were cloned into the Cas9-GFP expressing plasmid PX458 (Addgene #48138, gift from Feng Zhang). A combination of guides and the Cas9-expressing plasmid were transfected using lipofectamine 3000 (Thermo Fisher Scientific) according to manufacturer's instructions. Twenty-four hours post transfection a mixed population of haploid and diploid ES cells were stained with 10 lg/mL Hoechst 33342 (Thermo Fisher Scientific) for 30 min and sorted according to their DNA content and GFP expression on a FACSAria Fusion sorter (BD, Franklin Lakes, NJ, USA). Haploid GFP-positive cells were single cell sorted into 96-well plates. Emerging clones were transferred to 24-well plates 7 days later and genotyped following DNA extraction (DNA extraction solution, Epicentre Biotechnologies, Madison, WI, USA) using the following primers; F: GAGAC CCGCCTGTTTGTTTT, R: CCCAATTGATTCTCTGAG CCA. Sanger sequencing of PCR products was performed at Eurofins Genomics GmbH, Ebersberg, Germany. UBR5 knockout clones were selected for diploid cells (FACS after Hoechst staining) prior to further experiments.

Immunoprecipitation of UBR5 for interactome analysis
The immunoprecipitation was performed under the conditions described in [20]. Briefly, hESCs were lysed in RIPA buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton x-100, 1% sodium deoxycholate, 1 mM EDTA, 1 mM PMSF) supplemented with protease inhibitor cocktail (Roche) and centrifuged at 13 000 g for 15 minutes at 4°C. About 350 lg of protein was incubated with UBR5 Antibody (Cell Signaling, Frankfurt am Main, Germany, #8755, 1 : 50) or FLAG antibody as a control (Sigma, F7425, 4 lg). Subsequently, samples were incubated with 100 lL of lMACS Micro Beads for 1 h at 4°C and loaded to precleared lMACS column (#130-042-701). Beads were washed three times with 50 mM Tris (pH 7.5) buffer containing 150 mM NaCl, 5% glycerol, and 0.05% Triton and then washed five times with 50 mM Tris (pH 7.5) and 150 mM NaCl. Then, columns were subjected to in-column tryptic digestion containing 7.5 mM ammonium bicarbonate, 2 M urea, 1 mM DTT, and 5 ngÁmL À1 trypsin. Digested peptides were eluted using 50 lL of elution buffer 1 containing 2 M urea, 7.5 mM Ambic, and 5 mM IAA two times. Digests were incubated over night at room temperature and samples were stage-tipped the next day for label-free quantitative proteomics. All samples were analyzed on a Q-Exactive Plus (Thermo Scientific) mass spectrometer that was coupled to an EASY nLC 1200 UPLC (Thermo Scientific). All mass spectrometric raw data were processed with Maxquant and the resulting output was processed using Perseus. Significant differences between the groups were assessed using Student's t-test. A permutation-based FDR approach was applied to correct for multiple testing (FDR < 0.2 was considered significant).

Protein immunoprecipitation for interaction analysis
HEK293 cells were lysed in RIPA buffer supplemented with protease inhibitor cocktail. Lysates were homogenized by passing 10 times through a 27-gauge (27G) needle attached to a 1 mL syringe and centrifuged at 13 000 g for 15 min at 4°C. After preclearing the supernatant with Protein A agarose beads (Pierce, Schwerte, Germany), the samples were incubated overnight with UBR5 antibody (Cell Signaling, #8755, 1 : 50) on the overhead shaker at 4°C. Subsequently, samples were incubated with 30 lL of Protein A beads for 1 h at room temperature. After this incubation, samples were centrifuged 5 min at 5000 g and the pellet was washed three times with RIPA buffer. For elution of the proteins, the pellet was incubated with 2x Laemmli Buffer, boiled for 5 min and centrifuged 5 min at maximum speed. The supernatant was taken and loaded onto a SDS/PAGE gel for western blot analysis.

