ZIC
ZIC1 to ZIC5 are five-fingered C2H2 zinc-finger proteins whose DNA-binding domains are the closest relatives of GLI outside the GLIS subfamily, and which read a GLI-like site. In vitro selection gives an optimal ZIC site almost identical to the GLI consensus but bound with lower affinity; modern high-throughput measurements extend it to GACCCCCCGCTGTG, whose reverse complement, CACAGCAGGGGGT, is how several databases write the same motif. ZIC proteins also bind GLI zinc fingers directly and can raise or lower GLI-dependent transcription. They act mainly in dorsal neural tube, neural crest, somites and cerebellum, and human mutations cause holoprosencephaly (ZIC2), X-linked heterotaxy (ZIC3) and Dandy-Walker malformation (ZIC1 with ZIC4).
The factors
The ZIC family has five human members, ZIC1 to ZIC5, each with a tandem array of C2H2 zinc fingers that is more similar to the GLI fingers than to any other vertebrate zinc-finger class. ZIC1 and ZIC4 are a divergently transcribed pair on chromosome 3q24; ZIC2 and ZIC5 are a similar pair on 13q32; ZIC3 sits alone on Xq26 23443491. They bind as monomers. In addition to binding DNA, ZIC proteins associate physically with GLI zinc fingers and, depending on cell type, either enhance or suppress GLI-dependent transactivation 11238441,11053430, so part of their activity is exerted as a GLI cofactor rather than through their own site. Expression is largely neural and paraxial: dorsal neural tube and roof plate, neural crest, somites, and, postnatally, cerebellar granule cells. ZIC2 is also expressed in early epiblast and embryonic stem cells.
The motif
The first systematic definition of the ZIC site used gel-shift-based target selection with ZIC1, ZIC2 and ZIC3 and recovered a sequence almost identical to the GLI binding sequence, but bound with lower affinity than GLI 11053430; consistently, reporter activation by ZIC proteins depended less on the presence of a GLI site than GLI1 activation did, and ZIC proteins activated many promoters lacking one 11053430. Circular-dichroism comparisons of ZIC, GLI and GLIS finger domains on DNA support a shared recognition mode 18298960.
High-throughput SELEX extends the site to a 13-14 base element usually written GACCCCCCGCTGTG 23332764. Its 5' half, GACCCCCC, is the GLIS-like version of the GLI core, and the 3' extension GCTGTG is what distinguishes ZIC from GLI and GLIS in practice. Several databases write the same element on the other strand, as CACAGCAGGGGGT or the shorter CNCAGCAGG, and the ontology form CCTGCTGNG is the reverse complement of that short core; none of these is a different motif. The data consensus here, CNCAGCAGGGGGCC, is the reverse-strand form of the full site. One real subtlety is visible in the collection: the HT-SELEX and methyl-HT-SELEX versions of the ZIC motif differ at a central CpG (CAGCGGGGGG versus CAGCAGGGGG), so cytosine methylation shifts the measured preference at that position 28473536, and different database entries for the same protein can look inconsistent for that reason. No experiment distinguishes ZIC1 from ZIC2, ZIC3, ZIC4 or ZIC5 by sequence; the JASPAR Zic1::Zic2 entry is modeled as a dimer but its consensus is the same short core.
Where it acts
ZIC sites are enhancer elements. In mouse embryonic stem cells ZIC2 occupies thousands of enhancers, both active and poised, and works with the Mbd3/NuRD chromatin-remodeling complex to regulate genes linked to differentiation 25699711; later work describes ZIC2 acting first as a priming factor on closed chromatin during neural induction and then as an activator at established enhancers 42033229. In the cerebellum and dorsal neural tube ZIC occupancy overlaps with GLI-responsive elements, which is expected given the shared core, and ZIC and GLI compete or cooperate at the same sequences. Because the informative part of the site is roughly nine to thirteen bases with a GC-rich core, matches are common in GC-rich regulatory DNA, and the motif overlaps in scoring space with GLI, GLIS and, at the CAGCAGG core, with simple CAG-repeat and E-box-adjacent sequence.
