Control of Hes7 Expression by Tbx6, the Wnt Pathway and the Chemical Gsk3 Inhibitor LiCl in the Mouse Segmentation Clock

The mouse segmentation is established from somites, which are iteratively induced every two hours from the presomitic mesoderm (PSM) by a system known as the segmentation clock. A crucial component of the segmentation clock is the gene Hes7, which is regulated by the Notch and Fgf/Mapk pathways, but its relation to other pathways is unknown. In addition, chemical alteration of the Wnt pathway changes the segmentation clock period but the mechanism is unclear. To clarify these questions, we have carried out Hes7 promoter analysis in transgenic mouse embryos and have identified an essential 400 bp region, which contains binding sites of Tbx6 and the Wnt signaling effector Lef1. We have found that the Hes7 promoter is activated by Tbx6, and normal activity of the Hes7 promoter in the mouse PSM requires Tbx6 binding sites. Our results demonstrate that Wnt pathway molecules activate the Hes7 promoter cooperatively with Tbx6 in cell culture and are necessary for its proper expression in the mouse PSM. Furthermore, it is shown that the chemical Gsk3 inhibitor LiCl lengthens the oscillatory period of Hes7 promoter activity. Our data suggest that Tbx6 and the Wnt pathway cooperatively regulate proper Hes7 expression. Furthermore, proper Hes7 promoter activity and expression is important for the normal pace of oscillation.


Introduction
Vertebrate embryos are divided along the longitudinal axis into somites in a process called segmentation. During the vertebrate segmentation, a new pair of bilateral somites arises every two hours from the presomitic mesoderm (PSM). The pace of somite formation correlates with the periodic expression of genes of the Notch, Fgf and Wnt pathways [1].
The Hes7 gene, a crucial component of the segmentation clock, is downstream of the Notch and Fgf pathways and drives the oscillation of several cyclic genes of these pathways [2][3][4]. The Fgf pathway is mainly active in the posterior PSM, whereas Notch pathway activity is found in the PSM and budding somites [5,6]. This raises the question of how the domain of Hes7 expression is specified.
The oscillation period of the segmentation clock in vertebrates is changed after perturbation of the Notch and Wnt pathways [7][8][9][10]. The effect of Notch pathway perturbations on the segmentation clock period is understandable, because Notch target genes are critical components of the segmentation clock. By contrast, the mechanism of the Wnt pathway contribution to the segmentation clock period is unclear.
To investigate these questions, we have analyzed the Hes7 promoter and have found evidence that Tbx6 and the Wnt pathway regulate Hes7 expression in the PSM. Our results suggest that Tbx6 and the Wnt pathway are necessary for proper Hes7 expression. We have also found that treatment with the chemical Gsk3 inhibitor LiCl activates the Wnt pathway and lengthens the oscillatory period of Hes7 expression.

Transgenic embryos
Transgenic embryos were generated as previously described [15]. Embryos were genotyped and stained with X-gal using a standard protocol.
Embryo culture, inhibitor treatments and bioluminescence imaging Embryos were taken out and freed from extraembryonic structures in prewarmed PBS. For in situ hybridization assays, wild-type embryos were transferred to culture medium (DMEM, 10% FBS, 1% P/S) in the presence of solvent or inhibitor and cultured for the indicated times at 37uC with 5% CO 2 . For timelapse imaging experiments, tails of embryos were cut up to the second somite, transferred to 1 mM luciferin-containing culture medium and genotyped by observation of luminescence with a CCD camera. One positive tail was selected, transferred to inhibitor containing medium and immediately imaged (Conditions: 5% CO 2 , 85% O 2 and 37uC). The oscillation period was measured by two different methods by analyzing movies that spanned at least two oscillation cycles. In the first method, we counted the number of oscillation cycles and divided it by their duration. In the second method, we created a spatiotemporal plot of the oscillations with time in the x-axis and measured the distance between peaks ( Figure S6). Inhibitors were: 20, 40 and 100 mM LiCl (Nacarai Tesque), 10 mM Gsk3 Inhibitor IX (BIO) (EMD), 1 and 5 mM XAV939 (Sigma) and 100 mM CKI-7 (Sigma).

Luciferase assay
The luciferase assay was carried out as previously described using C3H10T1/2 cells in 10% FBS and 1% P/S in DMEM and Lipofectamin LTX/Plus (Invitrogen) transfection reagent [11].