RNA immunoprecipitation-quantitative RT-PCR
RIP experiments with anti-UBR5 (Cell Signaling, #8755) and control anti-FLAG antibodies (Sigma, F7425) in HEK293T cells overexpressing GFP-UBR5 were performed following the protocol described in ref [21] with some modifications [22]. Cells were pelleted by centrifugation at 1000 g for 10 min at 4°C and washed several times with ice cold PBS. The final cell pellet was resuspended with an equal volume of polysome lysis buffer [100 mM KCl, 5 mM MgCl 2 , 10 mM HEPES (pH 7.0), 0.5% NP40 (Sigma)] supplemented with 1 mM DTT, 100 per units RNase Out (Invitrogen, Schwerte, Germany), 400 lM VRC (New England BioLabs, Frankfurt am Main, Germany), and protease inhibitor. Cell lysates were incubated on ice for 10 min and homogenized through syringe needle. Antibody coating of protein A beads was prepared by preswelling Protein A sepharose beads (Thermoscientific) in NT2 buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1 mM MgCl 2 , 0.05% NP40) supplemented with 5% BSA to a final ratio of 1 : 5 for at least 1 h at 4°C prior to use. Then, the antibody was added to bead slurry and incubated for 2 h at 4°C. Immediately before use, antibody-coated beads were washed (five times) with ice cold NT2 buffer followed by resuspension of beads in ice cold NT2 buffer supplemented with 200 units of an RNase inhibitor, 400 lM vanadyl ribonucleoside complexes, 1 mM DTT, and 20 mM EDTA. The cell lysate was mixed with antibody-coated beads and incubated 2 h at room temperature. The beads were washed five times with ice cold NT2 buffer and finally washed with NT2 buffer with 1% Triton X-100. RNA extraction was done from the immunoprecipitated pellet using RNAbee (Tel-Test Inc., Friendswood, TX, USA). cDNA was generated using qScript cDNA Super-Mix (Quantabio, Beverly, MA, USA). SybrGreen real-time qPCR experiments were performed with a 1 : 20 dilution of cDNA using a CFC384 Real-Time System (Bio-Rad) following the manufacturer's instructions. Data were analyzed with the comparative 2DDCt method (RNA relative fold enrichment in anti-UBR5 RIP over RIP performed with control FLAG antibody) after normalization to the corresponding input values. See Table S1 for details about the primers used for this assay.

Bromodeoxyuridine proliferation assay
The mESCs were incubated with media containing 10 lMÁmL À1 bromodeoxyuridine (BrdU) for the indicated times. Cells were fixed with formaldehyde 4% in PBS. Then, cells were permeabilized with 0.2% Triton X-100 in PBS for 10 min and blocked with 3% BSA-PBS for 1 h at room temperature. 2N HCl was added for 15 min at room temperature. After this, cells were incubated in 0.1 M sodium tetra-borate for 15 min at room temperature. We performed overnight incubation with rabbit anti-BrdU (ABD Serotech, OBT0030, 1 : 1000) at 4°C followed by incubation with an anti-rat-AlexaFluor 546 Cross-Adsorbed (ThermoFisher Scientific, #A11081, 1 : 500) for 1 h. Hoechst 33342 was used to visualize nuclei.

UBR5 interacts with the H/ACA ribonucleoprotein complex
UBR5 is upregulated in both hESCs and mESCs compared with their differentiated counterparts [15,16,20], suggesting a role for this protein in ESC function. To gain insights into the function of UBR5, we performed co-immunoprecipitation experiments of endogenous UBR5 in hESCs followed by single-shot proteomic analysis. We identified 101 interactors of UBR5 (Data S1). Gene Ontology (GO) analysis of UBR5 potential interactors indicated enrichment for proteins involved in biological processes such as spliceosome assembly, small nuclear RNA (snRNA) modification, and rRNA pseudouridine synthesis (Data S1 and Fig. S1). Interestingly, we found that UBR5 interacts with the proteins (i.e., DKC1, GAR1, NOP10, and NHP2) of the nucleolar H/ACA RNP complex (Data S1).
