LEF1 — the Wnt/β-catenin DNA-binding factor of the T-lineage, named as an MLL-AF4-specific interactor in a T-ALL cell line and with essentially no B-lineage or KMT2A literature
also called LEF1 · LEF-1 · Lymphoid enhancer-binding factor 1 · T cell-specific transcription factor 1-alpha · TCF1-alpha · Q9UJU2 · LEF1_HUMAN
LEF1 is the protein that sits on DNA and lets the Wnt signal switch genes on; it is mostly a T-cell protein. It turns up in the one experiment that looked at what MLL fusion proteins bind — but that experiment was done in T-cells, which is exactly where LEF1 is abundant anyway, so the finding may be about the cells rather than the fusion.
The same thing, for a scientist
LEF1 (Q9UJU2, 399 aa, LEF1_HUMAN) is an HMG-box transcription factor and the DNA-binding partner of β-catenin in canonical Wnt signalling. UniProt's own protein name is *"T cell-specific transcription factor 1-alpha"* and its only subcellular-location annotation is nuclear by PROSITE rule inference (ECO:0000255), not by experiment [E1]. In leukaemia its literature is T-lineage: 20 PubMed records with "T-cell acute lymphoblastic", 2 with "B-cell precursor", 1 with KMT2A (off-topic) and 0 with AFF1 or MLL-AF4 [E2]. The most recent primary work shows a recurrent intragenic LEF1 deletion in T-ALL producing a dominant-negative isoform and a Wnt/β-catenin vulnerability [E3]. **This entry exists because LEF1 is the highest-spectral-count (10|8) of the six MLL-AF4-specific preys at p<0.05 in Skucha 2018 — and because that recovery is confounded three ways: the host cell is Jurkat T-ALL, LEF1 is the only one of the six with non-zero negative-control signal, and β-catenin and every other TCF/LEF family member are absent from all seven baits' lists.** The adjudication is in `interface-kmt2a-lef1-declined-a-t-lineage-factor-in-a-t-lineage-host`.
What LEF1 does
LEF1 belongs to the TCF/LEF family (LEF1, TCF7, TCF7L1, TCF7L2). Each binds a WRE (Wnt-responsive element) through an HMG box and, in the absence of nuclear β-catenin, recruits TLE/Groucho corepressors to keep the target silent; stabilised β-catenin displaces TLE and converts the same site to an activator 4. LEF1 has no intrinsic activation domain of its own — it is a DNA-binding adaptor, and its transcriptional output is entirely the property of whatever it recruits. That is the reason a physical association with a chromatin writer would be interesting if it were real, and also the reason such an association is hard to distinguish from co-residency on the same nucleosome.
In haematopoiesis LEF1 is a T-lineage and early-B factor; UniProt names it T cell-specific transcription factor 1-alpha 1. Its nuclear localisation carries only ECO:0000255 — a PROSITE profile rule — with no experimental citation, which is a rule-91 case: the compartment everyone assumes has no coded measurement behind it in the database, although the primary literature has demonstrated nuclear β-catenin–LEF1 complexes directly 4.
The one mechanistic route from LEF1 to KMT2A, and why it does not close
Sierra, Yoshida, Joazeiro and Jones (Genes Dev 2006) showed that on Wnt target genes β-catenin recruits MLL-family H3K4 methyltransferase activity and that APC counteracts both β-catenin occupancy and H3K4 methylation there 4. LEF1 is β-catenin's DNA anchor at those sites, so a KMT2A–LEF1 proximity has a published mechanism: LEF1 → β-catenin → MLL. That is the strongest argument available for taking the AP-MS row seriously, and it is an indirect, bridged one.
It does not close, for a measured reason. In the same seven-bait table, CTNNB1 is absent from the network entirely, and so are TCF7, TCF7L1, TCF7L2, APC, GSK3B, TLE1 and TLE3 5. The bridge protein the mechanism requires was not co-purified with any of the seven MLL fusions. Either the bridge is not present in this system, or it is present and the pulldown could not hold it — and nothing here distinguishes those.
Coverage note
Before this entry, LEF1 had no node in this graph and was named in 5 node files, none of them about LEF1: gene-hdac7, gene-mllt11, mark-h3r2me2s, a Q143 promoter-ranking analysis, and the King's own cycle-158 gap report. The gap was real. Whether it should have been closed with a substantive gene entry rather than a decline is the judgement this beat is making, and the argument for a gene entry is that LEF1's Wnt biology is worth having in the graph independently of whether the AP-MS row survives — the graph already holds gene-ctnnb1, and a Wnt node with no DNA-binding factor is incomplete.
