Tousled-like kinase 1 is a negative regulator of core transcription factors in murine embryonic stem cells

Although the differentiation of pluripotent cells in embryonic stem cells (ESCs) is often associated with protein kinase-mediated signaling pathways and Tousled-like kinase 1 (Tlk1) is required for development in several species, the role of Tlk1 in ESC function remains unclear. Here, we used mouse ESCs to study the function of Tlk1 in pluripotent cells. The knockdown (KD)-based Tlk1-deficient cells showed that Tlk1 is not essential for ESC self-renewal in an undifferentiated state. However, Tlk1-KD cells formed irregularly shaped embryoid bodies and induced resistance to differentiation cues, indicating their failure to differentiate into an embryoid body. Consistent with their failure to differentiate, Tlk1-KD cells failed to downregulate the expression of undifferentiated cell markers including Oct4, Nanog, and Sox2 during differentiation, suggesting a negative role of Tlk1. Interestingly, Tlk1 overexpression sufficiently downregulated the expression of core pluripotency factors possibly irrespective of its kinase activity, thereby leading to a partial loss of self-renewal ability even in an undifferentiated state. Moreover, Tlk1 overexpression caused severe growth defects and G2/M phase arrest as well as apoptosis. Collectively, our data suggest that Tlk1 negatively regulates the expression of pluripotency factors, thereby contributing to the scheduled differentiation of mouse ESCs.

and Nanog cooperatively regulate their target genes required for maintaining pluripotency and self-renewal and occupy the promoters of developmental genes associated with lineage specification whose expression is silenced in undifferentiated ESCs 2,16,17 .
The Tousled-like kinases (Tlk) are serine/threonine kinases that are evolutionarily conserved in both animals and plants 18 . Tousled, which was originally identified in the plant Arabidopsis thaliana, encodes a protein kinase that plays a role in both flower and leaf development 19 . TLK1 and TLK2 are mammalian homologs of Tousled that encode serine/threonine kinases that exhibit maximal activity in the S phase 20 . However, DNA damage induces the transient and rapid inactivation of TLKs via checkpoint kinase (Chk1)-dependent phosphorylation 21,22 . In Drosophila melanogaster and Caenorhabditis elegans, TLK depletion results in developmental arrest due to failures in proper chromatin organization and appropriate transcriptional regulation during development 23,24 . Existing data suggest that Tlk1 plays an important role in the regulation of development, but its functions in mESCs have not yet been investigated.
In this study, we investigated the roles of Tlk1 in mESC pluripotency and differentiation using gain-and loss-of-function approaches. Our results demonstrate that Tlk1-knockdown (KD) ESCs remained undifferentiated in the presence of LIF. In addition, Tlk1-depleted ESCs exhibited delayed silencing of pluripotency-related genes and maintained an undifferentiated state with high alkaline phosphatase (AP) activity even after the induction of differentiation. Conversely, the overexpression of Tlk1 in ESCs sufficiently abrogated the convex morphology and reduced AP activity. Interestingly, the ectopic expression of Tlk1 negatively controlled the expression of core ESC TFs and induced growth defects, most likely due to the arresting of ESCs in the G 2 /M-phase. Taken together, our data suggest that Tlk1 acts as a negative regulator of core pluripotency factors in mESCs.

Tlk1 deficiency does not affect mESCs in the undifferentiated state in the presence of LIF.
