GATA
GATA factors are zinc-finger transcription factors that bind the six-base core WGATAA (usually AGATAA or TGATAA) through a single C4-type zinc finger plus a basic tail. Humans have six canonical members, GATA1-6, in two expression groups: GATA1/2/3 in blood, lymphocytes and mammary epithelium, GATA4/5/6 in heart, gut and gonad. The atypical protein TRPS1 carries one GATA-type finger and binds the same core as a repressor. Sites are short and very common, and are often read in pairs or next to an E-box occupied by TAL1. GATA binding sites underlie globin gene switching, cardiac septation and T-helper-2 commitment, and the factors are mutated in congenital heart disease, HDR syndrome and breast cancer.
The factors
Six canonical human GATA proteins (GATA1-6) share a pair of C4-type zinc fingers. Only the C-terminal finger, with its adjacent basic tail, is required for sequence-specific binding; the N-terminal finger stabilizes binding to certain palindromic and double sites and is the docking surface for the cofactors ZFPM1/ZFPM2 (FOG1/FOG2) 14985365,23142663. GATA proteins bind DNA as monomers, not as obligate dimers, although two molecules can occupy adjacent sites on the same or opposite faces of the helix. Expression splits the family in two: GATA1 (erythroid, megakaryocytic, mast, eosinophil), GATA2 (hematopoietic stem and progenitor cells, endothelium) and GATA3 (T lymphocytes, mammary luminal epithelium, kidney, parathyroid, inner ear) versus GATA4, GATA5 and GATA6 (heart, endoderm-derived gut, lung, liver, gonad, pancreas). TRPS1 is an outlier: it has a single GATA-type finger, binds the same core, and represses rather than activates 11285235. ZGLP1 is annotated as GATA-like but its measured specificity in these databases (ATGATCGAT) is not the GATA core, and it clusters here only weakly.
The motif
The literature consensus is WGATAA, sometimes written (A/T)GATA(A/G) to include the less common 3' G 8321208,8321207. The NMR structure of the chicken GATA-1 finger-plus-tail bound to DNA showed that the core module contacts GAT in the major groove while the basic C-terminal tail wraps into the minor groove to read the 3' AA and the 5' W 8332909; this is why the invariant information sits on GATA and falls off sharply outside it. Only about 6 bp carry real information, so a GATA PWM is one of the least specific vertebrate models in routine use. Many database entries are written on the reverse strand as TTATCT or CTTATCT, which is the same site; the data consensus here, AGATAAg, is simply the forward-strand form of the same six-base core. Extended versions exist. Tandem sites in the arrangement GATAA-N(n)-TTATC are common in real regulatory elements and give the double-GATA archetype seen in footprint clustering. The GATA-E-box composite element places WGATAA 8-10 bp from a CAGNTG E-box and is bound by the TAL1/E47/LMO2/LDB1/GATA1 pentameric complex 9214632; JASPAR models this as a GATA1::TAL1 dimer motif. Systematic in vitro surveys find near-identical preferences for closely related DNA-binding domains 23332764,25215497, and the current in vitro and ChIP-derived models for GATA1-6 are effectively the same AGATAA matrix, so the primary site does not distinguish the paralogues. The one reported difference comes from random-site selection with chicken GATA1, GATA2 and GATA3, which recovered an alternative site, AGATCTTA, bound well by GATA2 and GATA3 but poorly by GATA1 8321208.
Where it acts
GATA sites are enhancer elements far more than promoter elements, and they are not associated with CpG islands. Because WGATAA occurs roughly once every few kilobases by chance, the human genome contains on the order of a million matches and only a small fraction are occupied in any cell type. Occupancy is set by chromatin: GATA1 and GATA2 mark erythroid enhancers, GATA3 marks luminal breast and Th2 enhancers, GATA4 and GATA6 mark cardiac and gut enhancers, and the same sequence is invisible in the wrong lineage. GATA2 and GATA1 exchange at many of the same sites during erythroid maturation, the so-called GATA switch. Frequent neighbors are the E-box of TAL1, KLF1 CACCC boxes, RUNX1, SP1, FOXA1 and estrogen receptor (in breast), and NKX2-5, TBX5 and MEF2 (in heart). GATA3 and GATA4 both behave as chromatin-opening factors at a subset of their sites.
