Grainyhead (GRHL)
The Grainyhead/CP2 family binds the palindrome AACCGGTT through an immunoglobulin-like DNA-binding domain that shares a fold with p53 but differs from it in important details. Six human members split into two branches: the epithelial GRHL1, GRHL2 and GRHL3, and the more broadly expressed TFCP2 (LSF/CP2), TFCP2L1 and UBP1 (LBP-1a). They bind as dimers, and many real elements are doubled, giving 16-26 bp tandem sites. GRHL factors control epidermal barrier formation, neural tube and palate closure, and epithelial identity, and they sit at the epithelial end of EMT regulation. Human mutations cause Van der Woude syndrome and nonsyndromic cleft palate.
The factors
Six human genes belong to this family. The grainyhead-like branch, GRHL1, GRHL2 and GRHL3, is epithelial: GRHL1 and GRHL3 in epidermis and periderm, GRHL2 in most simple and stratified epithelia including airway, gut and bladder. The CP2 branch, TFCP2 (also LSF or CP2), TFCP2L1 (CRTR-1) and UBP1 (LBP-1a), is more broadly expressed and includes the factor originally characterized on the alpha-globin promoter 8157699. All six share a roughly 250-residue region that carries both DNA binding and dimerization, inherited from Drosophila Grainyhead (also called NTF-1 and Elf-1), and the Drosophila and human domains are interchangeable enough that the family was recognized by that conservation 8196641,12175488. Crystal structures of the GRHL1 and GRHL2 DNA-binding domains showed a closely similar immunoglobulin-like core sharing the overall fold of the p53 DNA-binding domain while differing in the DNA-contacting elements 29309642. GRHL1 binds its recognition element as a dimer, and target-site selection is parsimonious: two conserved arginine residues make most of the base-specific contacts 29309642.
The motif
The consensus is the 8-bp palindrome AACCGGTT, often written with flanking A/T as AAACCGGTTT. Because selection rests on a small number of arginine contacts, the central CCGG is the most constrained part and the flanking AA/TT contribute the rest. The palindrome means the two halves are equivalent, and dimer binding is the normal mode. A conspicuous feature of this family's models is doubling: HOCOMOCO entries for TFCP2, TFCP2L1, UBP1 and GRHL3 span 15-26 columns and contain two AACCGGTT-like units, sometimes in the arrangement AACCGGTT-N(2)-AACCGGTT. That is a genuine property of Grh/CP2 elements rather than a modelling artifact 9668115, and it is why the highest information content in this family (about 20 bits) comes from CP2-branch models rather than from the single-site GRHL models. Degenerate variants are common and legitimate: AACAGGTT, AACTGGTT and AACCAGTT all appear as database consensuses, reflecting tolerance at the central positions. No member is reliably distinguishable from another by sequence. The GRHL and CP2 branches are similar enough that Vierstra collapses them into archetypes named GRHL/TFCP, and this family's cohesion (0.78) is limited mainly by the mixture of single-site and doubled models, not by disagreement about the core.
Where it acts
Grh/CP2 sites occur at both promoters and enhancers, with a strong bias toward epithelial differentiation loci: the epidermal differentiation complex, tight junction and desmosome genes, transglutaminases, and the genes of the cornified envelope. GRHL3 binding is highly dynamic: when progenitor keratinocytes differentiate it binds mainly at super-enhancers, whereas when they migrate it binds mainly at promoters, together with REST 28445475. GRHL2 also occupies enhancers in pluripotent cells during exit from naive pluripotency and in airway basal cells during mucociliary regeneration. Common neighbors are KLF4 and TP63 in keratinocytes, IRF6 in periderm (IRF6 binds an enhancer near GRHL3 and is needed for Grhl3 expression in periderm 22931925, although Irf6;Grhl3 double-heterozygous mice show no epistasis 24360809), and AP-1 at differentiation enhancers. GRHL factors recruit Trithorax-group complexes to switch on differentiation genes 22829784, and in cancer they sit in a mutually repressive loop with ZEB1 and the miR-200 family that keeps cells epithelial 26887977. The palindrome is 8 bp, so chance matches are frequent; doubled sites are much rarer and much more indicative.
