January 1, 2009;
Modulation of the beta-catenin signaling pathway by the dishevelled-associated protein Hipk1.
BACKGROUND: Wnts are evolutionarily conserved ligands that signal through beta-catenin-dependent and beta-catenin-independent pathways to regulate cell fate, proliferation, polarity, and movements during vertebrate development. Dishevelled
/Dvl) is a multi-domain scaffold protein required for virtually all known Wnt signaling activities, raising interest in the identification and functions of Dsh
METHODOLOGY: We conducted a yeast-2-hybrid screen using an N-terminal fragment of Dsh
, resulting in isolation of the Xenopus laevis ortholog of Hipk1
. Interaction between the Dsh
proteins was confirmed by co-immunoprecipitation assays and mass spectrometry, and further experiments suggest that Hipk1
also complexes with the transcription factor Tcf3
. Supporting a nuclear function during X. laevis development, Myc
localizes primarily to the nucleus
in animal cap explants, and the endogenous transcript is strongly expressed during gastrula
stages. Experimental manipulations of Hipk1
levels indicate that Hipk1
can repress Wnt/beta-catenin target gene activation, as demonstrated by beta-catenin reporter assays in human embryonic kidney
cells and by indicators of dorsal specification in X. laevis embryos at the late blastula
stage. In addition, a subset of Wnt-responsive genes subsequently requires Hipk1
for activation in the involuting mesoderm
during gastrulation. Moreover, either over-expression or knock-down of Hipk1
leads to perturbed convergent extension cell movements involved in both gastrulation and neural tube closure.
CONCLUSIONS: These results suggest that Hipk1
contributes in a complex fashion to Dsh
-dependent signaling activities during early vertebrate development. This includes regulating the transcription of Wnt/beta-catenin target genes in the nucleus
, possibly in both repressive and activating ways under changing developmental contexts. This regulation is required to modulate gene expression and cell movements that are essential for gastrulation.
[+] show captions
Figure 1. Hipk1 interacts with Dsh and Tcf3.(A) 35S-Met-labelled Dsh precipitates in a complex with the partial-length Hipk1 GST-tagged clone from the yeast-2-hybrid screen, but not with GST alone. (B) Immunoprecipitation (IP) of 35S-Met-labelled Dsh with Myc-tagged Hipk1 (left panel), and reciprocal 35S-Met-labelled Hipk1 IP with Myc-tagged Dsh (right panel), using in vitro translated proteins. (C) IP of Dsh with Hipk1 when both proteins are recombinantly expressed in X. laevis embryos. (D) Endogenous Tcf3 IP with recombinantly expressed Myc-(K/R) Hipk1 in ventral X. laevis embryos. An anti-Spectrin antibody was used as a negative control for the IP (top) and as a loading control for the input (bottom).
Figure 2. Hipk1 is expressed throughout vertebrate development.(A) Levels of the X. laevis hipk1 mRNA were measured by RT-PCR at the following developmental stages: Unfertilized egg (U), 2-cell (2), 4-cell (4), blastula (Stage 7, 8), gastrula (Stage 10), neurula (Stage 17), tail bud (Stage 25), and early tadpole (Stage 35). Histone H4 was used as a loading control, and reactions without reverse transcriptase (−RT) performed to rule out genomic DNA contamination. (B) WISH of the hipk1 transcript during X. laevis development. At Stage 10 (a, ventral aspect) signal is detected in the marginal zone and animal pole with greater staining above the dorsal lip of the blastopore (arrow). At Stage 17 (c, dorsal aspect; d, anterior aspect) signal is apparent at the margins of the lateral and anterior neural plate. At Stage 25 (f) and Stage 35 (h), signal is present in dorsal anterior regions and in the tail bud. Close-up of the Stage 35 head (i, lateral aspect) reveals expression in the branchial arches (black arrow), otic vesicle especially dorsally (white arrow), pronephros (black arrowhead), and retina (white arrowhead). Close-up of the Stage 35 head/torso (j, ventral aspect) shows high expression in the heart (arrow). Sense controls (b, e, g, k).
Figure 3. Hipk1 regulates Wnt/β-catenin targets in the early embryo.(A) Morpholinos against Hipk1 expand expression domains of the Wnt-responsive dorsal patterning genes siamois (a–c) and Xnr3 (d–f). Embryos were injected with CoMO (40 ng per embryo), Hipk1MO1 (40 ng per embryo), or Hipk1MO2 (20 ng per embryo) in the DMZ of 2 dorsal blastomeres at the 4-cell stage. (B) Quantification of results in A. (C, D) ShRNA plasmids targeted against HIPK1 reduce HIPK1 levels in human embryonic kidney cells (C) and potentiate the response of a β-catenin reporter construct to co-transfected Wnt1 (D). (E) Morpholinos against Hipk1 eliminate or reduce expression of the Wnt-responsive genes MyoD (a–c) Xbra (d–f), and otx2 (g–i), but do not similarly eliminate expression of Xnot (j–l), a marker for axial mesoderm. Embryos were injected as in A. (F) Quantification of results in E.
