The ubiquitin hybrid gene UBA52 regulates ubiquitination of ribosome and sustains embryonic development

Ubiquitination is a crucial post-translational modification; however, the functions of ubiquitin-coding genes remain unclear. UBA52 encodes a fusion protein comprising ubiquitin at the N-terminus and ribosomal protein L40 (RPL40) at the C-terminus. Here we showed that Uba52-deficient mice die during embryogenesis. UBA52-deficient cells exhibited normal levels of total ubiquitin. However, UBA52-deficient cells displayed decreased protein synthesis and cell-cycle arrest. The overexpression of UBA52 ameliorated the cell-cycle arrest caused by UBA52 deficiency. Surprisingly, RPL40 expression itself is insufficient to regulate cyclin D expression. The cleavage of RPL40 from UBA52 was required for maintaining protein synthesis. Furthermore, we found that RPL40 formed a ribosomal complex with ubiquitin cleaved from UBA52. UBA52 supplies RPL40 and ubiquitin simultaneously to the ribosome. Our study demonstrated that the ubiquitin-coding gene UBA52 is not just an ubiquitin supplier to the ubiquitin pool but is also a regulator of the ribosomal protein complex. These findings provide novel insights into the regulation of ubiquitin-dependent translation and embryonic development.

Ribosome biogenesis and protein synthesis are tightly regulated process linked to other fundamental cellular processes 20,21 . Targeted disruption of the ribosomal protein genes (e.g., Rps19 and Rps6) is lethal 22,23 . Dysfunctional ribosome biogenesis is associated with developmental defects and an increased risk of cancer 24,25 . Recent studies have shown that regulatory 40 S ribosomal ubiquitination is an important phase of translational control 26 . In addition, translation reactions terminate with ub-dependent removal of defective nascent chain 27,28 . Nevertheless, the association between ubiquitin-dependent regulation of translation and the ubiquitin-ribosomal hybrid gene UBA52 remains unclear. To determine the physiological functions of UBA52, we generated mice

Results and Discussion
UBA52 is required for embryonic development. To investigate the potential physiological roles of UBA52, we generated Uba52-deficient mice. We confirmed insertion of the tm1a cassette into Uba52 genomic fragment in embryonic stem cells by Southern blotting (Fig. 1B) and in DNA obtained from the tail by polymerase chain reaction (PCR; Fig. 1C). We found that the deletion of one Uba52 allele in mice did not affect the expression of Uba52 mRNA (Fig. 1D). To further confirm the UBA52 tm1a allele, we consider that aberrant UBA52 proteins may act as dominant-negative molecules. We analyzed the UBA52 protein expression by immunoblotting; the truncated protein was not detected in Uba52 tm1a/+ liver lysates (data not shown). However, no live-born Uba52-deficient (uba52 tm1a/tm1a ) mice were obtained from intercrossed Uba52 tm1a/+ parents (Fig. 1E). Because Ubc-deficient mice died at around E13.5, we analyzed Uba52-deficient mice at this time. Uba52 tm1a/+ normally developed, but we could not obtain uba52 tm1a/tm1a at E13.5 (Fig. 1E). Rps6-heterozygous embryos die during gastrulation 23 . Therefore, we analyzed at E10.5, but we could not obtain uba52 tm1a/tm1a at E10.5. Thus, these data indicate that one allele of the Uba52 gene is enough for development but UBA52 is required for embryonic development.
