TBP / TATA
The TATA box is the oldest known eukaryotic core promoter element and the binding site of TBP, the TATA-box-binding protein. TBP is a saddle-shaped, internally pseudo-symmetric protein that straddles about eight base pairs of AT-rich DNA, reads them entirely from the minor groove, kinks the duplex at both ends and bends it by roughly 80 degrees. It works within TFIID for RNA polymerase II and within SL1 and TFIIIB for polymerases I and III, so a single protein serves all three. Because minor-groove contacts barely distinguish A:T from T:A, the element carries little information: the canonical TATAWAWR is present in only a small minority of human promoters, most of which are instead GC-rich and TATA-less. A TATA match on its own is therefore very weak evidence of anything.
The factors
TBP is a single small protein, about 180 residues in its conserved core, encoded by one gene. Its fold is a pseudo-symmetric saddle: two structurally similar halves, each contributing a curved beta sheet that forms the concave underside, with helices on the convex top surface where general factors and activators dock 8413605,8413604. TBP does not act alone. For RNA polymerase II it is the anchoring subunit of TFIID, a roughly one-megadalton complex of thirteen or more TAF subunits 27007846; for polymerase I it sits in SL1 and for polymerase III in TFIIIB. Two paralogues exist, TBPL1 (TRF2/TLF) and TBPL2 (TRF3), which have the TBP fold but altered DNA-contacting surfaces and do not bind TATA boxes; neither has a motif in the databases used here. TAF1, the largest TFIID subunit, is included in this key because its motif models are attributed to a promoter-binding component of the same machinery, but TAF1 is not a sequence-specific DNA-binding protein in the way TBP is: its promoter contacts involve the initiator and downstream regions 27007846, and it carries histone-acetyltransferase and bromodomain activities.
The motif
The literature consensus is the eight-base TATAWAWR, most often realized as TATAAAAG or TATATAAG, positioned roughly 25-30 bp upstream of the transcription start site in a focused promoter 17123746,28808065. The co-crystal structures of plant (Arabidopsis) and yeast TBP on TATA DNA showed why the information content is low: TBP makes no major-groove base contacts at all 8413605,8413604,7634103. It binds the minor groove, where an A:T and a T:A base pair present nearly the same pattern of hydrogen-bond acceptors, and inserts pairs of phenylalanine side chains between the first two and last two base pairs, kinking the helix at both ends and bending the whole element by about 80 degrees while partially unwinding it 7634103,10617571. What TBP really reads is a short run of deformable, narrow-minor-groove AT-rich DNA, and later work showed that a wide range of AT-rich variants are bound with comparable affinity 10617571. The consequence for motif models is direct: every matrix in this family is a low-information AT-rich model. The representative here, the Vierstra archetype AC0057, is TATAAAAG; JASPAR MA0108.3 is TATAAAA; HOCOMOCO's TBP model is CTATAAAAGG 37971293. There is no half-site structure and no spacing rule, and because TBP is a monomer on a pseudo-symmetric site, orientation is only weakly determined by sequence, which is one reason TATA-box orientation prediction is unreliable. Family cohesion here is only 0.64, and two of the five motifs are flagged as belonging elsewhere, discussed below.
Where it acts
TATA boxes belong to focused, or sharp-initiation, promoters: a single dominant start site, AT-rich surroundings, no CpG island 16645617. These are a minority of mammalian promoters. A genome-scale analysis found that about 76% of human core promoters have no TATA-like element, are GC-rich and are enriched for SP1 sites; about 24% have something TATA-like, but only about 10% of those match the canonical TATAWAWR, which puts canonical TATA boxes at a few percent of human genes 17123746. Broad, dispersed-initiation promoters in CpG islands dominate the mammalian genome, and these use the initiator, downstream elements and TFIID-DNA contacts rather than a TATA box 16645617,28808065. Functionally the split is informative: TATA-less promoters skew toward housekeeping genes, while TATA-containing promoters are enriched for highly regulated, stimulus-responsive and tissue-specific genes, including many that are strongly induced by stress or hormones 17123746. The cryo-EM structure of promoter-bound human TFIID shows TBP on the TATA region while TAF subunits contact the initiator and downstream promoter, so the TATA box is one anchor point in an extended, multi-contact interaction rather than the sole determinant of where transcription starts 27007846.
