SEC — the AFF1/AFF4-scaffolded super elongation complex, the particle KMT2A::AFF1 is a permanent subunit of
also called super elongation complex · SEC · AEP (AF4-family/ENL-family/P-TEFb) · AFF1/AFF4 complex · ELL/P-TEFb complex
SEC is the machine that tells a stalled gene-copying enzyme to go, and it is the one the infant-leukaemia fusion protein is bolted onto. It is built on the scaffolds AFF1 and AFF4 and carries the kinase P-TEFb, the ELL speed-up subunits and the two mark-reading proteins ENL and AF9.
The same thing, for a scientist
SEC is the parent member of the super elongation complex family: a scaffold module of AFF1 and/or AFF4, the P-TEFb kinase (CDK9 with cyclin T1 or T2), an ELL/EAF elongation-stimulatory module (ELL, ELL2, ELL3 with EAF1/EAF2), and the two YEATS-domain acyl-lysine readers ENL/MLLT1 and AF9/MLLT3 [E1][E3]. It was purified in 2010 from the standpoint of leukaemia — as the complex common to several MLL chimeras [E1] — and independently as AEP, an AF4-family/ENL-family/P-TEFb higher-order complex recruited to MLL target chromatin physiologically and hijacked by MLL fusions [E2]. The reason this entry exists separately from `mechanism-sec-elongation-hijack` is that the hijack entry is a **step in a causal chain** and says what SEC does to a promoter; this entry is the **object**, with a subunit roster, accessions, the scaffold/reader/kinase/ELL module split, and — critically — the family frame that the 2012 SEC-family paper added and which the corpus has only ever held at abstract level: SEC is one of **three** AFF-scaffolded P-TEFb particles, and the two siblings SEC-L2 (AFF2) and SEC-L3 (AFF3) carry the **same two YEATS readers** while controlling **different gene sets** [E3][E4]. That framing is what makes the composition of SEC a therapeutic question rather than a diagram, and it is developed in `mechanism-q187-target-specificity-is-a-scaffold-property-while-the-reader-is-shared`.
The roster, by module, with what each module is and is not
| module | subunits | accession(s) | status in SEC |
|---|---|---|---|
| scaffold | AFF1, AFF4 | P51825 (1,210 aa), Q9UHB7 (1,163 aa) | defining; AFF4 is required for SEC stability and for SEC formation 13 |
| kinase | CDK9, cyclin T1, cyclin T2 | P50750 (372), O60563 (726), O60583 (730) | P-TEFb, the catalytic output |
| elongation stimulation | ELL, ELL2, ELL3 (+EAF1, EAF2) | P55199 (621), O00472 (640), Q9HB65 (397), Q96JC9 (268), Q96CJ1 (260) | present in SEC and NOT detected in SEC-L2 or SEC-L3 4 |
| YEATS readers | ENL/MLLT1, AF9/MLLT3 | Q03111 (559), P42568 (568) | shared with SEC-L2 and SEC-L3 4 |
The two right-hand rows are the whole argument of this beat. The ELL module is the part of SEC that is not in the siblings; the reader module is the part that is. An intervention aimed at one has a family-wide footprint and an intervention aimed at the other does not, and nothing in this corpus had stated that trade-off before 4.
AFF4 is not AF4. AF4 is AFF1 (P51825), the index fusion's 3' partner; AFF4 (Q9UHB7) is a different gene whose UniProt protein name is AF4/FMR2 family member 4, which is how the string "AF4" ends up inside the name of the protein that is not AF4. This is registry row AF4 / AFF4 in data/symbol-collisions.json and the coordinator has already been caught by it once.
One roster caveat the graph already carries. gene-ell3 warns that the strong form of "ELL3 does not occupy the SEC ELL slot" is not established, and that a protein can both mark enhancers and enter the complex. This table lists ELL3 in the ELL slot on the strength of the defining laboratory's own wording — SEC consists of "the elongation factors ELL1 to ELL3 and several other MLL translocation partners" 3 — and should be read as consistent with that entry rather than against it: ELL3 CAN be in SEC, and the ELL-negative result that matters here is the one in SEC-L2 and SEC-L3 4, which covers all three ELLs.
