EAF1 — the interactor with an actual subnuclear phenotype: an MLL fusion evicts it from Cajal bodies, and the bodies themselves fall apart

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.

• EAF1 is paralogue-buffered by EAF2 and has no KMT2A selectivity, so it is not a target; the value here is mechanistic and diagnostic, not therapeutic.
• A sequestration mechanism is hard to drug: restoring a mislocalised protein to a nuclear body is not something small molecules do, and the fusion would have to be removed to reverse it.
• Cajal bodies are required for snRNP biogenesis in all cells. Anything that restored or reinforced them would not be selective for the leukaemia.
• The subnuclear result is from 2001-2003 in murine MLL-ELL models and has not been replicated in a modern system or in human patient cells. Treating it as established for human KMT2A-rearranged leukaemia generally would over-read it.

1. Does KMT2A::AFF1 delocalise EAF1 from Cajal bodies as MLL-ELL does? The fusion retains no ELL-binding region of its own, so the prediction is no - and a negative result would be informative, because it would make the Cajal-body phenotype specific to ELL-partner fusions rather than general to KMT2A rearrangement.
2. Are Cajal bodies disrupted in KMT2A::AFF1 infant B-ALL blasts? This has never been looked at, is a coilin stain away, and would be the first subnuclear measurement in this disease.
3. Does Cajal body disruption contribute to leukemogenesis, or is it a bystander? Nothing in either 2001 or 2003 paper tests necessity.
4. The ECAS module is AFF4-ELL2-EAF1 and is distinct from P-TEFb-SEC. Does a KMT2A::AFF1 fusion recruit ECAS as well as SEC, and if so through which contact?
5. The 2003 paper carries a published correction on its fractionation result. What was corrected?

1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   EAF1 was isolated from a yeast two-hybrid screen with ELL; endogenous EAF1 and ELL co-immunoprecipitate from multiple cell lines and EAF1 also interacts with ELL2; EAF1 contains a serine/aspartate/glutamate-rich region homologous to the transcriptional activation domains of the MLL translocation partners AF4, LAF4 and AF5q31, and a similar activation domain is present in that region; endogenous EAF1 and ELL colocalise in a distinct nuclear speckled pattern by confocal microscopy; transfection of MLL-ELL delocalises EAF1 from the speckled distribution to a diffuse nucleoplasmic pattern; and in leukaemic cell lines derived from mice transplanted with MLL-ELL-transduced bone marrow, EAF1 speckles were not detected.

   EAF1, a novel ELL-associated factor that is delocalized by expression of the MLL-ELL fusion protein · Simone F, Polak PE, Kaberlein JJ, Luo RT, Levitan DA, Thirman MJ · Blood 98(1):201-9 2001
   PMID 11418481DOI
   ▸What this evidence cannot show

   Abstract fetched and read on 2026-09-19; full text not read. The fusion studied is **MLL-ELL**, not KMT2A::AFF1, so the delocalisation result does not transfer to this project's fusion without an experiment. Transfected MLL-ELL is over-expressed relative to a patient allele, and over-expression of any ELL-binding protein could in principle titrate EAF1. The murine leukaemia observation is qualitative (speckles 'not detected').

   ▸Evidence that argues against this (1)

   • The AF4-family homology the authors note cuts against a simple sequestration model. If EAF1 carries an activation domain resembling those of AF4, LAF4 and AF5q31, then in a KMT2A::AFF1 leukaemia the fusion itself supplies that function, and losing EAF1 from speckles might be inconsequential rather than pathogenic. The paper demonstrates mislocalisation, not that mislocalisation matters. Polak PE et al. 2003
2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   ELL and EAF1 are components of Cajal bodies: they colocalise with p80 coilin but do not exhibit a direct physical interaction with it; treatment with actinomycin D, DRB or alpha-amanitin disperses ELL and EAF1 from Cajal bodies, showing their localisation depends on active Pol II transcription; Cajal bodies are disrupted in MLL-ELL leukaemia, with EAF1 and p80 coilin delocalised in murine MLL-ELL leukaemia cells and in HeLa cells transiently transfected with MLL-ELL; and nuclear/cytoplasmic fractionation revealed diminished expression of p80 coilin and EAF1 in the nuclei of MLL-ELL leukaemia cells.

