IKZF (Ikaros)
The Ikaros family (IKZF1 Ikaros, IKZF2 Helios, IKZF3 Aiolos, IKZF4 Eos, IKZF5 Pegasus) are lymphoid zinc-finger proteins with an N-terminal C2H2 array that binds DNA and a C-terminal pair of fingers that mediates homo- and heterodimerization. They recognize a GGGAA core, present in lymphocyte-specific regulatory elements and also in pericentromeric satellite repeats, to which Ikaros localizes by direct DNA binding. IKZF1 is required for the development of all lymphoid lineages, is deleted in BCR-ABL1 and Ph-like B-cell acute lymphoblastic leukemia where it predicts poor outcome, and, with IKZF3, is the protein whose thalidomide-analog-induced degradation explains how lenalidomide works in myeloma. The GGGAA core is shared with ETS and NF-kB sites, which makes these motifs hard to trust.
The factors
Five human Ikaros-family proteins are encoded by IKZF1 to IKZF5: Ikaros, Helios, Aiolos, Eos and Pegasus. Each has an N-terminal array of C2H2 zinc fingers that binds DNA and a separate C-terminal pair of fingers that is not used for DNA contact but mediates homodimerization and heterodimerization among family members 7969165. Alternative splicing of IKZF1 produces isoforms (Ik-1 through Ik-5 in the original description) carrying different combinations of the N-terminal fingers; these differ in sequence specificity and affinity, with the full-length forms binding the whole set of recognition sequences found in lymphocyte-specific regulatory elements and the short forms binding only a subset 7969165. The short forms are also poorly nuclear and are weak activators, which is the basis of their dominant-negative behavior 7969165,7585946.
Expression is lymphoid and largely restricted: IKZF1 and IKZF3 across B and T lineages, IKZF2 in regulatory T cells and some other compartments, IKZF4 in regulatory T cells, IKZF5 more broadly. Only IKZF1, IKZF2 and IKZF3 have motif models in the databases behind this family. The proteins work in complexes rather than alone, associating with NuRD (Mi-2/CHD4) and with histone deacetylases, and they localize to pericentromeric heterochromatin 10970879,29278858.
The motif
The agreed core is GGGAA, written TTCCC on the other strand 7969165. Extended forms appear in the databases as TGGGAATAC and similar; the representative archetype for this family has that consensus. None of these carries much information beyond the five-base core, and that is the central problem for this family, because GGGAA is not distinctive. It is contained within the ETS core GGA(A/T) and within the NF-kB half-site GGGRNNYYCC, so an Ikaros matrix, an ETS matrix and an NF-kB matrix can all score the same DNA.
The collection makes this concrete. The JASPAR IKZF3 matrix has the consensus CAGGAAGTG, whose nearest neighbors across all four databases are ETS models rather than Ikaros models; the JASPAR IKZF2 matrix, AGGAAG, likewise resembles ETS more than it resembles the rest of its own family; and two Vierstra archetypes in this family are explicitly named for mixed groups, IKZF with ZNF and IKZF with NFKB. Family cohesion is correspondingly low. Ikaros also binds within centromeric DNA repeats, and targeting to pericentromeric foci depends on the DNA-binding finger domain together with the dimerization domain 10970879, so part of the genuine binding repertoire is repetitive sequence that motif scanners will report en masse.
Where it acts
Ikaros-family sites are found at promoters and enhancers of lymphoid genes, including the immunoglobulin and T-cell-receptor loci, and co-occur with EBF1, PU.1/SPI1, RUNX1 and E2A at B- and T-lineage regulatory elements 29278858,39017838. Because PU.1 is itself an ETS factor that prefers a purine-rich GGAA core, Ikaros and PU.1 sites are often indistinguishable in sequence at the same elements, and composite occupancy is common. The other distinctive location is pericentromeric heterochromatin, where Ikaros accumulates in foci; this was shown to require direct DNA binding rather than recruitment through Mi-2 10970879, and is consistent with Ikaros sites inside satellite repeats. Genome-wide, a five-base core occurs on the order of a million times, so scanning alone cannot define an Ikaros cistrome.
Biology and disease
Ikzf1-null mice lack B cells and have profound defects across lymphoid lineages, establishing IKZF1 as required for the development of all lymphoid lineages 7923373; mice expressing dominant-negative isoforms develop T-cell leukemia 7585946. In humans, focal deletions and point mutations of IKZF1 are a defining feature of BCR-ABL1-positive B-cell acute lymphoblastic leukemia and of Ph-like ALL 18408710, and IKZF1 deletion is associated with a substantially worse outcome 19129520, which is why it is now assayed in risk stratification. Germline dominant-negative IKZF1 variants cause a combined T-, B- and myeloid-cell immunodeficiency 29889099, and other IKZF1 variants cause common variable immunodeficiency 26981933.
