Genemoderateq166-the-akap95-dpy30-axis

ZNF384 - a C2H2 transcription factor that is the invariant 3' half of a whole B-ALL fusion subtype, including AKAP8::ZNF384

also called CIZ · NMP4 · CAGH1 · TNRC1 · Zinc finger protein 384

ZNF384 is a DNA-binding protein with eight zinc fingers. In one subtype of B-cell leukaemia it keeps its whole DNA-binding half and gets a new front end donated by any of about twenty different partner genes - one of which is AKAP8.

The same thing, for a scientist

**Written minimally and deliberately, to let the 3D domain viewer draw the 3' lane of `fusion-akap8-znf384`.** ZNF384 (UniProtKB **Q8TF68**, **577 aa**) is a C2H2 zinc-finger transcription factor with **eight tandem fingers spanning residues 228-455**, binding a [GC]AAAAA consensus and reported to regulate MMP1, MMP3, MMP7 and COL1A1 promoters; nuclear [E1]. It is the recurrent 3' partner of a defined BCP-ALL subtype with more than nineteen reported 5' partners, dominated by chromatin and scaffold proteins - EP300, CREBBP, TCF3, TAF15, EWSR1, ARID1B, SMARCA4/2, NIPBL, EHMT1 and **AKAP8** [E2]. In the UKALL14 whole-genome series, five of 52 adult B-other cases carried a ZNF384 fusion: **EP300::ZNF384 n = 4, AKAP8::ZNF384 n = 1** [E3]. **This entry makes no claim about ZNF384 biology beyond the database record and does not attempt the subtype's disease biology**, which belongs to a B-ALL beat, not to this one.

What it looks like

ZNF384 - a C2H2 transcription factor that is the invariant 3' half of a whole B-ALL fusion subtype, including AKAP8::ZNF384577 aa

loading AF-Q8TF68-F1…

AlphaFold AF-Q8TF68-F1 ↗

colourshape

This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.

very high (90+)confident (70–90)low (50–70)very low (<50)

I have never read one of these pictures before

A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.

Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.

Anything in yellow sticks is a bound small molecule, which is usually the drug.

Go deeper

Scope of this entry, stated so nobody mistakes it for a beat

This node exists because fusion-akap8-znf384 was the only new fusion entry this cycle and scripts/check-joins.mjs correctly reported it as one of two genuine join gaps in the whole graph - no protein entry for 3' gene ZNF384 - which means the 3D viewer could not draw its 3' lane. The standing brief permits creating a minimal but fully cited node with the canonical id so that references resolve, and that is all this is.

What is deliberately NOT here: the ZNF384-rearranged ALL subtype's immunophenotype, its characteristic CD13/CD33 myeloid-antigen expression, its outcome data beyond the one prognostic sentence, its transcriptional mechanism, or any partner-by-partner comparison. Those need a dedicated beat with primary sources, and quoting them from a review here would put unverified content behind a resolvable-looking citation.

And one thing a reader should notice: the 5' partner list is overwhelmingly transcriptional coactivators and chromatin remodellers, and the architecture is invariant - ZNF384 keeps its DNA-binding half and the partner supplies the N-terminus. That is the same grammar as KMT2A fusions, where the DNA-address half is retained and the partner supplies an activating module, except that here the retained half is the 3' gene. Whether that parallel is load-bearing or merely rhetorical is exactly the sort of question a B-ALL beat should take, and this entry does not settle it.

