EBF / COE
EBF1-EBF4 (COE1-COE4, also Olf-1 and O/E-1 to O/E-3) bind a 10-base palindrome, TCCCNNGGGA, as homodimers. Despite being filed with the bHLH proteins, their DNA-binding domain is a pseudo-immunoglobulin fold structurally similar to the Rel homology domains of NF-κB and NFAT, contacting DNA through two loops and a unique zinc-coordination motif, with each subunit reaching across both half-sites; the HLH they carry is used for dimerisation, not DNA contact. EBF1 is required at the earliest committed step of B-lymphocyte development and shows pioneer-like activity on closed chromatin; EBF2 establishes brown adipocyte identity; de novo EBF3 variants cause a hypotonia-ataxia-developmental-delay syndrome. The motif is long enough that scans for it are relatively clean.
The factors
EBF1, EBF2, EBF3 and EBF4 — also called COE1-COE4, and in the olfactory literature Olf-1 and O/E-1 to O/E-3 — make up the vertebrate COE (Collier/Olf-1/EBF) family. They were discovered twice in the same year: EBF1 as a regulator of the B-cell-specific *mb-1*/*Cd79a* promoter, and Olf-1 as the factor binding a conserved element in olfactory neuron gene promoters 8491377,8321284. The family is conventionally filed with the bHLH proteins, and the atlas inherits that placement, but structurally it is not one. Crystallography shows that the EBF DNA-binding domain adopts a pseudo-immunoglobulin fold with novel topology that is structurally similar to the Rel homology domains of NF-κB and NFAT, and that it contacts DNA through two loop-based modules plus a unique zinc-coordination motif, the "zinc knuckle", with each monomer reaching across to touch both halves of the palindrome 20876732. Biochemistry had already identified that zinc motif and multiple dimerisation and activation domains 7796816. The proteins also carry an IPT/TIG domain and an atypical helix-loop-helix with its own novel dimerisation motif — an HLH used for dimerisation, not for DNA contact 20592035. The Vierstra archetype covering this family is named AC0500:EBF:Rel_homology_region_(RHR)_factors, which reflects the structure better than the bHLH label does. EBF1 acts in B lymphocytes, olfactory neurons, adipocytes, striatal neurons and osteoblasts; EBF2 in brown adipocytes, olfactory epithelium and cerebellum; EBF3 broadly in the central nervous system; EBF4 is the least characterised. They function as homodimers and can heterodimerise with each other.
The motif
The site is a 10-base palindrome, TCCCNNGGGA: two inverted TCCC half-sites separated by two degenerate positions 8491377. The databases' versions fill those central positions in slightly different ways — HOCOMOCO EBF1 reads TCCCCTGGGA, HOCOMOCO EBF3 and JASPAR EBF1 read TCCCCAGGGA/TCCCCAGGGGA, the CIS-BP EBF1 model reads AKTCCCTWGGGAAT with extended flanks — but they are all the same element, and the representative model is the archetype consensus TCCCTGGGGA. Because the palindrome is read symmetrically by a dimer, with each subunit contacting both half-sites, the two central bases are the tolerant ones and the TCCC/GGGA blocks are essentially invariant 20876732.
Two practical consequences follow. First, this is an information-rich motif by the standards of a transcription factor site: eight effectively fixed positions rather than the four or five of an E-box. An exact TCCCNNGGGA is expected around 47,000 times in a 3.1-Gb genome, and because the pattern is its own reverse complement that count is not doubled — roughly one site per 65 kb, against roughly one per 4 kb for a six-mer. Scans for this motif are correspondingly cleaner. Second, the motif resembles nothing else in the atlas: the best match of any member to a motif outside the family is 0.93 and typically much lower, while within-family similarities run 0.91-1.00 (cohesion 0.91). The four paralogues are completely indistinguishable from sequence.
One provenance note: the CIS-BP EBF2 model M03721_2.00 is inherited from a Drosophila melanogaster protein (Collier/Knot) at 91% DNA-binding-domain identity. Because COE proteins are deeply conserved and the fly motif (AATTCCCTAGGGAATT) matches the human models closely, this transfer is benign, but it is a transfer.
Where it acts
EBF1 in B cells binds thousands of sites, mostly at enhancers of the B-lineage programme — Cd79a, Pax5, Cd19, Igll1, Vpreb, Blnk, Foxo1 — and acts by activation, by repression, and by a transcription-independent poising of chromatin at genes that only become active later 20451411. Its C-terminal domain confers pioneer-like activity, letting EBF1 engage closed and nucleosomal chromatin and open it for subsequent factors, which is unusual and consistent with the extended DNA contact area seen in the structure 26982363. In B cells EBF1 sites co-occur with E2A/TCF3, PAX5, FOXO1 and RUNX1 motifs. In brown fat, EBF2 occupies and activates the PPARG/PRDM16 thermogenic enhancer programme 23499423. In olfactory sensory neurons the same element sits in promoters of olfactory marker and signalling genes 8321284.