Biology and disease
Grhl3-null mice fail to form a skin permeability barrier 15831758 and develop spina bifida that is resistant to both inositol and folate supplementation, identifying Grhl3 as the gene behind the long-studied curly tail mouse 14608380. Grhl3 is also required for eyelid closure and urothelial differentiation. In humans, dominant GRHL3 mutations cause Van der Woude syndrome, the commonest syndromic form of cleft lip and palate, in families without IRF6 mutations 24360809; separate GRHL3 coding variants, including a common susceptibility allele and rare truncating variants, cause and predispose to nonsyndromic cleft palate 27018472. GRHL2 variants are associated with age-related hearing impairment 17921507, and GRHL2 has both oncogenic and tumor-suppressive effects depending on context, promoting the epithelial phenotype in breast, ovarian and prostate cancer while restraining metastatic dissemination. TFCP2 appears in cancer as a fusion partner: EWSR1-TFCP2 and FUS-TFCP2 fusions define an aggressive intraosseous spindle and epithelioid rhabdomyosarcoma with frequent ALK overexpression, often in the jaw and skull 33382123.
Reading a hit
An AACCGGTT match identifies the Grh/CP2 family and no individual member. That matters because the two branches have different biology: a hit could correspond to epithelial GRHL2 or to broadly expressed TFCP2, and only expression or ChIP data can tell you which. At 8 bp with a palindromic core, single matches occur often by chance, so treat a lone AACCGGTT the way you would treat any 8-mer: as a hypothesis. A doubled site, two AACCGGTT-like units within about 20 bp, is a much stronger signal and is the form in which many validated elements occur. Also note that models in this family differ in width by a factor of three, from 8 to 26 columns, so scores from different databases are not comparable and a "weaker" hit may simply reflect a narrower model. The evidence base is mixed but reasonable: several HOCOMOCO models for GRHL1, GRHL3, TFCP2, TFCP2L1 and UBP1 rest on in vitro selection (evidence codes SM, SB, S), which supports direct binding, whereas the GRHL2 models here are ChIP-only (P) and their degenerate variants (AACAGGTT, CCGGTTTGGCCAG) should be read as descriptions of GRHL2 peaks rather than of GRHL2's intrinsic specificity.
How this was worked out
Mapping of Drosophila Grainyhead (NTF-1) identifies a large DNA-binding domain and a separate dimerization domain that stabilizes DNA binding; both are conserved in the human alpha-globin factor CP2, which has similar activities, defining the Grh/CP2 family.
Still open then: The DNA-binding domain resembled no known fold.
LSF (CP2) and NTF-1 (Grainyhead) share a recognition motif but not a binding mode: dimeric NTF-1 binds a single half-site, whereas LSF cannot and instead binds directly repeated elements as a tetramer.
Still open then: Whether the mammalian grainyhead-like proteins behave like NTF-1 or like LSF.
Cloning of two new mammalian genes more closely related to grainyhead than CP2, LBP-1a and LBP-9 splits the family into a grainyhead-like branch and a CP2 branch.
Grhl3 is identified as the gene underlying the curly tail mouse; Grhl3-null embryos have neural tube defects that resist both folate and inositol.
Grhl3-null mice have a defective epidermal barrier and poor wound repair with reduced transglutaminase 1, mirroring Grainyhead's control of cuticle cross-linking in flies.
- 2014mediumFamily sequencing, zebrafish periderm assay, mouse genetics; microarrays and enhancer assays (2013)24360809 22931925
Dominant GRHL3 mutations cause Van der Woude syndrome in families lacking IRF6 mutations; double-heterozygous mice show no Irf6-Grhl3 epistasis, although a 2013 study had placed Grhl3 directly downstream of Irf6 in periderm.