| uniprot | Q9UJU2 |
|---|---|
| length AA | 399 |
| gene primary verified | LEF1, read back from UniProtKB on 2026-09-24; no gene synonyms listed; no row in data/symbol-collisions.json |
| entry name | LEF1_HUMAN |
| protein name | Lymphoid enhancer-binding factor 1 (LEF-1) (T cell-specific transcription factor 1-alpha) (TCF1-alpha) |
| family | TCF/LEF HMG-box transcription factors (LEF1, TCF7, TCF7L1, TCF7L2) |
| localization | Nucleus |
| localization evidence code | ECO:0000255 (PROSITE-ProRule:PRU00267) only - no ECO:0000269 experimental localisation citation in UniProt |
| localization shown for fusion | NO - nobody has looked at LEF1's localisation in any KMT2A fusion context. |
| shown for fusion | MLL-AF4 only, as one prey row in one AP-MS experiment, in a Jurkat T-ALL host with a tagged single-copy transgene. No B-lineage, no endogenous fusion, no second method. |
| epigenetic readout | none for LEF1 itself in any KMT2A context. The nearest epigenetic datum is H3K4 methylation at Wnt target genes under beta-catenin (PMID 16510874), which is a measurement about beta-catenin, not about LEF1. |
| absence class | measured-not-searched |
| absence class note | Whether beta-catenin was detected under MLL-AF4 below the rank cut is answerable from PXD009338 without wet work, and it is the single datum that would most change the reading of the LEF1 row. |
measurements — the numbers behind the chart
[
{
"quantity": "p_value",
"subject": "LEF1|MLL-AF4",
"value": 0.0261,
"units": null,
"n": 2,
"universe": "Skucha 2018 Suppl Table 5, GLM over spectral counts, top-300 per bait, Jurkat T-ALL host",
"null": null,
"note": "rank 62 of 291 in the MLL-AF4 column; the 4th of the six AF4-specific preys below p 0.05"
},
{
"quantity": "spectral_counts",
"subject": "LEF1|MLL-AF4",
"value": 18,
"units": "spectra (10|8 over two replicates)",
"n": 2,
"universe": "Skucha 2018 Suppl Table 5",
"null": 4,
"note": "null column is LEF1's own neg_ctrl (1|3) - the ONLY one of the six AF4-specific p<0.05 preys with non-zero control signal"
},
{
"quantity": "pubmed_records",
"subject": "LEF1 AND T-cell acute lymphoblastic",
"value": 20,
"units": "records",
"n": null,
"universe": "PubMed titles and abstracts 2026-09-24, one [tiab] per term",
"null": 2,
"note": "null column is LEF1 AND \"B-cell precursor\"; LEF1 AND AFF1 returns 0"
}
]wnt cofactors in the skucha network
{
"LEF1": [
"MLL-AF4"
],
"CTNNB1": null,
"TCF7": null,
"TCF7L1": null,
"TCF7L2": null,
"APC": null,
"GSK3B": null,
"TLE1": null,
"TLE3": null,
"note": "null means absent from the 960-protein network, i.e. below the top-300 cut in all seven baits; not absent from the raw data"
}- Wnt/β-catenin inhibition is an active therapeutic idea in leukaemia, and a LEF1 node in this graph could be read as supporting it for KMT2A-rearranged disease. It does not: the only KMT2A-adjacent evidence is one AP-MS row in the wrong lineage, and the recurrent LEF1 lesion in T-ALL is LOSS of function, so a Wnt-pathway inhibitor could in principle phenocopy an oncogenic event rather than oppose one.
- LEF1 is required for normal T- and early-B-lymphopoiesis, so systemic LEF1 or Wnt blockade in an infant carries a developmental-toxicity risk that has nothing to do with the fusion.
- Is LEF1 expressed at all in infant KMT2A::AFF1 B-ALL blasts, and at what level relative to Jurkat? If it is not expressed, the AP-MS row is closed outright. This is a lookup in any published infant KMT2A-r expression dataset and nobody in this corpus has done it.
- Was CTNNB1 detected under MLL-AF4 below the top-300 cut in PXD009338? Its presence would revive the beta-catenin bridge; its genuine absence from the raw data would close it.
- Does KMT2A::AFF1 occupy any canonical Wnt-responsive element in a B-lineage KMT2A-rearranged line? The deposited occupancy data this corpus already uses could answer this without new experiments, and would test the mechanism directly rather than through a co-purification.
- LEF1 has no intrinsic activation domain. If a KMT2A-LEF1 association were real, which direction would it run - LEF1 recruiting the fusion to WREs, or the fusion carrying LEF1 to HOXA? No experiment in any system distinguishes these.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
LEF1 is UniProt Q9UJU2 (entry name LEF1_HUMAN), primary gene symbol LEF1, 399 aa, protein name 'Lymphoid enhancer-binding factor 1 (LEF-1) (T cell-specific transcription factor 1-alpha) (TCF1-alpha)', with a single subcellular-location comment 'Nucleus' carrying evidence code ECO:0000255 (PROSITE-ProRule:PRU00267) and no experimental ECO:0000269 localisation citation.