Two independent purifications, and what each one licenses
- Lin et al., Mol Cell 2010 purified several commonly occurring MLL chimeras and found SEC associated with all of them; SEC includes ELL, P-TEFb, AFF4 and other factors, AFF4 is required for SEC stability, and AFF4 knockdown reduces MLL-chimera target gene expression 1.
- Yokoyama et al., Cancer Cell 2010 described the same biology from the other side as AEP, showing it is recruited to MLL-target chromatin physiologically, that MLL oncoproteins fused with AEP components form constitutive MLL/AEP hybrid complexes, and — the part usually dropped when this is cited — that MLL-AF6, a fusion with a cytoplasmic protein, recruits AEP by an unknown alternative mechanism without forming a hybrid complex 2.
That second clause matters for how this graph reasons. AEP recruitment is not synonymous with the partner being an AEP subunit; a fusion can arrive at the same endpoint by a route nobody has mapped. Any argument of the form "the partner is a SEC subunit, therefore the mechanism is SEC delivery" is a hypothesis with one documented exception already in the founding paper.
What a complex entry must and must not carry (rule 69b)
This entry carries no bare data.uniprot field. A single accession on a complex entry is a category error, and verify-accessions.py's fallback regex will happily report a subunit's accession as the entry's own. Accessions live in data.subunits, keyed by symbol, each with its verified length, exactly as the eight existing complex entries in this graph already do.
All twelve accessions and lengths in data.subunits were resolved live against UniProt on 2026-09-24 by gene_exact:<SYM> AND organism_id:9606 AND reviewed:true, keeping only the row whose gene_primary equalled the symbol asked for. Fifteen symbols were queried in that pass — the twelve here plus AFF2, AFF3 and MLLT6, which belong to sibling entries — and all fifteen returned exactly one reviewed record matching on primary name, so rules 62/65 did not fire. That is worth recording precisely because in this gene family they usually do: the registry already holds AF4 / AFF4 and MLLT3 / AF9 rows, and gene_exact:MLLT3 returns an entry named AF9_HUMAN whose protein name contains neither 'MLLT3' nor the queried symbol.
| complex name | SEC (super elongation complex); also AEP (AF4-family/ENL-family/P-TEFb) |
|---|---|
| scaffold | AFF1 and/or AFF4 |
| accession resolution | Twelve subunit accessions resolved live against UniProt on 2026-09-24 with query `gene_exact:<SYM> AND organism_id:9606 AND reviewed:true`, fields accession,id,gene_primary,protein_name,length; only the row whose gene_primary equalled the queried symbol was kept. Fifteen symbols were queried in the pass (these twelve plus AFF2, AFF3, MLLT6); each returned exactly one reviewed record and each matched on primary name. |
| epigenetic readout | SEC carries two YEATS-domain acyl-lysine readers (ENL/MLLT1, AF9/MLLT3) and no writer or eraser; the complex's chromatin readout is entirely the readers' and is therefore shared with SEC-L2 and SEC-L3 |
| shown for fusion | YES for KMT2A chimeras as a class — SEC was purified from several commonly occurring MLL chimeras and found associated with all of them (PMID 20159561), and MLL/AEP hybrid complexes were shown directly (PMID 20153263). NOT specifically re-established on endogenous KMT2A::AFF1 in a primary infant sample by either paper. |
| system | human cell lines with tagged or endogenous baits (293-derived and leukaemic); not primary infant blasts |
| geo accession | GSE34097 |
family members
[ "SEC (AFF1/AFF4)", "SEC-L2 (AFF2) — protein-sec-l2-complex", "SEC-L3 (AFF3) — protein-sec-l3-complex" ]
subunits — the numbers behind the chart
{
"AFF1": {
"accession": "P51825",
"length_aa": 1210,
"role": "scaffold; 3' moiety of the index fusion KMT2A::AFF1"
},
"AFF4": {
"accession": "Q9UHB7",
"length_aa": 1163,
"role": "scaffold; required for SEC stability and formation"
},
"CDK9": {
"accession": "P50750",
"length_aa": 372,