The H/ACA RNP box is formed by four different proteins bound to small nucleolar RNAs (snoRNAs) (Fig. 1A), and modulates the modification of rRNAs as well as other RNAs by catalyzing their pseudouridylation (i.e., conversion of uridine to pseudouridine) [23,24]. The noncoding H/ACA snoRNAs act as adaptors that link the catalytic protein DKC1 to its targets [25,26]. The three other proteins of the H/ACA RNP box -GAR1, NOP10, and NHP2act as scaffolding proteins and they are essential for the proper functioning of the complex [27,28]. We identified DKC1, GAR1, NOP10, and NHP2 as putative binding partners of UBR5 in our analysis, suggesting  a role of UBR5 in the regulation of the H/ACA RNP complex (Data S1). To further assess this interaction, we overexpressed UBR5 together with GAR1, NHP2, or DKC1 in HEK293 human cells and performed coimmunoprecipitation assays followed by western blot. These experiments confirmed the interaction of UBR5 with GAR1, NHP2, and DKC1 (Fig. 1B). Prompted by these results, we analyzed the intracellular localization of UBR5 and GAR1, NHP2, or DKC1 in HEK293. Notably, these components of the H/ACA RNP complex co-localized with UBR5 in the nucleus, further supporting our interactome analysis (Fig. 1C).     With the strong interaction between UBR5 and proteins of the H/ACA RNP complex, we examined whether UBR5 protein also pulls down snoRNAs by RNA immunoprecipitation (RIP) assays. We assessed four snoRNAs of the H/ACA RNP complex and found interaction of UBR5 protein with SNORA73A/ U17 and SNORA5A (Fig. 1D). In contrast, we did not observe significant interaction of UBR5 with snoRNAs of the C/D box RNP, a distinct nucleolar complex that catalyzes 2 0 O-methylation of rRNA (Fig. 1D). Since UBR5 protein pulled down snoRNAs of the H/ACA RNP, we asked whether the interaction of UBR5 with proteins of this complex is mediated by RNA. To assess this hypothesis, we performed co-immunoprecipitation assays treating the samples with RNase A prior to immunoprecipitation with UBR5 antibody. Notably, the interactions with GAR1, NHP2, and DKC1 proteins remained upon RNase A treatment (Fig. 1E), indicating that this binding is not due to secondary, RNA-mediated interactions. Given the role of UBR5 in polyubiquitination and proteasomal degradation of distinct substrates [15], we asked whether its interaction with H/ACA proteins triggers their degradation through the proteasome. For this purpose, we overexpressed GAR1, NHP2, or DKC1 with either wild-type UBR5 (UBR5 WT ) or a catalytically dead mutant of UBR5 (UBR5 ΔHECT ) which lacks the ubiquitin ligase activity. We assessed the levels of DKC1, NHP2, and GAR1 by western blot and found no differences upon overexpression of either UBR5 WT or UBR5 ΔHECT , indicating that UBR5 does not target them for proteasomal degradation ( Fig. 2A). In addition, we generated two independent hESC lines with reduced levels of UBR5 (hESC UBR5 KD ) and found no changes in the endogenous levels of DKC1 (Fig. 2B). Likewise, endogenous DKC1 levels were not affected in a mESC line in which UBR5 was knockedout using CRISPR/Cas9 (mESC Ubr5 KO ) (Fig. 2C). Thus, our results indicate that UBR5 interacts with distinct proteins of the H/ACA ribonucleoprotein complex but it does not induce their proteasomal degradation.

