DACH / SKI-family corepressors
These are transcriptional corepressors, not classical sequence-specific activators. DACH1 and DACH2 are the vertebrate Dachshund proteins, and SKI, SKIL/SnoN, SKOR1 and SKOR2 are the Ski family; all share a Dachshund homology domain with a winged-helix fold related to forkhead. They work mostly by being recruited: DACH proteins into SIX/EYA complexes and onto forkhead sites where they compete with FOXM1, the Ski proteins onto SMAD complexes to shut down TGF-beta signalling, and SKOR1/SKOR2 onto homeodomain factors such as LBX1. The motifs in this family therefore look borrowed. The best-supported models, TTAATTAC for SKOR1 and SKOR2, are indistinguishable from homeodomain dimer sites, which is what you would expect from proteins that arrive at DNA on someone else's back.
The factors
Two lineages sit in this family because they share one domain. DACH1 and DACH2 are the vertebrate homologues of Drosophila dachshund, a retinal determination gene. Their conserved N-terminal region, the Dachshund homology domain or DD1, was solved crystallographically and turned out to adopt a winged-helix fold of the kind used by forkhead proteins, with a surface consistent with DNA binding 12057194; a second conserved region, DD2, mediates protein-protein interaction. The Ski family, SKI, SKIL (SnoN), SKOR1 and SKOR2, carries a homologous Dachshund homology domain; the crystal structure of the human SKI domain confirmed the same fold and identified a SAND-like element within it as the likely DNA-contacting surface 15130471. None of these proteins has a homeodomain, a zinc finger or a leucine zipper. All of them function primarily as corepressors that recruit histone deacetylase and N-CoR complexes to sites chosen by a partner: DACH proteins into SIX-EYA complexes on SIX target promoters, SKI and SnoN onto SMAD2/3 and SMAD4 complexes at TGF-beta response elements, and SKOR1 and SKOR2, originally named Corl1 and Corl2, onto the homeodomain protein LBX1 and related factors in developing neurons 12215533,10485843,15528197.
The motif
There is no single well-founded consensus, and the two lineages point in different directions. For the DACH proteins, the functional claim from cell biology is that DACH1 binds forkhead-type sequences and thereby competes with FOXM1 for its targets 20351289; the winged-helix fold of DD1 makes that structurally plausible. The database models for DACH1 and DACH2, however, are HT-SELEX motifs reading CAATCGTCATAATT and TGACGATTG, which are not forkhead sites, are the weakest-matching members of this family internally, and should be treated as provisional 23332764. For SKOR1 and SKOR2 the models are much stronger and mutually consistent: in vitro selection and methyl-selection give TTAATTAC and CTTTAATTAA, around 15 to 16 bits, reproduced in CIS-BP as TTAATTACAGNT 23332764,28473536. Read as a homeodomain site, TTAATTAC is a TAAT/ATTA arrangement of exactly the type that LHX, ALX and ARX homeodomains bind as cooperative dimers, and that is what the data show: the nearest neighbours of these models outside the family are LHX, ALX and ARX homeodomain motifs, at similarities of 0.88 to 0.93. Whether the Ski/Dachshund winged helix genuinely prefers this sequence, or whether the selection experiments captured the preference of an associated homeodomain partner, is not settled. The family consensus offered here, TTAATTAC, is therefore the observed motif, not a validated statement about which protein reads it.
Where it acts
DACH1 occupies promoters and enhancers in complexes with SIX1 and EYA proteins, where the three together convert a SIX-bound site from repressed to activated via CREB-binding protein 12215533, and at forkhead sites where DACH1 opposes FOXM1-driven proliferative programs. SKI and SnoN are found wherever SMAD complexes are, so their apparent binding footprint is the SMAD-binding element landscape of TGF-beta and BMP signalling 10485843. SKOR1 and SKOR2 are restricted to specific neuronal lineages: SKOR1 in dorsal horn interneurons with LBX1 15528197, SKOR2 in cerebellar Purkinje cell precursors, where it is a selective marker and drives early differentiation 18522874,24491816. The practical consequence is that these factors' genome-wide occupancy tracks the expression and occupancy of their partners, so composite elements in the useful sense are the partner's elements: a SIX site, a SMAD-binding element, or an LBX1-type homeodomain site.
