Hbxip is essential for murine embryogenesis and regulates embryonic stem cell differentiation through activating mTORC1

Hepatitis B X-interacting protein (HBXIP, also known as LAMTOR5), was first identified as a binding factor to hepatitis B virus X protein (Melegari et al., 1998). In addition to its role in inhibiting the replication of hepatitis B virus, HBXIP has been extensively studied in various cancers. It has been shown that HBXIP promotes the proliferation, migration and tumorigenesis of breast, ovarian, liver, non-small-cell lung, bladder urothelial and esophageal squamous cell cancer (Hu et al., 2011; Liu et al., 2012, 2014; Xu et al., 2014; Zhang et al., 2014; Li and Liu, 2016; Shi et al., 2016; Wang et al., 2017b; Zhou et al., 2019b; Wu et al., 2020). Overexpression of HBXIP is associated with poor prognosis in breast, ovarian, liver, non-small-cell lung cancer and pancreatic ductal adenocarcinomas (Wang et al., 2017b,c; Li et al., 2017; Zhou et al., 2019a; Cheng et al., 2014). HBXIP also contributes to cisplatin and tamoxifen resistance in ovarian and breast cancers, respectively (Zou et al., 2017; Liu et al., 2018). However, the role of HBXIP in normal development remains poorly understood. Only a recent work reported that, owing to the pivotal role of Hbxip in activating the transcription of insulin by elevating the level of Pdx-1/Neurod1 complex, pancreatic β-cell-specific Hbxip knockout (KO) mice display higher fasting blood glucose levels and impaired glucose tolerance (Li et al., 2018).

The extensive studies of HBXIP in cancers have revealed its function as a transcriptional cofactor. HBXIP, cooperating with various transcription factors, such as c-MYB, SP1, cAMP response element-binding protein (CREB), TATA-binding protein (TBP) and E2F1, activates the expression of its downstream target genes, including YAP, FGF8, LIN28B, SKP2, PDGFB, S100A4 and LMO4, consequently promoting the proliferation and migration of cancer cells (Liu et al., 2012, 2014, 2013; Xu et al., 2014, 2013; Yue et al., 2013; Zhang et al., 2013; Wang et al., 2017a). In addition, HBXIP may interact with proteins other than transcription factors to fulfill its biological functions. HBXIP interacts with survivin to suppress caspase activation and hence apoptosis (Marusawa et al., 2003). HBXIP also associates with microtubules and centrosomes in dividing cells, and is required for the proper formation of centrosomes and spindles in HeLa cells (Wen et al., 2008; Fujii et al., 2006). Moreover, HBXIP, together with p18 (LAMTOR1), p14 (LAMTOR2), MP1 (LAMTOR3) and C7ORF59 (LAMTOR4), form a pentameric Ragulator complex, which is required for mTORC1 activation by amino acids (Bar-Peled et al., 2012). Yet, whether HBXIP regulates normal development or carcinogenesis through the mTORC1 pathway remains unexplored.

In this study, we demonstrated that KO of Hbxip is embryonic lethal, and retarded development of Hbxip^−/− embryos became obvious around E7.5. Using Hbxip KO embryonic stem cells (ESCs) as an in vitro model, we found that Hbxip is crucial for the differentiation of ESCs, particularly for ectodermal and mesodermal differentiation. Consistently, the epiblast of E8.5 Hbxip^−/− embryos remained undifferentiated. The differentiation defects were shared by ESCs lacking subunits of the Ragulator complex, including Hbxip, Lamtor3 and Lamtor4. Thus, we conclude that Hbxip regulates embryo development and ESC differentiation through activating mTORC1 signaling.

To study the function of Hbxip in normal development, we knocked out Hbxip in previously generated Hbxip floxed mice (Fig. 1A,B) (Li et al., 2018). Heterozygous Hbxip KO (Hbxip^+/−) mice were obtained through mating between Hbxip floxed mice and EIIa-cre mice. Hbxip^+/− mice are fertile; however, no Hbxip^−/− mice were born from Hbxip^+/− intercrosses (Table 1), indicating embryonic lethality of Hbxip null mice. To determine the timing of lethality, embryonic day (E)3.5, E6.5, E7.5 and E9.5 embryos from Hbxip^+/− intercrosses were genotyped, and Hbxip^−/− embryos were detected at all the analyzed time points (Table 1). However, at 7.5 days postcoitum, Hbxip^−/− embryos were smaller than wild-type (WT) and Hbxip^+/− embryos (Fig. 1C,D). At 8.5 and 9.5 days postcoitum, Hbxip^−/− embryos appeared to be developmentally arrested, compared with their WT and Hbxip^+/− littermates (Fig. 1C). These data suggest a crucial role of Hbxip in embryonic development.