Still open then: Whether IRF6 and GRHL3 act in one pathway or in convergent ones.
Crystal structures of the GRHL1 and GRHL2 DNA-binding domains show an immunoglobulin-like core with the p53 fold; GRHL1 binds its consensus element as a dimer and selects it through two conserved arginines, one of which is hit by a cancer-associated mutation that abolishes binding.
Still open then: No structure explains the tetrameric CP2-branch complex on repeated sites.
Still unresolved
- Whether IRF6 and GRHL3 act in one pathway in oral periderm is disputed: zebrafish, frog and mouse data place Grhl3 directly downstream of Irf6, whereas Irf6;Grhl3 double-heterozygous mice show no epistasis and were read as separate, convergent pathways.
- How far the doubled-site arrangement extends is unresolved: biochemistry shows the CP2-branch factor LSF needs directly repeated elements and binds as a tetramer while Grainyhead binds a single site as a dimer, yet the in vitro-supported HOCOMOCO GRHL1 and GRHL3 models are partly doubled too.
A hit names the family, not the member, and the family spans two branches with different expression: epithelial GRHL1-3 versus broadly expressed TFCP2, TFCP2L1 and UBP1. The 8-bp palindrome occurs frequently by chance, so single matches are weak; doubled AACCGGTT units within about 20 bp are much more convincing. Model widths in this family range from 8 to 26 columns, so scores across databases are not comparable. Degenerate variants such as AACAGGTT come mostly from ChIP-only GRHL2 models and describe peak composition rather than intrinsic specificity; the in vitro-supported models converge on AACCGGTT.
References
- Ming Q, Roske Y, Schuetz A, et al. Structural basis of gene regulation by the Grainyhead/CP2 transcription factor family. Nucleic Acids Res 2018;46:2082-2095. PMID 29309642
- Uv AE, Thompson CR, Bray SJ. The Drosophila tissue-specific factor Grainyhead contains novel DNA-binding and dimerization domains which are conserved in the human protein CP2. Mol Cell Biol 1994;14:4020-31. PMID 8196641
- Wilanowski T, Tuckfield A, Cerruti L, et al. A highly conserved novel family of mammalian developmental transcription factors related to Drosophila grainyhead. Mech Dev 2002;114:37-50. PMID 12175488
- Ting SB, Wilanowski T, Auden A, et al. Inositol- and folate-resistant neural tube defects in mice lacking the epithelial-specific factor Grhl-3. Nat Med 2003;9:1513-9. PMID 14608380
- Peyrard-Janvid M, Leslie EJ, Kousa YA, et al. Dominant mutations in GRHL3 cause Van der Woude Syndrome and disrupt oral periderm development. Am J Hum Genet 2014;94:23-32. PMID 24360809
- Leslie EJ, Liu H, Carlson JC, et al. A Genome-wide Association Study of Nonsyndromic Cleft Palate Identifies an Etiologic Missense Variant in GRHL3. Am J Hum Genet 2016;98:744-54. PMID 27018472
- Klein RH, Lin Z, Hopkin AS, et al. GRHL3 binding and enhancers rearrange as epidermal keratinocytes transition between functional states. PLoS Genet 2017;13:e1006745. PMID 28445475