Biology and disease
DACH1 restrains proliferation: it inhibits cyclin D1 and breast tumour growth 16980615, binds SMAD4 to attenuate TGF-beta signalling 14525983, and attenuates forkhead signalling 20351289. It is downregulated in several carcinomas, and transcriptome-wide association work identified DACH1 as a kidney disease risk gene contributing to fibrosis, making it a candidate target in chronic kidney disease 33998598. DACH1 and DACH2 act redundantly in Mullerian duct development. SKI and SnoN are proto-oncoproteins by virtue of repressing TGF-beta growth inhibition 10485843, and SKI mutations cause Shprintzen-Goldberg syndrome 23023332. SKOR2/Corl2 is required for early Purkinje cell differentiation 24491816. The SKOR1 locus, formerly LBXCOR1, was one of the three regions identified in the first genome-wide association study of restless legs syndrome, alongside MEIS1 and BTBD9 17637780.
Reading a hit
Be sceptical in a specific way. A TTAATTAC hit is a good homeodomain dimer site and a poor SKOR identification: if you scan with the SKOR1 or SKOR2 model you will recover LHX, ALX and ARX sites, and in this family's own data those motifs are the nearest neighbours. Conversely, if a ChIP experiment on any of these proteins yields a homeodomain or SMAD or forkhead motif, the most likely explanation is indirect binding through the partner that recruited them, which is the mechanism these proteins are known for. The DACH1 and DACH2 models are the weakest members of the family internally and do not match the forkhead-like specificity that the functional literature describes, so a DACH1 motif hit should not be used as evidence of DACH1 occupancy. In practice, use expression of the partner, not the motif, to decide whether a DACH or Ski family protein is plausibly at a locus.
How this was worked out
SKI is shown to bind Smad2, Smad3 and Smad4 and to repress TGF-beta target genes from the Smad complex, establishing that a protein with no classical DNA-binding domain reaches TGF-beta response elements by riding on a partner.
Still open then: Left open whether SKI contributes any sequence preference of its own at those elements.
The crystal structure of the Dachshund homology domain of DACH shows a winged-helix fold of the type used by forkhead proteins, with a surface compatible with DNA binding, which is the structural basis for every later claim that these corepressors can contact DNA directly.
Still open then: No DNA was present in the structure and no bound sequence was identified.
The crystal structure of the Dachshund homology domain of human SKI confirms the same winged-helix fold in the Ski lineage and identifies a SAND-like element as the probable DNA-contacting surface, unifying the two lineages on structure rather than on sequence preference.
Still open then: Again without DNA, so the predicted contacts remained untested.
- 2005mediumyeast two-hybrid, interaction mapping and expression analysis in mouse neural tube15528197
Corl1, later named SKOR1, is identified as a neuronal lineage-specific corepressor that binds the homeodomain protein Lbx1, which gives a concrete mechanism by which a Ski-family protein can be delivered to homeodomain sites.
- 2010mediumreporter assays, chromatin immunoprecipitation and expression analysis in cancer cell lines20351289
DACH1 is shown to attenuate forkhead signaling and to compete with FOXM1 at forkhead target genes, the main functional evidence that the DACH winged helix engages forkhead-type sequences.
Still open then: Occupancy at forkhead sites could also reflect displacement of FOXM1 through protein contact rather than direct DNA binding by DACH1.
HT-SELEX on recombinant proteins produces the first measured models for this family and contradicts the functional picture: the DACH1 and DACH2 models are not forkhead-like, while SKOR models select a TTAATTAC arrangement that is a classic homeodomain dimer site.
Still open then: Selection on a recombinant protein cannot exclude that a copurified or added homeodomain partner shaped the result.
Methylation-sensitive HT-SELEX reproduces the SKOR1 and SKOR2 preference for TTAATTAA-type sites, making the homeodomain-like SKOR motif the best-supported model in the family while leaving its attribution unresolved.
A transcriptome-wide association study in human kidney identifies DACH1 as a risk gene for chronic kidney disease whose reduced expression contributes to fibrosis, giving the DACH lineage a human genetic phenotype outside cancer.
Still unresolved
- It is unknown whether the Dachshund homology domain binds DNA directly in cells: both crystal structures show folds compatible with DNA binding, but neither contains DNA, and no protein-DNA complex or validated site has been reported for any member.
- Whether TTAATTAC is the SKOR proteins' own preference or the preference of an associated homeodomain is unsettled: the site is a canonical LHX, ALX and ARX dimer element and the nearest neighbours of the SKOR models across all four databases are those homeodomain motifs, but no experiment has separated the two possibilities.
- The functional literature and the database models disagree for DACH1: cell biology reports competition with FOXM1 at forkhead sites, while in vitro selection gives sequences that are not forkhead-like, and no study has tested the selected sequences for DACH1 occupancy or activity.
