THAP-type zinc finger
THAP-type zinc fingers are C2CH modules whose closest known relative is the site-specific DNA-binding domain of the Drosophila P-element transposase, making this a family of domesticated transposase DNA-binding domains. The atlas files six motifs for two proteins, THAP1 and THAP12. THAP1 recognizes a bipartite site - a GCCATCTT-like half and a GGGCA half separated by a short variable spacer - which the trimmed JASPAR model reduces to GCAGGGCA. Crystallography of the fly THAP domain explains the bipartite pattern: a beta-sheet inserts into the major groove while a basic C-terminal loop reads the adjacent minor groove. THAP1 is the DYT6 dystonia gene. THAP11 is deliberately not filed here; it shares the ZNF143 SBS element and is covered by the ZNF143 entry.
The factors
The THAP domain was defined in 2003 as a roughly 90-residue module whose sequence resembles the site-specific DNA-binding domain of the Drosophila melanogaster P-element transposase, carrying a putative metal-coordinating signature of the form CX(2-4)CX(35-53)CX(2)H 12575992,15863623. A survey across model animals catalogued about 100 distinct THAP proteins, including the human proapoptotic factors THAP1 and DAP4/THAP0, the transcriptional repressor THAP7, and the Caenorhabditis elegans chromatin protein HIM-17 and cell-cycle regulators LIN-36 and LIN-15B 15863623. The domain is therefore one of the clearest cases of molecular domestication: a transposase recognition module retained by the host after the mobile element that carried it was lost, a route to new transcription factors seen repeatedly in animal genomes 19132291,28844698.
The atlas family itself is small. Six motifs cover two named proteins, THAP1 and THAP12, plus two unnamed Vierstra archetypes labelled only THAP. THAP1 is a ubiquitously expressed nuclear protein encoded at 8p11.21. THAP12 is the protein also known as DAP4 or PRKRIR. THAP11, better known as Ronin, is excluded by design: it binds the same element as ZNF143 and is described in the ZNF143 entry, so a THAP11-labelled motif should be read there, not here.
The motif
In vitro binding-site selection with the THAP domain of human THAP1 yielded an 11-nucleotide consensus, and single-nucleotide substitutions within it abolished binding; chelating zinc with 1,10-o-phenanthroline also abolished binding, confirming the domain is a zinc-dependent member of the zinc-finger superfamily 15863623. The structural basis came from a crystal structure of the fly P-element transposase THAP domain bound to a natural ten-base-pair site 20010837. Unlike a C2H2 finger, the THAP domain docks a conserved beta-sheet into the major groove and a basic C-terminal loop into the adjacent minor groove, and sequence analysis of natural and selected sites suggested a shared bipartite TXXGGGX(A/T) consensus for the fly domain and for human THAP1 and THAP9 20010837.
The atlas models match that description at two different trimmings. The long models keep both halves: HOCOMOCO THAP1.H14CORE.0.P.B reads gccGCCATcTTbgntvcGGGCa over 22 positions, and CIS-BP M08020 gives CGCCATcTTggnTrhGGGCAg at 21.5 bits. The short models keep only the downstream half: JASPAR MA0597.3 and Vierstra archetype AC0141 are both GCAGGGCA, and a second Vierstra archetype AC0478 is the still shorter tnhGGGCA. Family cohesion is moderate at 0.58, driven entirely by this difference in trimming rather than by disagreement about the site. The remaining motif, CIS-BP M02538 for THAP12, carries 3.8 bits of information and is not usable. Label agreement across databases is very low, between 0 and 0.1, which is what an eight-base-pair GC-containing model will always produce: its nearest neighbours in other databases belong to other families.
Where it acts
THAP1 occupancy is concentrated at promoters and CpG islands rather than distal enhancers. The best-characterised direct target is TOR1A, the DYT1 dystonia gene, whose promoter THAP1 binds and represses 20976771. Because the site includes a GGGCA half that overlaps GC-box-like sequence, matches cluster in the same promoter-proximal, CpG-rich compartment occupied by SP/KLF factors, and work in rats and patient-derived neurons indicates THAP1 acts on that same regulatory space 35015830. Note the separation of labour within the wider domain family: the hyperconserved promoter element bound by Ronin/THAP11 together with HCF-1 at genes controlling embryonic stem cell growth 20581084 is the ZNF143 SBS, not the THAP1 site.
