ELP1 (IKBKAP/IKAP) - the Elongator scaffold that was first named for the IKK complex, is a paediatric-cancer predisposition gene, and has an RNA Pol II contact in IntAct

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.

• The direction of effect is probably wrong for a therapeutic. Both directional results in the Elongator field are tumour-suppressive: germline ELP1 loss predisposes to paediatric SHH medulloblastoma, and conditional Elp3 loss in mouse haematopoiesis causes bone marrow failure and then p53-mutated leukaemia/lymphoma. An agent that inhibits Elongator in a leukaemia is predicted by the only available data to be leukaemogenic rather than therapeutic.
• Loss of ELP1 function in humans causes familial dysautonomia, a neurodegenerative disease with sensory and autonomic neuron depletion. Systemic Elongator inhibition has a known and severe human phenotype rather than an unknown one, and it is a paediatric neurological phenotype in a paediatric disease.
• ELP1 is pan-essential at -1.02 across 1,208 DepMap lines, which is an essential-function toxicity signal and not a therapeutic window. The tRNA wobble-modification function is a general translational-fidelity requirement, so off-target consequences would be distributed across the whole proteome rather than concentrated at identifiable targets.
• The identifier hazard is itself a clinical hazard in this graph's context. ELP1's UniProt alternative names still assert an IKK-complex role and this project holds an IKK-inhibition hypothesis for menin-independent chromatin eviction. An agent joining an IKK-inhibitor rationale to an ELP1 hit off the protein name would be combining two different mechanisms on the strength of a 1998 name.
• No Elongator-directed chemical matter exists. The ChEMBL activity endpoint returned HTTP 500 on three attempts on 2026-09-22 and the row count could not be obtained, so the absence of chemistry is recorded here as UNMEASURED rather than as zero - this project's rule that an instrument failure is not a result.

1. Is any Elongator subunit present in a KMT2A-fusion proximity proteome? Elongator binds RNA polymerase II directly, ELP1 has a curated POLR2A interaction, this disease's mechanism is forced elongation, and the intersection is empty in the literature, in IntAct and in this repository's five fusion-interactome analyses. exp-turboid-map-of-kmt2a-aff1 would answer it as a by-product, at no marginal cost.
2. Which way does Elongator point in leukaemia? ELP1 loss predisposes to paediatric medulloblastoma and Elp3 loss drives p53-mutated leukaemia/lymphoma in mice, so the two available directional results both say Elongator is TUMOUR-SUPPRESSIVE. Before any beat writes Elongator as an elongation cofactor a KMT2A-fusion line needs an ELP1 knockdown with a growth readout, and the predicted sign is the inconvenient one.
3. Does an IKK-complex proteomics dataset contain ELP1? The IKAP assignment is still ECO:0000269 on the UniProt record, this graph holds mechanism-ikk-nfkb-retains-mll-on-chromatin, and nobody has asked whether the 1998 IKK preparation was capturing Elongator. A single search of a published IKK interactome for ELP1 through ELP6 would settle a 28-year-old ambiguity sitting on a protein name.
4. Is the ELP1 tRNA-binding region (1191-1209) the reason the complex is pan-essential? ELP1 is -1.02 across 1,208 lines while ELP4 is -0.46, and the difference between subunits within one obligate complex is itself a result nobody has explained.
5. Are the 87 curated ELP1 partners enriched for anything this project cares about? The list was counted for this entry and the members persisted, but not classified.

1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

   ELP1 (O95163, ELP1_HUMAN) is a 1,332-residue Elongator subunit whose only annotated features are a dimerization region (885-1332), a disordered stretch (1150-1208), a tRNA-binding stretch (1191-1209) and five phosphoserines; UniProt's recommended name is Elongator complex protein 1 while its alternative names remain IkappaB kinase complex-associated protein and p150, its gene synonyms are IKAP and IKBKAP, and it carries both a familial-dysautonomia and a medulloblastoma-susceptibility DISEASE comment.

