Abstract Ribosome biogenesis plays a pivotal role in maintaining stem cell homeostasis, yet the precise regulatory mechanisms governing this process in mouse embryonic stem cells (mESCs) remain largely unknown. In this investigation, we ascertain that DEAD-box RNA helicase 10 (DDX10) is indispensable for upholding cellular homeostasis and the viability of mESCs. Positioned predominantly at the nucleolar dense fibrillar component (DFC) and granular component (GC), DDX10 predominantly binds to 45S ribosomal RNA (rRNA) and orchestrates ribosome biogenesis. Degradation of DDX10 prevents the release of U3 snoRNA from pre-rRNA, leading to perturbed pre-rRNA processing and compromised maturation of the 18S rRNA, thereby disrupting the biogenesis of the small ribosomal subunit. Moreover, DDX10 participates in the process of liquid-liquid phase separation (LLPS), which is necessary for efficient ribosome biogenesis. Notably, the NUP98-DDX10 fusion associated with acute myelocytic leukemia (AML) alters the cellular localization of DDX10 and results in loss of ability to regulate pre-rRNA processing. Collectively, this study reveals the critical role of DDX10 as a pivotal regulator of ribosome biogenesis in mESCs. Subject terms: Pluripotency, Embryonic stem cells, RNA __________________________________________________________________ Ribosome biogenesis is crucial for maintaining stem cell homeostasis. Here, authors reveal that DDX10 plays a critical role in ribosome biogenesis by regulating 18S rRNA maturation, which is vital for the proliferation and maintenance of mESCs. Introduction ESCs, derived from the inner cell mass (ICM) of pre-implantation blastocysts, are characterized by rapid proliferation and pluripotency, enabling them to self-renew and differentiate into diverse cell types^[62]1,[63]2. The determination of ESC fate hinges upon intricate regulatory mechanisms that orchestrate gene expression across various tiers, encompassing chromatin, transcriptional, and post-transcriptional regulation. These regulatory processes have been extensively scrutinized in stem cells^[64]3–[65]8. Recently, emerging evidence has underscored the pivotal role of ribosome biogenesis in upholding ESC identity^[66]9,[67]10. ESCs exhibit robust rRNA transcription and heightened ribosome biogenesis^[68]11–[69]14. Safeguarding chromatin integrity at actively transcribed rDNA loci shields them from epigenetic silencing, thereby promoting rRNA transcription and ribosome biogenesis, pivotal for sustaining ESC self-renewal^[70]15. Maintaining steady-state ribosome biogenesis is imperative for the maintenance of ESC homeostasis^[71]9. In ESCs, there is pronounced expression of small subunit (SSU) processome genes, ensuring efficient processing of pre-rRNA and maturation of rRNA, which are vital for preserving pluripotency. Deletion of SSU processome genes leads to diminished protein synthesis and loss of pluripotency in ESCs^[72]13. Notably, recent investigations have unveiled a dichotomy: while undifferentiated ESCs demonstrate relatively lower polysome loading compared to differentiated progeny, they still expend considerable energy to sustain an abundant ribosome pool^[73]10. In contrast, there is an increase in polysome loading, protein synthesis, and protein content that occurs during differentiation^[74]16–[75]18. Thus, ribosome biogenesis is a linchpin for maintaining pluripotent stem cells and orchestrating their differentiation. Despite its pivotal role, the mechanisms governing ribosome biogenesis in ESCs remain unclear, warranting further exploration and study. The nucleolus, a highly prominent and extensively studied membraneless ribonucleoprotein (RNP) entity, is a pivotal hub for ribosome biogenesis^[76]19,[77]20. Comprising three distinct subregions – the fibrillar center (FC), the DFC, and the GC – the nucleolus plays a multifaceted role in orchestrating ribosome assembly. Within this intricate landscape, rDNA transcription occurs at the FC-DFC boundary, pre-rRNA processing takes place within the DFC, and the subsequent assembly of ribosomal subunits is executed within the GC^[78]20–[79]22. Ribosomal genes are transcribed by RNA polymerase I (Pol I) to produce the primary 47S rRNA precursor, which includes two external transcribed spacers (5’ETS and 3’ETS) and two internal transcribed spacers (ITS1 and ITS2) separating the mature 18S, 5.8S, and 28S rRNAs. To obtain these mature rRNAs, the transcribed spacers must be removed through a sequential series of endonucleolytic and exonucleolytic cleavages^[80]23. In mouse cells, the 47S rRNA transcript is first cleaved at site A0, generating the 46S rRNA, and then at site 6, producing the 45S rRNA. The processing of mouse 45S rRNA occurs primarily through two pathways. In pathway 1, sites A0 and 1 in the 5’ETS are successively cleaved successively to produce 43S and 41S rRNA. Subsequently, site 2c in ITS1 is cleaved to produce 20S rRNA (precursor of 18S rRNA) and 36S rRNA (precursor of 28S and 5.8S rRNA). In pathway 2, site 2c in ITS1 is firstly cleaved