Next, we utilized in vitro cultured mouse ESCs to understand the mechanism of Hbxip in embryonic development. Hbxip KO ESC lines were constructed using CRISPR/Cas9, with a sgRNA targeting the third exon of Hbxip (Fig. 2A). The disruption of Hbxip in two independent ESC clones (hereafter referred to as H^−/−-1 and H^−/−-2) with normal karyotype was validated by loss of the BanI site, Sanger sequencing and western blotting (Fig. 2A,B, Fig. S1A-D). Notably, deletion of Hbxip compromised the self-renewal of ESCs, demonstrated by reduced proliferation rate and colony-forming capacity (Fig. 2C,D). However, Hbxip KO did not change the cell cycle profile or induce apoptosis in ESCs (Fig. S1E,F). The mRNA levels of pluripotency genes, Nanog, Oct4 and Sox2, as well as the protein levels of Nanog and Oct4, were decreased in Hbxip KO ESCs (Fig. 2B,E). In addition, in undifferentiated ESCs, Hbxip KO suppressed the expression of ectodermal, and mesodermal markers, such as nestin, Celsr2, T and Dlx3, whereas the endodermal marker Gata6 was activated by Hbxip KO (Fig. 2F). All these data suggest an essential role of Hbxip in ESC self-renewal.

Transcriptomic analysis identified 787 upregulated genes and 1494 downregulated genes shared by H^−/−-1 and H^−/−-2 ESCs, compared with WT ESCs (Fig. 2G, Table S2). Gene ontology (GO) annotation revealed that several developmental terms, such as system development, nervous system development and ectoderm development, were enriched in the downregulated genes (Fig. 2H, Fig. S2A). These data imply the involvement of Hbxip in ESC and epiblast differentiation.

We then tested the role of Hbxip in the differentiation of ESCs. Two methods were used to differentiate ESCs, leukemia inhibitory factor (LIF) withdrawal or embryoid body (EB) formation in hanging drops. Under these two differentiation conditions, Hbxip^−/− ESCs failed to activate differentiation genes, including ectodermal markers, nestin and Celsr2, and mesodermal markers, T and Dlx3 (Fig. 3A,B). Through RNA-sequencing (RNA-seq) analysis, we identified 1028 downregulated genes and 482 upregulated genes in differentiated Hbxip^−/− cells induced by LIF withdrawal, compared with differentiated WT cells (Fig. 3C), and 719 downregulated genes and 655 upregulated genes in day 4 Hbxip^−/− EBs, compared with WT EBs (Fig. 3D, Table S3). Differentiated Hbxip^−/− cells by LIF withdrawal and day 4 Hbxip^−/− EBs shared 481 downregulated genes and 325 upregulated genes (Fig. 3E, Table S3). GO analysis revealed that genes involved in system development, mesoderm development, ectoderm development, embryo development and cell differentiation were enriched in the downregulated genes (Fig. 3F, Fig. S2B).

To further confirm the roles of Hbxip in pluripotency maintenance and ESC differentiation, rescue experiments were performed in H^−/−-1 ESCs by expressing either short-isoform (H-S) or long-isoform (H-L) Hbxip. Both H-S and H-L rescued the expression levels of Nanog and Oct4 proteins in undifferentiated ESCs (Fig. 3G, Fig. S2C), as well as the expression of differentiation genes, nestin, Celsr2, T and Dlx3, in differentiated cells (Fig. 3H). These data indicate that Hbxip is required for the proper differentiation of ESCs, particularly toward the ectodermal and mesodermal lineages.

