POU (octamer)
The octamer ATGCAAAT is the site read by the 16 human POU-domain factors, including OCT1/POU2F1, OCT2/POU2F2, PIT1/POU1F1, the BRN2/BRN1 group (POU3F1-4), the BRN3 group (POU4F1-3) and OCT4/POU5F1. What makes the family unusual is the binding unit: a POU-specific domain and a homeodomain joined by a flexible linker, so the two halves of the site can be re-arranged. The same proteins therefore also form dimers on the PORE and MORE elements and sit next to SOX partners in the SOX-OCT composite elements that control pluripotency genes. Hits matter because octamer elements drive immunoglobulin and snRNA promoters and because OCT4 occupancy of SOX-OCT sites is a defining feature of pluripotent cells. The motif is short and AT-rich, so it is common.
The factors
Sixteen human POU genes fall into six classes: POU1F1 (PIT1), POU2F1-3 (OCT1, OCT2, OCT11/SKN-1a), POU3F1-4 (OCT6, BRN2, BRN1, BRN4), POU4F1-3 (BRN3A/B/C), POU5F1 (OCT4) together with the retrogene-derived POU5F1B locus, and POU6F1/2. All of them use one composite DNA-binding unit, described from sequence in 1988 and named after PIT1, OCT1, OCT2 and C. elegans UNC-86 3215510,2905684: an N-terminal POU-specific domain (POU-S), a four-helix bundle whose helix-turn-helix resembles the bacteriophage lambda and 434 repressors, tethered by a flexible 15-27 residue linker to a C-terminal POU-homeodomain (POU-HD), a conventional 60-residue three-helix homeodomain 8156594,7622033. Expression is class-restricted. POU2F1 is essentially ubiquitous; POU2F2 is lymphoid 2905684; POU5F1 is confined to pluripotent and germ cells; POU3F and POU4F members mark defined neural, retinal and inner-ear lineages; POU1F1 is restricted to the anterior pituitary.
The motif
The core is the 8-bp octamer ATGCAAAT, first noted as a conserved element upstream of immunoglobulin genes 6425835. The 1994 crystal structure of the OCT1 POU domain on an octamer site showed why the site is bipartite: POU-S grips the 5' ATGCA half and POU-HD the 3' AAAT half, the two domains bind opposite faces of the duplex, and the linker is disordered and contacts no bases 8156594. The 3' half is a TAAT-type homeodomain site read in the reverse orientation; POU-S contributes most of the base-specific contacts, which is why the ATGC block is the least tolerant part of the site. Because the linker tethers rather than spaces, the half-sites can be rearranged: two POU domains dimerize on the PORE (ATTTGAAATGCAAAT) or on the MORE (ATGCATATGCAT) using two non-overlapping surface patches, and only the PORE arrangement recruits the coactivator OBF-1/POU2AF1 11136971,11583619. Pooled across JASPAR, HOCOMOCO, CIS-BP and the Vierstra archetypes, this family's consensus is tTATGCaAATt - the octamer plus its usual flanking T - but cohesion is only 0.68, because many HT-SELEX models for POU3F, POU4F and POU6F members report homeodomain-half arrangements (TGCATAAT, ATTAATTATGCA, TAATTA, TCATTA) instead of the full octamer 23332764. Those are configuration variants of one binding unit, not separate specificities 7622033.
Where it acts
Octamer sites are classically promoter-proximal, in immunoglobulin heavy- and light-chain promoters, the histone H2B promoter and snRNA promoters 6425835,7622033. The best-characterized POU enhancer elements are composite: an OCT4 octamer immediately adjacent to a SOX HMG site. OCT4 and SOX2 dimerize in different relative orientations on the FGF4 and UTF1 enhancers, and it is the arrangement of the two boxes in the DNA, not a switch in the proteins, that sets the geometry 12923055. The Vierstra archetype set recovers POU5F/NANOG and POU2F/POU3F clusters, reflecting how routinely these sites are found with partner motifs. Because the octamer is only 8 bp and AT-rich, an exact match is expected on the order of 10^5 times in 3.1 Gb, so the great majority of matches are unoccupied in any one cell type; occupancy tracks which POU class is expressed. Methyl-SELEX places several POU factors among those whose affinity is changed by CpG methylation in flanking positions, although the octamer core itself has no CpG 28473536.