Biology and disease
Gata1-null mouse embryonic stem cells arrest at the proerythroblast stage and die of apoptosis 1987478,7568185, and a megakaryocyte-selective knockdown causes thrombocytopenia. In humans, GATA1 exon 2 mutations producing the short GATA1s isoform cause transient abnormal myelopoiesis and Down syndrome-associated acute megakaryoblastic leukemia 12172547; other GATA1 mutations cause X-linked dyserythropoietic anemia and thrombocytopenia. GATA2 haploinsufficiency causes MonoMAC syndrome and familial MDS/AML 21670465. GATA3 is required for and sufficient to drive Th2 cytokine expression in CD4 T cells 9160750; heterozygous mutations cause HDR syndrome (hypoparathyroidism, deafness, renal dysplasia) 14985365, and GATA3 is one of the most frequently mutated genes in luminal breast cancer. GATA4 mutations cause septal defects and interact genetically with TBX5 12845333. TRPS1 loss causes tricho-rhino-phalangeal syndrome 11285235,12446778; TRPS1 immunostaining is now used as a breast-carcinoma marker.
Reading a hit
A WGATAA match tells you the family and nothing about the member. Treat any claim that a site is "the GATA1 site" rather than "a GATA site" as an inference from expression or ChIP data, not from sequence. Because the informative core is six bases, isolated matches in a random background are expected and carry almost no evidence by themselves; a hit becomes interesting when it is conserved, sits in open chromatin of a tissue that expresses a GATA factor, or is doubled or paired with an E-box. Two specific confusions are worth knowing. AT-rich homeodomain sites (TAAT-containing) and the reverse-complement TTATCT can score on loose GATA models, and vice versa. And the CIS-BP human TRPS1 motif is not measured on human TRPS1: it is transferred from a Drosophila protein at 63% DNA-binding-domain identity 25215497, so the model's fine structure should not be read as TRPS1-specific even though TRPS1 genuinely binds GATA cores. HOCOMOCO GATA models graded A are backed by both ChIP-seq and in vitro SELEX (evidence codes PSM), which is stronger support than ChIP alone.
How this was worked out
Disrupting the X-linked Gata1 gene in embryonic stem cells blocks their contribution to mature red cells in chimeric mice, showing that the factor binding erythroid GATA elements is required and that other GATA proteins cannot substitute.
Still open then: Which GATA elements carry the requirement, and why GATA2 present in the same precursors does not compensate.
Random-site selection with GATA1, GATA2 and GATA3 confirms the WGATAR consensus and AGATAA as the high-affinity core, but also recovers high-affinity non-consensus sites and partly distinct preferences, notably an AGATCTTA site bound well by GATA2 and GATA3 and poorly by GATA1.
Still open then: Whether the paralogue differences seen in vitro are used at genomic sites.
The NMR structure of the chicken GATA-1 C-terminal finger plus basic tail on DNA shows the zinc module in the major groove and the tail wrapping into the minor groove, with about 8 bp in contact; the tail is an essential determinant of specific binding.
Still open then: The N-terminal finger was absent, so its role in site choice remained unexplained.
In erythroid cells LMO2 bridges GATA1, TAL1, E2A and LDB1 into one DNA-binding complex that recognizes an E-box followed about 9 bp downstream by a GATA site, defining the GATA-E-box composite element.
Crystal structures of the full GATA3 DNA-binding domain show both zinc fingers wrapping a palindromic GATA site, which raises affinity and kinetic stability, or the two fingers binding GATA sites on separate DNA molecules, a possible basis for DNA looping.
Still open then: Whether the DNA-bridging mode occurs at endogenous loci.
High-throughput SELEX and protein-binding-microarray surveys extend measured models across the family and show that closely related DNA-binding domains have near-identical preferences; the GATA1-6 matrices now in JASPAR, HOCOMOCO and CIS-BP converge on AGATAA, and the CIS-BP TRPS1 model is inferred from a Drosophila protein rather than measured.
