Oncofusion

start / graph / mechanism-sec-elongation-hijack

Mechanismestablished confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.Q3-mechanism-core

Step 3 — The hijack: the fusion drags the super elongation complex to paused RNA polymerase II

also called SEC hijack · AEP recruitment · MLL fusion transcription elongation · P-TEFb capture

Genes usually keep a copying machine parked just past the start line, waiting for a green light. The fusion protein permanently attaches the green-light machinery — the super elongation complex — to its favourite genes, so the copier never stops.

The same thing, for a scientist

Most KMT2A fusion partners (AFF1/AFF4, MLLT1/ENL, MLLT3/AF9, ELL) are themselves components of the super elongation complex (SEC, also described as AEP: AF4-family/ENL-family/P-TEFb). Fusing KMT2A's chromatin-targeting module directly to a SEC subunit converts a regulated, transient recruitment into a constitutive one: P-TEFb (CDK9/CCNT1) is delivered permanently to fusion-bound promoters, phosphorylates the Pol II CTD and NELF/DSIF, and releases promoter-proximally paused polymerase into productive elongation. The fusion also keeps Pol II in an actively elongating state and prevents accumulation of inhibitory histone methylation, so target loci stay open even under differentiation pressure. The step is mechanistically central but built from broadly essential machinery, which is why it has yielded the most disappointing therapeutic index in the whole chain.

Go deeper
Why almost every partner is an elongation factor2 deeper

This is the field's single most clarifying observation. Purification of several commonly occurring KMT2A chimeras recovered the same complex from all of them: SEC, containing ELL, P-TEFb (CDK9/cyclin T1), AFF4 and additional factors. AFF4 is required for SEC stability and for proper transcription by poised RNA polymerase II, and AFF4 knockdown in leukemic cells reduces expression of KMT2A-chimera target genes 1.

The complementary study named the same module AEP (AF4-family/ENL-family/P-TEFb) and established the logic: AEP is normally recruited to KMT2A-target chromatin to facilitate transcription; KMT2A fused to an AEP component forms a constitutive KMT2A/AEP hybrid complex that causes sustained target expression and transforms haematopoietic progenitors. Critically, KMT2A::AFDN (MLL-AF6), whose partner is a cytoplasmic protein, does not form such a hybrid — yet it still constitutively recruits AEP by an alternative route 2.

That last fact is the load-bearing generalisation: the fusion partner's identity is a means, not the mechanism. More than 80 partners converge on 'get AEP/SEC onto KMT2A target chromatin and keep it there'.

What it does to paused polymerase

Mechanistically the complex is a tightly interlocked core stabilised by recursive protein-protein interactions, integrating H3K79 methylation, Pol II CTD phosphorylation and the fusion partner into one elongation-stimulating unit, demonstrated directly by RNA-tethering assays. Coimmunoprecipitation showed KMT2A fusions incorporate into it, causing constitutive recruitment of elongation activity to target loci. A time-resolved ChIP with a conditional fusion singled out H3K79 methylation as the parameter most closely correlated with target expression. The fusion also kept Pol II in an actively elongating state and prevented accumulation of inhibitory histone methylation, and Hox loci stayed open and productive even under forced differentiation. Finally, KMT2A-transformed cells were particularly sensitive to pharmacological inhibition of Pol II CTD phosphorylation 3.

That last sentence is the origin of the CDK9 idea — and also of its problem (see below).

SEC is not one complex, and DOT1L is not inside it

Two structural nuances change how this step should be drugged.

First, AFF1-SEC and AFF4-SEC are separate complexes with largely non-overlapping gene targets. AFF1 and AFF4 do not simply substitute for each other; a single amino-acid difference changes how efficiently each enhances P-TEFb affinity, and genome-wide the two regulate largely distinct gene sets 4. For a KMT2A::AFF1 leukemia this predicts that AFF4 is not a redundant backup — and that pan-SEC disruption and AFF1-selective disruption are different experiments with different toxicity.

Second, reconstitution of defined minimal complexes showed that DOT1 forms a stable complex with AF9 and does not interact with AF9-AF4-P-TEFb complexes, while P-TEFb provides the link that brings ELL plus ELL-associated factors EAF1/EAF2 to AF4 5. So the popular cartoon of one super-complex containing both SEC and DOT1L is wrong: they are alternative, partly mutually exclusive assemblies competing for the same partner surface. This matters directly for combination logic — menin inhibition and DOT1L inhibition may not be hitting two arms of one machine but two machines that trade places.

Why this step is the worst therapeutic target in the chain, despite being the mechanism

Every component here is general transcription machinery. In DepMap, CDK9, ELL and the PAF1c subunits are common-essential with strongly negative mean gene-effect scores (see the essentiality table in gene-men1). CDK9 inhibition kills KMT2A-fusion cells, as predicted 3 — and kills nearly everything else too. Flavopiridol and successor CDK9 inhibitors have a long history of dose-limiting toxicity in exactly this population.