Facts and figures

uniprot	Q8TF68
length AA	577
alphafold ID	AF-Q8TF68-F1
hgnc symbol	ZNF384

domains

PDB IDS

[]

domains — the numbers behind the chart

[
  {
    "name": "ZF1 (C2H2 zinc finger)",
    "start": 228,
    "end": 250
  },
  {
    "name": "ZF2 (C2H2 zinc finger)",
    "start": 256,
    "end": 278
  },
  {
    "name": "ZF3 (C2H2 zinc finger)",
    "start": 284,
    "end": 306
  },
  {
    "name": "ZF4 (C2H2 zinc finger)",
    "start": 317,
    "end": 339
  },
  {
    "name": "ZF5 (C2H2 zinc finger)",
    "start": 345,
    "end": 367
  },
  {
    "name": "ZF6 (C2H2 zinc finger)",
    "start": 373,
    "end": 397
  },
  {
    "name": "ZF7 (C2H2 zinc finger)",
    "start": 403,
    "end": 425
  },
  {
    "name": "ZF8 (C2H2 zinc finger)",
    "start": 433,
    "end": 455
  }
]

localization

{
  "compartment": "nucleus",
  "subnuclear": [],
  "shown_for_fusion": "NO - nobody has looked at any ZNF384 fusion protein's localisation in this beat's reading"
}

measurements

[
  {
    "quantity": "length",
    "subject": "ZNF384 Q8TF68 canonical sequence",
    "value": 577,
    "units": "aa",
    "n": 1,
    "universe": "UniProtKB reviewed human entry, retrieved live 2026-09-24 via an organism-restricted gene_exact search returning exactly one record",
    "null": null,
    "note": "the accession was fetched, not recalled"
  },
  {
    "quantity": "count",
    "subject": "annotated C2H2 zinc fingers in ZNF384",
    "value": 8,
    "units": "domains",
    "n": null,
    "universe": "UniProtKB Q8TF68 feature table; the array spans residues 228-455",
    "null": null,
    "note": "the array occupies 39.4% of the protein. CORRECTED cycle 158: it is NOT always retained intact - exon-8 exclusion produces an in-frame transcript lacking zinc fingers 4 and 5 (Dickerson KM et al., Blood Cancer Discov 2022;3(3):240-263, PMID 35247902, DOI 10.1158/2643-3230.BCD-21-0163, Fig 6B, read via the PMC rendering). The UniProt feature table above is correct as a record of the reference protein; only the retention claim needed qualifying. The finger COUNT is itself disputed in the literature (8 / 6 / 'five to eight'), and no case-level census has been published"
  }
]

How this could bite us back

This is a stub written to make a figure lane resolve. Anyone building an argument about ZNF384 on it is building on a UniProt record and one paragraph of one cohort paper, and the entry's own scope section says so.

Still open

Where does the ZNF384 breakpoint fall relative to the zinc-finger array at 228-455? The subtype's defining architecture is said to retain the complete ZNF384 coding sequence, which would place every breakpoint 5' of residue 228, and this beat verified that for no case.
Does the ZNF384-rearranged subtype deserve a beat of its own? Its 5' partner list is almost entirely chromatin coactivators and remodellers, which is the same partner grammar as KMT2A - and this graph has no ZNF384 coverage at all beyond this stub.

Every source behind this entry (3)