Biology and disease
*Ebf1*-null mice arrest B-cell development before the pro-B stage, with no immunoglobulin heavy-chain DJ rearrangement — EBF1 is required at the earliest committed step of the B lineage 7542362. EBF2 is required to establish and maintain brown adipocyte identity, and forced EBF2 expression imposes a brown-fat programme 23499423. In human genetics, de novo variants in EBF3, largely missense changes in the DNA-binding domain, cause a neurodevelopmental syndrome with hypotonia, ataxia, developmental delay and expressive language impairment, reported independently by two groups in the same issue; the mutations reduce DNA binding and transactivation, consistent with a dominant-negative or haploinsufficient mechanism 28017372,28017370.
Reading a hit
Because the site is long, palindromic and unlike any other family's, a TCCCNNGGGA hit is unusually trustworthy at the family level: if you find a good match, EBF-family binding is a reasonable hypothesis. What it will not tell you is which EBF, since all four share the motif, and in most tissues only one or two are expressed, so expression data resolve the question that sequence cannot. Two cautions. All three HOCOMOCO models here are ChIP-derived (evidence letter P, grade B), so the in vitro support for the recognition sequence comes from the original biochemistry and the crystal structure rather than from SELEX models in the current collection. And the family label "bHLH" is misleading: EBF has an HLH, but it is a dimerisation module, and an E-box search will neither find EBF sites nor be found by them.
How this was worked out
The same factor is cloned twice in one year from unrelated tissues: EBF as the regulator of the B-cell-specific mb-1/Cd79a promoter, and Olf-1 as the protein binding a conserved element in olfactory neuron gene promoters, both reading a palindromic TCCC...GGGA site.
Still open then: How a protein with no recognisable DNA-binding fold of a known class reads that palindrome.
Protein dissection identifies a novel zinc-coordination motif required for DNA binding plus separate dimerization and activation domains, and the Ebf-null mouse shows B-cell development arrested before the pro-B stage with no immunoglobulin heavy-chain DJ rearrangement.
Two structural studies show the EBF DNA-binding domain is a pseudo-immunoglobulin fold structurally similar to the Rel homology domains of NF-kB and NFAT, binding DNA through two loop modules and a zinc knuckle with each monomer contacting both halves of the palindrome, and that the atypical HLH serves dimerization rather than DNA contact — the family's bHLH label is structurally wrong.
Genome-wide analysis in B cells shows Ebf1 acts by activation, by repression, and by a transcription-independent poising of chromatin at genes expressed only at later stages; the poising works in T cells but not in fibroblasts, so the pioneer-like behaviour is context dependent.
Still open then: What limits the poising activity to a haematopoietic chromatin context.
EBF2 is shown to determine and maintain brown adipocyte identity by occupying and activating the PPARG/PRDM16 thermogenic enhancer programme, extending the family well beyond lymphocytes and olfactory neurons.
The pioneering activity is mapped to the EBF1 C-terminal domain, which is required for de novo gain of chromatin accessibility at a specific target set independently of the protein's activation and repression functions.
Still open then: Whether the C-terminal domain engages nucleosomal DNA directly or recruits remodelling activity.
- 2017mediumtrio exome sequencing with binding and reporter assays; structural modelling28017372 28017370
Two groups report in the same issue that de novo EBF3 variants, mostly missense changes at conserved DNA-binding residues including the zinc motif, cause a neurodevelopmental syndrome with hypotonia, ataxia and expressive language impairment, and reduce DNA binding and transactivation.
Still open then: Whether the variants act by haploinsufficiency or dominant-negatively through mixed dimers was not settled.
Still unresolved
- No comparative in vitro measurement separates EBF1, EBF2, EBF3 and EBF4: their models are near-identical, every HOCOMOCO model here is ChIP-derived, and whether the paralogues differ at all in sequence preference or only in expression domain is untested.
- The two central positions of TCCCNNGGGA are filled differently by different databases (TCCCCTGGGA in HOCOMOCO EBF1, TCCCCAGGGA in EBF3 and JASPAR EBF1), and whether those bases carry a real, measurable preference or are genuinely unconstrained has not been resolved.