Electrophoresis mobility shift assay (EMSA)
We subcloned the Tbx6 cDNA from pCS2-3xFlag-Tbx6 [12] into the T7 control plasmid of the TNT in vitro translation kit (Promega). The in vitro translation was carried out following the manufacturer's protocol. Oligonucleotides (15 bp) containing the wild-type and mutant T-box binding sites were annealed and labeled with gamma-32 P-ATP (GE Heathcare) using T4 polynucleotide kinase. The sense-strand sequences of the wild-type and mutant probes for T-box 21306 are: agc ctc acg tgc agg tga gaa aaa ctc aac and agc ctc acg tgc Cgg tTa gaa aaa ctc aac, respectively and for T-box 21350: cag ggg cgg ccc cac acc cgg gtg caa act g and cag ggg cgg ccc caA aAc cgg gtg caa act g, respectively. Purified proteins and labeled probes were incubated for 30 min at room temperature in 56 binding buffer (100 mM Tris-HCl, pH 8.0, 300 mM KCl, 5 mM EDTA, 60% Glycerol, 3.3 mg/ml BSA, 8.35 mM DTT, 12.5 mM MgCl 2 ). Samples were then electrophoresed on a 5% non-denaturing polyacrylamide gel (Acryl/bis-acryl 19:1 ratio, 0.56 Tris/Borate/EDTA buffer, 2.5% Glycerol).

In situ hybridization and real-time PCR
In situ hybridization was performed as previously described [2]. To quantify the in situ hybridizations after inhibitor culture, we classified control and treated embryos into categories of high and low staining and plotted the percentage of control and treated embryos with high staining. For real-time PCR, the PSM was cut until the first visible somite border, RNA was isolated using the NucleoSpin RNA II kit (Macherey-Nagel) and reverse transcribed with SuperScript II Reverse Transcriptase (Invitrogen). Two biological replicates were carried out. Real-time PCR was performed with the LightCycler 480 SYBR Green I Master mix and the Real-time PCR system (Roche). Primer sequences (forward and reverse) are Gapdh: atc ttc ttg tgc agt gcc agc ctc gtc ccg and agt tga ggt caa tga agg ggt cgt tga tgg; Axin2 (QT00126539, Quiagen) and Msgn1: cgg ctt agt cga gct gga tta and ctc cgc tgg aca gac atc ttg.

Statistical analysis
To calculate statistical significance, differences between control and treated populations were evaluated with the Student's t test.
Error bars indicate standard error of the mean.

Results and Discussion
A crucial 400 bp sequence in the Hes7 promoter To construct the Hes7 promoter reporter, we linked the 2.6 kb upstream promoter region of Hes7 (H7p2.6wt) to the lacZ gene. Transgenic embryos showed X-gal staining in the endogenous Hes7 expression region and somites ( Figure 1). The somitic staining was likely due to the stability of the lacZ mRNA and protein. We did not use destabilizing sequences, because we were mainly interested in finding promoter sequences required for the posterior onset of Hes7 promoter activity. Previous data showed posterior Hes7 expression in the absence of Notch signaling [4]. Therefore, we hypothesized that Rbpj binding sites are not required for the posterior onset of Hes7 expression. To test this hypothesis, we introduced point mutations into four Rbpj binding sites (H7p2.6dR). This construct also showed X-gal staining in the posterior PSM ( Figure 1). We created smaller 1.9 kb, (H7p1.9dR), 1.4 kb (H7p1.4dR) and 1.0 kb (H7p1.0dR) promoter reporters. H7p1.9dR and H7p1.4dR reporters showed X-gal staining, although H7p1.4dR showed either weak activity (4/8 embryos) or ectopic notochord X-gal staining (2/8 embryos) ( Figure 1). We speculate that T (Brachyury) exerts a repressive effect on a T-box binding site in the 1.9-1.4 kb region to prevent this ectopic notochord activity. Finally, the X-gal staining of the PSM was lost in the H7p1.0dR reporter, demonstrating that this 400 bp Hes7 promoter region is important for Hes7 promoter activity ( Figure 1). This 400 bp region is strongly conserved in humans and contains binding sites for interesting transcription factors such as Lef1 (Wnt pathway effector) [17], Ets (Fgf pathway effector) [18], the mesodermal factor Tbx6 [19], and the axial elongation factor Cdx2 [20] ( Figure S1).

