A6P10 mES cells (a gift from Dr. Chyuan-Sheng Lin, Columbia University, USA) and 46C mES cells (ES cell line in which EGFP was replaced into the open reading frame of Sox1 gene, provided by Dr. Qilong Ying, University of Southern California, USA) were cultured in ES medium (DMEM (Gibco) with 15% FBS, 2 mM GlutaMAX (Gibco), MEM nonessential amino acids, β-mercaptoethanol (Gibco), tylosin, 1% Pen/Strep (Gibco)) supplemented with LIF (Chemicon) on 0.2% gelatin-coated dishes. To induce neuronal differentiation, 46C cells were cultured in N2B27 medium (DMEM/F12 (Gibco), Neurobasal medium (Gibco), N2 supplement (Invitrogen), B27 supplement (Invitrogen), 1 mM GlutaMAX (Gibco), 0.1 μM β-mercaptoethanol (Gibco), 1% Pen/Strep (Gibco)) on 0.2% gelatin-coated tissue culture dish (Falcon). N2B27 medium was changed every 2 days. Embryoid body (EB) formation was induced by hanging drop method. Briefly, 20 μl drops (including 600 cells) of dissociated ES cells with ES medium plus 20% FBS were placed on inverted lids of petri-dish (Falcon), which was filled with 3 ml PBS. After incubation for 3 days, EB was plated on a 0.2% gelatin-coated dish in ES medium supplemented with 20% FBS. The medium was changed every 2 days.

RNA obtained from a mixture of undifferentiated and differentiated mES cells was used to clone Tcf1, Lef1, Tcf3, and Tcf4. Wild-type and dominant negative forms of Tcfs/Lef1 were inserted into the pCS2-HA3 vector. HA-tagged Tcfs and Lef1 were transferred into the pCAG-1 vector (modified from pPCAGIZ vector). The shRNA targeting sequences against mouse Lef1 were designed using the web tool from Promega.

Sense (5’-GATCCCCGACTTAGCCGACATCAAGTTTCA AGAGAACTTGATGTCGGCTAAGTCTTTTTGGAAA-3’) and antisense (5’-AGCTTTTCCAAAAAGACTTAGCCGA CATCAAGTTCTCTTGAAACTTGATGTCGGCTAAGTCGGG-3’) oligonucleotides were annealed and ligated into the BglII and HindIII sites of the pSUPER.retro.puro vector (Oligoengine). HA-tagged Tcfs and Lef1 plasmids were electroporated by Amaxa Nucleofector according to the manufacturer’s protocol and then selected with 50 μg/ml Zeocin (Invitrogen). The shLef1 plasmid was electroporated by Amaxa Nucleofector technology^TM and selected with 1 μg/ml puromycin (Sigma).

ES cells were lysed in lysis buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.5% Triton X-100, 50 mM NaF, 2 mM EDTA, 100 μM Na-orthovanadate, 1 mM PMSF, 5 μg/ml leupeptin, and 1 μM pepstatin A). The lysates were centrifuged at 13,000 rpm for 15 min at 4℃ and the supernatant was collected and used for Western blotting. Bradford (Bio-Rad) reagent was used to measure the quantity of protein. Equal amounts of protein were boiled in Laemmli sample buffer and resolved via SDS-PAGE followed by transfer to a PVDF membrane (Pall). Anti-β-actin (Sigma), anti-TCF1 (Cell Signaling), anti-LEF1 (Cell Signaling or Santa Cruz Biotechnology), anti-TCF3 (Santa Cruz Biotechnology), anti-TCF4 (Santa Cruz Biotechnology) antibodies were used as primary antibodies.

ES cells were plated layered on a 12-well plate and cultured with or without LIF. After washing twice with PBS, cells were fixed with 4% paraformaldehyde for 10 min at room temperature followed by PBS washing several times. AP staining was performed with NBT/BCIP (4-nitro blue tetrazolium chloride, Roche; 5-Bromo4-chloro-3-indolyl-phosphate, Roche) staining buffer (0.1 M Tris, pH 9.5, 100 mM NaCl, 5 mM MgCl₂) for 15 min in the dark.