Given the differentiation defects of Hbxip^−/− ESCs, we tested whether the epiblast properly differentiates into three germ layers, particularly the ectoderm and mesoderm. Immunohistochemical (IHC) analysis with anti-Hbxip antibody allowed us to distinguish Hbxip^−/− embryos from WT and Hbxip^+/− embryos (Fig. 4). Hematoxylin and Eosin (H&E) staining showed that the development of the amnion cavity was retarded in E7.5 Hbxip^−/− embryos (Fig. 4A). Immunofluorescent staining of Oct4 demonstrated that the epiblast in E7.5 Hbxip^−/− embryos was much smaller than that in E7.5 WT and Hbxip^+/− embryos, indicating defective epiblast formation (Fig. 4A). At 8.5 days postcoitum, the morphology of WT and Hbxip^+/− embryos changed drastically with the initiation of somitogenesis. By contrast, the development of Hbxip^−/− embryos did not progress further. Importantly, with the formation of three germ layers, Oct4 expression was diminished in E8.5 WT and Hbxip^+/− embryos, whereas Hbxip^−/− embryos retained Oct4 expression in the epiblast (Fig. 4B). We further examined the expression of germ layer markers in E7.5 embryos, and found that ectodermal marker nestin and mesodermal marker T were not expressed in E7.5 Hbxip^−/− embryos, while the expression of endodermal marker Gata4 was not affected by Hbxip KO (Fig. 4C). Taken together, these results suggest that Hbxip^−/− embryo lethality is due to the defects in epiblast proliferation, ectodermal and mesodermal differentiation.

Next, we investigated the molecular mechanism of Hbxip in embryonic development and ESC differentiation. It has been shown that HBXIP functions as a transcriptional cofactor in many cancers (Liu et al., 2012, 2014, 2013; Xu et al., 2014, 2013; Yue et al., 2013; Zhang et al., 2013; Wang et al., 2017a). We first determined the subcellular distribution of Hbxip in ECSs by western blot analysis of cytoplasmic and nuclear fractions of ESCs. The results showed that Hbxip was almost exclusively distributed in the cytoplasm (Fig. 5A), thus excluding its function as a transcriptional cofactor.

To understand the molecular mechanism of Hbxip, co-immunoprecipitation (co-IP) coupled with mass spectrometric analysis was performed to identify Hbxip-interacting proteins. We identified 128 proteins in the H-S and H-L co-IP samples, but not in the control co-IP sample (Fig. 5B, Table S4). GO analysis revealed that these Hbxip-interacting proteins were enriched in mTOR signaling, macroautophagy, PI3K cascade and insulin signaling (Fig. 5C). It has been shown that HBXIP is a member of the Ragulator complex, which recruits mTORC1 to the lysosome, therefore activating the mTORC1 signaling pathway (Bar-Peled et al., 2012). Consistently, three Ragulator components, Lamtor1 (p18), Lamtor2 (p14) and Lamtor3 (MP1), as well as an mTORC1 subunit, Rptor, were identified as Hbxip-interacting proteins (Fig. 5C, Table S4). Thus, we speculated that mTORC1 is the key downstream factor of Hbxip in ESCs and embryos. To test this hypothesis, we first showed that the activity of mTORC1 was reduced in Hbxip^−/− ESCs and embryos, as indicated by the level of phosphorylated S6K1 and 4EBP1 (Fig. 5D, Fig. S3). Next, we tried to reactivate mTORC1 to rescue the differentiation defects of Hbxip^−/− ESCs. However, even though knockout of Tsc1, or overexpression of Rheb or Rptor fused with the lysosome-targeting signal of Rheb (R-R15) (Sancak et al., 2010; Düvel et al., 2010; Long et al., 2005), successfully activated mTORC1 in WT ESCs, these strategies failed to activate the mTORC1 signaling in Hbxip^−/− ESCs (Fig. 5E,F, Fig. S4), demonstrating the essential role of Hbxip in mTORC1 activation.