The best-known pharmacology is degradation, not transcription. Thalidomide, lenalidomide and pomalidomide act as molecular glues that recruit a zinc finger of IKZF1 and IKZF3 to cereblon, the substrate receptor of the CRL4-CRBN ubiquitin ligase, causing their selective degradation; this accounts for the activity of these drugs in multiple myeloma 24292625,24292623. A systematic screen of the human C2H2 zinc-finger proteome identified eleven such degrons and predicted that more than 150 zinc fingers can engage the drug-CRBN interface 30385546, which turned the Ikaros story into a general strategy for targeting zinc-finger transcription factors.
Reading a hit
Treat an Ikaros hit as a weak, ambiguous call. Three problems stack. First, the informative core is five bases, so matches are everywhere. Second, that core is shared with ETS and NF-kB, and the databases themselves are contaminated: at least one JASPAR Ikaros matrix in this family is closer to ETS models than to Ikaros models, and two archetypes are named for IKZF together with ZNF or NFKB. If you get an Ikaros hit at a lymphoid enhancer, PU.1 or another ETS factor is at least as likely to be the direct binder. Third, no sequence argument distinguishes IKZF1 from IKZF2 or IKZF3, and IKZF4 and IKZF5 have no models here. Ikaros models are ChIP- and selection-derived in these collections rather than backed by a structure of the protein on DNA, and Ikaros works in NuRD and HDAC complexes, so indirect binding is a live possibility for any individual peak. A credible Ikaros call needs lineage-appropriate expression, a peak from a well-controlled experiment, and ideally loss of the peak on IKZF1 degradation rather than a motif match.
How this was worked out
- 1994mediumcDNA isolation and binding-site analysis of isoforms; germline gene targeting in mouse7969165 7923373
The Ikaros gene is shown to encode a set of alternatively spliced zinc-finger proteins whose N-terminal finger combinations differ in specificity and affinity, and Ikzf1-null mice are shown to lack all lymphoid lineages, so the family is defined simultaneously as a variable DNA-binding activity and as a requirement for lymphopoiesis.
Still open then: The recognition sequence was a short purine-rich core whose specificity relative to other GGAA-binding factors was not addressed.
Mice expressing a dominant-negative Ikaros isoform develop T-cell leukemia and lymphoma rapidly, linking loss of Ikaros DNA-binding function to malignancy rather than only to development.
- 2000mediumDomain mutagenesis with imaging of pericentromeric localization, binding assays on satellite repeats10970879
Ikaros is shown to be targeted to pericentromeric heterochromatin by its own DNA-binding fingers together with the dimerization domain, not by recruitment through Mi-2, so part of its genuine binding repertoire is repeat DNA.
Still open then: What fraction of genome-wide Ikaros signal comes from repeats rather than single-copy regulatory elements.
Genome-wide copy-number analysis finds focal IKZF1 deletions, often removing the N-terminal fingers and leaving a dominant-negative product, in most BCR-ABL1-positive B-cell acute lymphoblastic leukemias.
IKZF1 deletion is shown to mark a high-risk subgroup of B-ALL with substantially worse outcome, moving the gene into clinical risk stratification.
- 2014highQuantitative proteomics, CRBN binding and ubiquitination assays, cell-based degradation24292625 24292623
Two groups show that lenalidomide works by making cereblon bind IKZF1 and IKZF3 and degrade them, which reinterprets a decades-old drug as a targeted removal of these factors rather than a transcriptional modulator.
- 2016mediumExome sequencing in patient cohorts with functional assays of variant proteins26981933 29889099
Germline IKZF1 variants are shown to cause human immunodeficiency: haploinsufficiency presenting as common variable immunodeficiency with progressive B-cell loss, and, in a later series, dominant-negative variants causing combined T-, B- and myeloid-cell disease.
IKAROS is shown to assemble lineage-specific chromatin interactions over megabase distances in B-cell precursors, with contacts from IKAROS-bound enhancers crossing CTCF boundaries, which recasts the family as a genome-organizing activity rather than only a promoter and enhancer repressor.
Still open then: Whether these long-range interactions require the GGGAA core at the contacting elements or the dimerization and NuRD interfaces alone.
Still unresolved
- Whether the matrices labeled IKZF in these databases measure Ikaros specificity at all is unsettled: the JASPAR IKZF3 and IKZF2 models score closer to ETS matrices than to other Ikaros models, and PU.1 and other ETS factors occupy the same lymphoid elements, so co-occupancy and indirect binding cannot be excluded for any of them.
- No structure of an Ikaros-family protein on DNA has been reported, so which fingers read the GGGAA core, and why the splice isoforms differ in specificity, remain inferred from binding assays rather than observed.
- How much of Ikaros occupancy in vivo is at satellite and other repeat DNA rather than at single-copy regulatory elements has never been quantified, although pericentromeric targeting is known to require its own DNA-binding domain.
