Zinc finger of the cerebellum (Zic) proteins act as classical transcription factors to promote transcription of the Foxd3 gene during neural crest cell specification. Additionally, they can act as co-factors that bind TCF molecules to repress WNT/β-catenin-dependent transcription without contacting DNA. Here, we show ZIC activity at the neural plate border is influenced by WNT-dependent SUMOylation. In a high WNT environment, a lysine within the highly conserved ZF-NC domain of ZIC5 is SUMOylated, which decreases formation of the TCF/ZIC co-repressor complex and shifts the balance towards transcription factor function. The modification is critical in vivo, as a ZIC5 SUMO-incompetent mouse strain exhibits neural crest specification defects. This work reveals the function of the ZIC ZF-NC domain, provides in vivo validation of target protein SUMOylation, and demonstrates that WNT/β-catenin signaling directs transcription at non-TCF DNA binding sites. Furthermore, it can explain how WNT signals convert a broad domain of Zic ectodermal expression into a restricted domain of neural crest cell specification.

The ZIC proteins are a family of transcription regulators with a well-defined zinc finger DNA-binding domain and there is evidence that they elicit functional DnA binding at a Zic DnA binding site. Little is known, however, regarding domains within Zic proteins that confer trans-activation or-repression. to address this question, a new cell-based trans-activation assay system suitable for Zic proteins in HEK293T cells was constructed. This identified two previously unannotated evolutionarily conserved regions of ZIC3 that are necessary for trans-activation. These domains are found in all Subclass A Zic proteins, but not in the Subclass B proteins. Additionally, the Subclass B proteins fail to elicit functional binding at a multimerised Zic DnA binding site. All Zic proteins, however, exhibit functional binding when the Zic DnA binding site is embedded in a multiple transcription factor locus derived from Zic target genes in the mouse genome. this ability is due to several domains, some of which are found in all Zic proteins, that exhibit context dependent trans-activation or-repression activity. this knowledge is valuable for assessing the likely pathogenicity of variant Zic proteins associated with human disorders and for determining factors that influence functional transcription factor binding. The Zic genes encode a family of multi-functional transcription regulators required for a diverse range of biological processes in embryogenesis and adult homeostasis 1,2. The defining feature of the corresponding proteins (ZIC1, ZIC2, ZIC3, ZIC4 and ZIC5) is a zinc finger domain (ZFD) composed of five tandem Cys2His2 (C2H2)type zinc fingers (ZFs). The ZFD is most closely related to the GLI, GLIS and NKL families but the ZIC ZFD is distinguished by an atypical first ZF. Generally, one to five amino acids separate the two cysteines of a C2H2 ZF but, in the first ZF of ZIC proteins in species so far examined, this number ranges from 6 to 38 3. In addition, the first two ZFs of ZIC proteins contain a tryptophan residue between the canonical cysteines and structural analysis suggests these two ZFs may form a single structural unit called a CWCH2 motif 3,4. The ZIC ZFD was one of the first shown to participate in both DNA binding and protein binding 5 and this dual capability contributes to the myriad molecular roles of ZIC proteins (reviewed in 6). Outside of the ZFD, phylogenetic analysis has identified two other regions conserved amongst ZIC proteins. The N-terminal ZOC box is found within ZICs 1, 2 and 3 3 and participates in protein-protein interactions 7,8. Just N-terminal of the ZFD there is another short, highly conserved sequence, called the ZFNC, which is of unknown function. Each of the ZIC proteins also contain low complexity regions, including stretches of Alanine (ZICs 1, 2, 3, 4 and 5), Histidine (ZICs 1, 2 and 3), as well as Serine/Glycine (ZICs 2 and 5) and Proline (ZIC5) 2. On the basis of protein sequence conservation, ZIC proteins are classified into two Subclasses: Subclass A proteins (ZICs 1, 2 and 3) s...

The five murine Zic genes encode multifunctional transcriptional regulator proteins important for a large number of processes during embryonic development. The genes and proteins are highly conserved with respect to the orthologous human genes, an attribute evidently mirrored by functional conservation, since the murine and human genes mutate to give the same phenotypes. Each ZIC protein contains a zinc finger domain that participates in both protein-DNA and protein-protein interactions. The ZIC proteins are capable of interacting with the key transcriptional mediators of the SHH, WNT and NODAL signalling pathways as well as with components of the transcriptional machinery and chromatin-modifying complexes. It is possible that this diverse range of protein partners underlies characteristics uncovered by mutagenesis and phenotyping of the murine Zic genes. These features include redundant and unique roles for ZIC proteins, regulatory interdependencies amongst family members and pleiotropic Zic gene function. Future investigations into the complex nature of the Zic gene family activity should be facilitated by recent advances in genome engineering and functional genomics.

The mechanisms of neural crest cell induction and specification are highly conserved among vertebrate model organisms, but how similar these mechanisms are in mammalian neural crest cell formation remains open to question. The zinc finger of the cerebellum 1 (ZIC1) transcription factor is considered a core component of the vertebrate gene regulatory network that specifies neural crest fate at the neural plate border. In mouse embryos, however, Zic1 mutation does not cause neural crest defects. Instead, we and others have shown that murine Zic2 and Zic5 mutate to give a neural crest phenotype. Here, we extend this knowledge by demonstrating that murine Zic3 is also required for, and co-operates with, Zic2 and Zic5 during mammalian neural crest specification. At the murine neural plate border (a region of high canonical WNT activity) ZIC2, ZIC3, and ZIC5 function as transcription factors to jointly activate the Foxd3 specifier gene. This function is promoted by SUMOylation of the ZIC proteins at a conserved lysine immediately N-terminal of the ZIC zinc finger domain. In contrast, in the lateral regions of the neurectoderm (a region of low canonical WNT activity) basal ZIC proteins act as co-repressors of WNT/TCF-mediated transcription. Our work provides a mechanism by which mammalian neural crest specification is restricted to the neural plate border. Furthermore, given that WNT signaling and SUMOylation are also features of non-mammalian neural crest specification, it suggests that mammalian neural crest induction shares broad conservation, but altered molecular detail, with chicken, zebrafish, and Xenopus neural crest induction.

Quantitative reverse transcriptase PCR (RT-qPCR), a powerful and efficient means of rapidly comparing gene expression between experimental conditions, is routinely used as a phenotyping tool in developmental biology. The accurate comparison of gene expression across multiple embryonic stages requires normalisation to reference genes that have stable expression across the time points to be examined. As the embryo and its constituent tissues undergo rapid growth and differentiation during development, reference genes known to be stable across some time points cannot be assumed to be stable across all developmental stages. The immediate post-implantation events of gastrulation and patterning are characterised by a rapid expansion in cell number and increasing specialisation of cells. The optimal reference genes for comparative gene expression studies at these specific stages have not been experimentally identified. In this study, the expression of five commonly used reference genes (H2afz, Ubc, Actb, Tbp and Gapdh) was measured across murine gastrulation and patterning (6.5-9.5 dpc) and analysed with the normalisation tools geNorm, Bestkeeper and Normfinder. The results, validated by RT-qPCR analysis of two genes with well-documented expression patterns across these stages, indicated the best strategy for RT-qPCR studies spanning murine gastrulation and patterning utilises the concurrent reference genes H2afz and Ubc.