to produce 34S rRNA and 36S rRNA. Subsequently, sites A0 and 1 in 34S rRNA are cleaved in sequence to produce 20S rRNA. These rRNA precursors are ultimately processed into mature rRNA^[81]23,[82]24. Ribosome biogenesis is a multi-step process underpinned by the coordinated interplay of numerous proteins and non-coding RNAs (ncRNAs). These entities collectively govern the intricate choreography of events encompassing rRNA transcription^[83]25–[84]27, directed trafficking of nascent pre-rRNA^[85]28, and pre-rRNA processing^[86]29–[87]31. Some small nucleolar RNAs (snoRNAs) serve as scaffolds during snoRNPs formation and base pairing with pre-rRNA to guide the directional cleavage and folding of pre-rRNA^[88]32,[89]33, which is crucial for rRNA maturation. U3, U14, U22, U17/snR30, and snR83 affect the maturation of 18S rRNA^[90]34–[91]39, while U8 snoRNA is essential for the accumulation of mature 5.8S and 28S rRNAs^[92]40. DExD/H-box RNA helicases belong to the RNA-binding protein (RBP) family and are the largest consortium of RNA helicases, ubiquitously present across diverse organisms^[93]41. Their pivotal functions encompass remodeling RNA structures by harnessing the energy derived from ATP hydrolysis, thereby influencing multiple facets of cellular RNA metabolism, including transcription, splicing, ribosome biogenesis, RNA export, translation, RNA turnover, and organelle gene expression^[94]41–[95]43. Among these, DDX10 is a constituent of the DEAD-box RNA helicases^[96]44,[97]45. Existing literature underscores DDX10’s association with various tumors, observing abnormal expression within tumor tissues^[98]46–[99]49. Nonetheless, the specific physiological and molecular contributions of DDX10 within ESCs remain unknown. In this study, we unearthed that DDX10 is a crucial regulator of ribosome biogenesis and is essential for proliferation and maintenance of cell fate in mESCs. The degradation of DDX10 induces cell cycle arrest at the G1 phase while promoting apoptosis, potentially through a p53-dependent mechanism. Notably, the deficiency of DDX10 causes disruptions in pre-rRNA processing, manifesting as diminished 18S rRNA maturation and compromised ribosome biogenesis. Moreover, DDX10 interacts with the components of the SSU processome, and its absence hindering the liberation of U3 snoRNA from pre-rRNA. These insights collectively illuminate the integral role of DDX10 in maintaining proper ribosome biogenesis. Further, DDX10 undergoes LLPS, which is crucial for ribosome biogenesis. Finally, our findings show that NUP98-DDX10 fusion protein results in the loss of DDX10 function in regulating ribosome biogenesis. Results DDX10 is indispensable for the survival and maintenance of mESCs To delve into the role of DDX10 in mESCs, we initiated our investigation by assessing Ddx10 expression in mESCs and mouse embryonic fibroblasts (MEFs). Our findings unveiled heightened Ddx10 expression levels in mESCs, in contrast to MEFs (Fig. [100]1a). Throughout ESC differentiation, Ddx10 displayed robust expression in ESCs, which underwent rapid downregulation (Fig. [101]1b, c). Leveraging the auxin-inducible degron system^[102]50,[103]51, we aimed to degrade endogenous DDX10 in mESCs. Employing the CRISPR-Cas9 genome editing technique, we introduced the AID-eGFP sequence at the stop codon of Ddx10 (Fig. [104]1d and Supplementary Fig. [105]1a). Subsequent exposure of cells to the auxin analog indole-3-acetic acid (IAA) resulted in the rapid degradation of DDX10, becoming undetectable within 2 h of IAA treatment, while regaining initial levels post IAA removal, thereby confirming the efficacy of the degradation system (Fig. [106]1e and Supplementary Fig. [107]1b, c). We observed that the protein level of DDX10 in DDX10-AID (+ OsTir1) mESCs was lower compared to wild-type mESCs (Fig. [108]1e). However, this reduction could not lead to significant changes in cell morphology, cell cycle, and apoptosis (Supplementary Fig. [109]1d-f). We observed that DDX10 degradation resulted in smaller and flattened mESC clones, with morphological recovery upon IAA withdrawal (Fig. [110]1f and Supplementary Fig. [111]1g). Scrutinizing the influence of DDX10 degradation on mESC pluripotency, we evaluated the expression of pluripotency transcription factors (OCT4, NANOG, and SOX2) revealing minimal impact due to DDX10 degradation (Supplementary Fig. [112]1h). Moreover, our data spotlighted that DDX10 degradation significantly impeded mESC proliferation (Supplementary Fig. [113]1i, j), provoking cell cycle arrest at the G1 phase (Supplementary Fig. [114]1k, l), while instigating apoptosis (Supplementary Fig. [115]1m, n). Fig. 1. Disruption of gene expression upon acute degradation of DDX10 in mESCs. [116]Fig. 1 [117]Open in a new tab a RT-qPCR analysis for endogenous levels of Ddx10 mRNA in MEFs and mESCs. b RT-qPCR analysis of Ddx10 expression levels following LIF withdrawal. c RT-qPCR analysis of Ddx10 expression levels during embryoid body (EB) differentiation. d Schematic illustration of the generation of DDX10-AID