- Hopkin AS, Gordon W, Klein RH, et al. GRHL3/GET1 and trithorax group members collaborate to activate the epidermal progenitor differentiation program. PLoS Genet 2012;8:e1002829. PMID 22829784
- Chung VY, Tan TZ, Tan M, et al. GRHL2-miR-200-ZEB1 maintains the epithelial status of ovarian cancer through transcriptional regulation and histone modification. Sci Rep 2016;6:19943. PMID 26887977
- Van Laer L, Van Eyken E, Fransen E, et al. The grainyhead like 2 gene (GRHL2), alias TFCP2L3, is associated with age-related hearing impairment. Hum Mol Genet 2008;17:159-69. PMID 17921507
- Xu B, Suurmeijer AJH, Agaram NP, et al. Head and neck rhabdomyosarcoma with TFCP2 fusions and ALK overexpression: a clinicopathological and molecular analysis of 11 cases. Histopathology 2021;79:347-357. PMID 33382123
- Swendeman SL, Spielholz C, Jenkins NA, et al. Characterization of the genomic structure, chromosomal location, promoter, and development expression of the alpha-globin transcription factor CP2. J Biol Chem 1994;269:11663-71. PMID 8157699
- Shirra MK, Hansen U. LSF and NTF-1 share a conserved DNA recognition motif yet require different oligomerization states to form a stable protein-DNA complex. J Biol Chem 1998;273:19260-8. PMID 9668115
- Ting SB, Caddy J, Hislop N, et al. A homolog of Drosophila grainy head is essential for epidermal integrity in mice. Science 2005;308:411-3. PMID 15831758
- de la Garza G, Schleiffarth JR, Dunnwald M, et al. Interferon regulatory factor 6 promotes differentiation of the periderm by activating expression of Grainyhead-like 3. J Invest Dermatol 2013;133:68-77. PMID 22931925
23 motifs in this family — JASPAR 4 · HOCOMOCO 11 · CIS-BP 5 · Vierstra 3
| motif | database | consensus | bits |
|---|---|---|---|
| TFCP2.H14CORE.0.SM.BSMB | HOCOMOCO | CcgGttnrAaCCGGTTtaAACCGGTT | 19.8 |
| TFCP2.H14CORE.1.SM.BSMB | HOCOMOCO | AACCGGtTtgAACCGGTT | 19.3 |
| GRHL1.H14CORE.0.PSM.APSMA | HOCOMOCO | AACCGGTTtaaCcgG | 16.3 |
| TF2L1.H14CORE.0.PSM.APSMA | HOCOMOCO | CCGGttnrAACCGGTT | 15.5 |
| UBIP1.H14CORE.0.SM.BSMB | HOCOMOCO | AACCGGtThnAaCCgG | 15.1 |
| GRHL3.H14CORE.0.SB.ASBA | HOCOMOCO | CcgGthnAACCGGTT | 14.4 |
| GRHL1 (MA0647.2) | JASPAR | AAACCGGTTT | 13.9 |
| TFCP2.H14CORE.2.S.CSC | HOCOMOCO | AaAACCGGtT | 13.6 |
| AC0439:TFCP:Grainyhead | Vierstra | CTGGgThGdGC | 13.6 |
| GRHL1.H14CORE.1.PS.APSA | HOCOMOCO | AAACCGGTT | 12.7 |
| AC0473:TFCP/TFCP2L:Grainyhead | Vierstra | ACCGGTthdAACCGGt | 12.1 |
| M08129_2.00 | CIS-BP | AACTGGTTT | 12 |
| GRHL2.H14CORE.0.P.BPB | HOCOMOCO | CcgGTTtggCcaG | 11.9 |
| M04893_2.00 | CIS-BP | aCCGGttcgaaCCGGt | 11.8 |
| Tfcp2l1 (MA0145.2) | JASPAR | CCaGttnnArCCaG | 11.7 |
| M03074_2.00 | CIS-BP | AACCGGTTt | 11.7 |
| TFCP2 (MA1968.2) | JASPAR | aAACCGGTT | 11.4 |
| GRHL2 (MA1105.3) | JASPAR | AACAGGTT | 11 |
| M04887_2.00 | CIS-BP | AACCGGTT | 11 |
| GRHL2.H14CORE.1.P.BPB | HOCOMOCO | AACcGGTTTg | 10.6 |
| M04885_2.00 | CIS-BP | aCcgGttynaaCcgGt | 10.1 |
| AC0474:GRHL/TFCP:Grainyhead | Vierstra | AACCGGTT | 9.5 |
| TF2L1.H14CORE.1.P.BPB | HOCOMOCO | AaCCAGtt | 8.8 |
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.