   ELL and EAF1 are Cajal body components that are disrupted in MLL-ELL leukemia · Polak PE, Simone F, Kaberlein JJ, Luo RT, Thirman MJ · Molecular Biology of the Cell 14(4):1517-28 2003
   PMID 12686606DOI
   ▸What this evidence cannot show

   Abstract fetched and read on 2026-09-19; full text not read. Note that the PubMed record for this paper carries a '[corrected]' marker on the fractionation sentence, indicating a published correction whose content was not retrieved in this cycle - the fractionation claim should be checked against the corrected version before being relied on. Same laboratory as [E1], same fusion (MLL-ELL), and the leukaemia cells are murine.

   ▸Evidence that argues against this (1)

   • Because Cajal body localisation is abolished by Pol II inhibition, any perturbation that lowers global transcription will disperse ELL and EAF1 from Cajal bodies. An MLL-ELL-transformed leukaemic cell differs from a normal cell in many ways besides the fusion, so 'Cajal bodies are disrupted in MLL-ELL leukaemia' is compatible with the disruption being a consequence of the transformed state rather than a direct effect of the fusion protein - although the transient-transfection arm in HeLa argues against that. Simone F et al. 2001
3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   Native chromatin proteomics in human cells identified ECAS, a bipartite assembly linking an AFF4-ELL2-EAF1 elongation module to a U2 snRNP-associated intron binding complex; ECAS occupies promoter-proximal chromatin and is molecularly distinct from the canonical P-TEFb-containing super elongation complex; acute perturbation of AQR/XAB2-dependent ECAS functions impairs productive gene-body elongation without altering pause release, increases intron retention and preferentially affects long genes; and in mouse haematopoietic cells Aqr perturbation profoundly impairs leukaemia initiation, maintenance and leukaemia stem cell function in MLL-rearranged leukaemia.

   A dual-module ECAS complex couples transcription elongation and RNA processing · Xiang Y, Wang H, Yin Z, Xiao R, Song Z, Lu R, Dai R, Ye J, Li C, Cui Y, Wang M, Lv J, Li Q, Wang F, Li Z, Zhang K, Zhang Y, Wu D, Fang P, Liang K · Molecular Cell, online ahead of print 16 September 2026 2026
   PMID 42748909DOI
   ▸What this evidence cannot show

   Abstract fetched and read on 2026-09-19; full text not read. Very recent and not independently replicated. The leukaemia phenotype is reported for **Aqr**, not for EAF1: EAF1's membership in the ECAS elongation module is a proteomic assignment, and no EAF1 perturbation in leukaemia is described in the abstract. INTERACTION PROVENANCE (Q76, 2026-09-21, live IntAct REST), which is what 'a proteomic assignment' amounts to in the curated record: all six IntAct records for the AFF4-EAF1 pair are psi-mi:MI:1060 spoke-expanded and typed MI:0914 association, from PMIDs 21729782, 28514442 and 33961781, and no PDB entry contains both Q96JC9 and Q9UHB7 - EAF1 has two PDB entries and neither holds AFF4. So EAF1's link to the AFF4 scaffold has never been observed as a binary interaction; its well-evidenced binary partner is ELL, from the two-hybrid and co-IP work in E1, and the ELL2-EAF1 pair is additionally in 7OKX (E4). That distinction matters for the ECAS model: EAF1's place in an AFF4-ELL2-EAF1 module could be entirely ELL2-mediated, and nothing in the record distinguishes that from a direct AFF4 contact. See ana-two-thirds-of-the-curated-interactome-is-a-row-nobody-observed.

   ▸Evidence that argues against this (1)

   • This project's own genome-wide scan does not support AQR as a KMT2A-selective dependency and points the other way (rank 17,654, difference +0.244), and EAF1 itself scores +0.047. Neither half of the ECAS elongation module that this graph can measure looks like a KMT2A-rearranged dependency in cell lines, which weakens the inference from ECAS membership to relevance. Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026
4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   RCSB PDB entry 7OKX, 'Structure of active transcription elongation complex Pol II-DSIF (SPT5-KOW5)-ELL2-EAF1 (composite structure)' at 3.3 A, models EAF1 (UniProt Q96JC9) residues 1-268, ELL2 (O00472) residues 1-640 and SPT5 (O00267) residues 1-1087 together on RNA polymerase II; the primary citation is Chen et al., Molecular Cell 2021, PMID 34265249, DOI 10.1016/j.molcel.2021.06.019, which reports allosteric transcription stimulation by the RNA polymerase II super elongation complex.