   UniProtKB O95163 (ELP1_HUMAN) - reviewed human record, length 1332 aa · UniProt Consortium · UniProtKB 2026
   ▸What this evidence cannot show

   Accession resolved 2026-09-22 by gene_exact:ELP1 AND organism_id:9606 AND reviewed:true with gene_primary and length requested and read back (rules 62, 65, 66): exactly one reviewed human row, primary gene name ELP1, entry ELP1_HUMAN, 1,332 aa. The same query on the RETIRED symbol IKBKAP returns the identical single row, so the two symbols are not ambiguous at the database level even though they are very different queries at the literature level (see E8). Three annotation properties are recorded rather than copied: the FUNCTION comment's tRNA clauses are ECO:0000303 (an author statement taken from a review, not an experiment); the alpha-tubulin and tRNA-binding clauses are ECO:0000250 (by similarity); and the IKK-scaffold clause is ECO:0000269:9751059, an experimental code pointing at the 1998 Nature paper - see E4.

   ▸Evidence that argues against this (1)

   • An alternative-name field is not a curation error and should not be reported as one, but it is a retrieval and reasoning hazard: 'IkappaB kinase complex-associated protein' and 'IKK complex-associated protein' are both still on the record, and this graph has an NF-kB beat with a protein-ikk-complex entry. Reading the protein name back - the standard rule-65 defence - returns a pathway this protein is not now thought to work in. Hawkes NA et al. 2002
2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.review

   Elongator is conserved from yeast to humans and consists of two copies each of six core subunits ELP1-ELP6 organised into an ELP1/2/3 subcomplex and an ELP4/5/6 subcomplex, forming a stable assembly of about 850 kDa, whose catalytic subunit ELP3 contains a radical SAM domain and a PUTATIVE histone acetyltransferase domain while the ELP4/5/6 subcomplex possesses ATP-modulated tRNA-binding activity.

   Structural insights into the function of Elongator. · Dalwadi U, Yip CK · Cell Mol Life Sci 2018
   PMID 29332244DOI
   ▸What this evidence cannot show

   A REVIEW, cited as a map of the complex's architecture and nomenclature and not as primary evidence for any activity - this project's standing rule. Its own summary of the field says Elongator is 'implicated in multiple processes including transcription regulation, alpha-tubulin acetylation, and tRNA modification' and that how Elongator performs its multiple functions 'remains poorly understood'. The word PUTATIVE applied to the HAT domain is the review's own, verified in its abstract, and is the single most load-bearing word in this beat.

   ▸Evidence that argues against this (1)

   • A 2018 review's architecture has since been superseded in detail by cryo-EM of the fully assembled complex, which shows the Elp456 module as a hetero-hexamer and provides the molecular basis for assembly - so quote this paper for nomenclature and the putative-HAT judgement, and the structures for architecture. Jaciuk M et al. 2023
3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

   The C-terminal portion of human ELP1 forms a dimer that is sufficient for full-length ELP1 dimerization, the dimer contributes to ELP1 stability in vitro and in vivo and is required for assembly of both the human and the yeast Elongator complex, familial-dysautonomia-related ELP1 mutations affect Elongator assembly, and ELP1 dimerization is essential for yeast viability.

   Dimerization of elongator protein 1 is essential for Elongator complex assembly. · Xu H, Lin Z, et al. · Proc Natl Acad Sci U S A 2015
   PMID 26261306DOI
   ▸What this evidence cannot show

   Crystallography of the human ELP1 C-terminal domain and of the cognate yeast dimerization domain, with mutational validation split between human biochemistry and yeast genetics; the viability arm is yeast. Abstract read; full text not read. The structure is of a fragment, not of full-length ELP1, so 'sufficient for dimerization' is a statement about that fragment in isolation.

