ZNF410
ZNF410 is a five-finger C2H2 zinc-finger protein that recognizes a long, information-rich element, and it is the clearest known case of a mammalian transcription factor with essentially one target gene. Two conserved clusters near CHD4, which encodes the ATPase of the NuRD chromatin remodeling complex, together carry 27 copies of the motif; comparable clusters occur nowhere else in the genome. Removing ZNF410 lowers CHD4 protein by about 60 percent, which is enough to de-repress the fetal haemoglobin genes without obvious toxicity, and the factor was found in CRISPR screens for fetal haemoglobin silencers. Matches to the motif are therefore biologically meaningful mainly when they fall in those clusters, and isolated hits elsewhere rarely correspond to occupied, functional sites.
The factors
ZNF410 is a single, conserved protein with no close paralogue in the human genome: a tandem array of five C2H2 zinc fingers flanked by N- and C-terminal regions that are not themselves DNA-binding domains 36660822. It is broadly expressed. The pentadactyl array is the whole of the sequence-reading apparatus, and crystallographic and solution work shows it binds DNA as a monomer with 1:1 stoichiometry 33301730,36660822. An unusual regulatory feature sits outside the fingers: AlphaFold modeling combined with small-angle X-ray scattering indicates that an acidic, serine/threonine-rich hairpin loop in the N-terminal region folds back onto the DNA-binding face of the finger array, mimicking the charge and shape of DNA 36660822. Full-length ZNF410 binds DNA less tightly than the isolated finger array, both in vitro and in cells, consistent with the loop having to be displaced before binding 36660822. This is an autoinhibitory mechanism layered on top of ordinary sequence specificity.
The motif
The element is long for a five-finger protein: a 15-base-pair site, usually written from the CHD4 clusters as CATCCCATAATA with two or three further specified positions at the 5' end (reverse complement TATTATGGGATG) 23332764,33301730. All four databases in this family agree closely on it, and the family is among the most cohesive in the atlas. Five fingers contacting fifteen base pairs is the canonical three-base-pairs-per-finger arrangement 38754172, and the element contains no internal repeat or half-site: it is read in one pass, with no spacing rule and no dimerization. Information content is high and spread across the whole word, with the central CCCAT and the 3' AATA both contributing; the AT-rich 3' end is what distinguishes this motif from the many GC-rich promoter motifs it would otherwise resemble. Because ZNF410 has no paralogue, there is no paralogue ambiguity within the family: a strong match means ZNF410 or nothing.
Where it acts
In erythroid cells, ZNF410 chromatin occupancy is concentrated almost entirely at CHD4 33859416. Two regulatory elements, one upstream of the promoter and one intronic, carry 27 ZNF410 motifs between them 33859416; no other locus in the human or mouse genome has a comparable cluster, and the two clusters are conserved across mammals 33301730. Deleting them reproduces the effect of deleting ZNF410 itself, so essentially all of the factor's relevant output runs through this one gene 33859416. Occurrences of the isolated motif do exist elsewhere in the genome, and weak ZNF410 signal has been reported at some of them, but they do not measurably affect transcription 33301730. Mechanistically this is thought to be a matter of local concentration: a tight array of sites builds enough occupancy to activate, while single sites do not. The clusters sit in an open, CpG-island-adjacent promoter context rather than in tissue-specific enhancers.
Biology and disease
ZNF410 came out of CRISPR screens of transcription factors for regulators of fetal haemoglobin (HbF) 33301730,33859416. It is not a direct repressor of HBG1/HBG2; it activates CHD4, and CHD4/NuRD in turn silences the gamma-globin genes 33301730. Knocking out ZNF410, or its mouse homolog Zfp410, drops CHD4 protein by roughly 60 percent and raises HbF substantially in adult-type erythroid cultures and in xenotransplantation, without the cellular or organismal toxicity seen when CHD4 itself is removed 33859416. That partial, buffered effect is why ZNF410 attracted interest as a therapeutic target for sickle cell disease and beta-thalassaemia, alongside BCL11A and the HBG promoter BCL11A site 38728575,37924864. Outside erythropoiesis the phenotype is mild, consistent with a factor whose only strong target is a chromatin-remodeler subunit. There is no established human disease caused by ZNF410 mutation, and no recurrent cancer association.
Reading a hit
A strong match to this 15-base-pair word is good evidence for ZNF410 specifically, because the protein has no paralogue and the motif is long. What a hit does not license is the inference that the site is used. The lesson of this family is that a single occurrence of even a high-information motif is usually not a functional element; what mattered at CHD4 was 27 of them in two clusters. If your hit is a lone site, treat it as a sequence match rather than a regulatory element unless you have occupancy or perturbation data. Conversely, if you find a dense cluster of ZNF410 motifs at a locus, that is worth following up, and genome-wide scans for such clusters have been used to argue that CHD4 really is unique 33301730,33859416. The database matrices here are a mixture of in vitro (HT-SELEX/methyl-SELEX) and derived archetypes and agree well, so matrix choice matters little.
How this was worked out
High-throughput SELEX of human transcription factors yields a ZNF410 motif, the in vitro basis of the JASPAR matrix for this family, years before any target gene of the protein was known.
Still open then: Where ZNF410 bound in cells, and what it regulated, was unknown.
- 2021highCRISPR-Cas9 screen, ChIP-seq, RNA-seq, xenotransplantation, in vitro binding assays, X-ray crystallography33301730
A CRISPR screen for fetal hemoglobin regulators identifies ZNF410 as a factor that in erythroid cells directly activates a single gene, the NuRD subunit CHD4, through two conserved clusters of its binding sites with no counterpart elsewhere in the genome; crystallography defines how the five fingers read the 15-base-pair site.
