{
 "key": "CTCF",
 "name": "CTCF",
 "class": "C2H2 zinc finger",
 "aka": [
  "CCCTC-binding factor",
  "CTCF site",
  "CTCF-binding site (CBS)",
  "insulator element",
  "11-zinc-finger factor",
  "NeP1",
  "TFClass: more than 3 adjacent zinc finger factors"
 ],
 "definition": "19-bp core CCASYAGRKGGCRSYR with upstream module; the insulator/loop-anchor motif. CTCF.",
 "consensus": "CCASYAGRKGGCRSYR",
 "dataConsensus": "CCAccAGGgGGCgc",
 "motifCount": 10,
 "bySource": {
  "JASPAR": 3,
  "HOCOMOCO": 1,
  "CIS-BP": 1,
  "Vierstra": 5
 },
 "symbols": [
  "CTCF"
 ],
 "representative": "CTCF (MA0139.2)",
 "cohesion": 0.65,
 "short": "CTCF is a single, near-ubiquitously expressed protein whose eleven C2H2 zinc fingers read an asymmetric, roughly 15-20 bp GC-rich sequence, often with a second recognition module a few bases upstream. Tens of thousands of these sites act as insulators and as anchors for cohesin-extruded chromatin loops, so they delimit topologically associating domains and separate enhancers from promoters they should not reach. Because the motif is long, information-rich, and read by a well-characterized finger-to-base code solved in two crystal structures, a CTCF match is one of the few motif hits that predicts real occupancy reasonably often. The site is directional: which strand it lies on determines which way a loop is anchored, and convergent pairs of sites are the ones that form loops.",
 "long": "## The factors\n\nCTCF (CCCTC-binding factor) is one protein, not a family. The human gene encodes a 727-residue polypeptide with a central array of eleven Cys2-His2 zinc fingers flanked by N- and C-terminal regions that carry no DNA-binding domain. It is expressed in essentially every cell type examined, and it is extraordinarily conserved: the chicken and human proteins are 93% identical in amino acid sequence (Filippova et al. 1996) [PMID:8649389]. CTCF binds DNA as a monomer; no dimerization interface is required, and its long-range activity comes from partners rather than from self-assembly. The most important of these is cohesin: a short segment in the CTCF N terminus docks onto the SA2-SCC1 (STAG2-RAD21) surface of human cohesin, and a 2.7 A crystal structure of that complex showed the interaction is specifically needed for CTCF-anchored loops and for positioning cohesin at CTCF sites (Li et al. 2020) [PMID:31905366]. CTCF has exactly one paralogue, the testis-restricted CTCFL/BORIS, which carries an almost identical zinc-finger array and is treated as a separate family here.\n\n## The motif\n\nThe high-information core is 15 bp. JASPAR's core model (MA0139.2) reduces to RCCASYAGRKGGCRS in IUPAC, which is the sequence most people mean when they write the CTCF consensus; longer ChIP-derived models are also in use (JASPAR MA1929.2 and MA1930.2 span 31 and 33 positions), corresponding to the 33/34-mer that Schmidt et al. (2012) defined from six mammals [PMID:22244452]. The structural basis is unusually well understood. In the Hashimoto et al. (2017) crystal structure, zinc fingers 3-7 each contact three bases of the 15-bp core, with each conserved nucleotide making a base-specific hydrogen bond to a particular side chain; variable positions also make contacts, which is how CTCF tolerates deviation from consensus [PMID:28529057]. Yin et al. (2017) resolved CTCF on protocadherin sites and split the site into four modules: ZF3 reads module 4, ZF4-7 read modules 3 and 2, and ZF9-11 read an upstream module 1, with ZF8 acting as a spacer that absorbs variable distance between modules 1 and 2 [PMID:29076501]. Genome-wide, ZF4-7 anchor CTCF at roughly 80% of targets containing the core motif, while ZF1-2 and ZF8-11 read non-conserved flanks (Nakahashi et al. 2013) [PMID:23707059]. Two consequences matter in practice. The site is asymmetric, so it has an orientation, and a central asymmetric A is part of what makes recognition directional [PMID:29076501]. CTCF is also sensitive to cytosine methylation at position 2 of the core but insensitive at position 12, a difference the structure rationalizes [PMID:28529057]; across 19 human cell types, 41% of variable CTCF occupancy was attributable to differential methylation concentrated at two positions in the recognition sequence (Wang et al. 2012) [PMID:22955980].