   ▸Evidence that argues against this (1)

   • This paper's own framing is that the FD pathology follows from impaired Elongator activity, which is a model rather than a measurement - it does not show that the neuronal phenotype is caused by loss of any specific Elongator output. The disease mutation's mechanism is separately complicated by being tissue-specific at the RNA level rather than a uniform loss of protein. Slaugenhaupt SA et al. 2001
4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

   IKAP - the protein now named ELP1 - was reported in 1998 as a 150 kDa component of interleukin-1-inducible IKK complexes containing NIK, IKK-alpha, IKK-beta, IkappaB-alpha and NF-kB/RelA, able to bind NIK and the IKKs and assemble them into an active kinase complex, and described as a scaffold protein and regulator for three kinases in pro-inflammatory cytokine signalling.

   IKAP is a scaffold protein of the IkappaB kinase complex. · Cohen L, Henzel WJ, et al. · Nature 1998
   PMID 9751059DOI
   ▸What this evidence cannot show

   The FOUNDING paper for the name, cited here for what it claimed and for the naming consequence, NOT as current evidence that ELP1 functions in NF-kB signalling. Abstract read; full text not read. It is a complex-isolation and binding study from 1998 with no genetics. UniProt still codes this ECO:0000269 under FUNCTION ('May act as a scaffold protein that assembles active IKK-MAP3K1') and SUBUNIT, so the annotation is evidence-coded and the experiment is real - this entry's position is that the assignment was superseded, not that it was fabricated. Confidence is 'moderate' for the binding observation and this entry makes no claim at all about a physiological NF-kB role.

   ▸Evidence that argues against this (1)

   • Superseded in practice by the human Elongator purification, which identified IKAP as an Elongator subunit and proposed that familial dysautonomia is a transcription rather than an inflammation disorder. Neither paper retracts the other, and UniProt carries both, which is exactly why this entry records the conflict instead of choosing. Hawkes NA et al. 2002
5. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Familial dysautonomia is caused by mutations in IKBKAP/ELP1, the major Ashkenazi founder allele being an intron-20 donor-splice mutation that causes exon-20 skipping in a TISSUE-SPECIFIC pattern - patient lymphoblast RNA is primarily wild type while brain RNA shows only the deleted message - with a minor R696P missense allele predicted to disrupt a phosphorylation site.

   Tissue-specific expression of a splicing mutation in the IKBKAP gene causes familial dysautonomia. · Slaugenhaupt SA, Blumenfeld A, et al. · Am J Hum Genet 2001
   PMID 11179008DOI
   ▸What this evidence cannot show

   Positional cloning plus RNA analysis in patient material; abstract read, full text not read. The tissue-specific expression is the load-bearing detail for this beat and it cuts a specific way: a haematopoietic cell from an FD patient carries mostly WILD-TYPE ELP1 message, so FD is not a natural experiment in Elongator loss in blood, and cannot be read as evidence about Elongator in haematopoiesis. UniProt codes the FD disease comment to this paper plus PMIDs 11179021 and 12687659.

   ▸Evidence that argues against this (1)

   • The FD allele is a splicing hypomorph rather than a null, so the human phenotype bounds what PARTIAL Elongator loss does in the nervous system and says nothing about complete loss. The complete-loss phenotype in a mammal comes from conditional knockouts, and in haematopoiesis that experiment was done on Elp3 and produced bone marrow failure - a much more severe phenotype than anything in FD blood. Rosu A et al. 2021
6. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Rare germline loss-of-function variants in ELP1 occur in 14% of paediatric patients with Sonic-Hedgehog medulloblastoma, making ELP1 the commonest medulloblastoma predisposition gene and raising genetic predisposition to 40% in that subgroup; ELP1-associated tumours are restricted to the SHH-alpha subtype, show universal biallelic ELP1 inactivation through somatic loss of chromosome arm 9q, mostly carry additional somatic PTCH1 alterations, and are characterised by a destabilised Elongator complex, loss of Elongator-dependent tRNA modifications, codon-dependent translational reprogramming and induction of the unfolded-protein response.