Still open then: The 1993 report of paralogue-specific sites has not been reconciled with these identical matrices.
Still unresolved
- Whether GATA paralogues differ in intrinsic sequence preference is unsettled: 1993 random-site selection found an alternative AGATCTTA site bound well by GATA2 and GATA3 but poorly by GATA1, whereas the HT-SELEX, PBM and ChIP-derived matrices for GATA1-6 in the current collections are effectively identical.
- How the N-terminal finger contributes to site choice at genomic double-GATA sites is unresolved: GATA3 structures show it can either wrap a palindromic site together with the C-terminal finger or bind a second DNA molecule, and which mode operates in cells is not known.
- ZGLP1 is annotated as GATA-like, yet its in vitro-supported HOCOMOCO model (ATGATCGAT, grade A) does not contain the WGATAA core, so whether it should be read as a divergent GATA-family specificity or a separate one is not settled by the data.
The core is only 6 bp, so chance matches are abundant and a single hit is weak evidence. All six canonical paralogues share one matrix; sequence cannot assign a member, and neither can ChIP-seq in a cell expressing several of them. Reverse-strand renderings (CTTATCT) and AT-rich homeodomain sites are the two commonest misreadings. The CIS-BP TRPS1 model is inferred from a Drosophila protein, and the HOCOMOCO ZGLP1 model does not match the GATA core despite ZGLP1's GATA-like domain annotation. Vierstra archetypes labelled GATA/TAL describe a composite element, not a claim that GATA and TAL1 motifs are separable.
References
- Omichinski JG, Clore GM, Schaad O, et al. NMR structure of a specific DNA complex of Zn-containing DNA binding domain of GATA-1. Science 1993;261:438-46. PMID 8332909
- Chen Y, Bates DL, Dey R, et al. DNA binding by GATA transcription factor suggests mechanisms of DNA looping and long-range gene regulation. Cell Rep 2012;2:1197-206. PMID 23142663
- Pevny L, Simon MC, Robertson E, et al. Erythroid differentiation in chimaeric mice blocked by a targeted mutation in the gene for transcription factor GATA-1. Nature 1991;349:257-60. PMID 1987478
- Wadman IA, Osada H, Grütz GG, et al. The LIM-only protein Lmo2 is a bridging molecule assembling an erythroid, DNA-binding complex which includes the TAL1, E47, GATA-1 and Ldb1/NLI proteins. EMBO J 1997;16:3145-57. PMID 9214632
- Anderson KP, Crable SC, Lingrel JB. Multiple proteins binding to a GATA-E box-GATA motif regulate the erythroid Krüppel-like factor (EKLF) gene. J Biol Chem 1998;273:14347-54. PMID 9603943
- Zheng W, Flavell RA. The transcription factor GATA-3 is necessary and sufficient for Th2 cytokine gene expression in CD4 T cells. Cell 1997;89:587-96. PMID 9160750
- Garg V, Kathiriya IS, Barnes R, et al. GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. Nature 2003;424:443-7. PMID 12845333
- Nesbit MA, Bowl MR, Harding B, et al. Characterization of GATA3 mutations in the hypoparathyroidism, deafness, and renal dysplasia (HDR) syndrome. J Biol Chem 2004;279:22624-34. PMID 14985365
- Malik TH, Von Stechow D, Bronson RT, et al. Deletion of the GATA domain of TRPS1 causes an absence of facial hair and provides new insights into the bone disorder in inherited tricho-rhino-phalangeal syndromes. Mol Cell Biol 2002;22:8592-600. PMID 12446778