Biology and disease
POU5F1/OCT4 maintains the pluripotent state and is one of the four factors that reprogram human fibroblasts to induced pluripotent stem cells 18035408. POU2F2 with OBF-1 drives immunoglobulin promoter activity in B cells, and the PORE-versus-MORE distinction is the structural explanation for that coactivator dependence 11136971,11583619. POU3F2 and POU3F3 pattern the neocortex; POU4F1-3 specify sensory neurons, retinal ganglion cells and inner-ear hair cells; POU1F1 controls the growth hormone and prolactin genes. Mendelian loss of POU1F1 causes combined pituitary hormone deficiency 1509262, and mutations in POU3F4 and POU4F3 cause X-linked and autosomal dominant deafness respectively 7839145,9506947. In cancer, POU2F3 defines a tuft-cell-like subtype of small-cell lung carcinoma 29945888, and POU5F1 is re-expressed in germ cell tumors.
Reading a hit
A good octamer match tells you that a POU factor could bind; it does not tell you which one. The in vitro models for POU2F1, POU2F2, POU2F3, POU3F1-4 and POU5F1 are all the same 8-9 bp 23332764, so treat a hit as "POU family" unless expression data narrow it. Do not read a bare TAATTA or TCATTA match as POU: that is the homeodomain half alone and is shared with most of the homeodomain superfamily, which is exactly why two low-information CIS-BP and Vierstra entries in this family have nearest neighbours in the HOX and Q50 groups. Conversely, a full PORE, MORE or SOX-OCT composite is much stronger evidence, because the spacing constraint is specific. HOCOMOCO ids are informative here: PO2F1.H14CORE.0.PS.A and PO5F1.H14CORE.2.SM.B carry in vitro (S/M) support, whereas the P-only models such as PO3F2.H14CORE.0.P.B come from ChIP alone and may reflect co-binding.
How this was worked out
The octamer is found about 70 bp upstream of the transcription start of every immunoglobulin gene examined, as ATTTGCAT in light-chain genes and its exact inverse ATGCAAAT in heavy-chain genes, conserved while the surrounding sequence has diverged.
Still open then: That the element was a binding site for a tissue-specific factor was a proposal, not yet tested.
Cloning of OCT1 and comparison with OCT2, PIT1 and UNC-86 define the POU domain: a POU-specific box plus a homeobox subdomain, about 160 residues, that carries all the octamer-binding activity; ubiquitous OCT1 and lymphoid OCT2 read the octamer indistinguishably.
Still open then: How two subdomains in one polypeptide divide an 8-bp site between them was unknown.
The OCT1 POU domain-octamer structure shows POU-S on the 5' ATGC half, docked like the lambda and 434 repressors, and POU-HD on the 3' AAAT half, docked like other homeodomains, with an invisible linker and no protein-protein contact between the two.
Still open then: With no contact between the domains, the source of their cooperativity was unexplained.
Unlinked POU-S and POU-HD polypeptides still bind the octamer cooperatively (coupling energy about 1.6 kcal/mol), and the linker behaves as a tether with an effective concentration of about 3.6 mM rather than as a fixed spacer.
- 2000highBinding and transactivation assays, mutagenesis; X-ray crystallography (2001)11136971 11583619
POU dimers assemble in two configurations, on the PORE (ATTTGAAATGCAAAT) and the MORE (ATGCATATGCAT); only the PORE dimer recruits OBF-1, because the MORE dimer interface uses the same OCT1 residues that contact OBF-1. Crystal structures in 2001 showed the two non-overlapping interface patches.
A structure of the OCT1 POU domain and SOX2 HMG domain on the FGF4 enhancer element, plus modeling and mutational tests, shows that OCT4 and SOX2 assemble in different relative arrangements on the FGF4 and UTF1 enhancers, and that the enhancer DNA, not the proteins, sets the arrangement.
Systematic HT-SELEX of human POU proteins gives the octamer for POU2F, POU3F and POU5F1B proteins but homeodomain-led arrangements such as TGCATAATTAAT and CTCATTA for POU4F and POU6F proteins, so the in vitro preference of the family is a set of linker configurations rather than one octamer.
Still open then: Whether these alternative arrangements are what POU4F and POU6F proteins occupy in vivo was not tested.
Methyl-HT-SELEX across 542 human factors finds that many homeodomain proteins bind CpG-methylated sites differently from unmethylated ones; POU models from this kind of assay include CpG-containing MORE-like variants (for example POU2F2, GAATATGCGCAT) that the canonical octamer, which has no CpG, cannot capture.
Still unresolved
- Which DNA configuration POU3F proteins use in vivo is unsettled in the data: their JASPAR HT-SELEX models read the monomeric octamer (TATGCAAAT), whereas the HOCOMOCO models that incorporate ChIP-seq for POU3F1 and POU3F2 (PO3F1.H14CORE.1.PS.A, PO3F2.H14CORE.0.P.B) read the palindrome CATGAATATTCAT, a two-site arrangement the octamer does not describe.