   Germline Elongator mutations in Sonic Hedgehog medulloblastoma. · Waszak SM, Robinson GW, et al. · Nature 2020
   PMID 32296180DOI
   ▸What this evidence cannot show

   Germline analysis of all protein-coding genes with replication, plus tumour molecular characterisation; abstract read, full text not read. This is a BRAIN tumour and not a leukaemia, and the cooperating lesion is SHH-pathway activation, so nothing here transfers directly to KMT2A-rearranged disease - it is cited as PI-dimension-5 analogous biology and as the strongest available demonstration that an Elongator lesion can drive a paediatric cancer. The 14% and 40% figures are proportions within a molecular subgroup and should not be restated as properties of medulloblastoma generally.

   ▸Evidence that argues against this (1)

   • The DIRECTION is the counter-argument, and it is the central tension of this beat. ELP1 here is a TUMOUR SUPPRESSOR whose biallelic loss predisposes to cancer. A KMT2A-fusion framing would want Elongator as a positive elongation cofactor whose activity the fusion exploits; if Elongator loss is oncogenic in one paediatric lineage and Elp3 loss is oncogenic in haematopoiesis, then 'inhibit Elongator in leukaemia' is predicted to be the wrong sign, and the therapeutic reading of this beat is closed before it opens. Rosu A et al. 2021
7. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   No paper names ELP1 and a KMT2A fusion together: measured 2026-09-22 with one [tiab] tag per term and querytranslation read back, ELP1[tiab] AND KMT2A[tiab] = 0, ELP1[tiab] AND MLL[tiab] = 0, ELP1[tiab] AND leukaemia[tiab] = 0 and ELP1[tiab] AND leukemia[tiab] = 1 (a 15-colour flow-cytometry MRD assay in which ELP1 is not this gene), against positive controls through the identical shape AFF1[tiab] AND KMT2A[tiab] = 104, DOT1L[tiab] AND KMT2A[tiab] = 30 and KMT2A[tiab] AND histone[tiab] = 218; and in GEO DataSets, ELP1 AND leukemia = 0 against a control of KMT2A = 2,300.

   NCBI E-utilities esearch over PubMed and GEO DataSets, run 2026-09-22 by Q98 with querytranslation asserted non-empty on every call · Oncofusion knowledge-graph project (Q98-elongator-smyd3) · NCBI E-utilities 2026
   ▸What this evidence cannot show

   Zero in TITLES AND ABSTRACTS ONLY; supporting figures and supplementary tables not searched, which is this project's rule-80 exposure and matters here because the measurement a fusion beat would want (an Elongator subunit in a proximity proteome) is exactly the kind of result that lives in a supplementary table. The name list searched was ELP1, IKBKAP, IKAP, and at the complex level 'Elongator' - the complex form is mandatory for a subunit absence and it also returns 0 against KMT2A and MLL. INSTRUMENT ALARM RECORDED WITH THE RESULT. On the first pass of this beat's symbol sweep, ELP3[tiab] and ELP6[tiab] each returned a well-formed esearchresult with count 0, querytranslation EMPTY STRING, and errorlist and warninglist both {}. Re-run minutes later through the identical code path they return 202 and 43. Every count in this entry was therefore taken through a guarded wrapper that REFUSES any response whose querytranslation is empty and retries. A zero with an empty querytranslation is an instrument failure, not a result, and nothing else in the response says so. Symbol recall was measured rather than assumed and the answer is asymmetric: ELP1[tiab] = 214, IKBKAP[tiab] = 115, IKAP[tiab] = 126, union of all three = 360, and ELP1 intersect IKBKAP = only 28 records, so the two retired symbols contribute 146 of 360 records (40.6%) that the approved symbol cannot reach. ELP1[tiab] is also contaminated - 16 of its 214 titles are about elastin-like polypeptides, and the off-topic tail additionally contains the C. elegans EMAP-like protein ELP-1 and a human ERD-2-like protein called ELP-1 - so counts on this symbol need title-checking in both directions.