   Allosteric transcription stimulation by RNA polymerase II super elongation complex · Chen Y, Vos SM, Dienemann C, Ninov M, Urlaub H, Cramer P · Molecular Cell 81(16):3386-3399.e10 2021
   PMID 34265249DOI
   ▸What this evidence cannot show

   PDB entity descriptions, UniProt accessions and modelled reference ranges fetched from data.rcsb.org on 2026-09-19, and the PMID, DOI, title, authors, journal, volume and pages fetched from PubMed in the same session; nothing typed from recall. This is a **composite** structure, meaning the deposited coordinates combine more than one reconstruction, so the modelled ranges reported by RCSB are more generous than the experimental density in any single map. The Pol II subunits are Sus scrofa; the elongation factors are human.

   ▸Evidence that argues against this (1)

   • The composite nature of the deposition matters for anyone using these coordinates: 'EAF1 1-268' means the full sequence is present in the model, not that all 268 residues were resolved. UniProt annotates aa 106-268 of EAF1 as disordered, which cannot be simultaneously true of a fully ordered model, so the modelled range should not be read as an ordered range. UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

   In the project's genome-wide KMT2A-rearranged selectivity scan of 17,728 genes, EAF1 ranks 14,482 (81.690th percentile) with a difference of +0.0469 (KMT2A-rearranged mean -0.1203 against other-leukaemia mean -0.1672) and a pan-essentiality of -0.4086 across 1,208 lines. Its paralogue EAF2 scores +0.1395 (rank 17,372) and its partner ELL2 -0.0472 (rank 3,697), while ELL itself scores -0.2533 (rank 153) on a pan-essentiality of -1.3018.

   Genome-wide KMT2A-rearranged selectivity scan, 17,728 genes (DepMap Chronos via Open Targets) · Broad Institute DepMap / Open Targets Platform; analyses/depmap-kmt2a-selective-scan-full 2026
   ▸What this evidence cannot show

   Rows read out of analyses/depmap-kmt2a-selective-scan-full/results.json on 2026-09-19. Eleven KMT2A-rearranged against 43 other leukaemia lines, AML-heavy, none infant-derived. EAF1 and EAF2 are paralogues, so a single-gene knockout of either is buffered and neither number is informative about the EAF function as a whole.

   ▸Evidence that argues against this (1)

   • A cell-line proliferation assay cannot detect the phenotype this entry is actually about. The EAF1 result is a subnuclear mislocalisation in cells expressing MLL-ELL, not a requirement for EAF1; a protein that is evicted from Cajal bodies by a fusion would not be expected to score as a dependency, and its near-zero score neither supports nor undermines the localisation finding. Polak PE et al. 2003
6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   UniProtKB Q96JC9 (EAF1_HUMAN, ELL-associated factor 1) is 268 aa, with a disordered region annotated at aa 106-268 and a region at aa 182-262 annotated as necessary for transactivation activity.

   UniProtKB Q96JC9 (EAF1_HUMAN), ELL-associated factor 1 · UniProt Knowledgebase, reviewed (Swiss-Prot) 2026
   ▸What this evidence cannot show

   Record fetched 2026-09-19; primary gene name read back as EAF1 and length 268 confirmed from the same fetched record. UniProt annotates **no ELL-interaction region for EAF1**, which is notable given that the protein is defined by that interaction - the equivalent region is annotated for the paralogue EAF2 (aa 17-104). The transactivation region overlaps the annotated disordered region entirely.

   ▸Evidence that argues against this (1)

   • The absence of an annotated ELL-binding region means this entry cannot give a defensible interface coordinate for the interaction it is named after. Rather than transferring the paralogue's boundaries, the entry leaves the domain out, which is the rule this project adopted after invisible null-coordinate domains were found in the 3D viewer. Hahn J et al. 2007
7. E_q69_ct1weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

   In ZEBRAFISH, deletion of eaf1 or eaf2 reduces hypoxia tolerance and reduces erythropoiesis during embryonic haematopoiesis; eaf1-/- and eaf2-/- mutants show significantly reduced expression of scl, lmo2 and gata1a at 24 and 72 hours post fertilisation, increased enrichment of H3K27me3 at the gata1a promoter but not at the scl or lmo2 promoters, and reduced phospho-beta-catenin with reduced TCF4 binding at the scl and lmo2 promoters at 24 hpf.

   Zebrafish ELL-associated factors Eaf1/2 modulate erythropoiesis via regulating gata1a expression and WNT signaling to facilitate hypoxia tolerance. · Liu W, Lin S, Li L, Tai Z, Liu JX · Cell Regen 2023
   PMID 37002435DOI
   ▸What this evidence cannot show

   Abstract read in full; full text not read. ZEBRAFISH embryonic haematopoiesis - not mammalian, not human, not leukaemia, and the erythroid rather than the B lineage (brief rule 73). The mechanism is assembled from expression changes plus one ChIP mark plus a phospho-level, with no epistasis experiment reported at abstract level. Confidence is weak on system distance, not on execution.