\n\n## Where it acts\n\nThe first genome-wide survey reported 13,804 CTCF sites (Kim et al. 2007) [PMID:17382889]; current ChIP-seq gives tens of thousands per cell type. Sites are GC-rich but are not concentrated in CpG islands the way GC-box promoters are; they are spread across intergenic space, introns and promoter-proximal regions, and lineage-specific expansions have been driven by waves of retrotransposon activity (Schmidt et al. 2012) [PMID:22244452]. Occupied sites are enriched at topologically associating domain boundaries, where they often occur in clusters rather than singly, a redundancy that appears to stabilize boundaries over evolution (Kentepozidou et al. 2020) [PMID:31910870]. The essential partner is cohesin. Loops form preferentially between CTCF sites in convergent orientation (Rao et al. 2014) [PMID:25497547], and deleting or inverting a single site changes the loops formed (de Wit et al. 2015; Guo et al. 2015) [PMID:26527277,26276636] - the observation that motivated the loop-extrusion model (Sanborn et al. 2015; Fudenberg et al. 2016) [PMID:26499245,27210764] and that single-molecule work later explained as DNA-tension-dependent, orientation-specific blocking of extruding cohesin (Davidson et al. 2023) [PMID:37076620].\n\n## Biology and disease\n\nCTCF was defined functionally as the protein required for enhancer-blocking by vertebrate insulators (Bell et al. 1999) [PMID:10458613] and for methylation-sensitive imprinting control at the H19/Igf2 locus [PMID:10839546,10839547]. Acute degradation of CTCF abolishes CTCF-CTCF loops and TAD insulation while leaving A/B compartments intact, separating those two layers of genome folding (Nora et al. 2017) [PMID:28525758]. In disease, CTCF/cohesin-binding sites are recurrent somatic mutation hotspots in several cancers (Katainen et al. 2015) [PMID:26053496]; in IDH-mutant glioma, hypermethylation disrupts a CTCF boundary and allows an enhancer to activate PDGFRA (Flavahan et al. 2016) [PMID:26700815]; and germline monoallelic CTCF variants cause an autosomal dominant neurodevelopmental disorder affecting brain, growth and craniofacial development (Valverde de Morales et al. 2023) [PMID:36454652].\n\n## Reading a hit\n\nThis is the one family in the atlas where a position-weight-matrix match genuinely carries predictive weight: sites that match the consensus more closely are occupied more often and more constitutively (Plasschaert et al. 2014) [PMID:24121688]. Even so, only a fraction of genomic matches are bound in any one cell type, and occupancy is conditional on the methylation state of the CpG at core position 2. Report the strand: a CTCF hit without an orientation is half an answer, because loop anchoring depends on it. A hit cannot distinguish CTCF from CTCFL, which share the zinc-finger array. Finally, check which matrix matched. Two Vierstra archetypes labelled for CTCF are not the CTCF motif at all - AC0214:CTCF/ZNF is a poly-T model (tTTTTgTTTTT) and AC0174:CTCF/ZNF is a short C-rich model (tGCCCCCT, 8.4 bits) - and a hit to either says nothing about CTCF. The genuine CTCF models, including the ChIP-derived HOCOMOCO core model (CTCF.H14CORE.0.P.B, ChIP evidence, grade B), agree with each other and with two independent crystal structures [PMID:28529057,29076501], so the label itself is not in doubt.",