Still open then: Why isolated copies of the same motif elsewhere in the genome have no effect was not explained.
- 2021highCRISPR screen, ChIP-seq, enhancer deletion, mouse knockout, erythroid differentiation assays33859416 33859418
An independent screen reaches the same conclusion and quantifies it: the two CHD4 regulatory elements carry 27 ZNF410 motifs between them and account for the whole effect of ZNF410 on fetal globin, and knockout of ZNF410 or mouse Zfp410 lowers CHD4 by about 60%, enough to derepress fetal hemoglobin while avoiding the toxicity of removing CHD4 itself.
- 2023mediumSmall-angle X-ray scattering, AlphaFold modeling, binding assays in vitro and in cells36660822
Full-length ZNF410 binds DNA less tightly than its isolated finger array; small-angle X-ray scattering with AlphaFold modeling indicates an acidic, serine/threonine-rich hairpin loop in the N-terminal region lies on the DNA-binding face of the fingers, mimicking DNA and acting as an autoinhibitory switch that must be displaced.
Still open then: What displaces the hairpin loop in cells, and whether it is regulated, is unknown.
- 2024lowPatient transcriptome sequencing, eQTL and binding-site prediction, inhibition experiments in HUDEP-2 cells38325179 38728575
A distal CHD4 enhancer transcribed as an enhancer RNA is reported to differ between beta-thalassemia patients with high and low fetal hemoglobin, with a candidate variant at a predicted ZNF410 site, placing the ZNF410-CHD4 axis among the modifiers of fetal hemoglobin now under review as therapeutic targets.
Still open then: The ZNF410 site involvement was predicted rather than demonstrated by binding or editing.
Still unresolved
- Why a dense cluster of sites activates CHD4 while isolated copies of the same 15-base-pair motif elsewhere in the genome do not measurably affect transcription is unexplained; local concentration is the proposed reason but the threshold has not been measured.
- What relieves the autoinhibition is unknown: the N-terminal hairpin loop occupies the DNA-binding surface and lowers the affinity of the full-length protein, but no signal, modification or partner that displaces it has been identified.
- Whether ZNF410 does anything outside erythroid cells is unresolved: it is broadly expressed and highly conserved, yet its occupancy in the cells examined is concentrated at CHD4 and no phenotype outside erythropoiesis has been described.
- Whether partial loss of CHD4 is a safe way to raise fetal hemoglobin in patients is unsettled: the roughly 60% reduction caused by ZNF410 loss avoids toxicity in cultured cells, xenografts and mice, but no clinical data test this therapeutic window.
Trust a hit for the protein but not for function. ZNF410 has no close paralogue, the motif is 15 base pairs long, and the in vitro and ChIP matrices agree, so a strong match is genuinely ZNF410's word rather than a shared GC-rich site. The problem is the opposite of the usual one: isolated matches occur throughout the genome and are largely inert, because ZNF410 needs the dense clusters found only at CHD4 to produce an effect. Do not read a single hit as a regulatory element, and be wary of published ZNF410 peak lists that include weak non-CHD4 sites. Full-length ZNF410 also binds more weakly than its isolated finger array, so in vitro affinities overstate cellular occupancy.
References
- Lan X, Ren R, Feng R, et al. ZNF410 Uniquely Activates the NuRD Component CHD4 to Silence Fetal Hemoglobin Expression. Mol Cell 2021;81:239-254.e8. PMID 33301730
- Vinjamur DS, Yao Q, Cole MA, et al. ZNF410 represses fetal globin by singular control of CHD4. Nat Genet 2021;53:719-728. PMID 33859416
- Tumburu L, Thein SL. Targeting ZNF410 as a potential β-hemoglobinopathy therapy. Nat Genet 2021;53:589-590. PMID 33859418
- Kaur G, Ren R, Hammel M, et al. Allosteric autoregulation of DNA binding via a DNA-mimicking protein domain: a biophysical study of ZNF410-DNA interaction using small angle X-ray scattering. Nucleic Acids Res 2023;51:1674-1686. PMID 36660822
- Khandros E, Blobel GA. Elevating fetal hemoglobin: recently discovered regulators and mechanisms. Blood 2024;144:845-852. PMID 38728575
- Zhang X, Xia F, Zhang X, et al. C2H2 Zinc Finger Transcription Factors Associated with Hemoglobinopathies. J Mol Biol 2024;436:168343. PMID 37924864
- Zhang X, Blumenthal RM, Cheng X. Updated understanding of the protein-DNA recognition code used by C2H2 zinc finger proteins. Curr Opin Struct Biol 2024;87:102836. PMID 38754172
- Jiang Y, Ye Y, Zhang X, et al. Identification and characterization of CHD4-associated eRNA as a novel modulator of fetal hemoglobin levels in β-thalassemia. Biochem Biophys Res Commun 2024;701:149555. PMID 38325179
- Jolma A, Yan J, Whitington T, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327-39. PMID 23332764
4 motifs in this family — JASPAR 1 · HOCOMOCO 1 · CIS-BP 1 · Vierstra 1
| motif | database | consensus | bits |
|---|---|---|---|
| ZNF410 (MA0752.2) | JASPAR | cCATCCCATAATAntC | 24.4 |
| ZN410.H14CORE.0.SM.BSMB | HOCOMOCO | TATTATGGGATG | 21.3 |
| AC0327:ZNF:C2H2_ZF | Vierstra | TATTATGGGATG | 20.4 |
| M04435_2.00 | CIS-BP | TATTATGGGaTG | 18.9 |
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.