ZBTB18 (RP58)
ZBTB18, first described as RP58, is a BTB/POZ zinc-finger repressor that binds an E-box-like site, (A/C)ACATCTG(G/T)(A/C), whose CATCTG core reads CAGATG on the opposite strand. It is a C2H2 protein, not a bHLH one, so the E-box it occupies is a site that basic helix-loop-helix activators can also use - which makes competition, not just repression, part of how it works. RP58 concentrates in condensed chromatin and represses target promoters through its POZ domain. In the brain it controls the multipolar-to-bipolar transition of newborn cortical neurons and is required for cerebellar growth; human ZBTB18 mutations cause autosomal dominant intellectual disability and figure in 1q43-q44 deletion syndrome with corpus callosum malformation.
The factors
ZBTB18 is a one-member family in this atlas. The gene, also known as RP58 and ZNF238, encodes a 522-residue protein with an N-terminal POZ/BTB domain - the feature that first placed it alongside BCL6 - and four C-terminal C2H2 zinc fingers 9756912. The POZ domain mediates homodimerization and autonomous repression; it does not contact DNA. An early and still striking observation is that almost all RP58 protein localizes to condensed chromatin regions by immunogold electron microscopy 9756912, which was the basis for proposing that this factor links sequence-specific repression to chromosome organization. ZBTB18 is expressed most prominently in the developing and adult central nervous system, in cortical and cerebellar neurons, and also in skeletal muscle and other tissues 32959890; it is not a lineage-restricted factor in the way GFI1 or ThPOK are. In neurons it behaves as a differentiation-promoting repressor: it shuts down programs belonging to other cell identities rather than activating neuronal genes directly.
The motif
The binding site was determined by cyclic amplification and selection of targets with recombinant protein, giving the high-affinity consensus (A/C)ACATCTG(G/T)(A/C) 9756912. The internal CATCTG is a CANNTG E-box variant, and read on the opposite strand the same site presents CAGATG as its core. This is worth stating carefully because the two renderings look like different motifs: the JASPAR HT-SELEX matrix MA0698.2 reads TCCAGATGTT and the HOCOMOCO grade-A model reads CCAGATGTG, both of which are the reverse complement of the selected consensus, not a disagreement with it 23332764,37971293. Note also that the family definition used in this atlas describes the site as "E-box-like CACATG"; the experimentally selected core is CATCTG (equivalently CAGATG), with an A at the position where CACATG has a C, so the shorter label is slightly off and should not be used for scanning. Information content is 11-15 bits over nine or ten positions - better than a bare 6 bp E-box but still modest. Four adjacent fingers could cover more sequence than the consensus specifies, and the flanking A/C and G/T degeneracies suggest that only two or three fingers contribute strong base contacts. The family data flags one entry as inconsistent: the CIS-BP model M08257 for ZBTB18 reads CGCAGCTGCG, a canonical CAGCTG bHLH E-box whose nearest neighbors across databases fall in the NEUROD/ATOH family rather than with ZBTB18, and which matches no other ZBTB18 model. With only three matrices in the family, that single outlier drags the measured cohesion down to 0.79.
Where it acts
ZBTB18 acts mainly at promoters of genes belonging to alternative differentiation programs. In developing neurons it represses transcriptional programs linked to non-neuronal cell identity, including programs that correspond to glioblastoma subtypes 33903225, and the sites it uses are promoter-proximal E-boxes rather than distal enhancer clusters. Because the core is a CANNTG box with no CpG, matches are not concentrated in CpG islands and occur throughout the genome; a nine-position match with two degenerate positions still arises tens of thousands of times by chance in 3 Gb. The important co-occurrence to keep in mind is with bHLH factors: NEUROG2, NEUROD and ASCL1 bind overlapping CANNTG boxes in the same neuronal promoters, and the sequence alone cannot say which occupied a site.