UBA52 regulates protein synthesis. To better understand how UBA52 sustains embryonic development, we noted that UBC is essential for fetal development 16 . Given that UBA52 is a ubiquitin hybrid gene, we hypothesized that UBA52 regulates the total ubiquitin mRNA expression. To investigate this possibility, we used a short interfering RNA (siRNA) approach for reducing UBA52 expression in a colon cancer cell line (DLD-1). Acute knockdown of UBA52 did not affect the total ubiquitin mRNA levels. Conversely, knockdown of UBC reduced the total amount of ubiquitin ( Fig. 2A). Our finding that UBA52-deficient cells display similar levels of total ubiquitin mRNA as those displayed by wild-type cells indicates that changes in the level of ubiquitin mRNA can be compensated for by other ubiquitin-coding genes. Because previous reports have shown that UBA52 is rapidly cleaved 11,17 , we transfected a FLAG-UBA52-green fluorescent protein (GFP) vector into DLD-1 cells. We detected a band of approximately 34 kDa using anti-UBA52 and anti-GFP antibodies but did not detect this using anti-FLAG and anti-ubiquitin antibodies (Fig. 2B). The anti-FLAG antibody blot showed a high-molecular weight smear. These findings indicate that FLAG-UBA52-GFP is cleaved and that polyubiquitin is generated from FLAG-ubiquitin. In addition, we detected RPL40-GFP and endogenous RPL40 using an anti-UBA52 antibody. To confirm that ubiquitin was obtained from the cleavage of UBA52, we generated cleavage-resistant UBA52 (CR) expression vector, which was constructed by mutating the terminal region of ubiquitin (G75/76A) 17 . FLAG-UBA52-GFP (CR) expression vector showed decreased high-molecular weight smear as compared to FLAG-UBA52-GFP (WT) (Fig. 2C). FLAG-UBA52-GFP (CR) was detected at approximately 44 kDa by anti-FLAG antibody and anti-GFP antibody (Fig. 2C). Thus, we confirmed that ubiquitin was generated from UBA52. RPs are known to play many extra ribosomal roles 29 . For example, RPS3 regulates NF-κ B signaling 30 and RPL13a in macrophages resolves inflammation 31 . To examine the localization of UBA52, we transfected a FLAG-UBA52-GFP vector into DLD-1 cells. Microscopic analysis revealed that FLAG-GFP was expressed in both the cytosol and nucleolus. RPL40-GFP was not expressed in the nucleolus (Fig. 2D). To confirm the localization of endogenous RPL40, we lysed the cells and separated the soluble, ribosome-free cytosol (S100) from a particulate pellet containing ribosomes (P100) by ultracentrifugation 30 . RPL7a and RPS3 were expressed in the ribosomal fraction (P100) (Fig. 2E). In contrast, CDK6 and Actin were expressed in the cytosol (S100). We found that RPL40 was expressed not only in the ribosomal fraction (P100) but also in the cytosol (Fig. 2E). In addition, RPL40 expression in the ribosomal fraction (P100) gradually decreased over time after siRNA transfection. These findings showed that we could analyze ribosomal function by the siRNA approach. To better understand the roles of UBA52, we tested whether it regulates ribosomal function. A protein synthesis assay using O-propargyl-puromycin (OPP) revealed that cycloheximide (CHX) completely inhibited protein biosynthesis under these conditions. RPS3 and UBA52 knockdown decreased protein synthesis (Fig. 2F). To confirm the general role of UBA52, we tested Hela cells as well as DLD-1 cells (Fig. 2G). Along with DLD-1 cells, UBA52-deficient Hela cells showed decreased protein synthesis. Together, these data indicate that the loss of UBA52 does not regulate the total amount of ubiquitin mRNA but regulates protein synthesis functions in cells. Thus, Uba52-deficient lethality may be because of ribosomal dysfunction. Further investigation is required to clarify the mechanism by which UBA52 regulates embryonic development in vivo.