Cells were cross-linked with 1% formaldehyde (Sigma) at room temperature for 10 min with gentle shaking and then incubated in 0.125 M glycine for 5 min with gentle shaking. Cells were washed twice with ice-cold PBS before harvest. Re-suspended cells with hypotonic buffer (10 mM Hepes-KOH, pH 7.8, 10 mM KCl, 1.5 mM MgCl₂) were swollen on ice for 10 min and passed through a 26.5-gauge needle 6 times. After centrifugation at 1000 g for 5 min at 4℃, the pellets were incubated in nuclei lysis buffer (1% SDS, 50 mM Tris-HCl, pH 8.0, 10 mM EDTA) for 10 min on ice with occasional vortexing. Chromatin was sheared to an average length of 0.2∼1 kb by sonication on ice. Cellular debris was removed by centrifugation (13,000 rpm, 15 min, 4℃), and the concentration of supernatant was determined with a spectrophotometer. The appropriate volume of chromatin was diluted 1/10 in ChIP dilution buffer (0.01% SDS, 20 mM Tris-HCl, pH 8.0, 167 mM NaCl, 1.2 mM EDTA, 1.1% Triton X-100) followed by preclearance at 4℃ for 2 h with 10 μl protein A/G plus-agarose beads (Santa Cruz Biotechnology). For immunoprecipitation, goat, rabbit-IgG (Bethyl), anti-Lef1 (Santa Cruz Biotechnology or Cell Signaling), anti-Tcf3 (Santa Cruz Biotechnology), anti-Tcf1 (Cell Signaling) antibodies were used at 4℃ overnight. Immunoprecipitated chromatins were eluted, and then reverse cross-linked by the addition of 0.3 M NaCl at 65℃ overnight. Following phenol-chloroform extraction and ethanol precipitation, the DNA was dissolved in 50 μl TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA). For PCR, 2 μl DNA was used. A Nanog primer (forward: 5’-TCTGCTTATACACA GAAGCC-3’ and reverse: 5’-AAGTGCCTCAGCCGTCTA AG-3’) was used for PCR.

Total mRNA from ES cells were isolated by TRIzol (Invitrogen) according to manufacturer protocol. cDNA was made from total RNA using reverse transcriptase (Promega) with random primers. For PCR, taq polymerase (Cosmo genetech) was used and visualized by 1% of agarose gel. GAPDH or 18s rRNA were used for normalization of mRNA expressions. Primers used for RT-PCR can be found in Supplementary Table S1.

Wnt signal stimulates the self-renewal and differentiation of ES cells into various lineages. We investigated the varying response of ES cells in the presence of Wnt signal. As Tcfs and Lef1 act as the downstream effectors in Wnt signal transmission, we speculated that the differential response may result from differential function of each Tcfs and Lef1.

Prior to investigation of the function of Tcfs and Lef1, the change in Tcfs and Lef1 mRNA levels during differentiation was detected by RT PCR (Fig. 1A). Under all conditions of differentiation, the Tcf3 transcript was found to be increased during the early days of differentiation. This data is consistent with previous reports that TCF3 is involved in the early stage of differentiation by suppressing pluripotent genes (32). In addition, Tcf4 transcript was also gradually increased during differentiation. Interestingly, Lef1 transcript was transiently induced after the removal of LIF although this induction was maintained in N2B27 media. There was no significant change in Tcf1 expression following LIF depletion and N2B27. However, Tcf1 transcript decreased from day 6 post-EB formation. These results suggest that Lef1 might play a role in the initial stages of differentiation and Tcf4 might be associated with late differentiation.

It has been known that the function of Tcfs/Lef1 is affected by post-translational modifications such as phosphorylation and sumoylation (26). Thus, we performed western blotting analysis to investigate the protein levels of Tcfs and Lef1 (Fig. 1B). The protein levels of Tcfs and Lef1 were dynamically changed compared to the changes in Tcfs and Lef1 transcripts. Similar to RT PCR data, Lef1 protein was expressed transiently at an early stage of differentiation. In contrast, the expression of Tcf1 and Tcf3 proteins revealed different patterns of mRNA levels. Under the conditions of LIF depletion or N2B27 medium, the levels of Tcf1 and Tcf3 protein were elevated until the day Lef1 protein was expressed followed by a sharp decline during subsequent differentiation. The expression pattern of Tcf1 protein during EB formation was similar to that under other conditions, whereas the pattern of Tcf3 expression differed completely. After EB formation for 3 days, the loss of Tcf3 protein was inconsistent with RT PCR data, suggesting that the Tcf3 protein might be degraded by unknown mechanisms during EB formation. These results suggest that measurement of protein levels might be more appropriate for Tcfs/Lef1 studies in ES cells. The variable expression of Tcfs and Lef1 proteins under differentiated and undifferentiated conditions supports our hypothesis that the role of each Tcfs and Lef1 may differ in self-renewal and differentiation of ES cells.

