NF-Y (CBF)
NF-Y is a heterotrimer of NFYA, NFYB and NFYC that binds the CCAAT box, one of the most frequent elements in eukaryotic promoters. NFYB and NFYC carry histone-fold domains and pair like H2A/H2B, gripping the DNA backbone and bending it; NFYA supplies the only sequence-specific contacts, inserting an alpha helix into the minor groove to read CCAAT. All three subunits are required for binding, so the three gene symbols describe one molecular entity and one motif. NF-Y is ubiquitous, sits at a fixed distance upstream of transcription start sites, and helps open chromatin for cell-type-specific factors. NF-YA is overexpressed and alternatively spliced in several cancers.
The factors
NF-Y, also called CBF and historically CP1, is an obligate heterotrimer. NFYB and NFYC each contain a histone-fold domain (HFD) and first dimerize with each other; the NFYB/NFYC dimer then recruits NFYA, and only the complete trimer binds DNA 9153318. Yeast HAP2, HAP3 and HAP5 are the orthologues of NFYA, NFYB and NFYC respectively 22050321, and the conservation is high enough that subunits are interchangeable across large evolutionary distances. NFYA exists as two major splice isoforms, long and short, differing in the glutamine-rich activation region; the short isoform is enriched in embryonic stem cells. NFYB and NFYC are encoded by single genes with essentially constitutive expression, so NF-Y activity is regulated largely through NFYA levels and isoform choice 22050321. Because all three subunits are needed, a motif labelled NFYA, NFYB or NFYC in these databases refers to the same complex and the same site; the three-way split in this family's data (nine models across three symbols) reflects which subunit was tagged in an experiment, not three specificities.
The motif
The consensus is the 5-base CCAAT box, generally quoted with flanking preferences as CCAATCA or, on the reverse strand, TGATTGG 3476205,22050321; databases here report both renderings plus CCAATCAG and CATTGGCC. The crystal structure of NF-Y on a 25-bp CCAAT oligonucleotide explained the division of labor. The NFYB/NFYC histone-fold dimer binds the sugar-phosphate backbone over about 25 bp, mimicking the H2A/H2B-DNA interaction in a nucleosome and bending the DNA, but makes no base-specific contacts 23332751,12401788. NFYA contributes the specificity: a short alpha helix inserts deeply into the minor groove and reads the CCAAT bases directly 23332751. This architecture explains two features of the motif. First, information is concentrated on five bases with modest preferences a few positions to each side, giving models of 10-14 bits, and the flanking positions matter partly because they set the DNA's bendability rather than because they are read. Second, NF-Y has nucleosome-like, non-sequence-specific affinity in addition to its specific contacts, which is why it behaves like a chromatin protein as well as a transcription factor 23332751. Not every CCAAT box is an NF-Y site: the element was long used as a generic label, and several unrelated families were grouped under "CCAAT-binding protein", including the C/EBPs and NFI/CTF 3476205,3398920. Those are different proteins with different folds.
Where it acts
NF-Y is the most positionally constrained factor in this batch. Its CCAAT boxes sit overwhelmingly in core and proximal promoters, classically between 60 and 100 bp upstream of the major transcription start site, and frequently in bidirectional promoters 15647281. ChIP-on-chip and later genome-wide work showed that a substantial minority of NF-Y sites nonetheless lie outside core promoters 15647281. NF-Y is bound to a large fraction of cell cycle gene promoters (cyclin A, CDC25C, CDK1) often together with E2F factors 12771133, to the Y box of MHC class II promoters where it binds cooperatively with the RFX complex about 15-20 bp away 10779326, to the CCAAT boxes of heat shock gene promoters, and to the globin promoters. It recruits histone acetyltransferases including GCN5/PCAF and p300, contacts TBP 12771133,9153318, and imposes H3K4 trimethylation through Ash2L. In embryonic stem cells NF-Y is required to maintain identity and opens chromatin for lineage-determining factors 25132174, which makes it one of the clearest examples of a ubiquitous factor acting as a chromatin accessibility provider rather than a specificity determinant.