   ▸Evidence that argues against this (1)

   • A literature bound cannot see an unreported measurement, and the archive half of the bound is weaker than it looks: GEO DataSets indexes curated series metadata, so a gene name absent there can still be present in the processed matrices of a KMT2A dataset. The stronger negative is the interaction-database one at E9, because a fusion-interactome deposition would have to name the subunit. IntAct/EMBL-EBI 2026
8. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

   IntAct REST returns 130 interaction records for ELP1 (O95163) spanning 87 distinct molecules, which include POLR2A - the largest subunit of RNA polymerase II - and ELP1 through ELP5, and which include NONE of KMT2A, MEN1, AFF1, AFF4, MLLT1, MLLT3, MLLT10, DOT1L, PSIP1, CDK9, ELL, KAT7, WDR5 or HOXA9.

   IntAct molecular interaction database, REST endpoint findInteractions (not PSICQUIC) · IntAct/EMBL-EBI · IntAct 2026
   ▸What this evidence cannot show

   Queried 2026-09-22 through the IntAct REST findInteractions endpoint and NOT through PSICQUIC, because this project's rule 93 establishes that the PSICQUIC id: form cannot reach processed-chain identifiers and under-reports KMT2A by more than half. The control is the same one rule 93 uses: findInteractions/Q03164 returns 376 records here, reproducing the documented figure exactly, so the endpoint is working. The absence is bounded by what has been CURATED into IntAct - this project has already established that the index fusion's own AP-MS interactome (PMID 29777171) has never been entered into IntAct, so an IntAct absence is 'measured-not-curated' rather than 'never-measured' for any partner that appears in an uncurated proteome. Partner names were read from moleculeA/moleculeB across paged results, which are gene-symbol strings and not accessions.

   ▸Evidence that argues against this (1)

   • A POLR2A record is not evidence of a KMT2A-fusion contact and must not be read as one. Pol II is the most promiscuously reported partner in the database, Elongator's Pol II association is the founding observation of the whole complex, and the KMT2A fusion's own Pol II association is separately documented - so ELP1 and a KMT2A fusion sharing a Pol II partner is a statement about Pol II, not about either of them. Hawkes NA et al. 2002
9. E10established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

   Four cryo-EM structures of the human ELP123 subcomplex - which contains ELP1 - bound to tRNA and cofactors resolve four stages of the wobble-uridine modification reaction at resolutions to 2.9 Angstrom, show that tRNA binding exposes the conserved uridine at position 33 which triggers acetyl-CoA hydrolysis, identify conserved residues required for radical-based acetylation of U34, and profile the molecular effects of patient-derived mutations.

   Cryo-EM structures of the human Elongator complex at work. · Abbassi NE, Jaciuk M, et al. · Nat Commun 2024
   PMID 38750017DOI
   ▸What this evidence cannot show

   Human protein, but the substrate throughout is tRNA - there is no histone, no nucleosome and no chromatin template anywhere in the series. Abstract read; full text not read; the PDB identifiers 8PTX, 8PTY, 8PTZ and 8PU0 were taken from the UniProt cross-reference lists of ELP1, ELP2 and ELP3, which all three carry identically, and were NOT read off the paper. Per this project's PDB rules, residue spans have not been read off these deposits and none is asserted in this entry.

   ▸Evidence that argues against this (1)

   • Structures of ELP123 with tRNA cannot exclude a histone activity that nobody assayed in the same preparation, so this is decisive about what the human structural literature SHOWS and not about what Elongator does. The complementary gap is that no structure of any species' Elongator has been solved with a histone tail or a nucleosome. Dalwadi U et al. 2018
10. E11moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Human Elongator purified from HeLa cells contains histone acetyltransferase activity specific to histone H3 and much weaker on H4, interacts DIRECTLY with RNA polymerase II, and facilitates transcription by RNA polymerase II in a chromatin- and acetyl-CoA-dependent manner, while failing to interact with other factors that let Pol II traverse nucleosomes.