Hes7 promoter activity requires Tbx6 binding sites
Two T-box binding sites caught our attention (T-1306 and T-1350), because such binding sites are important in other PSM gene promoters, namely Dll1, Mesp2 and Msgn1 (Figures 1 and 2A) [12,21,22]. First, we co-transfected a luciferase reporter of the wild-type (WT) 2.6 kb Hes7 promoter (H7p2.6wt) with Notch intracellular domain (NICD) and Tbx6 expression plasmids. NICD upregulated the H7p2.6wt promoter reporter ( Figure 2B). Tbx6 induced a non-significant increase of the H7p2.6wt reporter activity alone and enhanced the positive effect of NICD ( Figure 2B). These data suggest that Tbx6 enhances NICD mediated activation of the Hes7 promoter in cultured cells. This is similar to the reported enhancement of Mesp2 reporter activity by NICD and Tbx6 [12].
Next, we checked whether in vitro synthesized Tbx6 can bind to oligos containing the two T-box binding sites by electrophoresis mobility shift assay (EMSA). The EMSA showed that in vitro synthesized Tbx6 protein can bind to WT ( Figure 2C, lanes 2, 3, 9, 10), but not mutant ( Figure 2C, lanes 5, 12), oligos containing these binding sites. The significance of the larger band in the presence of the mutant oligo is unclear ( Figure 2C, lane 5). The WT bands were supershifted by anti-Tbx6 antibody ( Figure 2C, lanes 4, 11), while they were antagonized by WT but not mutant unlabeled oligos ( Figure 2C, lanes 6, 7, 13, 14). The bands were very strong for the T-1306 binding site but weak for T-1350, suggesting that Tbx6 mainly binds to the T-1306 binding site.
To verify whether these T-box binding sites are also important in the PSM during embryonic development, we constructed a lacZ reporter with H7p1.4dRdT with point mutations in the two T-box binding sites of H7p1.4dR, and subsequently generated transgenic embryos. The H7p1.4dRdT lacZ reporter did not show X-gal staining in the tail bud and PSM in contrast to the control, suggesting that these T-boxes are important for WT Hes7 expression ( Figure 2D). In control embryos, the Hes7 promoter reporter activity is found in the paraxial mesoderm; however in embryos expressing the H7p1.4dRdT lacZ contruct, X-gal staining is mainly in the lateral plate mesoderm ( Figure 2E). A similar phenotype was reported for the Msgn1 promoter in the absence of T-box binding sites [22]. These data suggest that Tbx6 is very important for the initiation of Hes7 expression. The observation that Tbx6 alone did not significantly upregulate Hes7 in cultured cells may be due to the lack of factors that potentiate Tbx6 activity in the PSM such as the Wnt pathway ( Figure 3C) [22]. To investigate whether Tbx6 also defines the anterior limit of Hes7 expression, we examined the protein expression localization of Hes7 and Tbx6 proteins in the PSM by immunofluorescence. We found variable patterns of Hes7 protein that were mostly included within the Tbx6 protein domain ( Figure 2F). Specifically, in phases II/III of Hes7 protein expression, the anterior limit of Hes7 protein coincided or slightly exceeded that of Tbx6 (Phase II in Figure 2F). Hes7 gene is only transcribed in the Hes7 protein negative region [2], suggesting that Hes7 transcription occurs only in the Tbx6 protein domain. These results suggest that Tbx6 is important for controlling the proper expression of Hes7 in the PSM.
The Wnt pathway activates Hes7 expression Tbx6 regulates target genes synergistically with the Wnt pathway [21,22]. Therefore, we also investigated, whether the Wnt pathway activates the Hes7 promoter. To this end, we cotransfected the Rbpj mutated 2.6 kb Hes7 promoter reporter with expression plasmids of a constitutively active form of beta-Catenin  (Ctnnb1) and Lef1 in C3H10T1/2 cells and analyzed the luciferase luminescence. Our results show that the Hes7 promoter is activated by Ctnnb1 and Lef1( Figure 3A,B). Then, we co-transfected Tbx6 and T expression plasmids and found that these plasmids further upregulated the activation of the H7p2.6dR construct by Ctnnb1 and Lef1 ( Figure 3C). Previously, it was found that inactivation of Ctnnb1 in the PSM downregulates Hes7 expression [23]; therefore, we analyzed the hypomorph Wnt3a mutant mouse Vestigial (Vt) [24]. In Wnt3a mutant embryos at embryonic stage E10.5, Hes7 intronic expression was downregulated ( Figure 3D). Tbx6 mRNA expression was also downregulated in these embryos in agreement with results in PSM specific Ctnnb1 conditional KO mice ( Figure S2) [23]. These results suggest that Hes7 expression is positively regulated by the Wnt pathway.

