Unnamed zinc-finger archetypes

Not a motif but a label gap. The Vierstra footprint atlas groups redundant matrices from several databases into archetypes, each named after the factor groups whose motifs clustered into it. For 256 of the 693 archetypes in this atlas that name collapses to a generic token, ZNF, ZFP, ZBTB, ZSCAN or ZKSCAN, so the only claim being made is that some C2H2 zinc finger binds here. These 256 matrices are long, high-information and almost entirely unrelated to one another: not one has a close match to another member of this entry. A hit means you matched a specific, reproducible zinc-finger matrix whose protein was never named, and the useful next step is to find that matrix's nearest named neighbour.

How this was worked out

1. 2000mediumreview of structures and phage-display selections10940247

   A review of C2H2 recognition established the working rule that each finger contacts about three base pairs through a few helix residues, while concluding that no general recognition code allows a motif to be read off an array's sequence; interactions between neighbouring fingers and linkers complicate prediction.

   Still open then: Specificity for an uncharacterized zinc-finger protein could not be predicted, only measured.

2. 2014highprotein-binding microarrays, motif inference by DNA-binding-domain identity25215497

   A survey of more than 1,000 transcription factors across 131 eukaryotes showed that closely related DNA-binding domains almost always share sequence preferences, licensing motif inference for about a third of known eukaryotic factors and creating the inferred, cross-species labels that fill databases such as CIS-BP.

   Still open then: Inference was weakest exactly where zinc-finger arrays differ most, among recent paralogues.

3. 2015highbacterial one-hybrid, protein-binding microarray, ChIP-seq25690854

   A recognition code inferred from thousands of natural C2H2 domains, combined with protein-binding microarrays and ChIP, produced motifs for large numbers of previously unmeasured zinc fingers and showed that this class recognizes more motifs than all other human transcription factors combined.

4. 2016mediumChIP-seq, affinity purification-mass spectrometry, expression profiling after knockdown27852650

   Direct profiling of 131 human C2H2 proteins gave motif models for 78 that had none, and found that their protein-protein interactions are nearly as diverse as their motifs, so a zinc finger's effector function cannot be guessed from its accessory domain either.

5. 2020highDNase I footprinting in 243 cell and tissue types, motif clustering32728250

   The reference footprint atlas clustered redundant matrices from several collections into archetypes and named each after the factor groups it contained; clusters with no named majority took a generic token, which is the origin of the ZNF, ZFP and ZBTB labels in this entry.

   Still open then: An archetype's name reported which matrices clustered, not which protein binds, and the clustering was never revisited as new motifs appeared.

6. 2026highmultiple in vitro and in vivo assays, cross-validated42557334

   The Codebook effort determined specificities for 332 putative and poorly characterized human transcription factors in more than 4,000 experiments, producing motifs for 177 of them and adding roughly 130 distinct motifs, most tied to a single protein; in vitro motifs proved strongly enriched in cellular binding sites.

   Still open then: Half of the targeted factors still yielded no motif, and the archetype labels here were not rebuilt against the new models.

7. 2026mediumChIP-seq of tagged factors in living cells42557260

   Binding locations for 166 poorly characterized human transcription factors showed that about half associate with promoters and enhancers at shared conserved motif matches, while the other half bind largely unique sets of repeat sequences in closed chromatin, a compartment that footprint-derived archetypes systematically under-sample.

Still unresolved

• It is unresolved whether an unnamed archetype that matches a named matrix at 0.9 or better is that protein's motif or a motif shared by several zinc fingers, because similarity alone cannot distinguish the two and the archetype clustering has not been rebuilt since the recent motif surveys.
• How well any of these in vitro or footprint-derived models predict occupancy in living cells remains contested: the Codebook data report that in vitro motifs are strongly enriched in cellular binding sites, while in-cell profiling of poorly characterized factors finds half of them binding closed-chromatin repeats that accessibility-based atlases never sampled.
• Whether the long, high-information members of this entry are transposable-element sequence read by KRAB zinc fingers or genuine composite regulatory words has not been tested member by member; the base composition is consistent with the first reading but no repeat annotation has been published for the archetype set.

References

• Vierstra J, Lazar J, Sandstrom R, et al. Global reference mapping of human transcription factor footprints. Nature 2020;583:729-736. PMID 32728250
• Lambert SA, Jolma A, Campitelli LF, et al. The Human Transcription Factors. Cell 2018;172:650-665. PMID 29425488
• Schmitges FW, Radovani E, Najafabadi HS, et al. Multiparameter functional diversity of human C2H2 zinc finger proteins. Genome Res 2016;26:1742-1752. PMID 27852650
• Najafabadi HS, Mnaimneh S, Schmitges FW, et al. C2H2 zinc finger proteins greatly expand the human regulatory lexicon. Nat Biotechnol 2015;33:555-62. PMID 25690854
• Weirauch MT, Yang A, Albu M, et al. Determination and inference of eukaryotic transcription factor sequence specificity. Cell 2014;158:1431-1443. PMID 25215497
• Rauluseviciute I, Riudavets-Puig R, Blanc-Mathieu R, et al. JASPAR 2024: 20th anniversary of the open-access database of transcription factor binding profiles. Nucleic Acids Res 2024;52:D174-D182. PMID 37962376
• Vorontsov IE, Eliseeva IA, Zinkevich A, et al. HOCOMOCO in 2024: a rebuild of the curated collection of binding models for human and mouse transcription factors. Nucleic Acids Res 2024;52:D154-D163. PMID 37971293
• Imbeault M, Helleboid PY, Trono D. KRAB zinc-finger proteins contribute to the evolution of gene regulatory networks. Nature 2017;543:550-554. PMID 28273063
• Jolma A, Laverty KU, Fathi A, et al. An expanded codebook of human transcription factor DNA-binding specificity. Nature 2026;657:275-283. PMID 42557334
• Razavi R, Fathi A, Yellan I, et al. Extensive binding of poorly characterized human transcription factors to genomic dark matter. Nat Commun 2026;17. PMID 42557260
• Wolfe SA, Nekludova L, Pabo CO. DNA recognition by Cys2His2 zinc finger proteins. Annu Rev Biophys Biomol Struct 2000;29:183-212. PMID 10940247

Other C2H2 zinc finger families