"role": "P-TEFb kinase"
},
"CCNT1": {
"accession": "O60563",
"length_aa": 726,
"role": "P-TEFb cyclin"
},
"CCNT2": {
"accession": "O60583",
"length_aa": 730,
"role": "P-TEFb cyclin"
},
"ELL": {
"accession": "P55199",
"length_aa": 621,
"role": "elongation stimulation; NOT detected in SEC-L2 or SEC-L3"
},
"ELL2": {
"accession": "O00472",
"length_aa": 640,
"role": "elongation stimulation; NOT detected in SEC-L2 or SEC-L3"
},
"ELL3": {
"accession": "Q9HB65",
"length_aa": 397,
"role": "elongation stimulation; NOT detected in SEC-L2 or SEC-L3"
},
"EAF1": {
"accession": "Q96JC9",
"length_aa": 268,
"role": "ELL partner; enters SEC through ELL"
},
"EAF2": {
"accession": "Q96CJ1",
"length_aa": 260,
"role": "ELL partner; enters SEC through ELL"
},
"MLLT1": {
"accession": "Q03111",
"length_aa": 559,
"role": "YEATS-domain acyl-lysine reader; SHARED with SEC-L2 and SEC-L3"
},
"MLLT3": {
"accession": "P42568",
"length_aa": 568,
"role": "YEATS-domain acyl-lysine reader; SHARED with SEC-L2 and SEC-L3"
}
}module split
{
"unique_to_SEC": [
"ELL",
"ELL2",
"ELL3",
"EAF1",
"EAF2",
"AFF1",
"AFF4"
],
"shared_across_the_family": [
"MLLT1",
"MLLT3",
"CDK9",
"CCNT1",
"CCNT2"
]
}marks named
[ "H3K27ac", "H3K9ac", "H3K18cr" ]
measurements
[
{
"quantity": "aff_family_scaffolds_carrying_the_yeats_reader_module",
"subject": "SEC, SEC-L2, SEC-L3 | ENL and AF9 membership by MudPIT + endogenous co-IP + Superose 6",
"value": 3,
"units": "complexes",
"n": 3,
"universe": "the three AFF-scaffolded P-TEFb complexes purified in PMID 22547686 from HEK293 Flp-In-TRex / HEK293T; AFF1- and AFF4-scaffolded SEC taken from the same laboratory's PMID 20159561",
"null": null,
"note": "3 of 3. The comparator quantity is ELL membership, which is 1 of 3 — present in SEC, not detected in SEC-L2 or SEC-L3. The asymmetry between these two rows is the beat's finding."
}
]- SEC is general machinery. CDK9, AFF4 and the ELL module operate on most active genes in most cells, so any agent that disassembles or inhibits SEC has an on-target effect in every dividing tissue. This corpus's own `mechanism-sec-elongation-hijack` already calls this step 'the worst therapeutic target in the chain, despite being the mechanism', and nothing in the composition data changes that.
- The reader module is the part of SEC most often proposed as the druggable handle, and it is the part SEC does not own. ENL and AF9 sit in SEC, SEC-L2 and SEC-L3 (and, separately, in DotCom), so a reader-directed agent is the least fusion-selective choice that can be made from this roster on composition grounds alone. Whether that matters in practice is contested — see `ana-q187-the-scaffold-blindness-argument-does-not-survive-the-published-therapeutic-window`.
- Conversely, the ELL slot is the one place where the family is asymmetric, and it is the slot nobody is drugging. `protein-ell-eaf-module` already records that all three ELL/EAF members score positively or neutrally on selectivity in KMT2A-rearranged lines, so the one compositionally selective handle in SEC points at proteins the dependency data say are not dependencies. The two selectivity arguments — composition and dependency — point in opposite directions here and both are in this graph.
- Is the SEC of a KMT2A::AFF1 blast the same particle as the SEC of a HEK293 cell? Every roster in this entry comes from a non-haematopoietic or non-index system, and the fusion itself is a scaffold subunit, so the stoichiometry it imposes on its own particle is unmeasured.
- What fraction of cellular ENL and AF9 is in SEC versus SEC-L2, SEC-L3, DotCom and the free pool, in a leukaemic cell? This single partition decides whether a reader-directed drug is mostly acting on SEC or mostly acting somewhere else, and no experiment in any system reports it.
- Does a SEC particle ever contain both AFF1 and AFF4, or are AFF1-SEC and AFF4-SEC two populations? AFF1 and AFF4 hetero- and homodimerise through their C-terminal homology domains, so both are possible and the target-set consequences differ.