- How often POU5F1 binds the octamer on its own rather than as part of a SOX-OCT composite is unresolved: the in vitro POU5F1 models are the bare octamer, the ChIP-derived ones carry the adjacent SOX half-site, and genome-wide data have been read both as cooperative assembly and as largely independent binding of the two factors.
- Whether POU4F and POU6F proteins occupy the full octamer or the homeodomain-led arrangements their HT-SELEX models give (TGCATAATTAAT, CTCATTA) is not resolved; the latter are close to generic homeodomain sites, so the answer decides whether these proteins can be told apart from other homeodomain factors by sequence at all.
A hit identifies the POU family, not a paralogue: the octamer models for OCT1, OCT2, OCT4 and the BRN proteins are interchangeable at 8-9 bp. The family is only moderately cohesive (0.68) because HT-SELEX returns several linker configurations, and the lowest-information members reduce to TAAT cores indistinguishable from the rest of the homeodomain superfamily; three entries here have nearest neighbours in other homeodomain families. ChIP-derived (P-grade) POU models can capture partner motifs, particularly SOX sites in pluripotent cells. Exact octamers occur roughly 10^5 times genome-wide, so a match alone is weak evidence without expression or footprint support.
References
- Parslow TG, Blair DL, Murphy WJ, et al. Structure of the 5' ends of immunoglobulin genes: a novel conserved sequence. Proc Natl Acad Sci U S A 1984;81:2650-4. PMID 6425835
- Herr W, Sturm RA, Clerc RG, et al. The POU domain: a large conserved region in the mammalian pit-1, oct-1, oct-2, and Caenorhabditis elegans unc-86 gene products. Genes Dev 1988;2:1513-6. PMID 3215510
- Sturm RA, Das G, Herr W. The ubiquitous octamer-binding protein Oct-1 contains a POU domain with a homeo box subdomain. Genes Dev 1988;2:1582-99. PMID 2905684
- Klemm JD, Rould MA, Aurora R, et al. Crystal structure of the Oct-1 POU domain bound to an octamer site: DNA recognition with tethered DNA-binding modules. Cell 1994;77:21-32. PMID 8156594
- Klemm JD, Pabo CO. Oct-1 POU domain-DNA interactions: cooperative binding of isolated subdomains and effects of covalent linkage. Genes Dev 1996;10:27-36. PMID 8557192
- Herr W, Cleary MA. The POU domain: versatility in transcriptional regulation by a flexible two-in-one DNA-binding domain. Genes Dev 1995;9:1679-93. PMID 7622033
- Tomilin A, Reményi A, Lins K, et al. Synergism with the coactivator OBF-1 (OCA-B, BOB-1) is mediated by a specific POU dimer configuration. Cell 2000;103:853-64. PMID 11136971
- Reményi A, Tomilin A, Pohl E, et al. Differential dimer activities of the transcription factor Oct-1 by DNA-induced interface swapping. Mol Cell 2001;8:569-80. PMID 11583619
- Reményi A, Lins K, Nissen LJ, et al. Crystal structure of a POU/HMG/DNA ternary complex suggests differential assembly of Oct4 and Sox2 on two enhancers. Genes Dev 2003;17:2048-59. PMID 12923055