UBR5 is required for the correct maturation of ribosomal RNA
Given the interaction of UBR5 with the H/ACA RNP complex, we asked whether UBR5 has a role in the regulation of rRNA in ESCs. For this purpose, we assessed the maturation of rRNAs by northern blot analysis of their precursors in hESC UBR5 KD and mES-C UBR5 KO . In eukaryotes, the ribosome contains four distinct rRNAs: the 18S rRNA is present in the 40S small ribosomal subunit, while the large 60S subunit contains the 5.8S, 28S, and 5S rRNAs. The 18S, 5.8S, and 28S rRNA are transcribed by RNA polymerase I as a single rRNA precursor transcript, the 47S/45S pre-ribosomal RNA (pre-RNA) [29,30]. Ribosomal proteins and assembly factors bind co-transcriptionally to the pre-rRNA, which then undergoes a series of modifications and processing steps including endonucleolytic cleavage and exonucleolytic trimming [31]. The maturation process can be assessed by radioactive labeled probes (ITS1 and ITS2) which bind to internal sequences of the pre-RNA. These sequences flank the mature rRNA and are eliminated during the sequential cleavages of the pre-RNA [32] (Fig. 3A).
rRNA precursors of the 18S rRNA were detected with an ITS1 probe (Fig. 3B,C). In hESC UBR5 KD , the amount of 34/30S rRNA relative to the 47/45S pre-rRNA was significantly reduced compared with control hESCs. The 30S rRNA is further processed to 21S rRNA by additional cleavage steps at positions 01 and 1. Interestingly, the levels of 21S rRNA remained unchanged in hESC UBR5 KD when compared with control hESCs. Since the 47S pre-rRNA can also be processed involving an alternative pathway via the 41S rRNA, the reduced cleavage into 30S rRNA could potentially be compensated. In mESC Ubr5 KO , the reduction of 34/30S rRNA compared to 47/45S pre-rRNA was even more dramatic. Moreover, the levels of 20S rRNA were also affected in these cells. However, the amounts of 41S rRNA, the intermediate of the alternative pathway of cleavage at position 2, were unchanged. Using a probe targeting ITS2, we examined the 32S and 12S precursor rRNAs in hESC and mESCs (Fig. 3D,E). Although in hESCs UBR5 KD , the levels of 12S rRNA relative to 47/45S pre-rRNA were significantly increased, we did not observe these effects in mESC Ubr5 KO . In both hESCs and mESCs, the levels of 32S rRNA were not significantly changed upon loss of UBR5 (Fig. 3D,E). Taken together, our results indicate that dysfunction of UBR5 could affect rRNA maturation, particularly conversion into 34/30S rRNA.

UBR5 deficiency impairs translational profile and triggers a stress response in mESCs
Since alterations in rRNA might originate defects in the proper assembly of the ribosomes, we assessed ribosome pools in mESC Ubr5 KO (Fig. 4A). When compared with wild-type ESCs, these cells exhibited a decrease in 40S and 60S, as well as in the monosome and polysome fraction, suggesting lower translational rates in the absence of UBR5. The levels of 40S (i.e., RPS27) and 60S (i.e., RPL7A) proteins in mESC UBR5 KO cells were also altered in different ribosome fractions, although the decrease in RPL7A was more evident than RPS27 (Fig. 4B). Alterations in rRNA synthesis can trigger a cellular stress response [33,34], which, in turn, increases p53 levels. Notably, we found increased p53 levels in mESC Ubr5 KO , suggesting the activation of a stress response in these cells (Fig. 4C). p53-mediated response toward stress promotes cell cycle arrest to maintain cellular homeostasis [35]. Importantly, we observed that the proliferation rates in mESC Ubr5 KO cells were significantly diminished (Fig. 4D). We then asked whether the lack of UBR5 sensitized mESCs toward other stressors which inhibit translation. For this purpose, we exposed mESC Ubr5 KO cells to either heat stress (Fig. 4E) or cycloheximide (Fig. 4F) and determine the degree of apoptosis by measuring cleaved caspase-3 levels. However, we did not find notable differences in the apoptosis rates in the absence of UBR5. Altogether, our results suggest that UBR5 participates in the maturation of rRNA, a process necessary for proper ribosomal assembly. In the absence of UBR5, ESCs exhibit defects in their ribosomal content and increased p53-mediated stress response, resulting in lower proliferation rates.