Biology and disease
TBP is essential in every organism tested; there is no viable null. Its N-terminal region in humans contains a polyglutamine tract, and expansion of that tract to roughly 43 repeats or more causes spinocerebellar ataxia type 17, a rare autosomal dominant neurodegenerative disorder, placing TBP among the polyglutamine diseases 11448935. Mutations in individual TAFs cause developmental syndromes and intellectual disability, and TAF1 variants are associated with X-linked neurodevelopmental disease. TATA-box variants in individual promoters cause disease in their own right, the classic cases being promoter mutations that reduce beta-globin expression in beta-thalassemia and the TATA-box polymorphism in the UGT1A1 promoter that contributes to Gilbert syndrome.
Reading a hit
Treat a TATA match as almost no evidence. TATAWAWR is an eight-base AT-rich pattern with two degenerate positions, expected by chance roughly once every few kilobases on either strand, which is hundreds of thousands of matches in a human genome, the overwhelming majority of them nowhere near a promoter. A match is only interesting if it sits 25-30 bp upstream of an annotated or measured start site in a promoter that otherwise looks focused and AT-rich. It also says nothing about TBP itself, since TBP is present at essentially all active promoters, including TATA-less ones, where it is positioned by TFIID rather than by sequence. This family is also a good place to see what happens when a ChIP-derived matrix is assigned to the wrong protein. The HOCOMOCO model TAF1.H14CORE.0.P.B is AAGATGGCGGC, which is the YY1 motif, not a TBP-like or TAF1-specific sequence; YY1 sites are common at focused promoters, so a TAF1 ChIP experiment recovers them 22955990,29224777. Likewise the CIS-BP TBP entry M08231_2.00 is TATTTGCATA, whose core ATTTGCAT is the POU octamer read in reverse. Neither model should be used as a TATA box.
How this was worked out
A single T-to-G point mutation in the TATAAAA ('Goldberg-Hogness') box of the conalbumin promoter drastically reduces specific transcription in vitro, establishing the TATA box as a functional promoter element.
Still open then: The protein that recognizes the element was unknown.
Co-crystal structures of plant and yeast TBP on TATA DNA show a saddle-shaped protein binding a widened, underwound minor groove, with phenylalanines kinking the DNA at both ends and no major-groove base contacts.
Still open then: How TBP accommodates the variety of natural TATA sequences.
Structure of the TFIIB-TBP-TATA ternary complex: TFIIB recognizes the preformed TBP-DNA complex and its N-terminal domain forms the downstream face, where it could fix the start site.
Still open then: Whether TBP itself chooses an orientation on a nearly symmetric site.
Structures of TBP on ten natural TATA variants show the same DNA deformation in every case; A-to-T and T-to-A transversions leave the minor-groove face unchanged, explaining why many AT-rich sequences are bound.
Genome-wide CAGE mapping of mouse and human start sites splits promoters into a sharp, conserved, TATA-enriched class and a broad, CpG-rich class.
A genome-scale analysis estimates that about 76% of human core promoters lack any TATA-like element and that only about 10% of the TATA-like ones match the canonical TATAWAWR, overturning the textbook picture of the TATA box as a general core-promoter element.
Cryo-EM of human TFIID bound to promoter DNA shows every core promoter element contacted by TFIID subunits, with TAF1 and TAF2 holding the downstream promoter and TFIIA bridging TBP-TATA to lobe B.
Still open then: How TBP is placed at the majority of promoters that lack a TATA box.
Still unresolved
- How much orientation is intrinsic to TBP is unsettled: the TFIIB-TBP-TATA structure fixes the complex's polarity, but TBP alone reads a nearly symmetric minor-groove surface, and the share of orientation set by TBP, by TFIIB and by flanking sequence has not been agreed.
- How TBP is positioned at TATA-less promoters, which are most human promoters, is unresolved: TFIID contacts initiator and downstream elements, but whether TBP still engages a degenerate AT-rich sequence near -30 or is placed purely by TAF contacts is not settled.
- Two of the five matrices under this key are other factors' motifs, YY1 for the ChIP-derived HOCOMOCO TAF1 model and the POU octamer for CIS-BP M08231_2.00; whether these record indirect binding at promoters or misattribution in the source data is not stated by the databases.