The defensible read is that step 3 is the mechanism and not the intervention point. Its value to this project is diagnostic: it tells you that the fusion's output is an elongation phenotype, so the assays that matter are nascent-transcription assays (PRO-seq/TT-seq, pausing index) rather than steady-state RNA-seq, and that a good drug should be scored on release of paused polymerase at fusion targets, not on total target mRNA.

Facts and figures
step3
chainmechanism-kmt2a-fusion-causal-chain
biochemical outputP-TEFb-mediated Pol II CTD Ser2 / DSIF / NELF phosphorylation and release of promoter-proximally paused Pol II
partner count converging>80 KMT2A fusion partners
separate complexesAFF1-SEC and AFF4-SEC are distinct; DOT1L-AF9 is distinct from AF9-AF4-P-TEFb
rate limitingmechanistically central but built from common-essential machinery; poor therapeutic index
assay implicationscore interventions on nascent transcription / pausing index at fusion targets, not steady-state mRNA
complex names
[
  "SEC (super elongation complex)",
  "AEP (AF4-family/ENL-family/P-TEFb)"
]
components
[
  "gene-aff1",
  "gene-aff4",
  "gene-ell",
  "ELL2",
  "gene-cdk9",
  "CCNT1",
  "EAF1",
  "EAF2",
  "gene-mllt1",
  "gene-mllt3"
]
How this could bite us back
  • CDK9, ELL and PAF1c are DepMap common-essential; anything that disables SEC broadly is a general transcription poison, and the pediatric therapeutic index is likely unacceptable.
  • Partner redundancy: the KMT2A::AFDN precedent shows AEP can be recruited by an unknown alternative mechanism when the direct fusion route is absent, so blocking a specific partner interaction may be routed around.
  • SEC drives HSP70 and stress-response genes; inhibiting it removes a stress buffer in normal tissue as well as in the leukemia.
  • Because the fusion keeps Pol II elongating and blocks inhibitory marks, partial inhibition may reduce output without re-establishing the repressive state — producing transcriptional slowdown without differentiation, i.e. cytostasis and relapse rather than cure.
Still open
  1. Is AFF4 genuinely unable to substitute for AFF1 in KMT2A::AFF1 leukemia? The SEC-family specificity result was obtained in HeLa on HIV genes; the leukemia experiment has not been done and would decide whether AFF1 is a selective target.
  2. Are SEC and DOT1L-containing complexes truly mutually exclusive on chromatin at fusion targets, or does the in-vitro exclusion break down in vivo? Proximity-ligation or sequential ChIP at HOXA9 would settle it, and it determines whether menin + DOT1L inhibition is additive or redundant.
  3. Can pausing-release be blocked selectively at fusion-bound loci — for example by targeting the AFF1 pSER/ALF surface rather than CDK9's ATP site — and does that separate efficacy from the general transcription toxicity of CDK9 inhibitors?
  4. How does the KMT2A::AFDN (MLL-AF6) alternative AEP-recruitment route work? It has been unexplained since 2010 and is the clearest existing example of the mechanism routing around a blocked step.
Every source behind this entry (5)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Purification of several commonly occurring KMT2A chimeras recovered the super elongation complex from all of them, comprising ELL, P-TEFb, AFF4 and other factors; AFF4 is required for SEC stability and for transcription by poised RNA polymerase II, and AFF4 knockdown reduces KMT2A-chimera target gene expression in leukemic cells.

    AFF4, a component of the ELL/P-TEFb elongation complex and a shared subunit of MLL chimeras, can link transcription elongation to leukemia · Lin C, Smith ER, Takahashi H, et al. (Shilatifard A, senior author) · Molecular Cell 37(3):429-37 2010
    PMID 20159561DOI
    What this evidence cannot show

    Biochemical purification from overexpressing cell lines plus RNAi; the complex composition is an affinity-purification snapshot and stoichiometry is not established. AFF4 knockdown reduces but does not abolish target expression.

    Evidence that argues against this (1)
    • 'All of them' is true of the chimeras purified and false as a general claim, and the counterexample is not marginal. A 15-amino-acid tubulin-binding domain of GEPHYRIN - a protein with no transcriptional activity and no elongation machinery - is necessary and sufficient to convert MLL into an oncogene that transforms haematopoietic progenitors. Whatever that fusion does, it cannot be recruiting SEC through its partner, so SEC recruitment is what one partner CLASS does rather than the mechanism of KMT2A leukaemia. Blood 2004
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    AEP (AF4-family/ENL-family/P-TEFb) is normally recruited to KMT2A-target chromatin to facilitate transcription; KMT2A fused to an AEP component forms a constitutive KMT2A/AEP hybrid complex causing sustained target expression and progenitor transformation, while KMT2A::AFDN, whose partner is cytoplasmic, does not form such a hybrid yet still constitutively recruits AEP by an alternative mechanism.