UBA52 regulates the cell cycle. Ribosomal stress can regulate the cell cycle by p53-dependent and -independent pathways 32,33 . To understand the role of UBA52, we analyzed cell proliferation. We found that UBA52-deficient DLD-1 cells displayed decreased cell numbers and cell-cycle arrest at G1/S (Fig. 3A,B). To determine whether UBA52 expression is sufficient to regulate the cell cycle, we mutated UBA52 expression vectors to induce siRNA resistance. We found that Myc-UBA52 ameliorated the cell-cycle arrest caused by UBA52 deficiency (Fig. 3C). Together, these findings indicate that UBA52 regulates the cell cycle. Next, to understand the mechanism underlying this, we consider that cyclin D promotes cell cycle as a main regulator 34 . We analyzed cyclin D1 and D3 gene expressions. There were no differences in cyclin D1 and D3 mRNA expressions between control and UBA52-deficient cells (Fig. 3D). In contrast, the levels of cyclin D1 and D3 protein expression decreased in UBA52-deficient cells (Fig. 3E). To examine the possibility that UBA52 regulates cyclin D expression, we analyzed cell division protein kinase 6 (CDK6), which is known to be a major partner of cyclin D and a key molecule in the regulation of G1/S phase. CDK6 expressed in the cytosol fraction (Fig. 2E). Using a proximity ligation assay, we found that endogenous RPL40 co-localized with CDK6 (Fig. 3F). To confirm physical interaction, we performed immunoprecipitation assay. Endogenous CDK6 associated with RPL40-GFP in DLD-1 cells and HEK293T cells (Fig. 3G). These data suggested that RPL40 may affect the CDK6 complex formation. Nevertheless, cyclin D1 and D3 quickly degraded after CHX treatment in DLD-1 cells (Fig. 3H). Moreover, similar observations that the normal expression of mRNAs and the decreased protein levels of cyclins D1 and D3 in UBA52-deficient cells were found in Rps6 wt/del p53 −/− embryos 23 . These findings indicated that decreased levels of cyclin D1 and D3 were provoked mainly by the suppression of protein synthesis in UBA52-deficient cells. Taken together, UBA52 modulates cyclin D1 and D3 protein expression and regulates the cell cycle.
Ubiquitin cleaved from UBA52 forms a molecular complex with RPL40. To understand how the ubiquitin-ribosomal hybrid gene regulates protein synthesis, we generated cleavage-resistant (CR) and siRNA resistant UBA52 expression vectors 17 . Ubiquitin, Myc-RPL40, and Myc-UBA52 (CR) overexpression in DLD-1 cells did not affect cyclin D expression, but Myc-UBA52 (WT) ameliorated the decreased cyclin D1 and D3 expressions (Fig. 4A, Supple Fig. 1). These findings indicated that RPL40 expression itself is not sufficient to regulate protein synthesis. The RPL40 cleavage mechanism and localization of ubiquitin cleaved from UBA52 may influence the function of UBA52. Thus, we tested whether UBA52 and UBA52 (CR) formed a molecular complex. GFP immunoprecipitation revealed that RPL40 bound molecules that had been ubiquitinated using ubiquitin generated from UBA52 cleavage (Fig. 4B). Furthermore, FLAG immunoprecipitation revealed that ubiquitin cleaved from UBA52 ubiquitinated the ribosomal complex (Fig. 4C). These findings indicated that UBA52 supplies ubiquitin to the ribosomal complex. To confirm whether UBA52 regulates the ribosomal protein complex, we tested the ribosomal fraction. FLAG-UBA52-GFP transfected cell showed that both FLAG-ubiquitin from UBA52 and RPL40-GFP expressed in the ribosomal fraction (Fig. 4D). In addition, UBA52-deficient cell showed reduced ubiquitin smear in the ribosomal fraction (Fig. 4E). Taken together, these data suggested that RPL40 formed a ribosomal protein complex with ubiquitin cleaved from UBA52 (Fig. 4F). Recent studies have shown that the regulation of ubiquitin-dependent translation is an important feature of cell-fate determination 35 . Regulatory 40 S ribosomal ubiquitination events play a critical role in protein biogenesis 36 . Our findings demonstrated that ubiquitin, RPL40, UBA52 (CR), and a mixture of ubiquitin and RPL40 could not ameliorate cyclin D1 and D3 expressions (Fig. 4A). However, full-length UBA52 abolished the cell-cycle-related UBA52-deficient phenotype. These findings indicate that the RPL40 cleavage is critical for the functions of UBA52. Thus, we hypothesize that UBA52 regulates ribosomal function by two steps. RPL40 influences the ribosomal biogenesis as a ribosomal protein, at the same time, ubiquitin cleaved from UBA52 generates ubiquitination of the ribosomal protein complex. These data reveal a novel mechanism that ubiquitin regulates translation by the ubiquitin-ribosomal hybrid gene. Future studies will be needed to identify enzymes that cleave ubiquitin from UBA52 under physiological conditions. Moreover, it is critical to analyze UBA52's physiological functions in cell type-specific contexts.