The TATA box is the least informative element in common use: eight AT-rich bases read entirely from the minor groove, where A:T and T:A are nearly interchangeable, so the matrix cannot be much better than the physics allows. Expect hundreds of thousands of chance matches per genome and only a few percent of human promoters to carry a real one. A match means nothing without position 25-30 bp upstream of a start site. Two of the five motifs under this key are not TATA boxes: the HOCOMOCO ChIP-derived TAF1 matrix is the YY1 motif, and the CIS-BP TBP entry M08231_2.00 is the POU octamer. Family cohesion is correspondingly low at 0.64.
References
- Kim JL, Nikolov DB, Burley SK. Co-crystal structure of TBP recognizing the minor groove of a TATA element. Nature 1993;365:520-7. PMID 8413605
- Kim Y, Geiger JH, Hahn S, et al. Crystal structure of a yeast TBP/TATA-box complex. Nature 1993;365:512-20. PMID 8413604
- Kim JL, Burley SK. 1.9 A resolution refined structure of TBP recognizing the minor groove of TATAAAAG. Nat Struct Biol 1994;1:638-53. PMID 7634103
- Patikoglou GA, Kim JL, Sun L, et al. TATA element recognition by the TATA box-binding protein has been conserved throughout evolution. Genes Dev 1999;13:3217-30. PMID 10617571
- Nikolov DB, Chen H, Halay ED, et al. Crystal structure of a TFIIB-TBP-TATA-element ternary complex. Nature 1995;377:119-28. PMID 7675079
- Yang C, Bolotin E, Jiang T, et al. Prevalence of the initiator over the TATA box in human and yeast genes and identification of DNA motifs enriched in human TATA-less core promoters. Gene 2007;389:52-65. PMID 17123746
- Carninci P, Sandelin A, Lenhard B, et al. Genome-wide analysis of mammalian promoter architecture and evolution. Nat Genet 2006;38:626-35. PMID 16645617
- Vo Ngoc L, Wang YL, Kassavetis GA, et al. The punctilious RNA polymerase II core promoter. Genes Dev 2017;31:1289-1301. PMID 28808065
- Louder RK, He Y, López-Blanco JR, et al. Structure of promoter-bound TFIID and model of human pre-initiation complex assembly. Nature 2016;531:604-9. PMID 27007846
- Weintraub AS, Li CH, Zamudio AV, et al. YY1 Is a Structural Regulator of Enhancer-Promoter Loops. Cell 2017;171:1573-1588.e28. PMID 29224777
- Wang J, Zhuang J, Iyer S, et al. Sequence features and chromatin structure around the genomic regions bound by 119 human transcription factors. Genome Res 2012;22:1798-812. PMID 22955990
- Vorontsov IE, Eliseeva IA, Zinkevich A, et al. HOCOMOCO in 2024: a rebuild of the curated collection of binding models for human and mouse transcription factors. Nucleic Acids Res 2024;52:D154-D163. PMID 37971293
- Wasylyk B, Derbyshire R, Guy A, et al. Specific in vitro transcription of conalbumin gene is drastically decreased by single-point mutation in T-A-T-A box homology sequence. Proc Natl Acad Sci U S A 1980;77:7024-8. PMID 6938951
- Nakamura K, Jeong SY, Uchihara T, et al. SCA17, a novel autosomal dominant cerebellar ataxia caused by an expanded polyglutamine in TATA-binding protein. Hum Mol Genet 2001;10:1441-8. PMID 11448935
5 motifs in this family — JASPAR 1 · HOCOMOCO 2 · CIS-BP 1 · Vierstra 1
| motif | database | consensus | bits |
|---|---|---|---|
| M08231_2.00looks like HD_POU | CIS-BP | TATTTGCATA | 11.5 |
| TAF1.H14CORE.0.P.BPBlooks like YY1 | HOCOMOCO | AAgATGGCGGC | 11.3 |
| TBP.H14CORE.0.P.BPB | HOCOMOCO | cTATAAAAGg | 11.2 |
| AC0057:TBP/ZNF:TBP | Vierstra | TATAAAag | 9.3 |
| TBP (MA0108.3) | JASPAR | TATAAAa | 9 |
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.