- Takahashi K, Tanabe K, Ohnuki M, et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 2007;131:861-72. PMID 18035408
- Jolma A, Yan J, Whitington T, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327-39. PMID 23332764
- Yin Y, Morgunova E, Jolma A, et al. Impact of cytosine methylation on DNA binding specificities of human transcription factors. Science 2017;356. PMID 28473536
- Radovick S, Nations M, Du Y, et al. A mutation in the POU-homeodomain of Pit-1 responsible for combined pituitary hormone deficiency. Science 1992;257:1115-8. PMID 1509262
- de Kok YJ, van der Maarel SM, Bitner-Glindzicz M, et al. Association between X-linked mixed deafness and mutations in the POU domain gene POU3F4. Science 1995;267:685-8. PMID 7839145
- Vahava O, Morell R, Lynch ED, et al. Mutation in transcription factor POU4F3 associated with inherited progressive hearing loss in humans. Science 1998;279:1950-4. PMID 9506947
- Huang YH, Klingbeil O, He XY, et al. POU2F3 is a master regulator of a tuft cell-like variant of small cell lung cancer. Genes Dev 2018;32:915-928. PMID 29945888
82 motifs in this family — JASPAR 18 · HOCOMOCO 29 · CIS-BP 15 · Vierstra 20
| motif | database | consensus | bits |
|---|---|---|---|
| PO5F1.H14CORE.2.SM.BSMB | HOCOMOCO | TTATGCTAATtAgg | 21.4 |
| AC0091:POU3F:Homeodomain,POU | Vierstra | ATGcATTaATTCAT | 21.4 |
| POU4F2 (MA0683.2) | JASPAR | TGCATAATTAATGAG | 20.8 |
| POU2F1::SOX2 (MA1962.1)dimer | JASPAR | ATTTGCATaACAATg | 20 |
| POU1F1 (MA0784.3) | JASPAR | CTaATTTGCATAaT | 19.8 |
| PO3F4.H14CORE.0.SM.BSMB | HOCOMOCO | CTaATTAGCATAA | 19.5 |
| PIT1.H14CORE.0.SM.BSMB | HOCOMOCO | TAATGAGATGCg | 19.2 |
| P5F1B.H14CORE.0.S.BSB | HOCOMOCO | TgaATATGCAAATT | 18.5 |
| PO2F3.H14CORE.1.S.CSC | HOCOMOCO | aTATGCtAATTAG | 18.4 |
| PO3F1.H14CORE.3.M.CMC | HOCOMOCO | TTATGCTAATTAG | 18.2 |
| PO3F2.H14CORE.2.SM.BSMB | HOCOMOCO | CTcATTAGCATAa | 18.2 |
| POU2F2 (MA0507.3) | JASPAR | aTATGCAAATtAg | 18 |
| PO3F2.H14CORE.0.P.BPB | HOCOMOCO | cTCATGAATATTcAT | 17.9 |
| AC0006:POU3F/POU1F:Homeodomain,POU | Vierstra | TAATGAgATGC | 17.4 |
| PO4F2.H14CORE.0.S.BSB | HOCOMOCO | TcATTAATtATTcA | 17.2 |
| M05473_2.00 | CIS-BP | TAATGAgATGCA | 17.1 |
| PO4F1.H14CORE.1.SM.BSMB | HOCOMOCO | ATTAATTATgCA | 16.2 |
| M03310_2.00 | CIS-BP | TcATTAATTATgCA | 16 |
| PO2F2.H14CORE.1.M.CMC | HOCOMOCO | GAATATGCGCATa | 15.9 |
| M07976_2.00 | CIS-BP | cCtTTGTtATGCAaA | 15.9 |
| PO3F1.H14CORE.1.PS.APSA | HOCOMOCO | ctCATGaATAtTca | 15.4 |
| PO4F3.H14CORE.0.SM.BSMB | HOCOMOCO | TnAATAATTcA | 15 |
| Pou5f1::Sox2 (MA0142.1)dimer | JASPAR | CtTTGTtATGCAAAT | 14.8 |
| AC0090:POU3F/POU1F:Homeodomain,POU | Vierstra | TAATTTATGC | 14.2 |
| PO5F1.H14CORE.0.P.BPB | HOCOMOCO | caTTgTtATGCAaA | 14.1 |
| M07972_2.00 | CIS-BP | tATGCAAAT | 14 |
| POU3F2 (MA0787.1) | JASPAR | tTATGCaAATt | 13.8 |
| PO2F1.H14CORE.0.PS.APSA | HOCOMOCO | TATGCAAAT | 13.8 |
| AC0088:POU4F/POU5F:Homeodomain,POU | Vierstra | ATTAATTATGCA | 13.7 |
| PO2F2.H14CORE.2.S.BSB | HOCOMOCO | TtTGCATATT | 13.6 |