Biology and disease
Gli/Zic factors together pattern the vertebrate neural plate by defining domains of differentiation, with ZIC activity dorsal to the SHH-dependent ventral domains 9634234. Heterozygous ZIC2 loss-of-function causes holoprosencephaly type 5, characteristically with a middle-interhemispheric variant and distinctive facial features 29442327; ZIC3 mutations cause X-linked heterotaxy with complex congenital heart defects and laterality reversal 29442328; heterozygous deletion of the linked ZIC1 and ZIC4 pair causes Dandy-Walker malformation 15338008, and mouse work shows that combined Zic1 and Zic4 loss disturbs several developmental programs in the cerebellar anlage rather than one 21307096. ZIC5 was identified in Xenopus as a neural-crest regulator 23443491. Beyond development, ZIC family members are recurrently deregulated in tumors, most consistently by promoter hypermethylation and silencing; ZIC1, for example, is downregulated by promoter hypermethylation in gastric cancer 19135984.
Reading a hit
Treat a ZIC match as "ZIC family", never as a specific paralogue: all five members share one motif. Then treat it cautiously as "ZIC rather than GLI". The two families were defined on essentially the same 5' core, and separating them depends on the 3' GCTGTG extension, which carries relatively little information; a short or weak match can belong to either, and a ZIC hit inside a Hedgehog-responsive enhancer is as likely to be a GLI site. Note also that a real ZIC effect need not require a ZIC site at all, since ZIC proteins act as GLI cofactors and in reporter assays activate promoters lacking the site. The GC-rich core makes the motif frequent in GC-rich regulatory DNA, so isolated matches are weak evidence. In the data behind this family, one Vierstra archetype labelled ZIC (AC0536) has the consensus GACCACCC and matches the GLI family better than its own; that is the expected failure mode for this pair, not a claim that the label is wrong.
How this was worked out
zic is cloned from mouse cerebellum as a zinc-finger gene expressed in granule-cell precursors, naming the family and placing its zinc fingers closest to the GLI fingers.
Still open then: No binding site had been measured; the GLI similarity was a sequence inference.
Zic2 and Gli proteins are shown to pattern the frog neural plate into alternating domains, Zic2 inducing neural crest and inhibiting neurogenesis while Gli does the opposite, establishing the two as a functionally paired system.
Still open then: Whether the antagonism was exerted through shared DNA sites or through protein interaction.
- 2001mediumEMSA-based site selection with mutational analysis, reporter assays, co-immunoprecipitation11053430 11238441
Gel-shift target selection gives ZIC1, ZIC2 and ZIC3 a site almost identical to the GLI site but bound with lower affinity, and ZIC and GLI proteins are shown to associate physically through their zinc-finger domains and to modulate each other's activity.
Still open then: Because ZIC proteins also activated promoters lacking the site, how much of their activity is site-dependent was unclear.
Heterozygous deletion of the linked ZIC1-ZIC4 pair is identified as a cause of Dandy-Walker malformation, giving the family its first human cerebellar phenotype.
HT-SELEX extends the ZIC site to a 13-14 base element, GACCCCCCGCTGTG, whose 3' GCTGTG extension is what separates it from the GLI and GLIS versions of the shared GC-rich core.
Still open then: Whether the low-information 3' extension is used to partition sites in vivo.
ZIC2 is mapped genome-wide in mouse embryonic stem cells and found at thousands of enhancers, where it is required for normal differentiation, moving the family from a patterning gene to a measured enhancer-binding factor.
Methyl-HT-SELEX shows that the measured ZIC preference shifts at the central CpG, so the HT-SELEX and methyl-HT-SELEX models of the same protein differ (CAGCGGGGGG versus CAGCAGGGGG).
Still open then: Whether methylation-dependent ZIC binding occurs at methylated regulatory DNA in cells.
- 2026mediumGenome editing in mESC-derived neural progenitors, ChIP-seq, ATAC-seq, transcriptomics42033229
During neural induction from mouse embryonic stem cells, ZIC2 binds de novo to closed distal enhancers and raises their accessibility at pluripotency exit, then activates a subset of those primed enhancers, so one factor occupies the same sites in two mechanistically different roles; ZIC3 compensates for its loss in the pluripotent state.