E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
UniProtKB Q8TF68 (ZN384_HUMAN) is 577 aa, gene ZNF384, synonyms CAGH1, CIZ, NMP4 and TNRC1. Its feature table records eight C2H2-type zinc fingers at 228-250, 256-278, 284-306, 317-339, 345-367, 373-397, 403-425 and 433-455, and disordered regions at 171-225 and 501-550. Function: a transcription factor binding the consensus [GC]AAAAA, reported to bind and regulate the MMP1, MMP3, MMP7 and COL1A1 promoters (by similarity). Subcellular location: nucleus.
UniProtKB entry Q8TF68 (ZN384_HUMAN) - sequence, feature table, function and subcellular location; retrieved live 2026-09-24 by q166 · UniProt Consortium 2026
What this evidence cannot show
The accession was confirmed by a reviewed, organism-restricted gene_exact search returning exactly one human record, per rule 88 - it was NOT taken from memory. The promoter-target function is flagged 'By similarity' in UniProt, i.e. inferred from orthologue data rather than experimentally coded in human, and no paper behind it was opened here. The nuclear localisation comment carries no ECO:0000269 evidence code in this record.
Evidence that argues against this (1)
- Rule 91 has not been discharged for this entry: none of the papers behind the UniProt annotations was opened, so the function and localisation lines are being carried as database pointers rather than as read sources. That is acceptable for a node whose only job is to supply zinc-finger coordinates to a figure, and it is not acceptable for any claim about ZNF384's role in leukaemia - none of which this entry makes. UniProt Consortium 2026
E2moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review
More than 19 ZNF384 fusion partners have been reported in ALL, including EP300, CREBBP, TCF3, TAF15, EWSR1, ARID1B, SMARCA4, SMARCA2, SYNRG, CLTC, BMP2K, NIPBL, AKAP8, C11orf74, DDX42, ATP2C1, EHMT1 and TEX41; ZNF384 rearrangement in ALL was first reported in 2002 and patients commonly have a good prognosis.
ZNF384-Related Fusion Genes in Acute Lymphoblastic Leukemia. · Zhu L, Bai W, Cheng Q, Fang J · Cancer Control 2023
PMID 37306722 DOI
What this evidence cannot show
A REVIEW, cited as a map only, and abstract-level at that. The partner list was not traced to primary sources except for AKAP8 and EP300, both of which appear in the UKALL14 series cited in E3. No frequency, outcome or mechanistic claim should rest on this item.
Evidence that argues against this (1)
- A partner list compiled in a review flattens very different evidence: some partners rest on multi-case cohorts and others on single case reports, and the review's framing gives no way to tell which is which. Counting '19 partners' as though it described a uniform set is the error this item most invites, and the honest reading is that ZNF384 has one well-populated partner (EP300) and a long tail of thinly evidenced ones, of which AKAP8 is one. Leongamornlert D et al. 2023
E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In whole-genome sequencing of 52 adult B-other BCP-ALL cases from UKALL14, five carried a ZNF384 fusion: EP300::ZNF384 (n = 4) and AKAP8::ZNF384 (n = 1). Four of the five were aged 40 or under, against 18 of 47 in the remainder.
Diagnostic utility of whole genome sequencing in adults with B-other acute lymphoblastic leukemia. · Leongamornlert D, Gutierrez-Abril J, Lee S, Barretta E, et al. · Blood Adv 2023
PMID 36867579 DOI
What this evidence cannot show
Full text retrieved and this paragraph read directly; the rest of the paper was not read. The cohort is ADULT B-other ALL - a pre-selected residual group after standard-of-care genetics - so these proportions are not subtype frequencies in unselected ALL, and they are not paediatric, which is this project's population.
Evidence that argues against this (1)
- The denominator is the wrong one for almost any question a reader will bring. B-other is defined by the FAILURE of standard-of-care genetics to assign a subtype, so ZNF384 fusions are enriched in it by construction relative to unselected BCP-ALL. Quoting '5 of 52' as a frequency would overstate the subtype substantially, and quoting '1 of 52' for AKAP8::ZNF384 more so. Leongamornlert D et al. 2023