| POU3F3 (MA0788.1) | JASPAR | atTATGCtAATTt | 13.5 |
| PIT1.H14CORE.1.S.BSB | HOCOMOCO | ATGAATAAt | 13.5 |
| M03319_2.00 | CIS-BP | atTATGCtAATTt | 13.5 |
| POU3F4 (MA0789.1) | JASPAR | TATGCaAAT | 13.4 |
| PO2F3.H14CORE.0.PS.APSA | HOCOMOCO | GCAAATTA | 13.4 |
| POU2F3 (MA0627.3) | JASPAR | TATGCAAAT | 13.1 |
| POU4F3 (MA0791.2) | JASPAR | TGcATAATTaAT | 13.1 |
| M03312_2.00 | CIS-BP | TGaATAATTaAT | 13.1 |
| AC0230:POU2F/ZNF:Homeodomain,POU | Vierstra | tTgaTTgCATtCvAnT | 12.9 |
| POU4F1 (MA0790.2) | JASPAR | TGaATAATTaAT | 12.7 |
| PO6F1.H14CORE.0.SM.BSMB | HOCOMOCO | TAATGAG | 12.6 |
| PO6F2.H14CORE.0.S.BSB | HOCOMOCO | TAATTAG | 12.6 |
| AC0011:POU3F:Homeodomain,POU | Vierstra | CATbcaaAATgaa | 12.6 |
| PO2F2.H14CORE.0.PSM.APSMA | HOCOMOCO | tATGCAAAT | 12.5 |
| AC0100:POU2F:Homeodomain,POU | Vierstra | TGAATATTC | 12.5 |
| PO3F2.H14CORE.1.S.BSB | HOCOMOCO | aTTAATTA | 12.4 |
| M00362_2.00 | CIS-BP | cTAATTA | 12.3 |
| AC0007:POU2F:Homeodomain,POU | Vierstra | acATgtaAtTntG | 12.3 |
| M03315_2.00 | CIS-BP | tTATGcaAAT | 12 |
| AC0636:POU5F/NANOG:Homeodomain,POU | Vierstra | aTTGTtATGCAAA | 12 |
| AC0637:POU2F/POU3F:Homeodomain,POU | Vierstra | TATGCAAAT | 12 |
| POU3F1 (MA0786.2) | JASPAR | tTATGcaAAT | 11.6 |
| AC0103:POU3F/POU2F:Homeodomain,POU | Vierstra | TATGCGCAT | 11.6 |
| POU2F1 (MA0785.2) | JASPAR | TATGCAAAT | 11.5 |
| PO3F3.H14CORE.1.P.CPC | HOCOMOCO | caTGcATAtTnAt | 11.5 |
| POU6F2 (MA0793.2) | JASPAR | CTCATTA | 11.4 |
| PO3F3.H14CORE.2.S.BSB | HOCOMOCO | TAATTAa | 11.4 |
| PO5F1.H14CORE.1.P.BPB | HOCOMOCO | ttATGCAAAT | 11.4 |
| PO4F1.H14CORE.0.S.BSB | HOCOMOCO | ATtATgCAT | 11.2 |
| PO3F3.H14CORE.0.P.CPC | HOCOMOCO | ATGcAtATt | 11 |
| AC0089:POU4F:Homeodomain,POU | Vierstra | TgcATtATtcA | 11 |
| AC0661:POU6F:Homeodomain | Vierstra | tTATGCnnngCATAA | 10.9 |
| M03306_2.00 | CIS-BP | TATGCAAAT | 10.7 |
| POU5F1B (MA0792.1) | JASPAR | TATGCAAAT | 10.6 |
| POU6F1 (MA1549.2) | JASPAR | TAATGAG | 10.5 |
| M03300_2.00 | CIS-BP | ATGcATAaTTaA | 10.5 |
| PO3F1.H14CORE.0.P.BPB | HOCOMOCO | ATGCAaATt | 10.4 |
| M02094_2.00 | CIS-BP | TAATTA | 10.1 |
| AC0012:POU3F:Homeodomain,POU | Vierstra | AbTbTAATTCc | 9.9 |
| POU5F1 (MA1115.2) | JASPAR | ATGCAAA | 9.8 |
| PO3F1.H14CORE.2.S.CSC | HOCOMOCO | TAATTA | 9.4 |
| POU6F1 (MA0628.2) | JASPAR | TAATTA | 9.3 |
| AC0638:POU2F:Homeodomain,POU | Vierstra | ATGCTrAt | 8.8 |
| AC0017:POU2F:Homeodomain,POU | Vierstra | ATTAbcaTAAT | 8.2 |
| AC0101:POU3F/POU2F:Homeodomain,POU | Vierstra | TGCATAAt | 8.1 |
| AC0005:POU6F/NKX:Homeodomainlooks like HD_HOX_ANT | Vierstra | TCATTA | 7.8 |
| AC0102:POU2F/POU3F:Homeodomain,POU | Vierstra | TGaATATgCA | 7.5 |
| M00377_2.00 | CIS-BP | TGcATAa | 5.5 |
| AC0024:POU3F/HESX:Homeodomainlooks like HD_HOX_ANT | Vierstra | TAATt | 4.6 |
| M01509_2.00looks like HD_Q50 | CIS-BP | aatta | 1.9 |
| M00364_2.00 | CIS-BP | AT | 1.6 |
| M00360_2.00 | CIS-BP | a | 1.4 |
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.