To determine the role of Tlk1 in ESC function, we established Tlk1-KD mESCs using a shRNA-based RNAi method. We constructed two different shRNAs to avoid off-target effects and confirmed KD efficiency via qRT-PCR and Western blotting ( Fig. 1A and E). The Tlk1-KD cells were morphologically indistinguishable from the control cells. In addition, no notable changes in AP staining were observed in the Tlk1-KD cells compared to the control KD cells (shLuc) under LIF-supplementation, suggesting that Tlk1 is not required for the self-renewal of ESCs (Fig. 1B). Next, we investigated the effects of Tlk1 depletion on the expression of several genes involved in pluripotency or differentiation using qRT-PCR and found out that Tlk1 deficiency did not affect the expression of pluripotency-associated genes, including Oct4, Nanog, and Sox2 (Fig. 1C). Similarly, the expression of genes associated with early differentiation, namely Flk1 and Nkx2.5 for the mesoderm, Fgf5 and Tubb3 for the ectoderm, and Id2 and Hand1 for the trophectoderm, was not significantly changed in Tlk1-KD cells compared with control KD cells (Fig. 1D). However, the expression of other differentiation-associated genes (GATA4 and GATA6 for the endoderm) was moderately increased (Fig. 1D). Consistent with this mRNA expression profile, the Western blotting analysis revealed that the Oct4, Nanog, and Sox2 levels in Tlk1 KD cells were not significantly changed relative to the control KD cells ( Fig. 1E and F). Thus, these results suggest that, although it might not be necessary for mESC pluripotency and self-renewal, Tlk1 might regulate the expression of endoderm-associated genes.
Tlk1 is required for the proper induction of scheduled differentiation. Because some differentiation-associated genes were aberrantly expressed in Tlk1-KD mESCs, we investigated whether the Tlk1-depleted cells were resistant to differentiation cues using a commitment assay, as previously described 25 . Embryoid bodies (EBs) can be used as a differentiation assay to test ESC pluripotency 25 . Tlk1-KD or control cells were allowed to form EBs for 12 days. The cells were then transferred to and cultured in media containing LIF for 5 days, and their differentiation patterns were assessed by AP staining (Fig. 2A). The EB-dependent differentiation of control KD cells (shLuc) proceeded normally without detectable delays in the presence of LIF, as confirmed by a low number of AP-stained colonies (Fig. 2B). Conversely, AP-positive ESC-like colonies were highly enriched in the EBs derived from two Tlk1-KD cell lines that were maintained in LIF-supplemented culture, indicating a failure of Tlk1-depleted mESCs to differentiate into an EB (Fig. 2B). Moreover, the failure to induce Tlk1-KD cells to differentiate was also supported by the quantitative analysis of AP-stained colonies (Fig. 2C). Because the delayed differentiation of Tlk1-KD cells might affect the formation of EBs, we examined if Tlk1 depletion influenced EB formation and observed EB morphology using phase-contrast microscopy. We found that Tlk1 depletion decreased the size of EBs and caused them to form irregular shapes (Fig. 2D). In addition, we randomly selected 40 EBs and measured their sphericity and volume. Our results revealed that Tlk1 depletion significantly decreased the sphericity and volume of EBs, suggesting an impairment in the proper induction of differentiation into an EB (Fig. 2D, bottom panels).
Because some differentiation-associated genes were upregulated by Tlk1 depletion under LIF-supplemented conditions, we investigated whether Tlk1 depletion affected gene expression in response to differentiation cues. The expression of pluripotency-associated or differentiation-associated genes under three separate differentiation-inducing conditions including LIF-withdrawal, EB formation, and retinoic acid (RA)-treatment was assessed using qRT-PCR. The KD efficiency in the Tlk1-depleted mESCs during differentiation was likewise confirmed by qRT-PCR (Fig. 3A). The proper induction of differentiation was also confirmed by the rapid downregulation of Oct4 in control KD cells ( Supplementary Fig. S1). The differentiation-induced downregulation of pluripotency-related genes such as Oct4, Sox2, Nanog, Klf2, and Esrrb was delayed in Tlk1-KD cells relative to control KD cells (Fig. 3B,C and D). In contrast to pluripotency-associated gene expression, the induction of differentiation-associated genes was hindered during the differentiation of Tlk1-KD cells compared with control KD cells (Fig. 3B-D). In accord with the mRNA expression profiles, the Western blotting results confirmed that the Oct4, Nanog, and Sox2 levels were higher in differentiated Tlk1-KD ESCs compared with SCIeNtIfIC RePoRts | (2018) 8:334 | DOI:10.1038/s41598-017-18628-9 differentiated control KD cells (Fig. 4A-F). In addition, the immunostaining results revealed that Oct4 and Nanog levels were increased in Tlk1-KD cells compared with control cells (shLuc) during differentiation induced by LIF withdrawal (Fig. 4G). Thus, these results indicate that Tlk1 depletion leads to the aberrant expression of differentiation-associated genes and the failure to downregulate the expression of pluripotency-associated factors during differentiation. Collectively, our findings suggest that Tlk1 is required for the proper induction of scheduled differentiation.