In conclusion, the generation of mice lacking Uba52 has allowed us to unveil the physiological function of the ubiquitin hybrid gene Uba52. In particular, we have discovered that UBA52 regulates cell fate and embryonic development. RPL40, the ribosomal domain of UBA52, generated a ribosomal protein complex, at the same time, the UBA52 ubiquitin domain simultaneously generated ubiquitination of the ribosomal complex. Thus, UBA52 is a dual regulator of ribosomal protein complex. These findings provide unique molecular insights into ubiquitin-related protein synthesis and embryonic development.

Mice. Uba52 tm1a(EUCOMM)Wtsi embryonic stem cells were purchased from the European Conditional Mouse
Mutagenesis Program (EUCOMM) and microinjected into the blastocysts of an albino C57BL6 strain. The chimeric mice were backcrossed with the same strain of albino C57BL6 mice to generate heterozygous Uba52 mutant mice. All animal experiments were approved by the Institutional Animal Care and Use Committee of the Tokyo Medical and Dental University. Experiments were performed in compliance with Tokyo Medical and Dental University's Animal Facility regulations. Genotypes were initially confirmed by Southern blotting using embryonic stem cells. In addition, genotypes were confirmed by PCR using DNA derived from the tail and the following primers: Primer4, F 5′ -CTGCAGAGGGAGTTCAGGG-3′ and R 5′ -GTTTGGTAAGTAGGGGCAGC-3′ ; Primer5, F 5′ -FACAACCATGGAAGATCCCGT-3′ and R 5′ -CCGTTGCACCACAGATGAAA-3′ and Primer6, F 5′ -AGGAAGGAGTTGTGGCCAACCTGG-3′ and R 5′ -TGAACTGATGGCGAGCTCAGACC-3′ . Also, the following primers were used for long-range PCR: Primer1, F 5′ -TCCAGACAGAACGACTATTCTCGC-3′ and R 5′ -AACTGAAGGATCGGACAGCA-3′ ; Primer2, F 5′ -ACAACCATGGAAGATCCCGT-3′ and R 5′ -AACTGAAGGATCGGACAGCA-3′ and Primer3, F 5′ -TCCAGACAGAACGACTATTCTCGC-3′ and R 5′ -CCGTTGCACCACAGATGAAA-3′ . transfected with a UBA52 siRNA and lysed for ultracentrifugation at the indicated time. S100 (cytosol) and P100 (crude ribosome pellet) fractions were immunoblotted for the indicated proteins. Data are representative of more than three independent experiments. (F,G) Protein synthesis in UBA52-deficient cells. DLD-1 cells (F) or Hela cells (G) were transfected with UBA52 or ribosomal protein (RP) S3 siRNAs. The cells were treated with cycloheximide (100 μ g/ml) for 3 h. Cells were incubated with O-propargyl-puromycin (20 μ M) for 30 min and then harvested for the protein synthesis assay. Data are representative of more than two independent experiments. *P < 0.05, **P < 0.01 (one-way ANOVA followed by Tukey's test). We confirmed knockdown efficiency by quantitative real-time (RT) PCR and normalized to ACTB levels. Data are representative of two independent experiments.   . (D) Cell-cycle-related mRNA expression in UBA52-deficient cells. DLD-1 cells were transfected with a UBA52 siRNA. Twenty-four hours later, cells were harvested for quantitative real-time reverse-transcription PCR and normalized to ACTB levels. Data are representative of three independent experiments. (E) Cell cycle-related protein expression in UBA52-deficient cells. DLD-1 cells were transfected with a UBA52 siRNA. Twenty-four hours later, cells were harvested for immunoblotting. Data are representative of more than three independent experiments. (F) RPL40 co-localises with CDK6. DLD-1 cells were harvested for the in situ proximity ligation assay. Anti-UBA52 and anti-CDK6 antibodies were used. Data are representative