 "caveats": "Trust a strong CTCF hit more than almost any other motif in this atlas, but not unconditionally. Only some matches are occupied in a given cell type; occupancy depends on the methylation state of the CpG at core position 2, and a site bound in one lineage may be free in another. The hit cannot tell CTCF from CTCFL/BORIS. Orientation is part of the result, not an afterthought. A single site is weak evidence for a domain boundary, since boundaries are usually clusters of sites. And confirm which matrix matched: two Vierstra archetypes carrying the CTCF label are poly-T or short C-rich models that have nothing to do with the CTCF core.",
 "references": [
  {
   "cite": "Filippova GN, Fagerlie S, Klenova EM, et al. An exceptionally conserved transcriptional repressor, CTCF, employs different combinations of zinc fingers to bind diverged promoter sequences of avian and mammalian c-myc oncogenes. Mol Cell Biol 1996;16:2802-13.",
   "pmid": "8649389",
   "url": "https://pubmed.ncbi.nlm.nih.gov/8649389/"
  },
  {
   "cite": "Bell AC, West AG, Felsenfeld G. The protein CTCF is required for the enhancer blocking activity of vertebrate insulators. Cell 1999;98:387-96.",
   "pmid": "10458613",
   "url": "https://pubmed.ncbi.nlm.nih.gov/10458613/"
  },
  {
   "cite": "Kim TH, Abdullaev ZK, Smith AD, et al. Analysis of the vertebrate insulator protein CTCF-binding sites in the human genome. Cell 2007;128:1231-45.",
   "pmid": "17382889",
   "url": "https://pubmed.ncbi.nlm.nih.gov/17382889/"
  },
  {
   "cite": "Schmidt D, Schwalie PC, Wilson MD, et al. Waves of retrotransposon expansion remodel genome organization and CTCF binding in multiple mammalian lineages. Cell 2012;148:335-48.",
   "pmid": "22244452",
   "url": "https://pubmed.ncbi.nlm.nih.gov/22244452/"
  },
  {
   "cite": "Wang H, Maurano MT, Qu H, et al. Widespread plasticity in CTCF occupancy linked to DNA methylation. Genome Res 2012;22:1680-8.",
   "pmid": "22955980",
   "url": "https://pubmed.ncbi.nlm.nih.gov/22955980/"
  },
  {
   "cite": "Nakahashi H, Kieffer Kwon KR, Resch W, et al. A genome-wide map of CTCF multivalency redefines the CTCF code. Cell Rep 2013;3:1678-1689.",
   "pmid": "23707059",
   "url": "https://pubmed.ncbi.nlm.nih.gov/23707059/"
  },
  {
   "cite": "Plasschaert RN, Vigneau S, Tempera I, et al. CTCF binding site sequence differences are associated with unique regulatory and functional trends during embryonic stem cell differentiation. Nucleic Acids Res 2014;42:774-89.",
   "pmid": "24121688",
   "url": "https://pubmed.ncbi.nlm.nih.gov/24121688/"
  },
  {
   "cite": "Rao SS, Huntley MH, Durand NC, et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 2014;159:1665-80.",
   "pmid": "25497547",
   "url": "https://pubmed.ncbi.nlm.nih.gov/25497547/"
  },
  {
   "cite": "de Wit E, Vos ES, Holwerda SJ, et al. CTCF Binding Polarity Determines Chromatin Looping. Mol Cell 2015;60:676-84.",
   "pmid": "26527277",
   "url": "https://pubmed.ncbi.nlm.nih.gov/26527277/"
  },
  {
   "cite": "Hashimoto H, Wang D, Horton JR, et al. Structural Basis for the Versatile and Methylation-Dependent Binding of CTCF to DNA. Mol Cell 2017;66:711-720.e3.",
   "pmid": "28529057",
   "url": "https://pubmed.ncbi.nlm.nih.gov/28529057/"
  },
  {
   "cite": "Yin M, Wang J, Wang M, et al. Molecular mechanism of directional CTCF recognition of a diverse range of genomic sites. Cell Res 2017;27:1365-1377.",
   "pmid": "29076501",
   "url": "https://pubmed.ncbi.nlm.nih.gov/29076501/"
  },
  {
   "cite": "Li Y, Haarhuis JHI, Sedeño Cacciatore Á, et al. The structural basis for cohesin-CTCF-anchored loops. Nature 2020;578:472-476.",