Ectopic expression of Tlk1 is sufficient to induce the downregulation of core pluripotency factors. Because Tlk1 depletion caused the delayed differentiation of mESCs and we were unable to generate a mESC line that stably overexpressed Tlk1, which suggested that the overexpression of Tlk1 might cause lethality in mESCs, we investigated the effect of Tlk1 overexpression on mESC function. To test our hypothesis regarding the overexpression of Tlk1, we established mESCs that conditionally overexpressed Flag-tagged The morphology of control (shLuc) and Tlk1-KD (shTlk1 #1 and #2) mESCs was evaluated using phase-contrast microscopic images and AP staining. Scale bars represent 500 µm. (C and D) The mRNA expression of pluripotency-associated and development-associated genes were analyzed by RT-qPCR in control (shLuc) and Tlk1-KD (shTlk1 #1 and #2) mESCs cultured under undifferentiated self-renewal conditions. All data were normalized to Gapdh and plotted relative to the expression level in control cells. Data are means (n = 3) ± SEM. *Р < 0.05, **Р < 0.01, and ***Р < 0.001. (E) The protein levels of pluripotency factors in control (shLuc) and Tlk1-KD (shTlk1 #1 and #2) mESCs was analyzed by immunoblotting using antibodies specific to Oct4, Sox2, and Nanog. (F) Quantification based on densitometry of Western blotting data from (E). All data were normalized to α-tubulin. Data are means (n = 3) ± SEM. *Р < 0.05, **Р < 0.01, and ***Р < 0.001.
Tlk1 under the control of the Tet-On inducible expression system, which is a doxycycline-inducible promoter. We examined Oct4, Sox2, and Nanog levels by Western blotting, the results of which demonstrated that the steady-state levels of the core pluripotency factors were decreased following the overexpression of Flag-tagged wild-type Tlk1 ( Fig. 5A and B). To determine if the kinase activity of Tlk1 was associated with the downregulation of the core pluripotency factors following the overexpression of Tlk1, we examined the effects of the overexpression of a D607A Tlk1 mutant. In humans, TLK1 harboring the D607A mutation is catalytically inactive and considered a kinase-dead mutant 20,26 . In this study, we mutated the Asp607 residue of Flag-tagged Tlk1 to alanine (D607A; kinase-dead mutant) because the Asp607 residue within the catalytic domain of Tlk1 is completely conserved between mice and humans. Our data revealed that the overexpression of the Flag-tagged Tlk1-D607A mutant also resulted in decreased levels of the core pluripotency factors, similar to wild-type Tlk1  Fig. S4). Therefore, the results suggest that the overexpression of Tlk1 renders its kinase activity possibly unnecessary for the downregulation of the core pluripotency factors.
To assess the effects of Tlk1 overexpression on ESC self-renewal further, we examined the self-renewal capacity of mESCs overexpressing wild-type or D607A-mutant Tlk1 and found that the overexpression of Tlk1 led to morphological changes, including a diffuse epithelial appearance ( Supplementary Fig. S2). Consistent with the  apparent morphological changes, the AP staining of Tlk1-overexpressing cells was moderately reduced compared to control cells (Vector), even in the presence of LIF, which suggested a partial loss of self-renewal ability (Fig. 5C).

Tlk1 overexpression causes growth defects and an increase in the G 2 /M phase population.
