ETV2 / ETV6-type
ETV2 (ER71), ETV6 (TEL) and ETV7 (TEL2) are ETS-domain proteins that read the same GGA(A/T) core as the rest of the family. ETV2 is a transient activator required for the specification of blood and endothelium; ETV6 and ETV7 are repressors, and ETV6 spreads repression by polymerizing through its PNT/SAM domain. They are grouped here because their motif models cluster together, not because they share a distinct binding preference: ETV2's selected sequence is the ordinary class I ETS site and ETV6/ETV7 differ from it only marginally. What does recur among these models is arrangement - several are two GGAA cores in tandem or head-to-head. ETV6 is one of the most frequently rearranged genes in leukemia, most famously as ETV6-RUNX1 in childhood B-lineage acute lymphoblastic leukemia.
The factors
Three genes sit in this family. ETV2 (also ER71 or ETSRP) is a transient transcriptional activator expressed in a narrow window in nascent mesoderm and hemogenic endothelium and essentially silent in adult tissues. ETV6 (TEL) and ETV7 (TEL2) are transcriptional repressors; ETV6 is broadly expressed 9250681, ETV7 more narrowly and largely in hematopoietic cells. All three carry one ETS domain of about 85 residues, the family's winged helix-turn-helix fold, and bind DNA as monomers 21548782. ETV6 and ETV7 additionally carry a PNT domain (a SAM domain) that self-associates head-to-tail into an open-ended polymer 12193272,21548782, so one DNA-bound ETV6 molecule can nucleate a long repressive assembly; the polymer interface and its regulation have been mapped biophysically 33450226. ETV6 also has an autoinhibitory element that damps its DNA binding 21548782. It is worth stating plainly that this family is not a biochemical class. Systematic profiling of every human and mouse ETS protein put ETV6 and ETV7 in a small subclass of their own (class IIb) 20517297 and did not group ETV2 with them; ETV2's preference is the ordinary class I preference.
The motif
The consensus is ACCGGAAGT - the same nine-mer as ETS class I, read the same way, by two invariant arginines in recognition helix H3 that hydrogen-bond the two guanines of GGA, with wing and turn lysines contacting the flanks 21548782. The in vitro models differ only at the margins: JASPAR's HT-SELEX model for ETV2 (MA0762.2) is ACCGGAAAT, with A rather than G immediately 3' of the core, while the ETV6 (MA0645.2) and ETV7 (MA1708.2) HT-SELEX models are plain CCGGAAGT and CGGAAGTG. There is no sequence that belongs to this family and not to ETS class I. What is distinctive in part of the collection is arrangement rather than composition: three models here are doubled sites - CIS-BP M02990 (CCGGAAGCGGAAGT), CIS-BP M04712 (CGGAAGTACTTCCG) and the ETV7 in vitro model (CCCGGAAGTACTTC) - that is, two ETS cores in tandem or head-to-head. That is consistent with ETV6's ability to polymerize 21548782,33450226 and with the reported preference of these factors for repeated GGAA 21548782, but doubled and repeated GGAA matrices also occur in ETS class I, where the EWS-FLI1 fusion's GGAA microsatellite matrices live 18626011, so a tandem arrangement does not on its own assign a motif here.
Where it acts
ETV2 binds enhancers of endothelial and hematopoietic genes during a short developmental window, acting downstream of BMP, Notch and Wnt signals 18462699,21548782. ETV6 and ETV7 occupy ETS sites at promoters and enhancers and repress them, recruiting corepressors rather than activating 21548782. The functional point this family illustrates is competition: because ETV6, ETV7, ERF and ETV3 read the same sequence as ETS1, ERG, FLI1 and GABPA, the output of a given ETS site depends on which family member happens to occupy it, not on the sequence. That is also why sites bound by these repressors are not distinguishable in a scan from sites bound by activators. Notably, the genome-wide footprint archetypes derived from DNase data contain no archetype for this family 32728250: every archetype whose name includes ETV clustered instead with ELF, ELK or ETS archetypes that are assigned elsewhere in this atlas.
Biology and disease
Etv2-null mouse embryos fail to make blood and endothelium 18462699, placing ETV2 at the top of the hematoendothelial program; the gene is now widely used to force endothelial fate in reprogramming experiments. Mice lacking Tel/Etv6 die around mid-gestation with a yolk sac angiogenic defect and intra-embryonic apoptosis 9250681, and ETV6 is required for adult bone marrow hematopoiesis 9694803. ETV6 is among the most frequently rearranged genes in human leukemia: the t(12;21) fusion ETV6-RUNX1 (TEL-AML1) is the most common recurrent fusion in childhood B-lineage acute lymphoblastic leukemia 7761424, and ETV6 has many other partners including kinases 7761424,12193272. Germline heterozygous ETV6 mutations cause autosomal dominant thrombocytopenia with predisposition to hematologic malignancy 25581430.