   "pmid": "31905366",
   "url": "https://pubmed.ncbi.nlm.nih.gov/31905366/"
  },
  {
   "cite": "Chung JH, Whiteley M, Felsenfeld G. A 5' element of the chicken beta-globin domain serves as an insulator in human erythroid cells and protects against position effect in Drosophila. Cell 1993;74:505-14.",
   "pmid": "8348617",
   "url": "https://pubmed.ncbi.nlm.nih.gov/8348617/"
  },
  {
   "cite": "Bell AC, Felsenfeld G. Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene. Nature 2000;405:482-5.",
   "pmid": "10839546",
   "url": "https://pubmed.ncbi.nlm.nih.gov/10839546/"
  },
  {
   "cite": "Hark AT, Schoenherr CJ, Katz DJ, et al. CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus. Nature 2000;405:486-9.",
   "pmid": "10839547",
   "url": "https://pubmed.ncbi.nlm.nih.gov/10839547/"
  },
  {
   "cite": "Guo Y, Xu Q, Canzio D, et al. CRISPR Inversion of CTCF Sites Alters Genome Topology and Enhancer/Promoter Function. Cell 2015;162:900-10.",
   "pmid": "26276636",
   "url": "https://pubmed.ncbi.nlm.nih.gov/26276636/"
  },
  {
   "cite": "Sanborn AL, Rao SS, Huang SC, et al. Chromatin extrusion explains key features of loop and domain formation in wild-type and engineered genomes. Proc Natl Acad Sci U S A 2015;112:E6456-65.",
   "pmid": "26499245",
   "url": "https://pubmed.ncbi.nlm.nih.gov/26499245/"
  },
  {
   "cite": "Fudenberg G, Imakaev M, Lu C, et al. Formation of Chromosomal Domains by Loop Extrusion. Cell Rep 2016;15:2038-49.",
   "pmid": "27210764",
   "url": "https://pubmed.ncbi.nlm.nih.gov/27210764/"
  },
  {
   "cite": "Nora EP, Goloborodko A, Valton AL, et al. Targeted Degradation of CTCF Decouples Local Insulation of Chromosome Domains from Genomic Compartmentalization. Cell 2017;169:930-944.e22.",
   "pmid": "28525758",
   "url": "https://pubmed.ncbi.nlm.nih.gov/28525758/"
  },
  {
   "cite": "Davidson IF, Barth R, Zaczek M, et al. CTCF is a DNA-tension-dependent barrier to cohesin-mediated loop extrusion. Nature 2023;616:822-827.",
   "pmid": "37076620",
   "url": "https://pubmed.ncbi.nlm.nih.gov/37076620/"
  },
  {
   "cite": "Kentepozidou E, Aitken SJ, Feig C, et al. Clustered CTCF binding is an evolutionary mechanism to maintain topologically associating domains. Genome Biol 2020;21:5.",
   "pmid": "31910870",
   "url": "https://pubmed.ncbi.nlm.nih.gov/31910870/"
  },
  {
   "cite": "Katainen R, Dave K, Pitkänen E, et al. CTCF/cohesin-binding sites are frequently mutated in cancer. Nat Genet 2015;47:818-21.",
   "pmid": "26053496",
   "url": "https://pubmed.ncbi.nlm.nih.gov/26053496/"
  },
  {
   "cite": "Flavahan WA, Drier Y, Liau BB, et al. Insulator dysfunction and oncogene activation in IDH mutant gliomas. Nature 2016;529:110-4.",
   "pmid": "26700815",
   "url": "https://pubmed.ncbi.nlm.nih.gov/26700815/"
  },
  {
   "cite": "Valverde de Morales HG, Wang HV, Garber K, et al. Expansion of the genotypic and phenotypic spectrum of CTCF-related disorder guides clinical management: 43 new subjects and a comprehensive literature review. Am J Med Genet A 2023;191:718-729.",
   "pmid": "36454652",
   "url": "https://pubmed.ncbi.nlm.nih.gov/36454652/"
  }
 ],
 "timeline": [
  {
   "year": 1993,
   "finding": "A 5' element of the chicken beta-globin domain (the HS4 region) is shown to block the beta-globin LCR enhancer from a reporter directionally, and to protect a transgene from position effects in Drosophila, establishing the vertebrate insulator as an element that can be assayed.",
   "methods": "Stable reporter assays in K562 erythroid cells; transgenic Drosophila position-effect assay",
   "confidence": "medium",
   "pmids": [
    "8348617"
   ],
   "open": "No protein had been identified that bound the element or carried its activity."