To assess if the ectopic overexpression of Tlk1 induces growth defects in mESCs, we examined the growth rates using a CCK-8 assay. The mESCs stably expressing empty vector or either the Tet-On-Tlk1 or the Tet-On-Tlk1-D607A expression vector were cultured with or without doxycycline for specified times under undifferentiated self-renewal conditions. The growth rate of each cell line was evaluated using a CCK-8 assay. Our results indicated that the overexpression of Flag-tagged Tlk1 or Flag-Tlk1-D607A abolished the ability of mESCs to proliferate, suggesting that the precise control of Tlk1 expression is critical for mESC survival irrespective of the kinase activity of Tlk1 (Fig. 6A). To elucidate the growth defects in Tlk1-overexpressing cells, we investigated cell cycle progression using fluorescence-activated cell sorting (FACS) analysis. Our FACS data revealed that the proportion of wild-type Tlk1 or D607A overexpressing cells in the G 2 /M phase was significantly increased compared to the control mESCs (empty vector or without doxycycline), whereas the proportion of cells in the G 1 -and S-phases was decreased ( Fig. 6B and C). Notably, the changes in the cell cycle profile induced by the overexpression of Tlk1 was not observed in the mESCs cultured in the presence of doxycycline for 24 hrs (Supplementary Fig. S3). The data suggest that the defect in cell cycle progression in Tlk1-overexpressing cells could require the extended induction of Tlk1 expression for an additional 24 hrs. Thus, our data suggest that Tlk1 might play a role in cell cycle control in mESCs.
To support the accumulation of cells in the G 2 /M phase induced by the overexpression of Tlk1, we performed an immunoblotting analysis using antibodies against several cell-cycle regulators. We observed that the Wee1 levels and the phosphorylation of CDK1-Tyr15 were decreased in Tlk1 and Tlk1-D607A-overexpressing cells (Fig. 6D). These results were consistent with the observation that the Wee1-mediated phosphorylation of CDK1-Tyr15 is crucial for preventing the premature activation of CDK1 during interphase 27,28 . Moreover, our data indicated that histone H3-Ser10 phosphorylation, a hallmark of M-phase, was increased in Tlk1-overexpressing cells compared to control cells not treated with doxycycline (Fig. 6D). The H3-Ser10 phosphorylation levels in wild-type Tlk1-overexpressing cells were comparable to that of Tlk1-D607A overexpressing mESCs, suggesting that the increased phosphorylation of histone H3-Ser10 in Tlk1-or Tlk1-D607A-overexpressing mESCs might not be a direct consequence of its own Tlk1 kinase activity. Therefore, these data suggest that Tlk1 overexpression might induce the premature activation of CDK1 by the downregulation of Wee1 and the subsequent increase in the phosphorylation of histone H3-Ser10, causing a defect in cell-cycle progression in mESCs. Together, our data suggest that the precise and delicate control of Tlk1 expression levels is very important for proper cell-cycle progression, which likely contributes to the proper maintenance of mESC functions.

Forced expression of Tlk1 induces apoptosis in mESCs. Because our data demonstrated that Tlk1
overexpression caused an impairment in mESC proliferation, we next attempted to assess if growth defects in Tlk1-overexpressing mESCs might be correlated with the induction of cell death. Hence, we performed an immunoblotting analysis to detect the activation of an apoptotic marker in Tlk1-overexpressing cells. Our data showed that cleaved caspase-3, a hallmark of apoptosis, was clearly detectable 24 and 48 hrs after the induction of Tlk1 or D607A (Fig. 7). Thus, our data suggest that defects in cell cycle control and cell proliferation in mESCs overexpressing Tlk1 might cause apoptosis.

Discussion
In the present study, we aimed to determine the role of Tlk1 in ESC pluripotency and differentiation. The primary function of mESCs was only marginally affected in Tlk1-depleted cells, suggesting that Tlk1 is not required for mESC pluripotency and self-renewal in the undifferentiated state. Further, Tlk1 depletion caused resistance to differentiation cues and abnormal EB formation, suggesting an impairment in scheduled differentiation. As well, Tlk1 depletion induced the aberrant expression of both differentiation-associated and pluripotency-associated genes during differentiation. In particular, the downregulation of pluripotency factors during differentiation was blocked in Tlk1-depleted mESCs, demonstrating the failure of Tlk1-KD cells to differentiate. Conversely, the ectopic expression of Tlk1 was sufficient to induce the untimely downregulation of core pluripotency factors irrespective of kinase activity, thereby leading to a partial loss of self-renewal ability even in the undifferentiated state. Interestingly, we noted that Tlk1 overexpression causes growth defects and an increased number of cells in the G 2 /M phase, as well as apoptosis. The abnormal cell cycle profile was correlated with the increased phosphorylation of histone H3-Ser10 and the downregulation of Wee1 and CDK1-Tyr15 phosphorylation. Collectively, our current data suggest that precise and delicate control of Tlk1 expression levels is critical for proper cell-cycle progression and could contribute to scheduled differentiation.