To investigate the functions of Tcfs and Lef1 proteins, Tcfs and Lef1 constructs cloned from RNA expressed in ES cells were stably introduced into ES cells (Fig. 2A). First, we tested which Tcfs and Lef1 proteins regulated the stemness of ES cells using alkaline phosphatase (AP), a marker of self-renewal, staining (Fig. 2B and 2C). When LIF was removed from media, AP positive cells (AP⁺) were remarkably reduced and the morphology was lost in ES cells expressing empty vector. Likewise, AP⁺ decreased steadily in ES cells expressing Lef1, Tcf3, or Tcf4 (Fig. 2C). In contrast, AP⁺ was retained and the morphology remained intact in ES cells expressing Tcf1 but not dominant-negative Tcf1 (TCF1-DN) even after the removal of LIF for 12 days (Fig. 2B). Moreover, ES cells expressing Tcf1 were still stained with the antibody against SSEA1, a marker for proliferation (data not shown). These results suggested that Tcf1 may regulate the self-renewal of ES cells.

To further explore how Tcf1 regulates the self-renewal of ES cells, we checked the levels of core stem cell markers; Oct4, Sox2 and Nanog upon Tcf1 overexpression at ES cells stage and Day 4 of differentiation. We found that during ES cells differentiation, the levels of Oct4, Sox2 and Nanog were reduced. However, upon overexpression of Tcf1, the reduction of stem cell markers was hindered. These effects were most predominantly seen in Nanog levels compared to Sox2 and Oct4 (Fig. 3A). We speculated that Tcf1 is the major regulator for Nanog, therefore we focused on Nanog in subsequent experiments.

Nanog is vital for the self-renewal of ES cells (7, 8) and its promoter is regulated by Tcf3 (27, 28). Since all Tcfs and Lef1 proteins bind to the same consensus sequence, we tested whether Tcf1 regulates the expression of Nanog in days specific manner. We found that Nanog expression was reduced in control ES cells after day 4 of LIF depletion, whereas the expression of Tcf1 inhibited the reduction in Nanog expression (Fig. 3B). These data indicate that ectopic Tcf1 induced the self-renewal of ES cells by promoting Nanog expression.

To investigate whether Tcf1 binds to Nanog promoter in practice, we performed a chromatin immunoprecipitation (ChIP) assay using an endogenous antibody for Tcf1, Tcf3, and Lef1 (Fig. 3C). In the presence of LIF, Tcf1 bound to the promoter of Nanog. Consistent with previous reports (27, 28), Tcf3 also interacted with Nanog promoter, suggesting that Tcf3 acts as an limiter of ES cell self-renewal. Since Lef1 protein is not expressed in the presence of LIF, no interaction between Lef1 and Nanog promoter was observed. Based on these data, we hypothesized that the occupancy of Tcf1 on Nanog gene may be reduced during differentiation, which was demonstrated using a ChIP assay for Nanog gene after LIF removal for 2 days. As expected, the occupancy of Tcf1 was reduced relative to that of Tcf3 and Lef1 (Fig. 3B). The protein levels of Tcf1, Tcf3, and Lef1 were the highest 2 days after LIF removal or after incubation with N2B27 media (Fig. 1B). Thus, the reduced binding of Tcf1 with Nanog promoter might be attributed to competition with Tcf3 or Lef1 protein. To support this possibility, ES cells stably expressing Tcf1 were used in the ChIP assay. Interestingly, most of the Nanog promoter was occupied by Tcf1 (Fig. 3B). Taken together, these results indicate that Tcf1 promotes the transcription of Nanog by binding with its promoter. The binding of Tcf1 protein with Nanog promoter is prevented by the increasing levels of Tcf3 or Lef1 protein during differentiation.