Failure in activating mTORC1 in Hbxip^−/− ESCs indicated that Hbxip is essential for mTORC1 activation. Yet, it rendered the rescue experiment impossible. To prove that inactivation of mTORC1 is responsible for the self-renewal and differentiation defects in Hbxip^−/− ESCs, an alternative strategy was applied. Instead of the rescue experiment, we knocked out genes encoding other components of the Ragulator complex, such as Lamtor3 and Lamtor4, in ESCs (Fig. S5). Similar to Hbxip^−/− ESCs, Lamtor3^−/− and Lamtor4^−/− ESCs displayed self-renewal defects, including reduced protein levels of Oct4 and Nanog, downregulated Oct4, Nanog, and Sox2 mRNA expression, slower proliferation rate and decreased colony forming capacity (Fig. 6A-D). Moreover, upon differentiation, Lamtor3^−/− and Lamtor4^−/− ESCs also failed to activate ectodermal and mesodermal markers (Fig. 6E,F), just as we observed in Hbxip^−/− ESCs (Fig. 3A,B). All these data indicate that inactivation of mTORC1 leads to the self-renewal and differentiation defects in ESCs.

In this study, we demonstrated that Hbxip is essential for embryonic development and ESC differentiation. Hbxip^−/− ESCs, as well as Lamtor3^−/− and Lamtor4^−/− ESCs, with disrupted Ragulator complex, displayed differentiation defects toward ectodermal and mesodermal lineages (Fig. 3A,B, Fig. 6E,F). Moreover, Hbxip^−/−, p14^−/− and p18^−/− mice, lacking subunits of the Ragulator complex, are embryonic lethal, and retarded growth of these embryos was detected at the same developmental stage, around E7-E8 (Fig. 1C, Table 1) (Nada et al., 2009; Teis et al., 2006). These data indicate that the phenotypes of Hbxip^−/− embryos and ESCs are mainly caused by loss of function of the Ragulator complex, rather than lack of the transcriptional co-activator role of Hbxip.

The Ragulator complex is required for the activation of mTORC1 by amino acids (Sancak et al., 2010; Bar-Peled et al., 2012). Consistently, the mTORC1 activity was reduced in Hbxip^−/− ESCs (Fig. 5D, Fig. S3A). KO of Tsc1 and overexpression of Rheb or R-R15 failed to activate mTORC1 in the absence of Hbxip (Fig. 5E,F). In addition, the reduced mTORC1 activity by disruption of the Ragulator complex allowed us to investigate the function of mTORC1 in post-implantation embryo development. Deletion of mTORC1 components, mTor and Rptor, results in peri-implantation embryonic death around E5.5-E6.5 (Gangloff et al., 2004; Murakami et al., 2004; Guertin et al., 2006), thus preventing studying the role of mTORC1 in later stage of embryogenesis. Nevertheless, Hbxip^−/−, p14^−/− and p18^−/− embryos, in which mTORC1 activity is presumably attenuated, developed further after implantation, and retarded growth appeared around E7-E8. The reduced size of E7.5 Hbxip^−/− embryos might be attributed to slower cell proliferation rate caused by decreased mTORC1 activity. Moreover, the differentiation defects of Hbxip^−/− ESCs (Fig. 3A,B) imply that the Hbxip^−/− epiblast might be defective in differentiation. Indeed, the differentiation of the Hbxip^−/− epiblast was compromised, as indicated by the sustained expression of Oct4 in E8.5 Hbxip^−/− embryos, and the failure in expressing ectodermal marker nestin and mesodermal marker T in in E7.5 Hbxip^−/− embryos (Fig. 4), suggesting that mTORC1 activity is pivotal for epiblast differentiation. Consistent with this, inhibition of mTOR by INK128 induces a paused pluripotent state in the blastocyst, mimicking diapause (Bulut-Karslioglu et al., 2016), further supporting the role of mTORC1 in the exit from pluripotency.

Even though Hbxip^−/− epiblasts and Hbxip^−/− ESCs were both defective in activating the expression of ectodermal and mesodermal markers, Oct4 expression was sustained in E8.5 Hbxip^−/− embryos, while the downregulation of Oct4 was not affected during the differentiation of Hbxip^−/− ESCs, most likely due to in vitro differentiation of ESCs being unable to recapture all the features of in vivo embryo development, even though it mimics the in vivo development of embryos. It is possible that the downregulation of Oct4 in developing embryos might require the activation of differentiation genes, whereas the repression of Oct4 could be triggered by LIF withdrawal during ESC differentiation, regardless of whether differentiation genes are activated or not. Another seemingly conflicting point is that differentiation genes were not activated in Hbxip^−/− ESCs, in which pluripotency genes were downregulated. This is likely due to the unique effect of mTOR inhibition in ESCs. mTOR inhibition induces pausing of ESCs, which display global transcriptional suppression, including both pluripotency genes and differentiation genes (Bulut-Karslioglu et al., 2016). Thus, the reduced mTORC1 activity in Hbxip^−/− ESCs may be responsible for simultaneous downregulation of pluripotency genes and differentiation genes.