Discussion
Growing evidence indicates that UBR5 is involved in the regulation of distinct RNA-related pathways [4,12,13]. In mESCs, UBR5 is required for the proper function of the miRNA machinery by directing the scaffolding protein GW182 to the miRNA complex through a ubiquitin ligase-independent mechanism [13]. This process results in the recruitment of other downstream effectors like DDX6 and Tob1/2, which promote the deadenylation and consequent destabilization of target mRNAs [36][37][38]. Moreover, UBR5 phosphorylation by p90 ribosomal S6 kinase (p90RSK) is necessary for the effect of UBR5 on miRNA repression activity, adding an additional regulatory level [39]. The effects of UBR5 on the miRNA pathway could be evolutionary conserved among species as similar effects were also reported in Drosophila melanogaster [40]. Additionally, upon UV-induced DNA damage, UBR5 represses the transcription at the damaged sites by inhibiting the elongation of polymerase II [41]. UBR5 also interacts with the phosphorylated form of PIH1D1, which is a subunit of the co-chaperone complex R2TP [42]. R2TP is implicated in the assembly of large complexes such as the H/ACA RNP complex [43]. However, a direct link between UBR5 and the H/ACA RNP complex has not been previously reported. Notably, we find an interaction between UBR5 and the four proteins of the nucleolar H/ACA complex (i.e., DKC1, GAR1, NOP10, and NHP2). Besides these proteins, the H/ACA RNP complex contains small nucleolar RNAs (snoRNAs) that complement rRNA sequences and guide their pseudouridylation, a modification required for proper rRNA processing. As a further indication of the link of UBR5 with H/ACA RNP, RIP experiments showed that UBR5 protein also pulls down snoRNAs of this complex while it does not interact with snoRNAs of the nucleolar C/D Box RNP. Importantly, our data suggest that the interaction of UBR5 with proteins of the H/ACA RNP complex is RNA-independent. Thus, these direct protein-protein interactions raise the interesting hypothesis that increased levels of UBR5 could titrate out proteins from the H/ACA RNP complex, leading to a lower stability of this complex. However, our results do not support this hypothesis because we observe that loss of UBR5 impairs the maturation of rRNA and increases the levels of intermediates, a process that could ensue from impaired functioning of the H/ACA RNP complex. In addition, we show that the levels of the 40S and 60S subunits of the ribosome, as well as the monosome and polysomes fractions, are reduced upon loss of UBR5, correlating with impaired rRNA biogenesis. Taken together, these findings support that UBR5 is important for the function of the H/ACA RNP complex, but it does not act as a suppressor of its stability/activity. In support of this hypothesis, ESCs express intrinsic higher levels of UBR5 protein when compared with their differentiated counterparts, while they exhibit enhanced global translation rates [44].
Besides the nucleolar H/ACA RNP complex, small Cajal body-specific RNAs (scaRNAs) confined to Cajal bodies in the nucleus also contain similar box H/ ACA domains as well as box C/D domains [45]. As such, scaRNAs are involved in the pseudouridylation or 2 0 O-methylation required for maturation of not only small nuclear RNAs (snRNAs) but also nucleolar snoRNAs [45]. Since Cajal bodies contribute to the biogenesis of nucleolar RNPs [45], it will be fascinating to examine whether UBR5 interacts with nuclear RNPs in Cajal bodies and the impact of this interaction in nucleolar rRNA metabolism. Besides its effects in nucleolar RNP biogenesis, Cajal bodies are also involved in the proper maturation of snRNAs required for the assembly of distinct nuclear RNP complexes such as the spliceosome, which regulates RNA splicing [46]. Notably, the interactome of UBR5 is highly enriched in proteins involved in the assembly and activity of the spliceosome RNP complex (Fig. S1 and Data S1). Since these data support an interaction between UBR5 and spliceosome, it will be interesting to assess a role of UBR5 in the assembly and/or activity of the spliceosome RNP complex.
