We have used biochemical, transcriptional and developmental analyses to define the functional domains and a transcriptional cofactor required for FoxD3 function in Xenopus. A strong transcriptional repression domain was identified in the C-terminus of Xenopus FoxD3 that is required for biological activity. A consensus eh1/GEH Groucho corepressor interaction motif is present within the repression domain and this motif is conserved in all FoxD3 orthologs. The eh1/GEH motif is essential for the mesoderm induction and transcriptional repression activities of FoxD3, as well as for the direct physical interaction of FoxD3 and Grg4. In transcriptional assays, Grg4 synergistically enhances and Grg5 inhibits the repression activity of FoxD3, providing further support for a functional interaction of FoxD3 with Groucho corepressors. Taken together, the results demonstrate that FoxD3 interacts with Groucho corepressors via a conserved eh1/GEH motif, and that this interaction is required for the repression of transcription and induction of mesodermal cell types in the Xenopus embryo. Furthermore, the mechanistic studies we report here suggest that in the Xenopus gastrula, FoxD3 recruits Groucho corepressors to target gene promoters and the resulting transcriptional repression of FoxD3 target genes promotes the induction of mesoderm.
Cells of the Xenopus
blastula are competent to respond to FoxD3, and in these developmentally responsive embryonic cells we have shown that FoxD3 functions as a Groucho-dependent transcriptional repressor. Previous studies of FoxD3 in cell culture and in neural crest support this conclusion. In 293 and HeLa cells, mouse FoxD3 repressed transcription of a co-transfected reporter plasmid (9
). Similarly, chick FoxD3 strongly represses the transcription of a reporter in co-transfected chicken embryo fibroblasts, and the repression domain maps to a C-terminal region that contains the conserved eh1/GEH motif we identified in all FoxD3 orthologs (31
). In studies of Xenopus
mesoderm induction and neural crest development, an Engrailed-FoxD3 fusion protein containing the transcriptional repression domain of Drosophila
Engrailed and the WH DNA-binding domain of FoxD3 was functionally identical to native FoxD3 (17
). The Engrailed repression domain contains an eh1/GEH motif that mediates recruitment of Groucho corepressors (38
). This suggests that a heterologous eh1/GEH motif can recruit Groucho corepressors to FoxD3 target genes via the FoxD3 WH domain, and that this is sufficient to mimic the biological function of native FoxD3. Taken together, these studies argue strongly that FoxD3 regulates the development of mesoderm and neural crest by recruitment of Groucho corepressors to repress target gene transcription.
In contrast to our results, it has been reported that FoxD3 functions as a transcriptional activator in a context-dependent or lineage-specific manner. In co-transfected 293 cells, mouse FoxD3 was found to activate transcription of reporters containing regulatory elements of Osteopontin, FoxA1, or FoxA2 (57
). Interestingly, FoxD3 activates FoxA1 and FoxA2 in a narrow dose range, and transcriptional activation was not observed at higher doses of FoxD3. The complexity of the response of FoxA1 and FoxA2 to FoxD3 is further demonstrated by the ability of Oct4, a known transcriptional activator, to inhibit the transcriptional response of FoxA1 and FoxA2 to FoxD3. Given that mouse FoxD3 can also function as a repressor in 293 cells, as discussed above, these observations suggest that the transcriptional activity of FoxD3 may be dependent on promoter context and/or the availability of transcriptional cofactors that result in activation of certain targets and repression of others. It should also be noted that the mammalian FoxD3 proteins contain several polyalanine and polyglycine sequences that are not present in other FoxD3 orthologs, and these sequences may confer additional transcriptional functions on the mammalian proteins.
An activation function for FoxD3 has also been suggested in zebrafish somitogenesis (58
). In a yeast one-hybrid screen for regulators of the myogenic factor myf5
, zebrafish FoxD3 was identified as a protein that binds to a somite-specific regulatory element of myf5
. In cell culture studies, FoxD3 weakly activated (2–4-fold) a transcriptional reporter containing the myf5
regulatory element. Consistent with a role in myf5
regulation, FoxD3 is coexpressed with myf5
in somites and in presomitic mesoderm of the zebrafish. FoxD3 knockdown resulted in a loss of myf5
expression in somites during the 8–16-somite stage, but had no effect on myf5
expression in the presomitic mesoderm. At these stages, FoxD3 is also expressed in the zebrafish tailbud, but myf5
is not. Interestingly, ectopic myf5
expression was observed in the tailbud domain of FoxD3 knockdown embryos. These results indicate that the regulatory relation of FoxD3 and myf5
differs in distinct regions of the somite-stage zebrafish embryo. In newly formed somites FoxD3 activates myf5
expression, in the presomitic mesoderm FoxD3 has no apparent influence on myf5
expression, and in the tailbud FoxD3 inhibits myf5
expression. Therefore, the transcriptional function of FoxD3 may differ in distinct lineages of the somite-stage zebrafish, perhaps due to the lineage-specific expression of FoxD3 coactivators, corepressors, or other interacting factors.