Biology and disease
Heterozygous THAP1 mutations cause DYT6, an autosomal dominant primary torsion dystonia with prominent cranio-cervical and laryngeal involvement 19182804. Many of the pathogenic substitutions fall inside the THAP domain and reduce DNA binding directly 19182804, which is why this family is one of the few where a motif model maps onto a molecular disease mechanism. THAP1 represses TOR1A, linking the DYT6 and DYT1 loci in one pathway 20976771. DYT6-mutant THAP1 behaves as a cell-type-dependent regulator of the SP1 family 35015830, and mouse Thap1 loss produces transcriptional deficits that converge on pathways shared with other dystonic syndromes 30590536. Elsewhere in the domain family, Ronin/THAP11 is essential for mouse embryogenesis and for embryonic stem cell pluripotency 18585351.
Reading a hit
Use the long model if you can. GCAGGGCA is eight base pairs of partly GC-rich sequence and will match tens of thousands of times in a 3.1 Gb genome, most of them meaningless; the 22-base-pair bipartite HOCOMOCO model is far more specific because it requires both halves and the spacer. A hit identifies a THAP-domain site, not a protein: the bipartite consensus is shared by THAP1, THAP9 and the fly P-element transposase itself 20010837. Condition on a promoter or CpG island context and on THAP1 expression. If your hit instead resembles the ACTACA-containing SBS element, you are looking at THAP11 and should read the ZNF143 entry. Ignore the THAP12 model entirely.
How this was worked out
Sequence analysis defined the THAP domain as an N-terminal C2CH module resembling the site-specific DNA-binding domain of the Drosophila P-element transposase, proposing it as a DNA-binding domain shared between cellular proteins and a mobile element.
Still open then: Whether the cellular THAP domain binds DNA at all, and with what specificity, had not been tested.
- 2005mediumIn vitro binding-site selection, site-directed mutagenesis, zinc chelation, sequence survey15863623
The THAP domain of human THAP1 was shown to be a zinc-dependent, sequence-specific DNA-binding domain: in vitro binding-site selection gave an 11-nucleotide consensus in which single substitutions abolished binding, and about 100 THAP proteins were catalogued across model animals.
Still open then: How a single C2CH module of about 90 residues achieves base-specific recognition.
A THAP1 mutation segregating in three Amish-Mennonite families, and a second mutation in a German family, identified THAP1 as the DYT6 primary torsion dystonia gene; the missense change impaired DNA binding.
Still open then: Which THAP1 target genes mediate the dystonia.
The crystal structure of the Drosophila P-element transposase THAP domain on a natural 10 bp site showed a beta-sheet docked in the major groove and a basic C-terminal loop in the adjacent minor groove, and natural and selected sites gave a shared bipartite TXXGGGX(A/T) consensus for the fly domain, THAP1 and THAP9.
THAP1 was shown to bind the core promoter of TOR1A, the DYT1 dystonia gene, and repress it, with DYT6-mutant THAP1 repressing less, placing two dystonia genes in one pathway.
Still open then: Whether TOR1A repression accounts for DYT6 dystonia in vivo.
- 2017mediumConditional mouse knockout, oligodendrocyte progenitor assays, chromatin occupancy measurements28697333
Conditional deletion of Thap1 in the mouse CNS delayed myelination through a cell-autonomous requirement in the oligodendrocyte lineage, and loss of THAP1 reduced the DNA occupancy of YY1 at oligodendrocyte maturation genes, identifying YY1 as a working partner of THAP1.
Still open then: Whether THAP1 and YY1 bind adjacent sites in a composite element or separate elements within the same promoters.
Germline Thap1 deletion in mice was embryonic lethal, heterozygotes had no discernible phenotype because the intact allele was upregulated, and nestin-conditional deletion caused locomotor deficits with expression changes converging on pathways shared with other dystonias.
- 2022mediumEpigenomic and transcriptomic profiling across human and rat model systems, behavior and electrophysiology35015830
Across patient cortex, patient iPSC-derived midbrain dopaminergic neurons, a heterozygous knockout rat and knockout cell lines, THAP1 directly targeted only a minority of the genes dysregulated by its mutation; most changes ran through SP1 and SP4 in a cell-type-dependent way, revising the picture of THAP1 as a direct regulator of a few promoters such as TOR1A.
Still open then: How much of the disease depends on THAP1 binding its own site rather than on the SP1-family cascade.
Still unresolved
- The upstream GCCATCTT block of the long ChIP-derived THAP1 models (HOCOMOCO and CIS-BP) matches the YY1 core consensus GCCATnTT, and THAP1 is known to work with YY1; the in vitro selection and structural data define only the GGGCA-containing half, so whether the 22-position model is one THAP1 site or a THAP1-YY1 composite has not been tested.
- How much of THAP1's transcriptional effect requires its own DNA site is disputed: TOR1A repression was explained by direct promoter binding, whereas a multi-system study found that THAP1 directly binds only a minority of the genes its mutation dysregulates and acts mostly through SP1 and SP4.
