Kaiso does not affect β-catenin interaction with xTcf3 but can alter the nuclear localisation of TCF3 in mouse fibroblasts. (A) In vitro translated ³⁵S-methione labelled β-catenin and HA-tagged xTcf3 were incubated in the presence of an excess of xKaiso full-length (xKaisoFL) protein or GST. xTcf3 was immunoprecipitated with an anti-HA antibody. β-Catenin and xTcf3 proteins are indicated by arrows. 1/10 of inputs are shown. The interaction between β-catenin and xTcf3 is not impaired in the presence of GST-xKaiso or GST. (B,C) Localisation of myc-xTcf3 upon transient transfection into mouse MEFs in the absence and presence of Ha-xKaiso and T7tagged-dKaiso. (B) By itself, myc-xTcf3 either exhibits staining at nuclear foci (60% of cells) or homogenous nuclear staining (40%). (C) xKaiso by itself exhibits homogenous nuclear staining in 100% of cells. (D) In the presence of xKaiso or dKaiso the myc-xTcf3 protein exhibits only homogenous nuclear staining (100% of cells with double staining).

Models for mutual interference of DNA-binding functions by Kaiso and TCF3. (A) A schematic representation of β-catenin-dependent activation of TCF3 target genes. β-Catenin displaces the Groucho/CtBP complex from TCF3 leading to gene activation. (B) A model for Kaiso-mediated repression of Wnt target genes that was suggested by Park et al. (Park et al., 2005). Here, it was proposed that repression of the Siamois gene occurs by xKaiso binding to CTGCNA sequences in its promoter and the possible interaction pf Kaiso with xTcf3. Our results suggest this model is not operative (Ruzov et al., 2009). (C,D) Our model of non-DNA-dependent xKaiso displacement of xTcf3 from target genes proposes that Kaiso overexpression may result in either a block of repression by xTcf3/Groucho or a block of β-catenin/xTcf3-dependent activation. (E,F) TCF3 can potentially displace Kaiso from its methylated (filed circles) DNA-binding sites and interfere with its repression function.

Kaiso ZF1-3 interacts directly with the HMG domain of TCF factors. (A) Kaiso deletion constructs used in pull-down experiments. (B) GST pull-down using in vitro transcribed xTcf3 or xTcf3dn with Kaiso recombinant proteins indicated in A. GST was used as a control. (C) TCF4 constructs used for in vitro transcription in D. β-cat, β-catenin interaction domain; TLE, TLE/Groucho binding domain; CtBP, CtBP-binding sites; HMG, DNA-binding domain. (D) GST pull-down using TCF4 constructs and xKaiso GST-ZF domain (KaisoZFs). GST was used as a control; 1/10 of inputs (xTcf3 and xTcf3dn) are shown. (E) EMSA assay using in vitro translated xTcf3 and its DNA-binding site (TCF bs-oligo) in the absence and presence of GST fusions with ZF1-3 of xKaiso, full-length xKaiso or ZF1-3 of dKaiso. xZF2-3 and BTB/POZ fusions, which do not interact with xTCF3 were used as controls. Note absence of the xTcf3/DNA complex (arrow) in the presence of the Kaiso fusions. Increasing the xTcf3 concentration in the presence of xZF1-3 restores the binding of xTcf3 to DNA, supporting the hypothesis that sequestration is the mechanism of inhibition of xTcf3 binding to DNA by Kaiso.

Kaiso prevents xTcf3 localising to its target promoter and its subsequent activation or repression functions. (A) ChIP assay showing that, after transient transfection into A6 cells, myc-xTcf3 can be located at the Siamois promoter. Co-expression of xKaiso or dKaiso prevents xTcf3 binding at the promoter. Inputs are shown on the right and the myc-Tcf3 ChIP on the left with a non-specific antibody (Ig) control. (B) Overexpression of both dKaiso and xKaiso represses β-catenin-dependent transcription activity in transient transfection assays in HeLa cells. SuperTopflash (SuperTop) was used as a Wnt reporter and SuperFopflash (SuperFop) with mutated TCF3 sites as a control. (C) Expression from a methylated luciferase reporter can be enhanced by co-transfection with an xKaisoZFVP16 (KaisoZFVP16) expression plasmid. Targeting to the methylated reporter is provided by the Kaiso-ZFs and transcription activation function by VP16. Increasing amounts of xTcf3HMGVP16 (Tcf3VP16) reduce the activation potential of xKaisoZFVP16. In a reciprocal experiment, activation a Siamois luciferase reporter by xTcf3HMGVP16 is inhibited by the presence of xKaisoZFVP16. (D,E) Overexpression of myc-xKaiso in Xenopus embryos results in the formation of exogastrulae compared with controls at stage 12. Graph shows that by stage 12, over 80% of the embryos (n=55) exhibit exogastrulae. (F) Real-time RT-PCR analysis of Siamois expression in stage 10 control and embryos injected with myc-xKaiso mRNA (1 ng). Relative to histone, H4 Siamois expression is activated in the myc-xKaiso-injected embryos.