- MLL-AF6 recruits AEP without forming a hybrid complex (PMID 20153263). What is that alternative route, and does the index fusion use it in parallel with the direct one? A second, redundant recruitment path would make SEC-directed intervention weaker than the hybrid-complex model predicts.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Purification of several commonly occurring MLL chimeras identified a super elongation complex (SEC) associated with all of them; SEC includes ELL, P-TEFb and AFF4 among other factors, AFF4 is required for SEC stability and for proper transcription by poised RNA polymerase II in metazoans, and AFF4 knockdown in leukaemic cells reduces MLL-chimera target gene expression.
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, Lai KC, Martin-Brown S, Florens L, Washburn MP, Conaway JW, Conaway RC, Shilatifard A · Molecular Cell 2010What this evidence cannot show
The abstract names ELL, P-TEFb and AFF4 and says 'several other factors' without listing them; the full subunit roster in this entry's table is assembled from the 2012 family paper's own description of its 2010 result [E4] plus the independent AEP purification [E2], not from this abstract alone. The purifications are from human cell lines with tagged baits, not from patient blasts, and none is the index fusion KMT2A::AFF1 in a primary infant sample.
Evidence that argues against this (1)
- Composition claims for SEC in this corpus are overwhelmingly AFF4 data imported to AFF1 by paralogy — the symbol-collision registry records that AFF4 is crystallised against CDK9 and cyclin T1 while AFF1 has a single PDB entry covering 42 residues. An entry titled 'the complex the index fusion is a subunit of' therefore rests on a scaffold that is not the index fusion's scaffold for most of its structural content, and this table does not fix that. Yokoyama A et al. 2010
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
An independent purification described the same particle as AEP, a higher-order complex of AF4-family and ENL-family proteins with P-TEFb, showed that AEP is normally recruited to MLL-target chromatin to facilitate transcription and that MLL oncoproteins fused with AEP components form constitutive MLL/AEP hybrid complexes, and reported that MLL-AF6 — a fusion with a cytoplasmic partner — does NOT form such a hybrid complex yet still recruits AEP to target chromatin by an unknown alternative mechanism.
A higher-order complex containing AF4 and ENL family proteins with P-TEFb facilitates oncogenic and physiologic MLL-dependent transcription. · Yokoyama A, Lin M, Naresh A, Kitabayashi I, Cleary ML · Cancer Cell 2010What this evidence cannot show
AEP and SEC are described by two groups with partly different rosters and partly different names; this entry treats them as the same object, which is the field's consensus but is an editorial judgement, not a measurement. The MLL-AF6 result is an abstract-level statement about a fusion this beat did not read the full text for.
Evidence that argues against this (1)
- The rosters are not identical and the difference has been used as an argument. Lu and colleagues characterise the SEC family by tandem affinity purification and argue that SEC composition is combinatorial rather than fixed — different SEC species with different ELL and AFF content serve different genes, and HIV-1 Tat selects among them. On that reading a single 'SEC roster' table, including this one, is an average over a population of particles rather than a description of one. Lu H et al. 2015
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The laboratory that purified SEC describes it as 'the AFF1- and AFF4-containing super elongation complex' and as consisting of the elongation factors ELL1 to ELL3 together with several MLL translocation partners including AFF1, AFF4, ENL and AF9, with AFF4 directly interacting with P-TEFb and with AF9 or ENL and being required for SEC formation.
The super elongation complex family of RNA polymerase II elongation factors: gene target specificity and transcriptional output. · Luo Z, Lin C, Guest E, Garrett AS, Mohaghegh N, Swanson S, Marshall S, Florens L, Washburn MP, Shilatifard A · Molecular and Cellular Biology 2012What this evidence cannot show
This is the same laboratory describing its own earlier result in an introduction, so it is a secondary restatement of PMID 20159561 rather than an independent observation. Read from the PMC HTML full text (PMC3434493) because Europe PMC fullTextXML returns HTTP 500 for this article.