mTOR signaling is involved in growth and disease through regulating gene transcription, mRNA degradation, protein synthesis, autophagy, and lipid and carbohydrate metabolism (Giguère, 2018; Villa et al., 2021; Cho et al., 2021; Mossmann et al., 2018; Saxton and Sabatini, 2017). Further studies are required to elucidate how mTORC1 regulated by the Ragulator complex promotes the differentiation of the epiblast and ESCs. Hbxip^−/−, Lamtor3^−/− and Lamtor4^−/− ESCs provide an excellent in vitro experimental system for mechanistic investigation.

Nankai Animal Care and Use Committee approved the use of mice for this research (approval number: 2021-SYDWLL-000469). The care and use of experimental animals complied with relevant institutional and national animal welfare laws, guidelines and policies. EIIa-cre mice (Lakso et al., 1996), which express Cre at a very early stage of pre-implantation embryogenesis, most likely at the zygote stage, and are useful for whole-body and germ line deletion of floxed allele, were purchased from Shanghai Model Organisms. Hbxip^−/− mice were obtained by crossing Hbxip^+/− mice. All mice were maintained in the C57BL/6 background. Genotyping was performed as described previously (Qin et al., 2019). DNA was isolated from dissected embryos or the tail tips of 2-week-old mice, and then analyzed by PCR. The sequences of primers (illustrated in Fig. 1A) were as follows: F1, 5′-TTTTTGTCACTCTCGCCTTTG-3′; R1, 5′-GCTGGTATGTACTCACCCATT-3′; F2, 5′-GCTCTATGGCTTCTGAGGCGGAA-3′; R2, 5′-CTCATCCGACAGGGTACCACGG-3′.

Embryo manipulation experiments were carried out as described previously (Zhou et al., 2018). Female Hbxip^+/− mice (4-6 weeks) were induced to superovulate by intraperitoneal injection of 5 IU pregnant mare serum gonadotropin (Calbiochem) and, 48 h later, 5 IU human chorionic gonadotropin (hCG; Sigma-Aldrich). Then, females were mated with male Hbxip^+/− mice overnight. The next morning, females were checked for the presence of a vaginal plug, designated as E0.5. Zygote and E3.5 embryos were flushed from oviducts at day 0.5 or day 3.5 post-hCG. E6.5, E7.5, E8.5 and E9.5 embryos were collected at corresponding time points after successful mating.

Embryos were dissected from mice immediately after euthanasia, fixed in 4% paraformaldehyde for up to 24 h, stored in 70% ethanol and embedded in paraffin. Then, 5 μm-thick sections were prepared using a rotary microtome (Leica) and mounted on glass slides. After deparaffinization, sections were stained with H&E for histological analysis.

After deparaffinization, sections were boiled in 1× citrate buffer (ORIGENE, ZLI-9064) and then microwaved at low power for 15 min. After cooling to room temperature (RT), the sections were blocked with 5% bovine serum albumin at RT for 45 min, and then incubated with primary antibody (anti-Hbxip, Cell Signaling Technology, 14633S, 1:200) at RT for 1 h. After three washes with 1×PBS buffer and 3% H₂O₂ treatment for 10 min, the slides were incubated with goat anti-rabbit IgG conjugated to horseradish peroxidase (HRP) (ORIGENE, ZB-2301) at RT for 1 h. HRP activity was detected by a Vectastain ABC (avidin-biotin-peroxidase) kit (Vector Laboratories), as recommended by the manufacturer. The slides were examined using a Leica Microscope, and images were captured with a Leica DFC420C camera.