    Human Elongator facilitates RNA polymerase II transcription through chromatin. · Kim JH, Lane WS, et al. · Proc Natl Acad Sci U S A 2002
    PMID 11818576DOI
    ▸What this evidence cannot show

    Human, in vitro reconstituted chromatin transcription - and the decisive limitation is stated in the paper's own first sentence: the purified complex 'was found to be composed of THREE polypeptides'. A three-subunit preparation cannot place ELP1, ELP4, ELP5 or ELP6 in the HAT-active species, and the contemporaneous purification at E_HAT_HUMAN found that a three-subunit core has NO HAT activity. Abstract read; full text not read. This entry cites the paper for the Pol II contact and the chromatin dependence, which are its robust content, and not as evidence that any named subunit is required.

    ▸Evidence that argues against this (1)

    • Directly contradicted on the composition-versus-activity point by the six-subunit purification, which reports that the three-subunit core form has no HAT activity despite containing ELP3. Two human purifications from the same year disagree, and no later human paper this beat could find repeats either assay - so the human histone-acetyltransferase claim for Elongator rests on an unreplicated 2002 result that a second 2002 result contradicts. Hawkes NA et al. 2002
11. E12_HAT_YEASTestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In YEAST, intact Elongator is a histone H3 and H4 acetyltransferase whose predominant sites are H3K14 and H4K8, it acetylates core and nucleosomal substrates where isolated Elp3 is not specific, the three smallest subunits Elp4, Elp5 and Elp6 are REQUIRED for that HAT activity, and multiply-acetylated H3 and H4 are reduced in vivo in ELP3-null cells.

    Elongator is a histone H3 and H4 acetyltransferase important for normal histone acetylation levels in vivo. · Winkler GS, Kristjuhan A, et al. · Proc Natl Acad Sci U S A 2002
    PMID 11904415DOI
    ▸What this evidence cannot show

    Saccharomyces cerevisiae throughout - not human, not haematopoietic, not a fusion system (this project's rule that the system belongs in the claim, not only the caveat). Abstract read; full text not read. The in vivo arm is a chromatin immunoprecipitation of bulk acetylated H3/H4 in an ELP3 deletion, so it establishes that Elongator contributes to steady-state acetylation levels and not that Elongator is the writer at any named gene. The requirement for the small subunits is a requirement for HAT ACTIVITY OF THE COMPLEX and does not make ELP4, ELP5 or ELP6 catalytic.

    ▸Evidence that argues against this (1)

    • The same paper's own framing has been overtaken. UniProt's current human ELP3 record renames the protein 'tRNA uridine(34) acetyltransferase' and states that protein-lysine acetyltransferase activity is unclear in vivo; the 2018 structural review calls the ELP3 HAT domain putative. A 2002 yeast result is not withdrawn by a 2026 database opinion, but a beat that cites only this paper is quoting the high-water mark of a claim that has receded. Dalwadi U et al. 2018
12. E12_HAT_HUMANstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In HUMAN HeLa cells the purified Elongator complex exists in two forms and only the six-subunit holo-Elongator has histone acetyltransferase activity against H3 and H4: the three-subunit core form has NO HAT activity despite containing the catalytic ELP3 subunit, and Elongator is a component of early elongation complexes formed in HeLa nuclear extracts and interacts directly with RNA polymerase II in solution.

    Purification and characterization of the human elongator complex. · Hawkes NA, Otero G, et al. · J Biol Chem 2002
    PMID 11714725DOI
    ▸What this evidence cannot show

    Biochemical purification from HeLa; abstract read, full text not read. This is the paper that reassigned two proteins to Elongator - it identifies StIP1 (ELP2) and IKAP (ELP1) as subunits of the human complex - so it is simultaneously the source of the small-subunit requirement in human cells and the source of two of the identity traps recorded in this beat. It is ALSO the resolution of a disagreement this graph already records: gene-elp4 notes that a contemporaneous human purification recovered only three polypeptides WITH HAT activity (PMID 11818576) and leaves the disagreement open; this paper's three-subunit core has NO HAT activity. Both are 2002, both are human, and they disagree. Recorded as a live disagreement, not as a resolution in either direction.