  },
  {
   "year": 1996,
   "finding": "Cloning of human CTCF shows an 11-zinc-finger protein 93% identical between chicken and human that binds diverged c-myc promoter sites with different subsets of fingers (ZF2-7 on the chicken site, ZF3-11 on the human), the first evidence that CTCF is 'multivalent'.",
   "methods": "cDNA cloning, gel shifts with finger deletions, contact-base mutagenesis, proteolytic protection",
   "confidence": "medium",
   "pmids": [
    "8649389"
   ],
   "open": "Because each known site seemed to use a different set of fingers, no general consensus could be written."
  },
  {
   "year": 1999,
   "finding": "CTCF is identified as the protein required for enhancer blocking by HS4 and other vertebrate insulators; in 2000 two groups showed that CTCF binding at the H19 imprinting control region is abolished by CpG methylation, so the methylated paternal allele loses the boundary and expresses Igf2.",
   "methods": "Enhancer-blocking reporter assays, binding-site mutagenesis, in vitro methylation-sensitive binding, transgenic mice",
   "confidence": "high",
   "pmids": [
    "10458613",
    "10839546",
    "10839547"
   ],
   "open": "How a single DNA-bound protein blocks an enhancer acting from a distance was unexplained."
  },
  {
   "year": 2007,
   "finding": "The first genome-wide map finds 13,804 CTCF sites in human fibroblasts, mostly far from transcription start sites, fitting a conserved consensus motif, with occupancy largely invariant across cell types.",
   "methods": "ChIP-chip, de novo motif discovery",
   "confidence": "medium",
   "pmids": [
    "17382889"
   ],
   "open": "Why a minority of sites vary between cell types, and why most genomic matches are unbound, was not addressed."
  },
  {
   "year": 2014,
   "finding": "Kilobase-resolution Hi-C shows that chromatin loop anchors are usually CTCF-bound and that the two anchoring motifs are overwhelmingly in convergent orientation.",
   "methods": "In situ Hi-C, ChIP-seq, motif-orientation analysis",
   "confidence": "medium",
   "pmids": [
    "25497547"
   ],
   "open": "Whether orientation was causal, and what mechanism could read the orientation of two sites hundreds of kilobases apart, was unknown."
  },
  {
   "year": 2015,
   "finding": "Deleting or inverting single CTCF sites by CRISPR rewires loops (Sox2, protocadherin and beta-globin loci), showing that orientation is causal; the loop-extrusion model, in which cohesin extrudes DNA until it meets a CTCF site pointing toward it, explains why only convergent pairs form loops.",
   "methods": "CRISPR deletion and inversion with 4C-seq and Hi-C; polymer simulation",
   "confidence": "high",
   "pmids": [
    "26527277",
    "26276636",
    "26499245",
    "27210764"
   ],
   "open": "Cohesin had not yet been shown to extrude DNA, and how a bound CTCF stops it in one direction only was unknown."
  },
  {
   "year": 2017,
   "finding": "Two crystal structures of the CTCF finger array on DNA show ZF3-7 each reading three bases of the 15-bp core and explain methylation sensitivity at core position 2 but not 12; on protocadherin sites ZF9-11 read an upstream module, with ZF8 as a spacer.",
   "methods": "X-ray crystallography with binding assays",
   "confidence": "high",
   "pmids": [
    "28529057",
    "29076501"
   ],
   "open": "The two structures disagree about the C-terminal fingers: one places ZF9-11 in the major groove reading module 1, the other finds ZF8-9 spanning the backbone without base contacts and ZF10-11 not visible."