The link of UBR5 with rRNA biogenesis indicates a role in translational rates. Importantly, UBR5 contains a PABC domain at the C terminus, also known as MLLE domain, which mediates protein-protein interaction through the binding of the (PABP-interacting motifs) PAM2 peptide motif [47,48]. This PABC domain is equivalent to the one found in the poly(A) binding protein (PABPC1), a protein that recognizes 3 0 mRNA poly(A) tails and plays an essential role in eukaryotic translation initiation [49,50]. UBR5 and PABPC1 share common interacting partners, such as Paip1 [47] and Paip2 [51]. Paip2 is a repressor of translation initiation which can interact with both PABPC1 and UBR5. When bound to UBR5, Paip2 is ubiquitinated and degraded by the proteasome [51]. Conversely, the levels of the repressor increase in the absence of UBR5 and translation are consequently inhibited. Thus, UBR5 could modulate translational rates via both Paip2-mediated mechanisms and the maturation of rRNA.
As a consequence of perturbations in ribosomal biogenesis, cells enter into the so called 'nucleolar stress' [52]. Under normal conditions, the E3 ligase Mdm2 interacts with p53 and results in the ubiquitination and degradation of p53. During nucleolar stress, several assembly factors, such as Arf or Nucleophosmin1, and ribosomal proteins bind to Mdm2 preventing its interaction with p53, a process that results in stabilization of p53 levels [53,54]. In addition, the 5S rRNA also interacts and blocks Mdm2 [55]. As a consequence of p53 accumulation, cell enters in cell cycle arrest to ensure cellular homeostasis until the stress is over. Remarkably, we find that the levels of p53 levels are increased in mES-C UBR5 KO cells. This could indicate that UBR5-lacking cells undergo nucleolar stress, as suggested by the altered rRNA precursors and polysome levels. Moreover, loss of UBR5 induces a decline in the proliferation rates of ESCs, which could be ensue from increased p53 levels. In these lines, UBR5 is deregulated in many cancer types and alters p53 levels [56][57][58]. However, it is unclear whether UBR5 would act as a tumor suppressor or oncogene, since both amplifications as well as loss-offunction mutations have been linked to tumorigenesis. Thus, the cellular context might be determinant for the role of UBR5 in p53 regulation and cellular proliferation [59][60][61]. Whereas previous studies reported that the role of UBR5 in the regulation of mRNA machinery is necessary for cell proliferation [13,62], here we provide data indicating that UBR5 also modulates cell proliferation via modulation of rRNA maturation. Thus, UBR5 could impinge upon cellular proliferation through multiple pathways.
Taken together, our results indicate that UBR5 participates in the maturation of rRNA through regulating the nucleolar H/ACA RNP complex. This regulatory pathway is required for proper rRNA processing and ribosomal assembly. Upon loss of UBR5, rRNA intermediates accumulate, decreasing the pool of ribosomes. This perturbation may activate the nucleolar stress response, triggering the accumulation of p53 and consequent proliferation arrest. Taking into consideration the multiple roles of UBR5 in cell cycle and cellular homeostasis, our findings add an additional regulation layer to this process. interpretation through discussions with NHG and DV. JVG assessed rRNA maturation. SK performed western blot experiments. HJL performed RIP experiments. MH, VK and MSD generated mESCubr5 KO line and contributed to experiment design. NHG and CD contributed with their expertise on RNA metabolism and performed data analysis. IS, JVG, NHG and DV wrote the manuscript. All the authors discussed the results and commented on the manuscript.

Supporting information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Fig. S1. Gene Ontology analysis of UBR5 potential interactors indicated enrichment for proteins involved in RNA metabolism. Table S1. List of primers used for qPCR. Data S1. Analysis of proteomics data from co-immunoprecipitation experiments with UBR5 and FLAG antibodies in hESCs (FDR (q-value) <0.2 was considered significant).