IF analysis of paraffin sections was carried out as described previously (Qin et al., 2019). After deparaffinization, the sections were blocked with 5% goat serum at RT for 45 min, then incubated with primary antibody [anti-Oct4, Abcam, ab181557, 1:200; anti-nestin, Absin, abs137231, 1:100; anti-Brachyury (T), Abmart, T58977, 1:100; anti-Gata4, Abcam, ab84593, 1:200; anti-p-S6K1, Sigma-Aldrich, SAB4301518, 1:100] at RT for 1 h. After washing three times in 1×PBS buffer, the slides were incubated with Alexa Fluor 594-conjugated goat anti-rabbit IgG (Invitrogen, A11037) at RT for 1 h. The nuclei were stained with 4′,6-diamidino-2-phenylindol (DAPI; Thermo Fisher Scientific). All images were captured with a Leica DM3000 microscope with DFC420C camera.

Mouse ESCs (E14 and its derivatives) were cultured in high-glucose Dulbecco's modified Eagle medium (GIBCO) supplemented with 15% fetal bovine serum (Hyclone), 1% penicillin/streptomycin, 1% L-glutamine, 1% β-mercaptoethanol (Sigma-Aldrich), 1% non-essential amino acids (Invitrogen) and 1000 U/ml LIF (ESGRO, Chemicon), plated on gelatin-coated tissue culture dishes and grown in humidified 5% CO₂ at 37°C. Transfection was carried out with Lipofectamine^® 3000 (Invitrogen) according to the manufacturer's instructions. For construction of stably transfected clones, puromycin (1.25 μM) selection was applied for 5-7 days, starting from 24 h after transfection. All ESC lines were routinely tested for mycoplasma contamination.

pX330-U6-Chimeric_BB-CBh-hSpCas9 (pX330) vector (Addgene plasmid #42230) was used to construct gene knockout plasmids targeting Hbxip, Tsc1, Lamtor3 and Lamtor4. sgRNA was designed in Zhang Feng's online website (http://crispr.mit.edu/). A pair of sgRNA oligonucleotides containing the guide sequence were annealed and ligated to the BbsI-linearized vector pX330. The sgRNA oligonucleotides targeting Hbxip were 5′-CACCGTCTCACTATTCTCTAGGCCG-3′ and 5′-AAACCGGCCTAGAGAATAGTGAGAC-3′. The sgRNA oligonucleotides targeting Tsc1 were 5′-CACCGGTGGTCAGGATGTGCAATAC-3′ and 5′-AAACGTATTGCACATCCTGACCACC-3′. The sgRNA oligonucleotides targeting Lamtor3 were 5′-CACCGTCCTGTACAAAAAGTTGCCA-3′ and 5′-AAACTGGCAACTTTTTGTACAGGAC-3′. The sgRNA oligonucleotides targeting Lamtor4 were 5′-CACCGTGTGCTGACCAACTCCGAGA-3′ and 5′-AAACTCTCGGAGTTGGTCAGCACAC-3′; 5′-CACCGCTGCTTGCTCATCGTTCTCA-3′ and 5′-AAACTGAGAACGATGAGCAAGCAGC-3′.

ESCs (1×10⁵) were plated in 12-well plates in triplicates. Cells were passaged continuously and counted every 2 days.

ESCs were collected and dissociated in cold lysis buffer (20 mM HEPES, pH 7.7; 5 mM KCl; 1.5 mM MgCl₂; 2 mM DTT; 2 mM PMSF) for 30 min on ice. Then, the cell lysate was centrifuged at 12,000 g, for 15 min at 4°C. Protein concentration was measured using a BCA Protein Assay Kit (Beyotime). The samples were resolved with SDS-PAGE gel and transferred onto PVDF membrane. Then, the PVDF membrane was incubated in blocking buffer for 1 h at RT, followed by incubating with primary antibodies (anti-Hbxip, Cell Signaling Technology, 14633S, 1:2000; anti-Nanog, Bethyl Laboratories, A300-397A, 1:2000; anti-Oct4, Santa Cruz Biotechnology, sc-5279, 1:1000; anti-tubulin, Sigma-Aldrich, ASB4800087, 1:2000; anti-S6K1, Sigma-Aldrich, SAB4502686, 1:1000; anti-p-S6K1, Sigma-Aldrich, SAB4301518, 1:1000; anti-H3, Abcam, ab1791, 1:2000; anti-Tsc1, Absin, abs131176, 1:2000; anti-MP1/Lamtor3, Novus, NBP1-50631, 1:2000; anti-C7orf59/Lamtor4, Cell Signaling Technology, D4P6O, 1:2000; anti-p-4EBP1, Cell Signaling Technology, 2855, 1:2000; anti-4EBP1, Cell Signaling Technology, 9452, 1:2000) at 4°C overnight. After washing three times, the membrane was incubated with HRP-conjugated secondary antibodies at RT for 1 h. Immunoreactivity was detected by ECL Plus (Beyotime). Digital images were taken with an automatic chemiluminescence imaging analysis system (Tanon).