Reading a hit
This is the family in this atlas where the label deserves the least weight. The motifs assigned here are not separable from ETS class I, and the data say so directly: for all 16 models in this family, most of the nearest neighbours across the four databases are labelled with a different family, usually ETS class I, and for 14 of them the single best match anywhere is a motif from another family rather than one from this family - including the CIS-BP model that this atlas picks as the family's representative. One JASPAR entry, the heterodimer model ETV2::FIGLA (MA1941.2), is flagged outright as belonging with class I. Practical consequences: read a hit as "an ETS site"; use expression to decide the factor, remembering that ETV2 is effectively absent outside early mesoderm and engineered endothelial systems, so an ETV2 call in an adult tissue dataset is far more likely a generic ETS site; and do not try to tell ETV6 from ERF or ETV3, all repressors reading the same sequence. Four of the seven JASPAR models here are CAP-SELEX heterodimers (ETV2::HOXB13, ETV2::FIGLA, ETV2::DRGX, ETV2::FOXI1) in which much of the matrix is a HOX, bHLH or FOX site rather than an ETS site 26550823.
How this was worked out
Cloning of the t(12;21) breakpoint in childhood pre-B acute lymphoblastic leukemia shows the amino-terminal HLH domain of TEL (later recognized as the PNT domain) fused to the DNA-binding and transactivation domains of AML1 (RUNX1); TEL had already been found fused to the PDGFRB kinase in chronic myelomonocytic leukemia.
Still open then: What normal TEL does, and which DNA sites it binds, was unknown.
Tel-null embryos die at E10.5-11.5 with a yolk sac angiogenic defect and intra-embryonic apoptosis while yolk sac hematopoiesis is spared; chimeras built from Tel-null ES cells (1998) show TEL is needed to establish hematopoiesis of all lineages in the bone marrow, though not in yolk sac or fetal liver.
The structure of the native TEL SAM (PNT) domain interface confirms the head-to-tail helical polymer first seen in a mutant, and shows flat interacting surfaces without obvious pockets.
Still open then: How polymerization contributes to DNA binding and repression at genomic sites was not addressed.
ER71 (ETV2) acts downstream of BMP, Notch and Wnt in specifying FLK1+ mesoderm: enforced expression induces FLK1+ mesoderm and blood and endothelium from ES cells, and Er71-deficient embryos die early with severe blood and vessel defects resembling Flk1-null mice.
- 2010highHigh-throughput microwell binding assay, protein-binding microarrays, site-directed mutagenesis20517297
Systematic in vitro profiling splits ETV6 and ETV7 off from the ELF proteins as class IIb, marked by a strong preference for A at the fourth core position that tracks a tyrosine-to-histidine substitution in the ETS domain.
Still open then: ChIP-seq was done for one member of the other classes but not for ETV6 or ETV7, so whether the class IIb preference matters in cells was untested.
Consecutive-affinity SELEX of transcription factor pairs recovers 618 heterodimeric motifs, most of them composite sites markedly different from either partner's motif; this is the assay behind the ETS-HOX, ETS-bHLH and ETS-FOX heterodimer models that make up four of the seven JASPAR entries here.
Still open then: How many of these DNA-dependent pairs form in cells was not measured.
Germline ETV6 missense variants segregating with dominant thrombocytopenia and diverse hematologic cancers in three kindreds abolish DNA binding, mislocalize the protein and reduce repression in a dominant-negative manner.
- 2021highNMR spectroscopy, X-ray crystallography, molecular dynamics, surface plasmon resonance33450226
The two ETV6 PNT self-association interfaces are mapped in detail: monomeric variants keep a stable helical-bundle fold on association, and alanine scanning identifies hydrophobic and salt-bridge hot spots.
Still unresolved
- Whether this is a motif family at all is doubtful: its models cluster with ETS class I in every database, the published classification puts ETV6 and ETV7 in their own class IIb, and ETV2's in vitro models are indistinguishable from class I, so the grouping reflects a clustering outcome rather than a shared sequence preference.
- Whether ETV6 and ETV7 polymers select clustered ETS sites in the genome is unresolved: PNT-domain self-association supports cooperative binding to neighboring sites in vitro, and three models here are doubled sites, but genome-wide evidence that the repressors prefer site clusters is not reported in the sources cited here.