In our study of FoxD3, we find no evidence of transcriptional activation function in the Xenopus
gastrula. No domains of FoxD3 were identified that were capable of activating reporter transcription, and the biological activity of FoxD3 was completely dependent on the eh1/GEH Groucho interaction motif. Furthermore, a VP16-FoxD3 fusion protein containing the strong activation domain of HSV VP16 and the WH DNA-binding domain of FoxD3 not only failed to mimic the activity of native FoxD3, but dominantly inhibited the activity of FoxD3 in both mesoderm induction and neural crest specification (17
). It is important to emphasize that our experiments were performed in the Xenopus
gastrula, providing a cellular and embryonic context that is equivalent to that of endogenous FoxD3. So although FoxD3 may have distinct transcriptional functions in other lineages or at other stages of development, we can strongly conclude that FoxD3 functions as a Groucho-dependent transcriptional repressor to regulate Xenopus
FoxD3 is an essential transcriptional regulatory protein for multiple developmental processes in vertebrates. Studies of neural crest development in Xenopus
, zebrafish and chick indicate that FoxD3 regulates the determination, survival, migration, and/or differentiation of neural crest lineages (13
), while in the mouse embryo, FoxD3
is essential for maintenance of embryonic and trophoblast stem cells (21
). Embryonic stem cells, trophoblast stem cells and neural crest progenitors are multipotent cell types, and the requirement for FoxD3 in each of these cell types suggests a conserved role for FoxD3 in maintaining cellular multipotency. The functional interaction of FoxD3 and Groucho corepressors we describe in Xenopus
mesoderm induction may provide mechanistic insight into FoxD3 function in neural crest and stem cells. Mouse Grg3 and Grg4 and Xenopus
Grg4 are coexpressed with FoxD3 in neural crest lineages (42
), consistent with a potential functional interaction of FoxD3 and Groucho corepressors in neural crest cells, although a direct role for Groucho corepressors in neural crest development has yet to be demonstrated. In addition, microarray analysis has shown that mouse Grg3 and Grg4 are enriched in embryonic stem cells when compared to differentiated cell types (62
). The coexpression of FoxD3 and Groucho corepressors in these progenitor cells suggests that FoxD3 repression of target gene transcription may be essential for maintaining multipotency in diverse progenitor cell populations. This is a compelling idea given the evidence that transcriptional repression of differentiation genes is a key mechanism of stem cell and progenitor cell maintenance (63
). Further work will be necessary to determine if FoxD3 has similar transcriptional activity and common target genes in neural crest and stem cells, and whether FoxD3 function in these progenitor populations is dependent on an interaction with Groucho corepressors.
The interaction of FoxD3 and Groucho we describe appears to reflect a conserved functional characteristic of the entire FoxD subclass. Xenopus
FoxD1 and FoxD5, and chick FoxD2 each function as developmentally important transcriptional repressors and contain a C-terminal repression domain (31
). In fact, every vertebrate and invertebrate member of the FoxD subclass (27 proteins) contains a C-terminal sequence with high similarity to the eh1/GEH Groucho interaction motif, including ancestral FoxD proteins in marine sponge, Ciona, and Amphioxus (68
). While the identification of Groucho orthologs in primitive species (72
) suggests that the eh1/GEH-like sequences may be functional, it remains to be determined for many of the FoxD subclass proteins whether the eh1/GEH sequences mediate Groucho interaction and transcriptional repression. Taken together, however, these data suggest an ancient origin for the eh1/GEH motif-dependent recruitment of Groucho corepressors by FoxD subclass proteins.
FoxD3 is an essential transcriptional regulator of multiple developmental processes, and our results demonstrate that in Xenopus mesoderm induction FoxD3 functions as a Groucho-dependent transcriptional repressor. Among a number of questions for future work, it will be important to determine the transcriptional activity and identify the transcriptional targets of FoxD3 in distinct lineages. Whether FoxD3 is a context-dependent regulator that can activate or repress transcription in a cell type-specific or target-specific manner will require further functional analyses in a number of FoxD3-expressing lineages. Furthermore, identification of FoxD3 target genes in the organizer, neural crest and stem cells will reveal if a common regulatory pathway is utilized in each of these cell types, or if there are lineage-specific mechanisms of FoxD3 function. Ongoing studies of FoxD3 in multiple embryonic settings are likely to provide further insight into the developmental and molecular mechanisms of vertebrate embryogenesis.