  },
  {
   "year": 2020,
   "finding": "A 2.7 A structure shows a segment of the CTCF N terminus docking onto the SA2-SCC1 subunits of cohesin, an interaction required specifically for CTCF-anchored loops. Single-molecule imaging in 2023 then showed CTCF blocks extruding cohesin asymmetrically but in a DNA-tension-dependent way, and can redirect extrusion or shrink loops: a regulator of extrusion, not the simple wall the 2015 model assumed.",
   "methods": "X-ray crystallography, Hi-C of interface mutants; in vitro single-molecule imaging of CTCF and cohesin",
   "confidence": "high",
   "pmids": [
    "31905366",
    "37076620"
   ],
   "open": "How permeable CTCF boundaries are to cohesin in living cells, and what sets DNA tension there, was not measured."
  }
 ],
 "openQuestions": [
  "What CTCF's C-terminal fingers contribute to recognition is unsettled: the protocadherin-site structures place ZF9-11 in the major groove reading the upstream module, whereas the other crystal structure finds ZF8-9 spanning the backbone without base contacts and ZF10-11 disordered, and genome-wide mutant data assign ZF8-11 to non-conserved flanks.",
  "Why most genomic matches to the CTCF core are unbound in a given cell type is only partly explained: CpG methylation at core position 2 accounts for about 41% of variable occupancy across 19 cell types, and the contribution of chromatin accessibility, flanking sequence and cofactors to the remainder has not been quantified.",
  "Whether CTCF acts in cells as a near-absolute barrier to loop extrusion, as the convergent-orientation rule suggests, or as a leaky, tension-dependent one, as single-molecule assays show, is not resolved, because the rate at which cohesin passes CTCF sites in vivo has not been measured.",
  "How much gene regulation depends on CTCF-anchored loops in general is open: acute CTCF degradation dissolves loops and TAD insulation genome-wide, while the clearest transcriptional consequences come from disrupting individual boundaries such as the PDGFRA boundary in IDH-mutant glioma, and which boundaries matter for expression cannot yet be predicted."
 ],
 "members": [
  {
   "id": "AC0383:CTCF:C2H2_ZF",
   "source": "Vierstra",
   "consensus": "AaACAGAaaTCCcCTGTGT",
   "bits": 22.5
  },
  {
   "id": "CTCF (MA1930.2)",
   "source": "JASPAR",
   "consensus": "cTGCAgTkccnnchnnnggCCActAGggGGCg",
   "bits": 20.7
  },
  {
   "id": "CTCF.H14CORE.0.P.B",
   "source": "HOCOMOCO",
   "consensus": "gCCACyAGggGGCGCt",
   "bits": 19,
   "prov": {
    "evidence": "P",
    "grade": "B"
   }
  },
  {
   "id": "AC0194:CTCF:C2H2_ZF",
   "source": "Vierstra",
   "consensus": "TGCAGTtCCmvndntgGCCA",
   "bits": 18.8
  },
  {
   "id": "CTCF (MA1929.2)",
   "source": "JASPAR",
   "consensus": "TGCAgTtccnnnnnyngCCAccAGGgGGCg",
   "bits": 17.8
  },
  {
   "id": "CTCF (MA0139.2)",
   "source": "JASPAR",
   "consensus": "CCAccAGGgGGCgc",
   "bits": 16.2
  },
  {
   "id": "M07550_2.00",
   "source": "CIS-BP",
   "consensus": "tGCCcCCTggTGG",
   "bits": 14.6,
   "prov": {
    "inferred": false
   }
  },
  {
   "id": "AC0214:CTCF/ZNF:C2H2_ZF",
   "source": "Vierstra",
   "consensus": "tTTTTgTTTTT",
   "bits": 14.1
  },
  {
   "id": "AC0495:CTCF/CTCFL:C2H2_ZF",
   "source": "Vierstra",
   "consensus": "cGCCccCTgGTGG",
   "bits": 14
  },
  {
   "id": "AC0174:CTCF/ZNF:C2H2_ZF",
   "source": "Vierstra",
   "consensus": "tGCCCCCT",
   "bits": 8.4
  }
 ],
 "url": "https://motif.zhoulab.io/families/ctcf/",
 "generated": "2026-10-21",
 "provenance": "Entry text drafted by a language model reading the primary literature; every cited PMID verified against PubMed. Computed fields derive from JASPAR 2024, HOCOMOCO v14, CIS-BP and the Vierstra archetypes. See https://motif.zhoulab.io/families/methods/"
}