   ▸Evidence that argues against this (1)

   • A closely related claim in this literature has been withdrawn: 'Eafs control erythroid cell fate by regulating c-myb expression through Wnt signaling' (PLoS One 2013) is a RETRACTED PUBLICATION, verified by NCBI pubtype on 2026-09-21. The Eaf-erythroid-Wnt axis therefore has one surviving report rather than two, and this entry does not treat the surviving one as replicated. PLoS One 2013

• binds →ELL — rank 7 on selectivity and the deepest common-essential in the top ten, with one NMR fragment to its name

  isolated in a yeast two-hybrid screen with ELL; endogenous EAF1 and ELL co-immunoprecipitate from multiple cell lines

• binds →ELL2 — a SEC subunit that almost no cell needs, next to a paralogue almost every cell needs

  EAF1 aa 1-268 modelled with ELL2 aa 1-640 on Pol II in PDB 7OKX (3.3 A composite)

• part of →The ELL/EAF activation module (ELL2, EAF1, EAF2) — a general elongation stimulator, not a KMT2A-fusion dependency

  the EAF half of the ELL/EAF activation module the graph already carries

• part of →AQR (Aquarius) - a spliceosomal RNA helicase whose loss impairs MLL-rearranged leukaemia stem cells

  ECAS: an AFF4-ELL2-EAF1 elongation module joined to the AQR/XAB2 intron binding complex, distinct from P-TEFb-containing SEC

• part of →AFF4 — the textbook SEC scaffold that KMT2A::AFF1 cells need LESS than other leukemias

  AFF4-ELL2-EAF1 is the elongation module of ECAS

• related to →KMT2A::ELL (MLL-ELL), t(11;19)(q23;p13.1)

  MLL-ELL delocalises EAF1 from Cajal bodies to diffuse nucleoplasm; EAF1 speckles undetectable in murine MLL-ELL leukaemia cells - a sequestration mechanism, not recruitment

• related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)

  NOT a demonstrated interactor: EAF1's contacts are with ELL and ELL2, and the fusion phenotype is MLL-ELL, a different partner

• co-occurs with →SUPT5H (SPT5) — the pausing factor the KMT2A field has never studied, and the counted absence that undercuts 'the fusion hijacks pause release'

  EAF1, ELL2 and SPT5 are modelled together on Pol II in PDB 7OKX - the one structure placing DSIF and a SEC-family activation module in the same complex

• ← related toEAF2 (U19) — the EAF paralogue whose ELL-binding domain is a mapped tumour-suppressor module, described entirely in prostate cancer

  paralogues; UniProt annotates an ELL-interaction region for EAF2 and none for EAF1, and only EAF1 has a Cajal-body phenotype

• ← related toELL is in Cajal bodies - or nuclear speckles, depending which paper you read - and the MLL-ELL fusion empties them

  EAF1 is the protein actually scored in every MLL-ELL delocalisation experiment

• ← regulatesATM phosphorylates ELL at T551, S561 and S589 - inside the OCEL domain this graph already calls essential for MLL-ELL immortalisation

  phosphorylation increases ELL-EAF1 binding

• ← bindsELL-EAF1: a phosphorylation-gated interface reported to drive phase separation - and the assays used cannot distinguish that from clustered binding

  EAF1 side of the interface

• regulates →Megakaryocyte-erythroid progenitor (MEP) — the branch KMT2A fusions almost never take, which is why it was a usable control

  zebrafish eaf1 deletion reduces erythropoiesis with loss of scl, lmo2 and gata1a expression and H3K27me3 gain at the gata1a promoter (PMID 37002435). Zebrafish, erythroid - four steps from this disease.

• related to →The hypoxic, glycolytic state — the one state that appears independently in leukaemia-initiating cells, in measurable residual disease, and in KMT2A-rearranged B-ALL as a class

  WEAK and labelled so: eaf1/eaf2 deletion reduces whole-organism hypoxia tolerance in zebrafish through reduced red cell production (PMID 37002435), which is not the same quantity as a leukaemic cell's metabolic adaptation in a marrow niche

• regulates →Erythroid progenitor - the committed red-cell precursor, distinct from the bipotent MEP that two entries had to borrow, and the compartment that pays for any haem-pathway or fetal-globin intervention

  eaf1 or eaf2 deletion in ZEBRAFISH reduces erythropoiesis via gata1a, scl and lmo2 with H3K27me3 gain at the gata1a promoter. Weak by design - non-mammalian, and the obvious prior paper on Eafs and erythroid fate is retracted