- Whether SKOR1 is the gene through which the restless legs syndrome association acts is not established: the associated variants are non-coding, and the locus was named for SKOR1 by proximity rather than by a demonstrated effect on the protein.
This family's motifs probably belong to other proteins. DACH and Ski family members have no homeodomain, no zinc finger and no leucine zipper; they are corepressors recruited by SIX/EYA, SMAD or homeodomain partners, and their conserved Dachshund homology domain has only a plausible, not a demonstrated, sequence preference. The strongest models here, SKOR1 and SKOR2 reading TTAATTAC, sit among LHX, ALX and ARX homeodomain motifs at 0.88 to 0.93 similarity, which is the expected signature of partner carry-over. The DACH1 and DACH2 models do not match the forkhead-like specificity the functional literature reports and are the weakest members internally. Do not call occupancy for these factors from a motif hit alone.
References
- Kim SS, Zhang RG, Braunstein SE, et al. Structure of the retinal determination protein Dachshund reveals a DNA binding motif. Structure 2002;10:787-95. PMID 12057194
- Wilson JJ, Malakhova M, Zhang R, et al. Crystal structure of the dachshund homology domain of human SKI. Structure 2004;12:785-92. PMID 15130471
- Ikeda K, Watanabe Y, Ohto H, et al. Molecular interaction and synergistic activation of a promoter by Six, Eya, and Dach proteins mediated through CREB binding protein. Mol Cell Biol 2002;22:6759-66. PMID 12215533
- Zhou J, Wang C, Wang Z, et al. Attenuation of Forkhead signaling by the retinal determination factor DACH1. Proc Natl Acad Sci U S A 2010;107:6864-9. PMID 20351289
- Wu K, Li A, Rao M, et al. DACH1 is a cell fate determination factor that inhibits cyclin D1 and breast tumor growth. Mol Cell Biol 2006;26:7116-29. PMID 16980615
- Luo K, Stroschein SL, Wang W, et al. The Ski oncoprotein interacts with the Smad proteins to repress TGFbeta signaling. Genes Dev 1999;13:2196-206. PMID 10485843
- Mizuhara E, Nakatani T, Minaki Y, et al. Corl1, a novel neuronal lineage-specific transcriptional corepressor for the homeodomain transcription factor Lbx1. J Biol Chem 2005;280:3645-55. PMID 15528197
- Minaki Y, Nakatani T, Mizuhara E, et al. Identification of a novel transcriptional corepressor, Corl2, as a cerebellar Purkinje cell-selective marker. Gene Expr Patterns 2008;8:418-423. PMID 18522874
- Nakatani T, Minaki Y, Kumai M, et al. The c-Ski family member and transcriptional regulator Corl2/Skor2 promotes early differentiation of cerebellar Purkinje cells. Dev Biol 2014;388:68-80. PMID 24491816
- Winkelmann J, Schormair B, Lichtner P, et al. Genome-wide association study of restless legs syndrome identifies common variants in three genomic regions. Nat Genet 2007;39:1000-6. PMID 17637780
- Doke T, Huang S, Qiu C, et al. Transcriptome-wide association analysis identifies DACH1 as a kidney disease risk gene that contributes to fibrosis. J Clin Invest 2021;131. PMID 33998598
- Doyle AJ, Doyle JJ, Bessling SL, et al. Mutations in the TGF-β repressor SKI cause Shprintzen-Goldberg syndrome with aortic aneurysm. Nat Genet 2012;44:1249-54. PMID 23023332
- Wu K, Yang Y, Wang C, et al. DACH1 inhibits transforming growth factor-beta signaling through binding Smad4. J Biol Chem 2003;278:51673-84. PMID 14525983
- 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
7 motifs in this family — HOCOMOCO 5 · CIS-BP 2
| motif | database | consensus | bits |
|---|---|---|---|
| DACH1.H14CORE.0.S.BSB | HOCOMOCO | CAaTcGTCATaATt | 16.9 |
| SKOR1.H14CORE.0.M.CMC | HOCOMOCO | AnCttTAATTAA | 16.4 |
| SKOR2.H14CORE.0.SM.BSMB | HOCOMOCO | CktTAATTAA | 16.1 |
| M05739_2.00 | CIS-BP | TTAATTACagnT | 15.6 |
| SKOR1.H14CORE.1.S.CSC | HOCOMOCO | cgTTAATTAc | 15.1 |
| M05829_2.00 | CIS-BP | tTAATTACcgnt | 13.7 |
| DACH2.H14CORE.0.S.BSB | HOCOMOCO | TgACgATTG | 10 |
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.