    ▸Evidence that argues against this (1)

    • THE DIRECT CONTRADICTION, and it is not resolved. A human Elongator purified from HeLa the same year was composed of THREE polypeptides and DID have HAT activity specific to histone H3, facilitating Pol II transcription in a chromatin- and acetyl-CoA-dependent manner. Complex composition determined by purification depends on the tag, the salt and the cell type, so 'three subunits, no HAT' and 'three polypeptides, HAT' are both properties of preparations. No later human paper this beat could find repeats either HAT assay. Kim JH et al. 2002
13. E12_TRNA_NOWstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The current curated and structural position is that Elongator is a tRNA wobble-uridine modification enzyme: UniProt names human ELP3 'tRNA uridine(34) acetyltransferase' (EC 2.3.1.311) and states that its protein-lysine acetyltransferase activity is unclear in vivo, and every high-resolution structure of human Elongator is of the ELP123 subcomplex bound to tRNA and cofactors, resolving four stages of the U34 acetylation reaction with no histone in the series.

    Cryo-EM structures of the human Elongator complex at work. · Abbassi NE, Jaciuk M, et al. · Nat Commun 2024
    PMID 38750017DOI
    ▸What this evidence cannot show

    Cryo-EM to 2.9 Angstrom of human ELP123 with tRNA; abstract read, full text not read. An absence of histone in a structural series is not evidence that histones are not substrates - nobody put one in - so this bounds what the human structural work SHOWS and not what Elongator can do. The UniProt FUNCTION clause quoted here is coded ECO:0000269 to PMID 29415125 for the tRNA half, while the replacement catalytic activity and the EC number are ECO:0000250 BY SIMILARITY from a non-human ortholog: UniProt renamed the human protein on an inferred activity, which is worth knowing before quoting the new name as settled.

    ▸Evidence that argues against this (1)

    • ELP3 does demonstrably acetylate at least one non-histone PROTEIN in vivo: in Drosophila neurons ELP3 is necessary and sufficient to acetylate the ELKS-family protein Bruchpilot at presynaptic densities, with a structural and a neurotransmitter-release phenotype. So 'primarily a tRNA enzyme' is a statement about the dominant activity, not a demonstration that the GNAT pocket cannot take a protein substrate. Note also that UniProt cites this same paper for 'Not a major acetyltransferase of tubulin', a clause its abstract does not contain - read the body before quoting that negative. Miśkiewicz K et al. 2011
14. E15weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.dataset

    ELP1 shows no KMT2A-selective dependency in this project's own DepMap scan: mean Chronos -0.9603 in 11 KMT2A-rearranged lines against -0.9985 in 43 other leukaemia lines, difference +0.0381, at a pan-essentiality of -1.0221 across 1208 lines and selectivity rank 13710 of 17,728.

    DepMap CRISPR (Chronos) gene-effect scan of 11 KMT2A-rearranged versus 43 other leukaemia lines (this project's own genome-wide scan, analyses/depmap-kmt2a-selective-scan-full) · Oncofusion knowledge-graph project · analyses/depmap-kmt2a-selective-scan-full/results.json 2026
    ▸What this evidence cannot show

    Joined 2026-09-22 from analyses/depmap-kmt2a-selective-scan-full/results.json (generated 2026-09-19T10:12:06Z). NOT re-fetched and NOT independent of that scan: every caveat it carries applies, including that the 11 KMT2A-rearranged lines come from nine individuals and that the arm is myeloid-enriched, so a difference there mixes fusion status with lineage. The scan records a generation timestamp and NOT a DepMap RELEASE, which is this project's rule-83 exposure carried forward rather than fixed here. Read the numbers by this project's own DepMap rules: a rank is not an effect size, a difference is not a dependency, the anti-selective half is 50 percent of any gene set by construction with this graph's base rate at 45.3 percent, and a gene depleted this deeply across all lines is FLOOR-COMPRESSED - there is almost no dynamic range left in which selectivity could appear, so a small arm-to-arm difference is uninterpretable rather than small. The screen is also leukaemia lines only.