Treat the family label as provisional. Its separation from ETS class I is largely a clustering outcome: all 16 models here have most of their cross-database nearest neighbours labelled with another family, usually class I, 14 match a motif outside the family better than any inside it, and one JASPAR entry (ETV2::FIGLA) is flagged as belonging with class I. The published classification also does not support the grouping - ETV6 and ETV7 form their own small subclass while ETV2 behaves like a class I factor. No genomic-footprint archetype corresponds to this family. Four of seven JASPAR models are CAP-SELEX heterodimers whose matrices are partly a HOX, bHLH or FOX site. A hit therefore means "an ETS site"; which of ETV2, ETV6, ETV7, ERF or ETV3 is bound has to come from expression and ChIP data.
References
- Wei GH, Badis G, Berger MF, et al. Genome-wide analysis of ETS-family DNA-binding in vitro and in vivo. EMBO J 2010;29:2147-60. PMID 20517297
- Hollenhorst PC, McIntosh LP, Graves BJ. Genomic and biochemical insights into the specificity of ETS transcription factors. Annu Rev Biochem 2011;80:437-71. PMID 21548782
- Lee D, Park C, Lee H, et al. ER71 acts downstream of BMP, Notch, and Wnt signaling in blood and vessel progenitor specification. Cell Stem Cell 2008;2:497-507. PMID 18462699
- Wang LC, Kuo F, Fujiwara Y, et al. Yolk sac angiogenic defect and intra-embryonic apoptosis in mice lacking the Ets-related factor TEL. EMBO J 1997;16:4374-83. PMID 9250681
- Golub TR, Barker GF, Bohlander SK, et al. Fusion of the TEL gene on 12p13 to the AML1 gene on 21q22 in acute lymphoblastic leukemia. Proc Natl Acad Sci U S A 1995;92:4917-21. PMID 7761424
- Zhang MY, Churpek JE, Keel SB, et al. Germline ETV6 mutations in familial thrombocytopenia and hematologic malignancy. Nat Genet 2015;47:180-5. PMID 25581430
- Tran HH, Kim CA, Faham S, et al. Native interface of the SAM domain polymer of TEL. BMC Struct Biol 2002;2:5. PMID 12193272
- Gerak CAN, Cho SY, Kolesnikov M, et al. Biophysical characterization of the ETV6 PNT domain polymerization interfaces. J Biol Chem 2021;296:100284. PMID 33450226
- Jolma A, Yin Y, Nitta KR, et al. DNA-dependent formation of transcription factor pairs alters their binding specificity. Nature 2015;527:384-8. PMID 26550823
- Vierstra J, Lazar J, Sandstrom R, et al. Global reference mapping of human transcription factor footprints. Nature 2020;583:729-736. PMID 32728250
- Gangwal K, Sankar S, Hollenhorst PC, et al. Microsatellites as EWS/FLI response elements in Ewing's sarcoma. Proc Natl Acad Sci U S A 2008;105:10149-54. PMID 18626011
- Wang LC, Swat W, Fujiwara Y, et al. The TEL/ETV6 gene is required specifically for hematopoiesis in the bone marrow. Genes Dev 1998;12:2392-402. PMID 9694803
16 motifs in this family — JASPAR 7 · HOCOMOCO 6 · CIS-BP 3
| motif | database | consensus | bits |
|---|---|---|---|
| M02990_2.00 | CIS-BP | CCGGAAgCGGAAGT | 20.4 |
| M04712_2.00 | CIS-BP | cGGAAGTACTTCCg | 19.3 |
| ETV2::HOXB13 (MA1943.2)dimer | JASPAR | CCGGAAacATAAA | 15.7 |
| ETV2::FIGLA (MA1941.2)dimerlooks like ETS_CLASS1 | JASPAR | cCGGAAaCAGGTG | 15.3 |
| ETV7.H14CORE.0.SM.BSMB | HOCOMOCO | CCCGGAAGTacttC | 14.9 |
| ETV2 (MA0762.2) | JASPAR | ACCGGAAAT | 14.1 |
| ETV2::DRGX (MA1940.2)dimer | JASPAR | CGGAAgtAATTA | 14 |
| ETV2.H14CORE.1.PM.APMA | HOCOMOCO | ACAGGAAGT | 13.9 |
| ETV2::FOXI1 (MA1942.2)dimer | JASPAR | AAACAGGAAG | 12.3 |
| ETV6 (MA0645.2) | JASPAR | gCGGAAGT | 12.2 |
| ETV6.H14CORE.1.P.BPB | HOCOMOCO | aagAGGAAGTg | 12.2 |
| ETV6.H14CORE.0.PS.APSA | HOCOMOCO | CCGGAAGTG | 11.6 |
| M04720_2.00 | CIS-BP | ACCGGAAgT | 11.1 |
| ETV7 (MA1708.2) | JASPAR | CGGAAGTg | 10.9 |
| ETV2.H14CORE.0.S.BSB | HOCOMOCO | CCGGAAa | 10 |
| ETV7.H14CORE.1.P.CPC | HOCOMOCO | AGGAAg | 8.7 |
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.