Biology and disease
Mouse Rp58 mutants show defective cortical development: newborn neurons fail to complete the multipolar-to-bipolar transition needed for radial migration 23395638, and cerebellar growth and patterning are disrupted, with abnormal development of both glutamatergic and GABAergic neurons 22513377. In humans, heterozygous ZBTB18 loss-of-function and missense variants cause autosomal dominant non-syndromic intellectual disability 22 (MRD22), with developmental delay, seizures and frequent corpus callosum abnormalities 38215144; ZBTB18 lies in the critical region of 1q43-q44 deletion syndrome, and its loss is a principal candidate for the corpus callosum and microcephaly phenotypes in that contiguous gene syndrome 36192753. Missense variants have been shown to differ in how much they impair DNA binding versus repression, so the allelic series is informative rather than uniform 35083747. In glioblastoma, ZBTB18 behaves as a tumor suppressor that is silenced epigenetically, and restoring it represses mesenchymal and lipid-metabolic programs 28512252.
Reading a hit
The CATCTG/CAGATG core makes this an E-box family, with all the attendant ambiguity. Candidates for any CANNTG match include the bHLH factors of the NEUROD/ATOH, ASCL, NEUROG, TCF3/4/12, MYOD and TWIST groups, and among C2H2 proteins the SNAIL, ZEB and INSM-related repressors. An E-box hit is not evidence of a bHLH factor, and equally it is not evidence of ZBTB18. What tilts the balance toward ZBTB18 is the specific internal composition: CATCTG rather than CACCTG, CAGCTG or CACGTG, with the flanking A and G/T. The best-supported matrix here is JASPAR MA0698.2, which is HT-SELEX-derived and thus direct in vitro evidence 23332764, and the HOCOMOCO model is grade A with PSM evidence, so "ZBTB18 binds this sequence" is well established 37971293. Do not use the CIS-BP entry M08257 (CGCAGCTGCG): it looks like a NEUROD/ATOH-class site, and scanning with it will return bHLH sites labelled as ZBTB18. Finally, since ZBTB18 is one gene, there is no paralogue problem inside the family - but the outside-the-family problem is severe, and a hit should be reported as "E-box consistent with ZBTB18" rather than as a ZBTB18 site.
How this was worked out
RP58 was cloned as a POZ-domain zinc-finger protein, its high-affinity site was selected as (A/C)ACATCTG(G/T)(A/C) - an E-box core read by zinc fingers - and the protein was shown to repress from that site and to sit almost entirely in condensed chromatin.
Still open then: Whether the protein competes with bHLH activators for the same boxes in cells.
Neural deletion of Rp58 in mice caused severe cerebellar hypoplasia with failure of Purkinje neuron, Bergmann glia and granule neuron development, showing that the factor is required for both glutamatergic and GABAergic lineages rather than for one neuronal class.
RP58 was shown to be required cell-autonomously for the multipolar-to-bipolar transition of newborn cortical neurons, acting by repressing Ngn2 and the Ngn2-Rnd2 pathway, which put a proneural bHLH gene directly downstream of this repressor.
HT-SELEX returned the same site from the other strand, TCCAGATGTT, independently confirming the 1998 selection and giving the collections an in vitro model for ZBTB18.
- 2017mediummethylation and expression profiling of tumors, re-expression and xenograft assays28512252
ZBTB18 was found to be silenced by promoter hypermethylation specifically in the mesenchymal subtype of glioblastoma, and re-expression suppressed the aggressive phenotype, framing it as an epigenetically inactivated tumor suppressor.
- 2021mediummouse conditional knockouts with RNA sequencing and gene set enrichment, glioma stem cell experiments33903225
Transcriptome analysis of Rp58 mutant cortical neurons showed that the repressed program consists largely of genes belonging to other cell identities, including mesodermal, pluripotency and glioma signatures, so the factor maintains neuronal identity by exclusion rather than by activating neuronal genes.
- 2022mediumvariant functional assays in neural cells, array CGH and clinical correlation35083747 36192753
Functional dissection of patient missense alleles showed that they differ in whether they impair DNA binding or repression, and reviews of 1q43q44 deletions placed ZBTB18 as the principal candidate for the corpus callosum and microcephaly components of that contiguous gene syndrome.
Still open then: Whether any single gene in the 1q43q44 interval accounts for the callosal phenotype on its own.
Still unresolved
- The CIS-BP entry M08257 reads CGCAGCTGCG and its cross-database neighbors are NEUROD/ATOH motifs, and it is unresolved whether ZBTB18 has a genuine secondary preference for a CAGCTG box or the model simply carries the wrong label; nothing in the literature reports CAGCTG binding by this protein.