Still open then: What converts a primed ZIC2-bound enhancer into an active one is not identified.
Still unresolved
- How much of ZIC activity runs through the ZIC site itself is unsettled: ZIC proteins activate reporters lacking the site and bind GLI zinc fingers directly, yet genome-wide maps show them occupying thousands of enhancers in a sequence-dependent way.
- Whether the 3' GCTGTG extension is actually used in cells to partition sites between ZIC and GLI/GLIS is unknown; it is the only feature separating the two families in vitro, and it carries little information.
- Whether the methylation-dependent shift at the central CpG, seen when HT-SELEX and methyl-HT-SELEX models of the same ZIC protein are compared, has any counterpart at methylated enhancers in vivo has not been tested.
- Nothing in the measured specificities separates ZIC1 from ZIC2, ZIC3, ZIC4 or ZIC5, yet their loss-of-function phenotypes differ sharply (holoprosencephaly, heterotaxy, Dandy-Walker malformation); whether that reflects expression domains alone or unmeasured differences in site preference or partner use is open.
The five ZIC paralogues are indistinguishable by sequence, so a hit names the family only. The larger problem is the GLI boundary: ZIC and GLI were defined on nearly the same core, ZIC binds the GLI consensus in vitro (with lower affinity), and the two are separated only by a weakly informative 3' extension, so weak matches cannot be assigned to one family. One Vierstra archetype labelled ZIC in these data in fact scores closest to GLI. In vitro entries for the same ZIC protein can also disagree at a central CpG because the methyl-SELEX and standard SELEX preferences differ there. Finally, ZIC can act on GLI-dependent transcription without occupying its own site.
References
- Mizugishi K, Aruga J, Nakata K, et al. Molecular properties of Zic proteins as transcriptional regulators and their relationship to GLI proteins. J Biol Chem 2001;276:2180-8. PMID 11053430
- Koyabu Y, Nakata K, Mizugishi K, et al. Physical and functional interactions between Zic and Gli proteins. J Biol Chem 2001;276:6889-92. PMID 11238441
- Brewster R, Lee J, Ruiz i Altaba A. Gli/Zic factors pattern the neural plate by defining domains of cell differentiation. Nature 1998;393:579-83. PMID 9634234
- Sakai-Kato K, Ishiguro A, Mikoshiba K, et al. CD spectra show the relational style between Zic-, Gli-, Glis-zinc finger protein and DNA. Biochim Biophys Acta 2008;1784:1011-9. PMID 18298960
- Jolma A, Yan J, Whitington T, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327-39. PMID 23332764
- Yin Y, Morgunova E, Jolma A, et al. Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science 2017;356. PMID 28473536
- Luo Z, Gao X, Lin C, et al. Zic2 is an enhancer-binding factor required for embryonic stem cell specification. Mol Cell 2015;57:685-694. PMID 25699711
- Grinberg I, Northrup H, Ardinger H, et al. Heterozygous deletion of the linked genes ZIC1 and ZIC4 is involved in Dandy-Walker malformation. Nat Genet 2004;36:1053-5. PMID 15338008
- Blank MC, Grinberg I, Aryee E, et al. Multiple developmental programs are altered by loss of Zic1 and Zic4 to cause Dandy-Walker malformation cerebellar pathogenesis. Development 2011;138:1207-16. PMID 21307096