ESC extracts were prepared using lysis buffer (20 mM Tris-HCl, pH 8; 137 mM NaCl; 10% glycerol; 1% NP-40; 2 mM EDTA, supplemented with protease inhibitor and PMSF). FLAG antibody (M2)-conjugated magnetic beads were washed in TBS buffer (50 mM Tris-HCl, 150 mM NaCl, pH 7.4) three times. ESC extracts were incubated with 15 μl M2 beads at 4°C overnight. The immunoprecipitates were washed three times using TBS buffer, eluted in 2× SDS-PAGE loading buffer and subjected to mass spectrometric analysis.

Total RNA was purified using TRIZOL (Ambion) according to the manufacturer's instructions. cDNA was synthesized using a Reverse Transcription Kit (Roche). PCR reactions were performed with FastStart Universal SYBR Green Master (Roche) in a Bio-Rad IQ5 system. PCR cycling conditions were 95°C for 2 min, 40 cycles of 95°C for 15 s, 58°C for 15 s and 72°C for 30 s, and then a melting curve of the amplified DNA was acquired. Quantification of target genes was normalized with β-actin. Primers are listed in Table S1.

Total RNA was purified using an RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Construction of an RNA-seq library, sequencing with BGISEQ-500 and bioinformatic analysis were performed by BGI Tech (Shenzhen, China).

After washing once with PBS, mouse ESCs were stained with AP Substrate Kit III (Vector Laboratories), according to the manufacturer's instructions.

Mouse ESCs were suspended in ESC medium and cultured in a six-well plate (800 cells per well). After 4-6 days of culture, ESCs were stained for AP. The AP-positive colonies were counted under a microscope.

Mouse ESCs were suspended in LIF-free ESC medium (40,000 cells/ml) and cultured in 25 μl hanging drops. EBs were collected on day 4.

Cells were collected and fixed with 70% ethanol, then stained with staining buffer (5 μg/ml propidium iodide, 1 mg/ml RNase A) for 30 min at 37°C and analyzed by FACSCalibur flow cytometer (BD Biosciences). Modfit software was used to analyze the results.

Apoptosis was detected by an Annexin V-EGFP Apoptosis Detection Kit (Beyotime, C1067S) as recommended by the manufacturer. Cells were harvested and washed in 1× PBS once. Then they were resuspended in 195 μl Annexin V- EGFP binding buffer supplemented with 5 μl Annexin V-EGFP and 10 μl propidium iodide, incubated at RT for 20 min and analyzed by a FACSCalibur flow cytometer (BD Biosciences). Flow Jo software was used to analyze the results.

Cells were incubated with 0.3 μg/ml nocodazole for 3-5 h to enrich cells at metaphase. These metaphase-enriched cells were harvested and exposed to 0.075 M KCl solution at 37°C for 25 min, then fixed with cold methanol: glacial acetic acid (3:1) buffer at least four times and spread onto clean slides. Finally, the cells were stained with DAPI. Cells with more than 39 chromosomes were counted.

All images were processed with Photoshop 2021 (Adobe) and ImageJ (National Institutes of Health). All data were analyzed using Excel (Microsoft) and Prism (GraphPad Software). All experiments were confirmed with at least three independent experiments, and at least three control or mutant embryos (the exact sample size is shown in figure legends) were used for immunostaining. Statistical analyses were performed with χ² test, unpaired two-tailed Student's t-test or two-way ANOVA, as indicated in the figure legends. The quantitative results were presented as the mean±s.d. P<0.05 was considered statistically significant.

We thank the core facility at College of Life Sciences, Nankai University.

The peer review history is available online at https://journals.biologists.com/dev/article-lookup/doi/10.1242/dev.200527.