mESCs. e Western blot analysis of DDX10 protein levels in DDX10-AID cells with or without IAA treatment. WT: wild-type E14 mESCs. The red asterisk indicates endogenous DDX10-AID-eGFP and DDX10 proteins. β-ACTIN serves as the loading control. Experiments were repeated three times independently with similar results. f Brightfield images of DDX10-AID (+ OsTir1) mESC colonies with or without IAA treatment. Experiments were repeated three times independently with similar results. Scale bar, 200 µm. g Principal component analysis (PCA) plot displays RNA-seq data from DDX10-AID (+ OsTir1) mESCs treated with IAA at different time points or treated with IAA for 48 h followed by 48 h of washing. h Bar plots showing the number of differentially expressed genes (DEGs) upon DDX10 degradation at different time points with IAA treatment (orange: upregulated genes, green: downregulated genes). i Line chart illustrating gene expression patterns of 24 different clusters of DEGs. j Heatmap presenting gene ontology results for genes in each cluster. For (a–c) transcription levels were normalized against Gapdh. Data are presented as mean values ± SD with the indicated significance from two-sided t-test. Exact p-values are reported in the figure. n = 3 independent experiments. Source data are provided as a Source Data file. To investigate the molecular anomalies ensuing from DDX10 degradation, we executed RNA-seq experiments on DDX10-AID mESCs, both in the presence and absence of IAA treatment. Through correlation analysis, we validated the high reproducibility among replicates. Gradual changes in gene expression patterns were discernible beginning at 2 h post-IAA treatment. Yet, the pattern closely resembled that of untreated cells at 48 h after IAA withdrawal (Fig. [118]1g). This temporal dynamic suggests that gene expression alterations resulting from DDX10 degradation are reversible. Compared with untreated cells, the number of differentially expressed genes (DEGs) gradually increased during IAA treatment (Fig. [119]1h and Supplementary Data [120]1). Subsequently, we clustered the DEGs at all time points into 24 groups based on their expression patterns (Fig. [121]1i). Notably, downregulated gene clusters predominantly associated with cell division, energy metabolism, and RNA metabolism (Fig. [122]1j). To further investigate the effect of DDX10 degradation on cell fate determination, we analyzed the RNA-seq data from primed (cultured in medium containing serum) and naive (cultured in medium containing 2i) mESCs^[123]52, and integrated these data with our RNA-seq data. PCA results showed that mESCs following DDX10 degradation were neither close to the primed nor to the naive mESCs (Supplementary Fig. [124]2a). Then we perform Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analysis on the upregulated genes, which revealed significant enrichment of the p53 signaling pathway (Supplementary Fig. [125]2b). Concretely, genes like Mdm2 and cyclin-dependent kinase inhibitor 1 A (Cdkn1a/p21) that are associated with the p53 pathway, as well as pro-apoptotic genes Bbc3 and Pmaip1 exhibited notable upregulation post DDX10 degradation (Supplementary Fig. [126]2c, d). Collectively, these findings underscore that DDX10 degradation prompts cell cycle arrest and propels apoptosis by activating the p53 signaling pathway, consequently impeding mESC growth. Previous studies have shown that p53 activation can induce the transition of mESCs to 2-cell-like cells (2CLCs)^[127]53,[128]54. Therefore, we compared our RNA-seq data with the results from previously published 2-cell data^[129]55, and observed that 2-cell specific genes, such as Zscan4b and Zscan4d, were significantly activated after 24 h of IAA treatment (Supplementary Fig. [130]2e, f). Together, these results indicate that DDX10 degradation promotes the transition of mESCs to 2CLCs. DDX10 localizes to the nucleolar DFC and GC and primarily binds to 45S rRNA Given DDX10’s classification as an RNA binding protein, we executed crosslinking immunoprecipitation followed by high-throughput sequencing (CLIP-seq) to unveil its downstream targets in mESCs. Despite initial failures with both commercial and self-made anti-DDX10 antibodies for CLIP-seq, we engineered mESCs overexpressing FLAG-tagged DDX10, revealing nuclear localization of DDX10-FLAG (Supplementary Fig. [131]3a). Subsequently, we conducted CLIP-seq experiments by using anti-FLAG M2 magnetic beads to capture DDX10-bound RNAs (Supplementary Fig. [132]3b). Significantly, our findings illuminated that DDX10 strongly binds to 45S rRNA, with a particular preference for the 18S rRNA sequence (Fig. [133]2a, b). Furthermore, our data unveiled an association of DDX10 with a subset of snoRNAs, including U22 (Fig. [134]2a and Supplementary Data [135]2), recognized for guiding site-specific pre-rRNA cleavage and influencing 18S rRNA processing^[136]36. These results were further validated by RNA immunoprecipitation (RIP)-qPCR (Fig. [137]2c). Fig. 2. Localization and binding preferences of DDX10 in the nucleolus.