- Barratt KS, Arkell RM. ZIC2 in Holoprosencephaly. Adv Exp Med Biol 2018;1046:269-299. PMID 29442327
- Bellchambers HM, Ware SM. ZIC3 in Heterotaxy. Adv Exp Med Biol 2018;1046:301-327. PMID 29442328
- Houtmeyers R, Souopgui J, Tejpar S, et al. The ZIC gene family encodes multi-functional proteins essential for patterning and morphogenesis. Cell Mol Life Sci 2013;70:3791-811. PMID 23443491
- Aruga J, Yokota N, Hashimoto M, et al. A novel zinc finger protein, zic, is involved in neurogenesis, especially in the cell lineage of cerebellar granule cells. J Neurochem 1994;63:1880-90. PMID 7931345
- Wang LJ, Jin HC, Wang X, et al. ZIC1 is downregulated through promoter hypermethylation in gastric cancer. Biochem Biophys Res Commun 2009;379:959-63. PMID 19135984
- Mariner-Faulí M, Sánchez-Gaya V, Robert SM, et al. Dual role of ZIC2 during neural induction: from priming transcription factor to enhancer activator. Nucleic Acids Res 2026;54. PMID 42033229
24 motifs in this family — JASPAR 6 · HOCOMOCO 11 · CIS-BP 5 · Vierstra 2
| motif | database | consensus | bits |
|---|---|---|---|
| M04491_2.00 | CIS-BP | GACCCCCCGCTGTGa | 25 |
| ZIC5.H14CORE.1.M.CMC | HOCOMOCO | CACAGCAGGGGGTC | 22.5 |
| ZIC5.H14CORE.2.S.CSC | HOCOMOCO | CgCAGCGGGGGGTc | 20.2 |
| ZIC3.H14CORE.2.M.BMB | HOCOMOCO | CACAGCAGGGGGt | 20 |
| ZIC5 (MA1584.2) | JASPAR | GACCCCCCGCTGyG | 19 |
| M04471_2.00 | CIS-BP | GACCCCCCGCTGyG | 19 |
| ZIC1 (MA0696.1) | JASPAR | GACCCCCCGCTGtG | 18.8 |
| ZIC1.H14CORE.1.M.BMB | HOCOMOCO | CACAGCAGGGGGt | 16.9 |
| ZIC1.H14CORE.2.S.BSB | HOCOMOCO | CaCAGCGGGGGGT | 16.2 |
| M04485_2.00 | CIS-BP | aCCCCCCGCtGcGc | 16.2 |
| ZIC5.H14CORE.0.P.BPB | HOCOMOCO | CnCAGCAGGGGGcC | 16.1 |
| AC0494:ZIC/GLIS:C2H2_ZF | Vierstra | ACCCCCcGCTGtG | 14.7 |
| ZIC3.H14CORE.1.S.BSB | HOCOMOCO | CaCAGCGGGGGG | 13.8 |
| M02912_2.00 | CIS-BP | gaCCCCCCGCtGnG | 13.6 |
| ZIC4 (MA0751.2) | JASPAR | gaCCCCCCGCtGnG | 13.5 |
| ZIC3.H14CORE.0.P.BPB | HOCOMOCO | ChCAGCAGGrg | 13.2 |
| Zic2 (MA1629.2) | JASPAR | CaCAGCAGG | 12.8 |
| ZIC2.H14CORE.0.P.BPB | HOCOMOCO | ChCAGCAGGgg | 12.8 |
| ZIC4.H14CORE.0.SM.BSMB | HOCOMOCO | CaGCgGGGGGT | 12 |
| Zic1::Zic2 (MA1628.2)dimer | JASPAR | CAGCAGG | 11.5 |
| ZIC1.H14CORE.0.P.BPB | HOCOMOCO | CnCAGCaGG | 11.2 |
| Zic3 (MA0697.3) | JASPAR | CAGCAGG | 10.6 |
| M08233_2.00 | CIS-BP | CCTGCTGdG | 10.4 |
| AC0536:ZIC:C2H2_ZFlooks like GLI | Vierstra | gACCaCCC | 5.5 |
HOCOMOCO tags read as assay + quality grade: P ChIP-seq, S HT-SELEX, M methyl-HT-SELEX, G GHT-SELEX, I SMiLE-seq, B PBM; grade A means two or more assay types agreed. A P-only matrix is a record of where the protein was crosslinked, which need not be a sequence it binds itself.
The factors, elsewhere
These entries are written from the primary literature and every citation is checked against PubMed, but this one can still be out of date, too confident, or simply wrong — especially where a motif’s published site and its database matrices disagree. Corrections and additions are welcome, and they are the fastest way to make this page better for everyone.