Evidence that argues against this (1)
- An introduction's summary of the authors' own prior work is the least scrutinised sentence in a paper and this project has been burned by exactly that class of restatement. The primary purification (PMID 20159561) should be preferred wherever the two differ; they do not appear to differ here, but nobody has diffed the two rosters subunit by subunit and this entry does not claim to have. Lin C et al. 2010
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
SEC is one of three AFF-scaffolded P-TEFb complexes: SEC-L2 (AFF2) and SEC-L3 (AFF3) contain the same two YEATS-domain readers ENL/MLLT1 and AF9/MLLT3 and the same P-TEFb, but neither MudPIT of the purified AFF2 and AFF3 complexes nor Western analysis of endogenous immunoprecipitates from 293 cells detected any ELL protein in SEC-L2 or SEC-L3, and the three complexes control different subsets of genes.
The super elongation complex family of RNA polymerase II elongation factors: gene target specificity and transcriptional output. · Luo Z, Lin C, Guest E, Garrett AS, Mohaghegh N, Swanson S, Marshall S, Florens L, Washburn MP, Shilatifard A · Molecular and Cellular Biology 2012What this evidence cannot show
HEK293/293T only, with over-expressed tagged baits, and the authors themselves state they cannot rule out ELL association with AFF2 or AFF3 under other conditions or in other cell types, offering the roughly 10-fold lower AFF2/AFF3 expression in HEK293T as one reason for the failure to detect. The ELL negative therefore constrains SEC-L2 and SEC-L3 in that one cell type at that one abundance and nothing more.
Evidence that argues against this (1)
- The same paper's Figure 1F schematic predicts the opposite of its own blots: it states that the presence of ELL- and EAF-related proteins in SEC-L2 and SEC-L3 'has not been experimentally confirmed but is expected based on sequence conservation among AFF family members'. A reader taking the cartoon rather than the data would record ELL as present in all three complexes, which would delete the entire ELL-versus-reader asymmetry this entry's table is built on. Luo Z et al. 2012
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
The mRNA-Seq and ChIP-Seq data underlying the SEC-family target-set comparison are publicly deposited as GEO series GSE34097, which records PMID 22547686 as its source publication and has been public since 2 May 2012.
GEO Series GSE34097 — The SEC family of RNA Polymerase II elongation factors: gene target specificity and transcriptional output · Luo Z, Lin C, Shilatifard A · NCBI Gene Expression Omnibus 2012What this evidence cannot show
Fetched 2026-09-24 via the SOFT brief record (acc.cgi targ=self form=text view=brief), which is not behind bot detection. Attribution is by the deposit's own `!Series_pubmed_id` field, not by topic match. This beat confirmed the series exists and carries the right PMID; it did NOT download the data, enumerate the samples or verify that the per-gene differential-expression calls are recoverable from it.
Evidence that argues against this (1)
- A deposit existing is not the deposit containing what you want. GEO series routinely carry processed matrices rather than the differential-expression gene lists a paper reports, and the 226/423/1,116 sets in this paper are Cufflinks/TopHat calls that may have to be recomputed rather than read off. The claim here is only that the raw material is public — treating that as 'the gene lists are available' is the step that would be wrong, and this beat did not take it. Luo Z et al. 2012
- supports →Step 3 — The hijack: the fusion drags the super elongation complex to paused RNA polymerase II
the complex object behind the hijack step: subunit roster, accessions and module split
AFF1 is a scaffold subunit of SEC
AFF4 is a scaffold subunit and is required for SEC stability and formation
ENL is a YEATS-reader subunit of SEC, and of SEC-L2 and SEC-L3 as well
AF9 is a YEATS-reader subunit of SEC, and of SEC-L2 and SEC-L3 as well
P-TEFb is the catalytic module of SEC
the ELL/EAF activation module is in SEC and is the module NOT detected in SEC-L2 or SEC-L3
SEC was purified from several commonly occurring MLL chimeras and found associated with all of them
- ← related toSEC-L2 — the AFF2-scaffolded super-elongation-like complex: P-TEFb plus ENL and AF9, and NO ELL
SEC-L2 is the AFF2-scaffolded paralogue of SEC; shares the readers and P-TEFb, lacks detectable ELL
SEC-L3 is the AFF3-scaffolded paralogue of SEC; shares the readers and P-TEFb, lacks detectable ELL
MYC is reported as a direct target of AFF4/SEC in the family paper
AFF4/SEC covers 72.5% of the three-way DEG union
the targeted surface (YEATS) is shared across the family; the scaffold surface (AHD) is what differs
the axis controls abundance of the SEC scaffolds, so it is a dosage control on the complex itself
measures how much of the cellular reader pool SEC actually holds