Biology and disease
NF-Y is essential: loss of NFYA or of the histone-fold subunits is lethal in cells and in mice, and NF-Y regulates cell cycle progression, cholesterol and lipid metabolism genes, and the endoplasmic-reticulum stress and heat shock programs. Because the CCAAT box is so widespread, the downstream consequences are broad rather than tissue-specific. In disease, the clearest links are quantitative: NF-YA is overexpressed, and its isoform ratio shifted, in breast cancer 31506469 and in lung adenocarcinoma, and NF-YA is required for proliferation of epithelial ovarian cancer cells through EZH2 upregulation 23360797. NF-YA splice isoform status has been reported as a prognostic variable in low grade endometrial cancer. There is no well-established Mendelian disease caused by NFYA, NFYB or NFYC mutation, which is consistent with the complex being required in every cell.
Reading a hit
A CCAAT match is 5 bp of information and occurs roughly a million times in the human genome, so a hit in isolation is not evidence. What makes a CCAAT hit credible is position: an NF-Y site 60-100 bp upstream of an annotated transcription start site, ideally with a preferred flank (CCAATCA or its reverse complement TGATTGG), is a reasonable prediction, while the same 5-mer in an intron or a distal enhancer is usually noise. The paralogue problem is unusual here and easy: there are no paralogues to separate, because NFYA, NFYB and NFYC are three subunits of one complex. Treat all three labels as synonyms for NF-Y. The main real confusion is with other families historically called CCAAT-binding. C/EBP factors bind ATTGC-containing palindromes, not CCAAT, and NFI/CTF, whose old name is literally "CCAAT box transcription factor", binds TTGGC(N5)GCCAA 3398920; its GCCAA half-site contains a CCAAT-like string and will cross-score on loose CCAAT models. If a scan returns both NF-Y and NFI matches at the same position, they are competing explanations. Finally, note that every HOCOMOCO model in this family is ChIP-only (evidence code P, grade B) and the label agreement scores are correspondingly moderate; the structural and biochemical evidence for NFYA reading CCAAT is strong 23332751, but these particular matrices are peak-derived.
How this was worked out
- 1987mediumGel retardation, methylation interference, saturation mutagenesis, cross-competition3476205
NF-Y, the factor binding the Y box of MHC class II promoters, is shown to recognize the reversed CCAAT box within it and to be distinct from the other CCAAT-binding activities then known, CBP and CTF/NF-I, plus a further thymidine kinase promoter factor.
Still open then: Which CCAAT boxes are NF-Y sites rather than sites for the other CCAAT-binding proteins.
- 1997mediumImmunopurification, sedimentation, in vitro binding with deletion and point mutants9153318
All three subunits are required for DNA binding, and NF-YB and NF-YC contact TBP through short conserved stretches adjacent to their histone-fold motifs.
The NF-YB/NF-YC structure shows a histone-fold pair closely related to H2A/H2B and NC2, and modeling suggests nonspecific DNA contact by the dimer within the full NF-Y/CCAAT complex.
Still open then: How NF-YA reads the CCAAT bases remained unseen.
On cell-cycle promoters NF-Y binding is timed and lost on repression, with PCAF/GCN5 and p300 recruited in a hierarchy, contrary to the picture of NF-Y as a constitutive factor that presets promoters.
ChIP-on-chip on CpG-island arrays places 40-50% of NF-Y sites outside core promoters, in introns or distant 5' and 3' positions, and finds 40% of targets in divergent, convergent or tandem promoter units.
Still open then: Whether the non-promoter sites are functional.
The structure of the NF-Y trimer on a 25-bp CCAAT oligonucleotide shows the histone-fold dimer binding the backbone like H2A/H2B in a nucleosome, while an NF-YA helix inserts into the minor groove and makes the CCAAT-specific contacts.
In embryonic stem cells and neurons NF-Y binds cell-type-invariant promoters and also cell-type-specific enhancers together with master factors, where it promotes a permissive chromatin conformation; it is required to maintain ES cell identity.
Still unresolved
- How promoter-restricted functional NF-Y binding is remains unsettled: the classical CCAAT box sits 60-100 bp upstream of the start site, but ChIP-on-chip placed 40-50% of NF-Y sites outside core promoters and genome-wide data in ES cells and neurons find NF-Y at cell-type-specific enhancers, so distal CCAAT matches may be real more often than the promoter-position rule assumes.
- Whether NF-Y is a constitutive promoter-presetting factor or a regulated one is not settled: its histone-fold subunits are ubiquitous and it occupies promoters broadly, yet on cell-cycle promoters its binding is timed and lost on repression.
- Why the flanks matter is unclear: the trimer structure shows base-specific contacts concentrated on CCAAT, yet every model here carries flank preferences (CCAATCA, CCAATCAG), which could reflect DNA shape and bendability or contacts outside the crystallized site.