- Why heterozygous THAP1 loss causes dystonia in humans but no discernible phenotype in mice is unresolved; upregulation of the intact mouse allele is one explanation, and it is not known whether human THAP1 is buffered in the same way.
- Whether THAP12 (DAP4) binds DNA sequence-specifically at all is unknown: its only model carries 3.8 bits, and no in vitro selection or structure exists for it.
The family carries only six motifs for two proteins, and one of them, CIS-BP M02538 for THAP12, has 3.8 bits and no usable information. The HOCOMOCO THAP1 model is ChIP-only (evidence code P, grade B), so within the atlas there is no in vitro support for it, although independent binding-site selection and a crystal structure of the homologous fly domain support the same bipartite site in the literature. Short GCAGGGCA-style matches are common and largely uninformative on their own. Nothing in the motif separates THAP1 from THAP9 or from the ancestral P-element transposase domain. THAP11 is intentionally absent from this family: its site is the ZNF143 SBS and belongs to that entry.
References
- Roussigne M, Kossida S, Lavigne AC, et al. The THAP domain: a novel protein motif with similarity to the DNA-binding domain of P element transposase. Trends Biochem Sci 2003;28:66-9. PMID 12575992
- Clouaire T, Roussigne M, Ecochard V, et al. The THAP domain of THAP1 is a large C2CH module with zinc-dependent sequence-specific DNA-binding activity. Proc Natl Acad Sci U S A 2005;102:6907-12. PMID 15863623
- Sabogal A, Lyubimov AY, Corn JE, et al. THAP proteins target specific DNA sites through bipartite recognition of adjacent major and minor grooves. Nat Struct Mol Biol 2010;17:117-23. PMID 20010837
- Fuchs T, Gavarini S, Saunders-Pullman R, et al. Mutations in the THAP1 gene are responsible for DYT6 primary torsion dystonia. Nat Genet 2009;41:286-8. PMID 19182804
- Kaiser FJ, Osmanoric A, Rakovic A, et al. The dystonia gene DYT1 is repressed by the transcription factor THAP1 (DYT6). Ann Neurol 2010;68:554-9. PMID 20976771
- Cheng F, Zheng W, Barbuti PA, et al. DYT6 mutated THAP1 is a cell type dependent regulator of the SP1 family. Brain 2022;145:3968-3984. PMID 35015830
- Frederick NM, Shah PV, Didonna A, et al. Loss of the dystonia gene Thap1 leads to transcriptional deficits that converge on common pathogenic pathways in dystonic syndromes. Hum Mol Genet 2019;28:1343-1356. PMID 30590536
- Dejosez M, Krumenacker JS, Zitur LJ, et al. Ronin is essential for embryogenesis and the pluripotency of mouse embryonic stem cells. Cell 2008;133:1162-74. PMID 18585351
- Dejosez M, Levine SS, Frampton GM, et al. Ronin/Hcf-1 binds to a hyperconserved enhancer element and regulates genes involved in the growth of embryonic stem cells. Genes Dev 2010;24:1479-84. PMID 20581084
- Sinzelle L, Izsvák Z, Ivics Z. Molecular domestication of transposable elements: from detrimental parasites to useful host genes. Cell Mol Life Sci 2009;66:1073-93. PMID 19132291
- Jangam D, Feschotte C, Betrán E. Transposable Element Domestication As an Adaptation to Evolutionary Conflicts. Trends Genet 2017;33:817-831. PMID 28844698
- Yellajoshyula D, Liang CC, Pappas SS, et al. The DYT6 Dystonia Protein THAP1 Regulates Myelination within the Oligodendrocyte Lineage. Dev Cell 2017;42:52-67.e4. PMID 28697333
6 motifs in this family — JASPAR 1 · HOCOMOCO 1 · CIS-BP 2 · Vierstra 2
| motif | database | consensus | bits |
|---|---|---|---|
| M08020_2.00 | CIS-BP | CGCCATcTTggnTrhGGGCAg | 21.5 |
| THAP1.H14CORE.0.P.BPB | HOCOMOCO | gccGCCATcTTbgntvcGGGCa | 11.8 |
| THAP1 (MA0597.3) | JASPAR | GCAGGGCA | 10.7 |
| AC0141:THAP:['C2CH_THAP-type_zinc_finger_factors'] | Vierstra | GCAGGGCA | 10.7 |
| AC0478:THAP:THAP_finger | Vierstra | tnhGGGCA | 6.3 |
| M02538_2.00 | CIS-BP | tcgnnbGg | 3.8 |
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.