LiCl lengthens the oscillation period
LiCl has been reported to activate the Wnt pathway targets by inhibiting Gsk3b, a kinase that targets Ctnnb1 for degradation [25,26]. First we tested whether LiCl can activate the Wnt pathway in the PSM. We cultured E9.5 embryos for 6 h in the presence of 0, 20 and 40 mM LiCl, cut the PSM until the first somite and synthesized cDNA. Real-time PCR for the Wnt target genes Axin2 and Msgn1 showed an increased expression of these genes in the presence of LiCl ( Figure S3A,B). Second, we also determined whether the Ctnnb1 protein is stabilized by LiCl in the PSM. Our results show higher levels of Ctnnb1 in the PSM of E10.5 embryos cultured with 40 mM LiCl for 6 h ( Figure S3C), which led us hypothesized that LiCl may activate Hes7 expression. To test it, we cultured E10.5 WT mouse embryos with 20 mM LiCl for 6 h. Our results show a non-significant tendency to higher Hes7 expression in the presence of LiCl ( Figure S4A). Culture of E9.5 embryos with a higher LiCl concentration (100 mM) also showed a non-significant tendency to higher Hes7 mRNA expression after 2 h culture ( Figure S4B). Treatment of E9.5 embryos with Gsk3 Inhibitor IX for 2 h also showed a nonsignificant increase of Hes7 mRNA expression ( Figure S4C). These results suggest that LiCl activates the Wnt pathway. The low effect of LiCl on Hes7 expression might be due to a compensatory effect of the autoinhibitory feedback of Hes7 protein on itst promoter. Therefore, we hypothesized that the effect of LiCl on Hes7 would manifest as a change of the oscillatory period. Therefore, we monitored the oscillations of E10.5 transgenic embryos carrying a Hes7 promoter luciferase reporter by timelapse microscopy in the presence of LiCl [15]. In the control experiment, we found luminescence bands spreading anteriorly ( Figure 4A, Movie S1). In contrast, 20 and 40 mM LiCl treatments induced an abnormal band of Hes7 promoter activity in the posterior PSM ( Figure 4A, red arrows, Movies S2, S3). Furthermore, in some embryos, 40 mM LiCl treatment resulted in arrest of the oscillations of Hes7 promoter activity by locking the reporter in an active state ( Figure  S5, Movie S4). Quantification of the oscillation period in the presence of LiCl demonstrated that treatment with 40 mM LiCl increased the oscillation period ( Figure 4B). To exclude the possibility that the longer period was a consequence of the arrested oscillation, we determined the duration of individual cycles after addition of LiCl ( Figure S6). Our results show that the largest period difference occurred during the second oscillation cycle after inhibitor addition suggesting that the period change is not a consequence of the arresting oscillation.
Our data indicate that in the PSM, as in other tissues, LiCl increases the rate of transcription of target genes through stabilization of Ctnnb1 ( Figure S3). Interestingly, a higher rate of transcription might be responsible for the longer period of Hes7 promoter activity oscillations in the presence of LiCl in agreement with numerical models of Hes7 that predict a positive correlation between higher rate of transcription and longer period [27,28]. We then hypothesized that an inhibitor of the Wnt pathway would shorten the oscillation period. The small molecule XAV939 stabilizes Axin and increases Ctnnb1 degradation [29]. Timelapse imaging of Hes7 promoter activity with low concentration (1 mM) of XAV939 did not cause apparent effects the reporter activity, but high concentration (5 mM) downregulated promoter activity ( Figure S7, Movies S5, S6). We measured the oscillation period for embryos treated with 1 mM XAV939 but did not observe any effect ( Figure S7).
A previous study found a correlation between lower Hes7 expression and longer oscillation period due to treatment with the chemical agent CKI-7, which targets the Wnt pathway [7]. That result is unclear, because to our understanding, low Hes7 expression was observed with high concentration (200 mM) of CKI-7, whereas the longer period was found with a different lower concentration (100 mM) of CKI-7 [7]. Therefore, we also examined the oscillatory activity of the Hes7 promoter in the presence of 100 mM CKI-7. Measurement of the oscillation period showed an increase of the period from 2.5 h to 3.3 h ( Figure S8, Movie S7), which agreed with the previous data [7]. By contrast, in situ hybridization of embryos cultured with 100 mM CKI-7 for 3 h did not show a clear change of Hes7 expression levels (data not shown). These results suggest that CKI-7 inhibitor lengthens the oscillatory period although the mechanism for this effect requires further investigation.
Our results suggest that the Wnt related chemicals LiCl and CKI-7 lengthen the period, whereas another Wnt related chemical, XAV939, does not. Our data that LiCl upregulates the expression of the Wnt target genes Axin2 and Msgn1 and the levels of Ctnnb1 suggest that the effects of LiCl on the period is mediated by the Wnt pathway. However, a mutant mouse where Ctnnb1 is stabilized does not change the oscillation period [30]. Therefore, alternative possibilities could be explored in the future. For instance LiCl could act through the Mapk pathway [31], which is also important for Hes7 expression [32].
Here, we have investigated the transcriptional activation of Hes7 expression, in an effort to better understand the mechanism of the mouse segmentation clock. We found evidence that Tbx6 binding sites are required for normal Hes7 promoter activity ( Figure 5) and also that Hes7 expression is positively regulated by the Wnt pathway ( Figure 5). Furthermore, we showed that proper Wnt activity is important for normal expression of Hes7, and that different Wnt-related chemicals lengthen the oscillatory period of Hes7 promoter activity ( Figure 5). Tbx6, the Wnt pathway and Hes7 are very important molecules for the murine segmentation  clock. Therefore, our analysis of their regulation provides new insights into the mechanism of the segmentation clock.