Figure 4. Hipk1 is nuclear and acts at the level of transcription during X. laevis development.(A) Confocal micrographs of animal pole explants recombinantly expressing Myc-tagged Hipk1 and visualized with anti-Myc antibody. Hipk1 localizes strongly to the nucleus and nuclear speckles, with low-levels of apparent extra-nuclear signal near the plasma membrane. (B–D) Hipk1 over-expression blocks activation of the Wnt-responsive genes siamois and Xnr3 downstream of β-catenin (B), Lef/Tcf (C), and Dsh (D). X. laevis embryos were uninjected (lanes furthest on left in all panels) or injected with synthetic RNAs encoding β-catenin (B), LEF1δN-VP16 (C) or Dsh (D). Nanogram quantities of RNAs encoding Myc-Hipk1, Myc-(K/R)Hipk1 or β-galactosidase were co-injected as indicated. RT-PCR was used to assess target gene activation; histone H4 was assessed as a loading control, and reactions without reverse transcriptase (−RT) performed to rule out genomic DNA contamination. (B, inset) Myc-tagged Hipk1 and (K/R)Hipk1, injected at 1 ng per blastomere, are expressed at comparable levels in X. laevis as demonstrated by western blot. (E) Over-expression of Dsh in HEK293T cells leads to an increase in cytoplasmic β-catenin that is not affected either positively or negatively by over-expression of Hipk1. (F) Induction of Xanf by Noggin in X. laevis animal caps is not affected by over-expression of Hipk1.
Figure 5. Hipk1 over-expression and knock-down both produce gastrulation and neural tube defects.(A) Over-expression of Hipk1 dorsally results in shortened embryos with gastrulation defects and an open neural tube both anteriorly and posteriorly (a). When Hipk1 is over-expressed ventrally, the neural tube closes, but embryos are shortened and fail to complete blastopore closure (b). Synthetic Hipk1 RNA was injected into the marginal zone of 2 dorsal or 2 ventral blastomeres at the 4-cell stage. An uninjected (Uninj) control embryo is shown at the bottom of each panel for comparison. (B) Representative phenotypes used for quantification in C and E. (C) Graphical representation of phenotype distribution resulting from increasing quantities of Hipk1 RNA injected into the DMZ. (D) Injection of Hipk1MO results in gastrulation defects and an open neural tube similar to the over-expression phenotypes in A–C. CoMO (a, c) or Hipk1MO (b, d) was injected either dorsally (a, b), or ventrally (c, d). Embryos injected with the control morpholino did not differ from uninjected embryos (Uninj) shown at the bottom of each panel. (E) Hipk1MO produces gastrulation defects in a dose-dependent manner. Anterior is to the left in A, D, and at the top in B.
Figure 6. Hipk1 is required for gastrulation and convergent extension movements.Morpholino-mediated knock-down of Hipk1 perturbs initiation and progression of gastrulation. (A) Hipk1 morphant embryos exhibit delay of dorsal blastopore lip formation and do not close the blastopore compared to controls (see text). Frames from movies of gastrulating embryos in parallel are presented, from embryos that were either uninjected (Uninj) or injected in the DMZ at the 4-cell stage with 80 ng of CoMo, 40 ng of Hipk1MO1, or 80 ng of Hipk1MO1 as labeled. Elapsed time after appearance of the dorsal lip (black arrowhead) in controls (0 hrs) is indicated. See text for details. (B and C) Keller explant sandwiches from Hipk1 morphant embryos exhibit dose-dependent failure of convergent extension movements compared to controls. DMZ explants were cultured as sandwiches to Stage 19 and average LWR determined. LWR was significantly decreased in Hipk1 morphants compared to controls, and this was dose-dependent (*Pr(T>t) = 0.00001 using 2-sample t test with equal variances).
Figure 7. Hipk1 knock-down and Xdd1 over-expression synergize to perturb gastrulation.(A) Reduction of Hipk1 levels by morpholino, or interference with Dsh function by expression of a dominant negative protein (Xdd1), each produce no gastrulation defects at low doses, with only mild neural tube closure defects (b, c vs. control a). Combining these low-dose treatments produces gastrulation defects plus more severe neural tube closure defects (d, e). (B) Gastrulation defects are present in 68% of embryos in the Xdd1+Hipk1MO group, but not at all in groups injected with either Xdd1 RNA or Hipk1MO alone.
hipk1 (homeodomain interacting protein kinase 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 11, vegetal view, dorsal up.
hipk1 (homeodomain interacting protein kinase 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 17, anterior view, dorsal up.
hipk1 (homeodomain interacting protein kinase 1) gene expression in Xenopus laevis embryo, assayed via in situ hybridization, NF stage 35, lateral view, anterior left, dorsal up.