    PMID 20153263DOI
    What this evidence cannot show

    The KMT2A::AFDN alternative-recruitment mechanism was explicitly left 'unknown' in this paper, so the universality claim rests on an unexplained observation. Mouse progenitor transformation and cell-line biochemistry.

    Evidence that argues against this (1)
    • The AEP model is built from complex composition and transformation assays, and the KMT2A::AFDN comparison is the interesting exception rather than the confirming case: a cytoplasmic partner that nonetheless produces leukaemia is evidence that the elongation account is not the whole mechanism, and this graph's non-elongation partner entries record several more such partners. Oncofusion knowledge graph (internal analysis) 2026
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    KMT2A fusions capture an ENL-associated elongation complex that integrates H3K79 methylation, Pol II CTD phosphorylation and the fusion partner to stimulate elongation; time-resolved ChIP with a conditional fusion identified H3K79 methylation as the parameter best correlated with target expression, the fusion kept Pol II actively elongating and prevented accumulation of inhibitory histone methylation, Hox loci stayed open under forced differentiation, and KMT2A-transformed cells were particularly sensitive to inhibition of Pol II phosphorylation.

    Misguided transcriptional elongation causes mixed lineage leukemia · Mueller D, Garcia-Cuellar MP, Bach C, Buhl S, Maethner E, Slany RK · PLoS Biology 7(11):e1000249 2009
    PMID 19956800DOI
    What this evidence cannot show

    The Pol II-phosphorylation inhibitor used is flavopiridol-class and not selective for CDK9 in cells, so 'particularly sensitive' is a relative statement in a small panel. Conditional fusion systems are engineered and overexpressed. Correlation of H3K79me with expression does not establish direction of causation.

    Evidence that argues against this (1)
    • 'H3K79 methylation is the parameter best correlated with target expression' is a correlation that broke when it was tested as a dependency. Paediatric KMT2A::AFF1 ALL cells driven to pinometostat resistance became largely INDEPENDENT of DOT1L-mediated H3K79 methylation while still requiring the physical presence of DOT1L, HOXA9 and the fusion protein, losing only a subset of fusion target genes and upregulating myeloid genes. The mark can therefore be removed while the hijack continues, which places H3K79me2/3 downstream of, rather than load-bearing for, the elongation hijack. Schneider P et al. 2023
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    AFF1 and AFF4 reside in separate super elongation complexes with largely distinct, non-overlapping gene target specificities; a single amino-acid difference between them changes the efficiency with which each enhances P-TEFb affinity.

    Gene target specificity of the Super Elongation Complex (SEC) family: how HIV-1 Tat employs selected SEC members to activate viral transcription · Lu H, Li Z, Zhang W, Schulze-Gahmen U, Xue Y, Zhou Q · Nucleic Acids Research 43(12):5868-79 2015
    PMID 26007649DOI
    What this evidence cannot show

    Work done in HeLa/HEK-class cells on HIV-1 Tat transactivation and heat-shock genes, NOT in leukemia and NOT with a KMT2A fusion present. The extrapolation that AFF4 cannot substitute for AFF1 in a KMT2A::AFF1 leukemia is an inference and should be tested directly before it is relied on.

    Evidence that argues against this (1)
    • Largely distinct target specificities for AFF1-SEC and AFF4-SEC were mapped in cell lines chosen for biochemical tractability rather than in leukaemia cells, and "non-overlapping" is a threshold on a continuum of occupancy. The single-amino-acid difference in P-TEFb affinity is an in vitro measurement whose consequence in a nucleus crowded with both complexes is not established. Nature 2012
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In reconstituted minimal complexes, AF4 and AFF4 independently mediate P-TEFb association with AF9 through direct interactions with AF9 and cyclin T1; P-TEFb links ELL and ELL-associated factors EAF1 and EAF2 to AF4; and DOT1 forms a stable complex with AF9 that does NOT interact with AF9-AF4-P-TEFb complexes, indicating SEC and DOT1L-containing assemblies are distinct rather than one super-complex.

    Function of leukemogenic mixed lineage leukemia 1 (MLL) fusion proteins through distinct partner protein complexes · Biswas D, Milne TA, Basrur V, Kim J, Elenitoba-Johnson KSJ, Allis CD, Roeder RG · Proceedings of the National Academy of Sciences USA 108(38):15751-6 2011
    PMID 21896721DOI
    What this evidence cannot show

    In-vitro reconstitution with recombinant proteins (insect-cell expressed); 'does not interact' is an absence of detectable interaction under those conditions and could be permissive in chromatin with additional bridging factors. The immortalisation readout is MLL-AF9, not KMT2A::AFF1.

    Evidence that argues against this (1)
    • Reconstituted minimal complexes report what purified components do at the concentrations supplied. The negative result carried here — that the DOT1-AF9 complex does not interact with AF9-AF4-P-TEFb — is the kind of finding most sensitive to that limitation, since a weak or transient bridge would be lost on reconstitution while still operating on chromatin. eLife 2021
Connected to