Both the pluripotency and self-renewal abilities of ESCs are maintained by the expression of specific genes including the transcriptional regulatory circuitry controlled by core TFs such as Oct4, Sox2, and Nanog 29 . The regulatory circuitry serves as a master switch in the establishment and maintenance of the pluripotency state via the positive regulation of several undifferentiated cell markers and the silencing of lineage commitment genes 29 . More specifically, the transcriptional inactivation of a large set of lineage-specific markers by core pluripotency factors is critical for the maintenance of pluripotency and self-renewal 2,16 . In contrast, the pluripotency factors are repressed in a rapid and timely manner in response to differentiation cues, thereby leading to the induction of scheduled differentiation 2 . However, the properties of normal mESCs are largely abrogated in Tlk1-depleted cells. Our data indicate that the expression of pluripotency factors are comparatively maintained in Tlk1-depleted cells relative to those of control cells during differentiation, concomitant with the aberrant expression of developmental genes (Figs 3 and 4). Consistent with their failure to downregulate the expression of undifferentiated cell markers, Tlk1-deficient cells formed irregularly shaped EBs and induced resistance to differentiation cues (Fig. 2), indicating a failure of Tlk1-deficient mESCs to differentiate in the context of an EB. The failure of Tlk1-depleted mESCs to differentiate is very similar to BRPF2-and Mbd3-deficient mESCs, in which the scaffold protein BRPF2/BRD1 is a key component of histone acetyltransferase complexes and Mbd3 functions in nucleosome remodeling and the histone deacetylation (NuRD) complex, respectively 25,30 . Together, these data suggest that Tlk1 is required for the scheduled differentiation of mESCs but not for the maintenance of pluripotency and self-renewal in the undifferentiated state.
Tlk1, a serine/threonine kinase, plays important roles in chromatin assembly, DNA repair, and cell cycle progression and phosphorylates Asf1, Rad9, Aurora B kinase, and histone H3 26,[31][32][33][34][35][36] . In C. elegans, although TLK-1 kinase activity is not required for enhancing the kinase activity of Aurora kinase B (AIR-2), AIR-2 phosphorylates TLK-1 and, in turn, the phosphorylated TLK-1 reinforces AIR-2 kinase activity, suggesting that TLK-1 is a substrate and activator of the Aurora kinase B 33 . Furthermore, both human TLK1 and AURKB are required for the phosphorylation of histone H3 34,[37][38][39] . These studies could indicate that mouse Tlk1 is potentially associated with the highly conserved mouse Aurkb. Very recently, an elegant study demonstrated that the Aurkb/PP1-mediated resetting of Oct4 during the cell cycle is crucial for determining the identity of mESCs 40 . The phosphorylation of Oct4 by Aurora kinase B during the G 2 /M phase caused Oct4 to dissociate from chromatin, whereas PP1-mediated Oct4 dephosphorylation is required for Oct4 to reoccupy chromatin during exit from the M phase 40 . Interestingly, an Oct4 phosphomimetic mutant that mimicked the Aurkb-mediated phosphorylation of Oct4 caused the cells to lose pluripotency 40 . The authors proposed that the Aurkb/PP1-mediated Oct4 phosphorylation/dephosphorylation cycle plays a potential role in the cell-cycle-dependent control of pluripotency and self-renewal. In the present study, we showed that Tlk1 overexpression caused growth defects, the accumulation of cells in the G 2 /M phase, and apoptosis in mESCs (Figs 5-7). As well, the G 2 /M arrest in Tlk1-overexpressing cells was correlated with increased histone H3 H3-Ser10 phosphorylation, a mitotic marker, and with the downregulation of Wee1 and CDK1-Tyr15 phosphorylation (Fig. 6). These results suggest that the Tlk1-mediated control of scheduled differentiation in mESCs might