- Whether the long-range chromatin interactions attributed to IKAROS depend on sequence-specific binding at the contacting elements, or on its dimerization and NuRD interfaces, is open, and this determines whether a motif hit is informative about genome organization at all.
The informative core is only GGGAA, which is shared with ETS (GGAA) and NF-kB (GGGAA) sites, so an Ikaros match is rarely attributable to an Ikaros protein. The databases show the contamination directly: the JASPAR IKZF3 matrix here has ETS models as its nearest neighbors across all four sources, the IKZF2 matrix also resembles ETS more than its own family, and two Vierstra archetypes are named IKZF/ZNF and IKZF/NFKB. Family cohesion is low as a result. No model distinguishes IKZF1 from IKZF2 or IKZF3, and IKZF4 and IKZF5 have none. Ikaros additionally binds satellite repeats and acts within NuRD and HDAC complexes, so peaks need not reflect its own specificity.
References
- Molnár A, Georgopoulos K. The Ikaros gene encodes a family of functionally diverse zinc finger DNA-binding proteins. Mol Cell Biol 1994;14:8292-303. PMID 7969165
- Georgopoulos K, Bigby M, Wang JH, et al. The Ikaros gene is required for the development of all lymphoid lineages. Cell 1994;79:143-56. PMID 7923373
- Molnár A, Wu P, Largespada DA, et al. The Ikaros gene encodes a family of lymphocyte-restricted zinc finger DNA binding proteins, highly conserved in human and mouse. J Immunol 1996;156:585-92. PMID 8543809
- Cobb BS, Morales-Alcelay S, Kleiger G, et al. Targeting of Ikaros to pericentromeric heterochromatin by direct DNA binding. Genes Dev 2000;14:2146-60. PMID 10970879
- Mullighan CG, Miller CB, Radtke I, et al. BCR-ABL1 lymphoblastic leukaemia is characterized by the deletion of Ikaros. Nature 2008;453:110-4. PMID 18408710
- Mullighan CG, Su X, Zhang J, et al. Deletion of IKZF1 and prognosis in acute lymphoblastic leukemia. N Engl J Med 2009;360:470-80. PMID 19129520
- Boutboul D, Kuehn HS, Van de Wyngaert Z, et al. Dominant-negative IKZF1 mutations cause a T, B, and myeloid cell combined immunodeficiency. J Clin Invest 2018;128:3071-3087. PMID 29889099
- Krönke J, Udeshi ND, Narla A, et al. Lenalidomide causes selective degradation of IKZF1 and IKZF3 in multiple myeloma cells. Science 2014;343:301-5. PMID 24292625
- Lu G, Middleton RE, Sun H, et al. The myeloma drug lenalidomide promotes the cereblon-dependent destruction of Ikaros proteins. Science 2014;343:305-9. PMID 24292623
- Sievers QL, Petzold G, Bunker RD, et al. Defining the human C2H2 zinc finger degrome targeted by thalidomide analogs through CRBN. Science 2018;362. PMID 30385546
- Heizmann B, Kastner P, Chan S. The Ikaros family in lymphocyte development. Curr Opin Immunol 2018;51:14-23. PMID 29278858
- Kastner P, Chan S. IKAROS Family Transcription Factors in Lymphocyte Differentiation and Function. Adv Exp Med Biol 2024;1459:33-52. PMID 39017838
- Winandy S, Wu P, Georgopoulos K. A dominant mutation in the Ikaros gene leads to rapid development of leukemia and lymphoma. Cell 1995;83:289-99. PMID 7585946
- Kuehn HS, Boisson B, Cunningham-Rundles C, et al. Loss of B Cells in Patients with Heterozygous Mutations in IKAROS. N Engl J Med 2016;374:1032-1043. PMID 26981933
- Hu Y, Salgado Figueroa D, Zhang Z, et al. Lineage-specific 3D genome organization is assembled at multiple scales by IKAROS. Cell 2023;186:5269-5289.e22. PMID 37995656
8 motifs in this family — JASPAR 3 · CIS-BP 1 · Vierstra 4
| motif | database | consensus | bits |
|---|---|---|---|
| Ikzf3 (MA1992.2)looks like ETS_EHF | JASPAR | CAGGAAGTG | 11.6 |
| IKZF2 (MA2326.1) | JASPAR | AGGAAG | 11.2 |
| IKZF1 (MA1508.2) | JASPAR | AACAGGAA | 10.9 |
| AC0544:IKZF:C2H2_ZF | Vierstra | tCTCCCAAA | 10.7 |
| M08320_2.00 | CIS-BP | TTCCCnncc | 10.4 |
| AC0692:IKZF:C2H2_ZF | Vierstra | tGGGAAtac | 10 |
| AC0540:IKZF/ZNF:C2H2_ZF | Vierstra | cTCCCAgvangc | 9.9 |
| AC0693:IKZF/NFKB:Rel | Vierstra | TTCCCc | 7.6 |
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.