of four independent experiments. **P < 0.01 (two-tailed Student's t-test). (G) UBA52 interacts with CDK6. DLD-1 cells and HEK293T cells were transfected with FLAG-UBA52-GFP. Twenty hours later, cells were lysed and protein extracts were immunoprecipitated with GFP antibody and immunoblotted for the indicated proteins. Data are representative of more than three independent experiments in DLD-1 cells and HEK293T cells. (H) CHX chase experiment for the cyclin D protein. DLD-1 cells were treated with 100 μ g/ml CHX and cell lysates were harvested for immunoblotting at the indicated time points. Data are representative of more than three independent experiments. Ultracentrifugation. Cells were incubated in lysis buffer without glycerol [20 mM HEPES (pH 7.5), 150 mM NaCl, 0.5% TritonX-100, and 0.5% CHAPS] with added Halt protease and phosphatase inhibitor cocktail on ice for 20 min and centrifuged at 14,000 × g for 20 min to remove the nuclear fraction. Then, the supernatant was ultracentrifuged at 4 °C and 100,000 × g for 1 h. The supernatant was preserved as the S100 fraction. The pellet was washed with PBS and ultracentrifuged again at 4 °C and 100,000 × g for 20 min. The supernatant was discarded and the pellet, which was the P100 fraction, was refused in lysis buffer using an ultrasonic homogenizer.
Protein synthesis assay. To analyze protein synthesis in cells, a Click-iT ® Plus OPP Protein Synthesis Assay Kit (Alexa Fluor ® 488; C104567570; Molecular Probes, Eugene, OR, USA) was used in accordance with the manufacturer's protocol. The GloMax ® Discover System (Promega Corp.) was used to scan the plates.
Cell viability assay. CellTiter-Glo Luminescent Cell Viability Assay (G7570; Promega Corp.) was used to count viable cells in wells, as indicated by the manufacturer.
In situ proximity ligation assay. To detect protein interactions in cells, a Duolink ® PLA in-situ kit (92101; Sigma-Aldrich Corp.) was used according to the manufacturer's instructions. The primary antibodies were as follows: a rabbit anti-UBA52 antibody (EPR4547; ab109230; Abcam) and a mouse anti-CDK6 antibody (ab54576; cells were transfected with FLAG-UBA52-GFP vector or vehicle, and 24 h later, lysed for ultracentrifugation at 10,000 × g for 1 h. S100 (cytosol) and P100 (crude ribosome pellet) fractions were immunoblotted for the indicated proteins. Data are representative of two independent experiments. (E) Subcellular fractions of UBA52-deficient DLD-1 cells. DLD-1 cells were transfected with UBA52 siRNA or control siRNA, and 24 h later, ultracentrifugation was performed. Data are representative of three independent experiments. (F) Schematic showing that UBA52 is a dual regulator of ribosomal function. RPL40 influences the ribosomal biogenesis as a ribosomal protein, at the same time, ubiquitin cleaved from UBA52 generates ubiquitination of the ribosomal protein complex.
Scientific RepoRts | 6:36780 | DOI: 10.1038/srep36780 Abcam). The primary antibody rabbit anti-Immunoglobulin G was used as a control (ab125938; Abcam). Images were acquired with a confocal laser microscope using a × 60 oil immersion objective lens. Dots within cells were counted using Duolink ImageTool (OLINK Bioscience, Uppsala, Sweden). Cells were only counted when the whole cell was in the field of vision.
Statistical Analysis. Data were analyzed using the two-tailed unpaired Student t-test or one-way ANOVA followed by Tukey's test using Prism software (GraphPad Software, Inc.).