Nanog has been shown to be necessary for self-renewal and to prevent differentiation of ES cells in the absence of LIF (7, 8). Since ectopic expression of Tcf1 induced Nanog expression even after LIF depletion, we investigated whether overexpression of Tcf1 inhibited normal differentiation. For differentiation into neural precursors, we used Sox1-GFP knock-in (46C) ES cells in which the differentiated cells are easily detected based on the GFP signal when cultured with N2B27 medium (33). In control 46C ES cells, GFP-positive cells appeared after 6 days of incubation in N2B27 medium. However, the ectopic expression of Tcf1, but not TCF1-DN, inhibited the emergence of GFP-positive cells (Fig. 3D, upper panel). Consistent with the GFP signal, the expression of Sox1, a marker of neural precursors, was not induced in 46C ES cells expressing Tcf1 (Fig. 3D, lower panel). In addition, Nanog expression was not declined compared to the control cells after 6 days of incubation in N2B27 medium. These data indicate that overexpression of Tcf1 results in the inhibition of differentiation into neural precursors via maintaining the levels of Nanog.

To determine whether the inhibition of differentiation by Tcf1 overexpression occurred only in the case of differentiation into neural precursors, we further analyzed the cardiomyocyte differentiation via EB formation (Fig. 3D). While the expression of Nanog was diminished starting from day 5 in control EB, it persisted even after 8 days in EB expressing Tcf1. Moreover, the expression of marker genes related to cardiomyocyte differentiation was delayed in EB-expressing Tcf1. These data show that the ectopic expression of Tcf1 leads to delayed differentiation into cardiomyocytes. Overall, the overexpression of Tcf1 prevented normal differentiation via maintaining the level of Nanog although it is possible that Tcf1 may still regulate the genes related to differentiation.

AP staining revealed that the ectopic expression of Lef1 did not induce self-renewal of ES cells. Interestingly, we also found that the expression of dominant-negative Lef1 (Lef1-DN) maintained ES cell self-renewal even after 12 days of LIF removal (Fig. 2C). In addition, endogenous Lef1 protein was transiently expressed following differentiation (Fig. 1B). These results suggested that Lef1 may be related to differentiation. To test this possibility, the level of Lef1 was reduced by shRNA against β-catenin-binding domain of Lef1 to knock down full-length Lef1 but not dominant-negative Lef1. Two days after LIF removal, the expression of Lef1 protein was remarkably reduced by shRNA for Lef1 (Fig. 4A). In order to identify the effect of Lef1 reduction, 46C ES cells expressing shControl or shLef1 were differentiated into neural precursors in N2B27 medium. Notably, 46C ES cells expressing shLef1 did not express Sox1-GFP compared with control after differentiation (Fig. 4B). Nevertheless, the level of Nanog in 46C ES cells expressing shLef1 was reduced in the control after 7 days of differentiation (Fig. 4B, lower panel). These results indicate that knockdown of Lef1 inhibits differentiation of ES cells into neural precursors without upregulation of Nanog expression. Similar to shLef1, the overexpression of dominant negative Lef1 blocked the differentiation into neural precursors. We further tested whether ectopic expression of Lef1 enhanced the differentiation of ES cells. Surprisingly, overexpression of full-length Lef1 did not enhance differentiation, which was blocked, instead (Fig. 4C). This result may be explained in part by the transient expression of endogenous Lef1 protein at the onset of differentiation (Fig. 1B). The effect of Lef1 reduction was also evaluated in the differentiation into cardiomyocytes. The expression of α-MHC, a marker for cardiomyocytes, was attenuated in ES cells expressing shLef1. Since EB can also be partially differentiated into endoderm, we investigated the effect of shLef1 on endoderm differentiation. Knockdown of Lef1 reduced the expression of GATA4, an endoderm marker (Fig. 4D). These data show that the reduced Lef1 expression attenuated the differentiation into cardiac mesoderm and endoderm. Taken together, these results suggest that the transient expression of Lef1 protein at the onset of differentiation is required for differentiation to three germ layers.