- How EBF1 opens closed chromatin is unknown: the activity maps to the C-terminal domain and works in haematopoietic and T-cell but not fibroblast chromatin, leaving open whether the domain engages nucleosomal DNA directly or depends on a lineage-restricted cofactor.
- Whether pathogenic EBF3 variants act by haploinsufficiency or as dominant negatives through heterodimers with wild-type EBF3 and other COE proteins is unresolved, and the two mechanisms predict different consequences for occupancy at the palindromic site.
The site is long, palindromic and unlike any other family's, so a good match is reasonable evidence for EBF-family binding — unusual in this atlas. It is no evidence at all about which EBF: all four paralogues share one motif (cohesion 0.91) and only expression can separate them. All three HOCOMOCO models here are ChIP-derived (evidence letter P, grade B), so the sequence preference rests on the original biochemistry and the crystal structure rather than on in vitro models in the current collection. One CIS-BP EBF2 model is inherited from a Drosophila protein at 91% domain identity, which is benign but is a transfer. The "bHLH" class label is structurally wrong; do not expect E-box searches to find these sites.
References
- Hagman J, Belanger C, Travis A, et al. Cloning and functional characterization of early B-cell factor, a regulator of lymphocyte-specific gene expression. Genes Dev 1993;7:760-73. PMID 8491377
- Wang MM, Reed RR. Molecular cloning of the olfactory neuronal transcription factor Olf-1 by genetic selection in yeast. Nature 1993;364:121-6. PMID 8321284
- Hagman J, Gutch MJ, Lin H, et al. EBF contains a novel zinc coordination motif and multiple dimerization and transcriptional activation domains. EMBO J 1995;14:2907-16. PMID 7796816
- Lin H, Grosschedl R. Failure of B-cell differentiation in mice lacking the transcription factor EBF. Nature 1995;376:263-7. PMID 7542362
- Treiber N, Treiber T, Zocher G, et al. Structure of an Ebf1:DNA complex reveals unusual DNA recognition and structural homology with Rel proteins. Genes Dev 2010;24:2270-5. PMID 20876732
- Siponen MI, Wisniewska M, Lehtiö L, et al. Structural determination of functional domains in early B-cell factor (EBF) family of transcription factors reveals similarities to Rel DNA-binding proteins and a novel dimerization motif. J Biol Chem 2010;285:25875-9. PMID 20592035
- Treiber T, Mandel EM, Pott S, et al. Early B cell factor 1 regulates B cell gene networks by activation, repression, and transcription- independent poising of chromatin. Immunity 2010;32:714-25. PMID 20451411
- Rajakumari S, Wu J, Ishibashi J, et al. EBF2 determines and maintains brown adipocyte identity. Cell Metab 2013;17:562-74. PMID 23499423
- Boller S, Ramamoorthy S, Akbas D, et al. Pioneering Activity of the C-Terminal Domain of EBF1 Shapes the Chromatin Landscape for B Cell Programming. Immunity 2016;44:527-541. PMID 26982363
- Chao HT, Davids M, Burke E, et al. A Syndromic Neurodevelopmental Disorder Caused by De Novo Variants in EBF3. Am J Hum Genet 2017;100:128-137. PMID 28017372
- Sleven H, Welsh SJ, Yu J, et al. De Novo Mutations in EBF3 Cause a Neurodevelopmental Syndrome. Am J Hum Genet 2017;100:138-150. PMID 28017370
10 motifs in this family — JASPAR 4 · HOCOMOCO 3 · CIS-BP 2 · Vierstra 1
| motif | database | consensus | bits |
|---|---|---|---|
| M03721_2.00Drosophila melanogaster 0.91 | CIS-BP | aATTCCCtaGGGAATt | 22 |
| M02963_2.00 | CIS-BP | akTCCCtwGGGAaT | 16 |
| COE1.H14CORE.0.P.BPB | HOCOMOCO | TCCCCtGGGA | 14.7 |
| COE2.H14CORE.0.P.BPB | HOCOMOCO | TCCCyaGGGa | 13.1 |
| EBF1 (MA0154.5) | JASPAR | TCCCCAGGGGA | 12.9 |
| COE3.H14CORE.0.P.BPB | HOCOMOCO | TCCCCAGGGA | 12.9 |
| Ebf4 (MA2122.1) | JASPAR | TCCCCAGGGGA | 12.7 |
| Ebf2 (MA1604.2) | JASPAR | CCCAAGGGA | 12.3 |
| AC0500:EBF:['Rel_homology_region_(RHR)_factors'] | Vierstra | TCCCTgGGGA | 12.1 |
| EBF3 (MA1637.2) | JASPAR | CCCAAGGGA | 11.9 |
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.