Connected to

fuses with →AKAP8::ZNF384 - the same protein that MLL-rearranged leukaemia depends on is itself a rare B-ALL fusion partner, on the 5' side, donating its condensate-forming IDR
the 3' partner, retaining its complete eight-finger C2H2 array at 228-455
related to →AKAP8 (AKAP95) - the only nuclear A-kinase anchoring protein, and the one that turned out to be required for MLL-rearranged leukaemia
the two halves of AKAP8::ZNF384; written as a minimal stub so the 3D viewer can draw the fusion's 3' lane
← fuses withTAF15 (RBP56, TAF2N) - the FET-family, prion-like partner that donates a low-complexity N-terminus to ZNF384
TAF15::ZNF384, t(12;17); 17 of 193 partner-typed Ponte di Legno cases (9%)
← analogous toZNF362 - the paralogue that makes the same leukaemia with the same partners, and that explains the six-zinc-finger convention
paralogue; ZNF362 fusions produce the same expression signature and share the partners TAF15 and SMARCA2 with similar breakpoints
← part ofZNF384::TEX41 - the case that breaks the family's architecture: ZNF384 is the 5' partner, the 3' partner is a lncRNA, and the fusion is out of frame
ZNF384 is the 5' partner here - the one reported case that inverts the family's architecture
← depends onLast-exon indels that truncate the ZNF384 C-terminus phenocopy the fusion - so the fusion partner may not be the necessary element
in-frame indels in the last exon of ZNF384, p.393_397 KHNPP>RAG and p.483_484 CS>YFEWDPND
← depends onThe ZNF384-fusion immunophenotype - weak CD10 with aberrant CD13/CD33 - and the one marker that separates it from KMT2A-rearranged disease runs the OPPOSITE way to expectation
weak CD10 with aberrant CD13/CD33 is common to the whole ZNF384 fusion family
← depends onZNF384 fusions make lineage-ambiguous leukaemia by acting in a haematopoietic stem cell - and the founding lesion is in the HSC itself, unlike Ph-like B-ALL
ZNF384 fusion oncoproteins drive haematopoietic expansion, myeloid skewing and self-renewal from HSPCs
← depends onFLT3 in ZNF384-rearranged leukaemia - the fusion builds its own enhancer 25 kb upstream, and gilteritinib works in xenografts
CRISPR knockout of EP300-ZNF384 cuts FLT3 expression 48% and reduces gilteritinib sensitivity ~100-fold
← part ofEP300::ZNF384 - the commonest ZNF384 fusion, and it throws away the acetyltransferase
the complete ZNF384 protein is retained, joined at exon 3, the first coding exon
← part ofTCF3::ZNF384 - the partner whose contribution is now actually demonstrated: steroid resistance, and it is not shared with EP300::ZNF384
almost all of the ZNF384 protein is retained
← part ofTAF15::ZNF384 - the FET-partner version, where the naming order in the literature is a trap and the founding experiment says the transactivation domain is not doing the work
the entire ZNF384 coding sequence is retained, joined in the intron between exons 2 and 3
← part ofEWSR1::ZNF384 - the same EWSR1 exon 7 fragment that drives Ewing sarcoma, fused instead to a zinc-finger protein, in B-ALL
the complete ZNF384 amino acid sequence, joined at exon 2 or exon 3
← part ofBMP2K::ZNF384 - a PROTEIN KINASE as the 5' partner, with its catalytic domain retained: the counterexample to 'the partner is always a chromatin coactivator'
BMP2K exon 14/15 to ZNF384 exon 3 at nucleotide 266; the complete ZNF384 coding sequence is retained
← related toThe ZNF384 family is not the inverse of KMT2A::AFF1 - it is the SAME grammar written in the opposite gene order, and three of the four claims in the inverse framing are false
scores the four claims of the inverse-architecture framing; three fail
← related toFour sources give ZNF384 three different zinc-finger counts - 8, 6 and 'five to eight' - and the paralogue's feature table explains where the 6 comes from
the entry records eight fingers 228-455 and calls the array retained intact in fusions; exon-8 exclusion makes the retention clause unsafe as written
← fuses withBMP2K (BIKe) - a nuclear serine/threonine kinase, and the ZNF384 fusion partner that is not a chromatin protein
BMP2K::ZNF384, exon 14/15 to ZNF384 exon 3; the kinase domain is retained
← related toHow common is ZNF384-rearranged leukaemia? Published answers span 1.6% to about 20% and the spread is almost entirely denominator, not biology
1.6-1.9% of consecutively diagnosed childhood BCP-ALL/B-myeloid MPAL; every larger published figure is a narrower denominator