- Whether ZBTB18 and proneural bHLH factors contest the same boxes in neurons is not established: ZBTB18 represses the Ngn2 gene, which is a regulatory relationship rather than site-level competition, and no experiment has shown the two protein classes exchanging at individual CANNTG sites.
- Which fingers of the four carry the specificity is unknown, since there is no structure of the ZBTB18 zinc-finger array on DNA and the degeneracy at the flanking positions of the selected consensus only suggests that two or three fingers make strong base contacts.
- It remains unsettled how patient missense alleles cause disease, because the tested variants fall on a spectrum between loss of DNA binding and loss of repression, and whether some act as dominant negatives through the dimerizing POZ domain has not been determined.
The site is an E-box variant, so the crowd of bHLH factors plus SNAIL, ZEB and INSM-related C2H2 repressors all compete for interpretation of a match; report an E-box, then argue for ZBTB18 from the internal CATCTG composition and the cell type. The commonly quoted "CACATG" shorthand for this family is not the selected consensus and should not be scanned. Trust JASPAR MA0698.2 (HT-SELEX) and the HOCOMOCO grade-A model, which are reverse-complement renderings of the same site; discard the CIS-BP entry CGCAGCTGCG, which the cross-database analysis places with NEUROD/ATOH and which accounts for this family's low internal cohesion. Single gene, so no paralogue ambiguity within the family.
References
- Aoki K, Meng G, Suzuki K, et al. RP58 associates with condensed chromatin and mediates a sequence-specific transcriptional repression. J Biol Chem 1998;273:26698-704. PMID 9756912
- Ohtaka-Maruyama C, Hirai S, Miwa A, et al. RP58 regulates the multipolar-bipolar transition of newborn neurons in the developing cerebral cortex. Cell Rep 2013;3:458-71. PMID 23395638
- Baubet V, Xiang C, Molczan A, et al. Rp58 is essential for the growth and patterning of the cerebellum and for glutamatergic and GABAergic neuron development. Development 2012;139:1903-9. PMID 22513377
- Xiang C, Frietze KK, Bi Y, et al. RP58 Represses Transcriptional Programs Linked to Nonneuronal Cell Identity and Glioblastoma Subtypes in Developing Neurons. Mol Cell Biol 2021;41:e0052620. PMID 33903225
- Okado H. Nervous system regulated by POZ domain Krüppel-like zinc finger (POK) family transcription repressor RP58. Br J Pharmacol 2021;178:813-826. PMID 32959890
- Heng JI, Viti L, Pugh K, et al. Understanding the impact of ZBTB18 missense variation on transcription factor function in neurodevelopment and disease. J Neurochem 2022;161:219-235. PMID 35083747
- Khadija B, Rjiba K, Dimassi S, et al. Clinical and molecular characterization of 1q43q44 deletion and corpus callosum malformations: 2 new cases and literature review. Mol Cytogenet 2022;15:42. PMID 36192753
- Yang F, Ding Y, Wang Y, et al. A de novo variant in ZBTB18 gene caused autosomal dominant non-syndromic intellectual disability 22 syndrome: A case report and literature review. Medicine (Baltimore) 2024;103:e35908. PMID 38215144
- Fedele V, Dai F, Masilamani AP, et al. Epigenetic Regulation of ZBTB18 Promotes Glioblastoma Progression. Mol Cancer Res 2017;15:998-1011. PMID 28512252
- Jolma A, Yan J, Whitington T, et al. DNA-binding specificities of human transcription factors. Cell 2013;152:327-39. PMID 23332764
- 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
3 motifs in this family — JASPAR 1 · HOCOMOCO 1 · CIS-BP 1
| motif | database | consensus | bits |
|---|---|---|---|
| ZBTB18 (MA0698.2) | JASPAR | TCCAGATGTt | 15.3 |
| M08257_2.00looks like NEUROD_ATOH | CIS-BP | CGCAGCTGCG | 14 |
| ZBT18.H14CORE.0.PSM.APSMA | HOCOMOCO | CCAGaTGtg | 10.9 |
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.