- Malik TH, Shoichet SA, Latham P, et al. Transcriptional repression and developmental functions of the atypical vertebrate GATA protein TRPS1. EMBO J 2001;20:1715-25. PMID 11285235
- Jolma A, Yan J, Whitington T, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327-39. PMID 23332764
- Weirauch MT, Yang A, Albu M, et al. Determination and inference of eukaryotic transcription factor sequence specificity. Cell 2014;158:1431-1443. PMID 25215497
- Ko LJ, Engel JD. DNA-binding specificities of the GATA transcription factor family. Mol Cell Biol 1993;13:4011-22. PMID 8321208
- Merika M, Orkin SH. DNA-binding specificity of GATA family transcription factors. Mol Cell Biol 1993;13:3999-4010. PMID 8321207
- Weiss MJ, Orkin SH. Transcription factor GATA-1 permits survival and maturation of erythroid precursors by preventing apoptosis. Proc Natl Acad Sci U S A 1995;92:9623-7. PMID 7568185
- Wechsler J, Greene M, McDevitt MA, et al. Acquired mutations in GATA1 in the megakaryoblastic leukemia of Down syndrome. Nat Genet 2002;32:148-52. PMID 12172547
- Hsu AP, Sampaio EP, Khan J, et al. Mutations in GATA2 are associated with the autosomal dominant and sporadic monocytopenia and mycobacterial infection (MonoMAC) syndrome. Blood 2011;118:2653-5. PMID 21670465
31 motifs in this family — JASPAR 8 · HOCOMOCO 11 · CIS-BP 7 · Vierstra 5
| motif | database | consensus | bits |
|---|---|---|---|
| M07965_2.00 | CIS-BP | aGATAagaATCT | 15.1 |
| GATA3.H14CORE.1.SM.BSMB | HOCOMOCO | AGATAAGAT | 14.8 |
| GATA1::TAL1 (MA0140.3)dimer | JASPAR | TTATCtgnnnnnnvCAG | 14.2 |
| AC0008:GATA:GATA | Vierstra | GATAAtgncaTTATC | 13.8 |
| ZGLP1.H14CORE.0.SGB.ASGBA | HOCOMOCO | ATGATCGaT | 13.2 |
| AC0247:GATA:GATA | Vierstra | AGATnntATC | 13.1 |
| GATA1.H14CORE.0.P.BPB | HOCOMOCO | cAGATAAG | 12 |
| M08124_2.00 | CIS-BP | AGATAAGA | 11.9 |
| GATA6.H14CORE.0.PSM.APSMA | HOCOMOCO | AGATAAG | 11.5 |
| GATA5 (MA0766.3) | JASPAR | aGATAAg | 11.3 |
| M03761_2.00Drosophila melanogaster 0.63 | CIS-BP | cGATAAg | 11.2 |
| M07963_2.00 | CIS-BP | cAGATAAG | 11.2 |
| M04875_2.00 | CIS-BP | GATAAc | 11 |
| GATA4 (MA0482.3) | JASPAR | CCTTATCT | 10.9 |
| M07966_2.00 | CIS-BP | AGATAA | 10.8 |
| GATA6 (MA1104.3) | JASPAR | TCTTATCT | 10.7 |
| GATA1.H14CORE.1.PSM.APSMA | HOCOMOCO | AGATAAG | 10.7 |
| GATA2 (MA0036.4) | JASPAR | CTTATCT | 10.6 |
| GATA5.H14CORE.0.SM.BSMB | HOCOMOCO | GATAA | 10.5 |
| GATA2.H14CORE.0.PSM.APSMA | HOCOMOCO | AGATAAg | 10.4 |
| Gata3 (MA0037.5) | JASPAR | TCTTATCT | 10.3 |
| GATA2.H14CORE.1.P.BPB | HOCOMOCO | AGATAAG | 10.3 |
| TRPS1 (MA1970.2) | JASPAR | TCTTATCT | 10.2 |
| GATA4.H14CORE.0.PSM.APSMA | HOCOMOCO | AGATAAg | 10 |
| GATA1 (MA0035.5) | JASPAR | CTAATCT | 9.8 |
| AC0246:GATA/TAL:GATA | Vierstra | AGATAAg | 9.5 |
| GATA3.H14CORE.0.PS.APSA | HOCOMOCO | GATAA | 9.4 |
| TRPS1.H14CORE.0.P.BPB | HOCOMOCO | AGATAa | 8.6 |
| M03065_2.00 | CIS-BP | GATAA | 8.3 |
| AC0245:GATA:GATA | Vierstra | GATAA | 8.1 |
| AC0453:GATA:['Other_C4_zinc_finger-type_factors'] | Vierstra | GAT | 5.4 |
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.