    ▸Evidence that argues against this (2)

    • A difference of +0.0381 at a pan-essentiality of -1.022 across 1208 lines is floor-compressed and uninterpretable, not small: a gene required in nearly every line has no dynamic range left in which KMT2A selectivity could show. Reporting the rank would be reporting the null. Oncofusion knowledge-graph project 2026
    • The comparison that carries information is within the complex, not against the genome: ELP5 -1.209, ELP1 -1.022, ELP6 -0.981, ELP3 -0.965, ELP2 -0.627 and ELP4 -0.459 across 1,208 lines. Five of six subunits are deeply required in nearly every line, which is what a core tRNA-modification complex should look like and is also exactly the regime in which no selectivity claim can be made. This project's own complex-level analysis scores the Elongator holoenzyme (IntAct CPX-1949) at median coherence 0.5571 and median KMT2A-selectivity rank 7,272 of 17,728 - internally co-dependent, and mid-table for selectivity. Oncofusion knowledge-graph project 2026

• part of →ELP4 (PAXNEB) - a required subunit of Elongator, the one partner of the seven with a genuine histone-acetyltransferase connection, and the one whose fusion supplies only its last 43 residues

  both are subunits of the same six-protein Elongator complex (ELP1/2/3 lobe and ELP4/5/6 lobe); ELP1 is the assembly scaffold, ELP4 is in the accessory hexamer

• related to →Plain-English: what transcription elongation is, and why a traffic light explains this leukemia

  Elongator was named for its association with hyperphosphorylated elongating RNA Pol II; ELP1 carries a curated POLR2A interaction in IntAct

• tests →Hypothesis: transfer-RNA and ribosome biology is a genuinely empty cell in KMT2A-rearranged leukemia, and translational capacity may be the unmeasured dependency

  ELP1 is the scaffold of the tRNA wobble-modification complex that hypothesis is about; ELP1-mutant medulloblastoma is the strongest demonstration that loss of U34 modification can drive a paediatric cancer

• related to →KMT2A / MLL1 — the gene that breaks

  BOUNDED ABSENCE, not an interaction: ELP1 x KMT2A = 0 records in titles/abstracts, 0 IntAct records among 130, 0 GEO leukaemia datasets, and absent from all five KMT2A-fusion interactome analyses in this repo

• related to →The IKK complex — the unit the eviction experiment actually resolves, because the chemistry that produced the result cannot see the subunits

  IDENTIFIER AND MECHANISM TRAP: ELP1 was discovered as IKAP, reported as the IKK scaffold (PMID 9751059), and UniProt still carries that clause at ECO:0000269 and the name in its alternative-names field

• related to →TP53 - named in 84 entries of this graph, absent from the index genome study of the disease, and four different things depending on when you look

  the Elongator-p53 axis, via the complex: conditional Elp3 loss in mouse haematopoiesis activates a p53-dependent antitumour checkpoint and p53-mutant leukaemia/lymphoma follows

• ← part ofELP2 (STATIP1/StIP1) - the double beta-propeller of the Elongator core, whose informal name collides with a different human gene and whose STAT3 function is a by-similarity annotation

  both in the catalytic ELP1/2/3 lobe of Elongator; ELP2 double beta-propeller required for assembly, ELP1 the scaffold

• ← part ofELP3 - the catalytic subunit, the only Elongator protein with a real acetyltransferase domain, and the one whose loss in mouse blood causes marrow failure and then p53-mutant leukaemia

  catalytic subunit and scaffold of the same ELP1/2/3 lobe; UniProt records a ELP1 FD allele with reduced interaction with ELP3