be demonstrated via Tlk1-dependent cell cycle control. Notably, the Oct4 protein levels were noticeably reduced in Tlk1-overexpressing cells (Fig. 5), suggesting a negative role of Tlk1 in Oct4 expression. After 2 hr release from G 2 /M-arrest by nocodazole, Oct4 phosphorylation was slightly increased in Tlk1-overexpressing cells treated with doxycycline compared to control cells without doxycycline ( Supplementary Fig. S5). Based on our study and other reports, we hypothesize that mouse Tlk1 is a substrate and activator of Aurora kinase B as in C. elegans; thus, the ectopic and persistent expression of Tlk1 might activate Aurkb, possibly leading to enhanced phosphorylation of Oct4 and the subsequent persistent dissociation of Oct4 from the target promoters. Further, this persistent dissociation of Oct4 from chromatin might cause the loss of self-renewal in mESCs. These findings may raise the possibility that Tlk1 may interact with Aurora kinase B (Aurkb) to control mESC function.
The ectopic expression of Tlk1 leads to the inhibition of cell proliferation in mESCs, which is likely due to G 2 /M phase arrest (Fig. 6). The growth defects and G 2 /M arrest in Tlk1-overexpressing mESCs occur irrespective of its kinase activity. However, in the case of human Tlk1, the overexpression of the wild-type Tlk1 resulted in a normal diploid karyotype, whereas a dominant negative Tlk1 mutant (kinase dead) caused chromosome missegregation and aneuploidy 41 . This discrepancy between human and mouse Tlk1 might be due to species-related differences or cell line-specific differences.
In summary, our results indicate that a Tlk1 deficiency in mESCs disturbs scheduled differentiation. As well, we found that only a moderate increase in the Tlk1 level is sufficient to downregulate the expression of core pluripotency factors, even under undifferentiated self-renewal conditions, which subsequently leads to markedly reduced cell proliferation, an increased number of cells in the G 2 /M phase, and apoptosis. The G 2 /M arrest observed in Tlk1-overexpressing mESCs correlated with the enrichment of histone H3-Ser10 phosphorylation, a mitotic marker, and with a reduction in the Wee1 and CDK1-Tyr15 phosphorylation levels. Accordingly, we propose that the Tlk1-mediated cell cycle control functions prominently in the lineage commitment of mESCs, as Tlk1 acts as a negative regulator of core pluripotency factor expression.
Alkaline Phosphatase (AP) staining. AP staining was performed following the manufacturer's instructions (Cat. #SCR004; Merck Millipore). Briefly, cells were cultured for 5 days, fixed with 4% paraformaldehyde, washed in a rinse buffer and stained with Alkaline Phosphatase Detection Kit. Western Blot. Western blot analysis was done as described 46  Flow Cytometry. Cell cycle analysis was performed as previously described 46 . Briefly, cells were harvested and fixed with cold 70% ethanol at -20 °C for about 3 hrs. And then the fixed cells were stained with propidium iodide and subjected to flow cytometry analysis.
Cell growth rate assay. Cell growth rate analysis were performed with the Cell Counting Kit-8 (CCK-8 assay kit; Dojindo Corporation, Kumamoto, Japan) as previously described 46 . Briefly, 24 hrs prior to experiments, 1200 cells per well were plated onto each well of a 96-well plate (100 µl medium of cell suspension) and from the next day medium in the absence or presence of doxycycline (final concentration 100 ng/ml) was changed for every day. And after adding CCK-8 solution, absorbance was measured four times at an interval of 24 hrs. Statistical analysis. Data are presented as the means ± S.E.M. or means ± S.D. Two-tailed student's t-tests were performed to analyze the data between controls and experimental groups. Statistical significance (P value) is indicated for each graph as asterisks (*P < 0.05, **P < 0.01, ***P < 0.001).