CCAAT is 5 bp and occurs about a million times in the genome, so only position makes a hit credible: NF-Y sites cluster 60-100 bp upstream of transcription start sites. NFYA, NFYB and NFYC are subunits of one obligate trimer, so the three labels denote one motif, not three specificities, and only NFYA makes base contacts. The recurring confusion is with other families once called CCAAT-binding: NFI/CTF's GCCAA half-site and C/EBP sites will cross-score on loose CCAAT models and are competing explanations, not corroboration. All HOCOMOCO models here are ChIP-derived (grade B), though the structure independently confirms CCAAT reading.
References
- Dorn A, Bollekens J, Staub A, et al. A multiplicity of CCAAT box-binding proteins. Cell 1987;50:863-72. PMID 3476205
- Nardini M, Gnesutta N, Donati G, et al. Sequence-specific transcription factor NF-Y displays histone-like DNA binding and H2B-like ubiquitination. Cell 2013;152:132-43. PMID 23332751
- Romier C, Cocchiarella F, Mantovani R, et al. The NF-YB/NF-YC structure gives insight into DNA binding and transcription regulation by CCAAT factor NF-Y. J Biol Chem 2003;278:1336-45. PMID 12401788
- Dolfini D, Gatta R, Mantovani R. NF-Y and the transcriptional activation of CCAAT promoters. Crit Rev Biochem Mol Biol 2012;47:29-49. PMID 22050321
- Testa A, Donati G, Yan P, et al. Chromatin immunoprecipitation (ChIP) on chip experiments uncover a widespread distribution of NF-Y binding CCAAT sites outside of core promoters. J Biol Chem 2005;280:13606-15. PMID 15647281
- Oldfield AJ, Yang P, Conway AE, et al. Histone-fold domain protein NF-Y promotes chromatin accessibility for cell type-specific master transcription factors. Mol Cell 2014;55:708-22. PMID 25132174
- Caretti G, Salsi V, Vecchi C, et al. Dynamic recruitment of NF-Y and histone acetyltransferases on cell-cycle promoters. J Biol Chem 2003;278:30435-40. PMID 12771133
- Bellorini M, Lee DK, Dantonel JC, et al. CCAAT binding NF-Y-TBP interactions: NF-YB and NF-YC require short domains adjacent to their histone fold motifs for association with TBP basic residues. Nucleic Acids Res 1997;25:2174-81. PMID 9153318
- Villard J, Peretti M, Masternak K, et al. A functionally essential domain of RFX5 mediates activation of major histocompatibility complex class II promoters by promoting cooperative binding between RFX and NF-Y. Mol Cell Biol 2000;20:3364-76. PMID 10779326
- Santoro C, Mermod N, Andrews PC, et al. A family of human CCAAT-box-binding proteins active in transcription and DNA replication: cloning and expression of multiple cDNAs. Nature 1988;334:218-24. PMID 3398920
- Dolfini D, Andrioletti V, Mantovani R. Overexpression and alternative splicing of NF-YA in breast cancer. Sci Rep 2019;9:12955. PMID 31506469
- Garipov A, Li H, Bitler BG, et al. NF-YA underlies EZH2 upregulation and is essential for proliferation of human epithelial ovarian cancer cells. Mol Cancer Res 2013;11:360-9. PMID 23360797
9 motifs in this family — JASPAR 3 · HOCOMOCO 3 · CIS-BP 2 · Vierstra 1
| motif | database | consensus | bits |
|---|---|---|---|
| M08178_2.00 | CIS-BP | gaCCAATCAG | 14.5 |
| NFYC.H14CORE.0.P.BPB | HOCOMOCO | agCCAATCAGaa | 13.8 |
| NFYB.H14CORE.0.P.BPB | HOCOMOCO | agCCAATcAgag | 13.6 |
| NFYA.H14CORE.0.P.BPB | HOCOMOCO | agCCAATcAgAa | 13 |
| M08104_2.00 | CIS-BP | cTGATTGG | 12.1 |
| NFYB (MA0502.3) | JASPAR | CATTGGCC | 11.9 |
| NFYA (MA0060.4) | JASPAR | CCAATCAg | 11.6 |
| AC0240:NFYA/NFYB:CBF/NF-Y | Vierstra | gCCAATCAg | 11.4 |
| NFYC (MA1644.2) | JASPAR | CCAATCA | 10 |
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.