UniProtKB Q9UJU2 (LEF1_HUMAN), Lymphoid enhancer-binding factor 1 — accession, primary gene name, length and subcellular-location comment retrieved 2026-09-24 · UniProt Consortium, EMBL-EBI/SIB/PIR 2026What this evidence cannot show
Resolved by accession, and the primary gene name was read back and equals LEF1 (rules 62, 65). LEF1 has no gene synonyms listed and no row in data/symbol-collisions.json (216 rows, checked with the corrected dict-aware command before the first query). The expected_absent_terms entry records the absence of an ECO:0000269 localisation code as part of the claim - rule 91's point is that this absence is informative, not that the protein is not nuclear.
Evidence that argues against this (1)
- The absence of a coded experimental localisation in UniProt says nothing about LEF1's actual compartment, which is not in doubt: nuclear beta-catenin-LEF1 complexes have been visualised directly and a nuclear LEF1 is the premise of the entire Wnt literature. This is a database-completeness observation, and quoting it as though it weakened the nuclear assignment would be a misreading of rule 91. Molecular and Cellular Biology 1999
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
LEF1's leukaemia literature is T-lineage and it has essentially no KMT2A literature: on 2026-09-24 PubMed returns 20 records for LEF1[tiab] AND "T-cell acute lymphoblastic"[tiab], 152 for LEF1[tiab] AND leukemia[tiab], 37 for LEF1[tiab] AND "acute lymphoblastic leukemia"[tiab], 2 for LEF1[tiab] AND "B-cell precursor"[tiab], 2 for LEF1[tiab] AND MLL[tiab] (both off-topic), 1 for LEF1[tiab] AND KMT2A[tiab] (a bladder-cancer PRMT5-WDR5 paper), and 0 for both LEF1[tiab] AND AFF1[tiab] and LEF1[tiab] AND "MLL-AF4"[tiab].
PubMed esearch via scripts/pubmed_safe.py, run 2026-09-24 by Q180; one [tiab] tag per term (rule 94), querytranslation read back and non-empty for every query (rule 97), cytogenetic terms quoted (rule 95). · NCBI PubMed E-utilities 2026What this evidence cannot show
Bounded by titles and abstracts only - a LEF1 blot in a supplementary figure of a KMT2A paper is invisible (rule 80), and this is exactly the class of measurement a mechanism graph wants. Name list tried: LEF1, LEF-1. NOT tried: TCF1-alpha, TCF1A, Lef1 (mouse), so the counts are floors and the older literature is under-represented (the retired-symbol rule). Querytranslations were non-empty throughout, e.g. LEF1[tiab] AND KMT2A[tiab] -> "LEF1"[Title/Abstract] AND "KMT2A"[Title/Abstract], so no zero here is an instrument failure.
Evidence that argues against this (1)
- Searching LEF-1 rather than LEF1 changes the answer: LEF-1[tiab] AND MLL[tiab] returns PMID 16510874, the beta-catenin/MLL H3K4 paper, which the LEF1 spelling misses. That is a live instance of the alias rule and it means the '2 records, both off-topic' figure for LEF1 AND MLL is an artefact of spelling, not a real emptiness. The corrected statement is that the LEF1-MLL literature is one paper, indirect and in colon-cancer cells, not zero. Sierra J et al. 2006
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
In T-ALL, a recurrent intragenic LEF1 deletion produces a dominant-negative LEF1 isoform and exposes a Wnt/β-catenin vulnerability — establishing that LEF1 loss-of-function, not gain, is the recurrent lesion in the lineage where LEF1 is most studied.
LEF1 intragenic deletion induces a dominant-negative isoform and unveils a Wnt/β-catenin vulnerability in T-ALL. · Delafoy M, Bonnet MF, Lengliné E, Cieslak A, et al. (17 authors; Asnafi V senior author) · Blood 2025What this evidence cannot show
Read from the PubMed record (title, authors, journal, volume 146 pages 3036-3049, DOI) fetched 2026-09-24; the full text was NOT read and no figure or n was extracted, so this entry states only what the title asserts. The expected_absent_terms record that the paper is not about KMT2A or the index fusion - it is cited here as the best current evidence on what LEF1 does in leukaemia, in a lineage that is not the index disease (rule 73).
Evidence that argues against this (1)
- A T-ALL result constrains nothing about infant B-lineage KMT2A::AFF1, and citing it here risks exactly the generalisation rule 73 exists to prevent. It is included because the DIRECTION matters for the interface adjudication: if LEF1's recurrent role in leukaemia is loss-of-function in T-cells, an argument that KMT2A::AFF1 co-opts LEF1 as a gain-of-function partner in B-cells has no supporting precedent in either lineage. No paper was found that reports a LEF1 lesion, expression change or dependency in KMT2A-rearranged leukaemia of any lineage. NCBI PubMed E-utilities 2026
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
β-catenin recruits MLL-family H3K4 methyltransferase activity to Wnt target genes and the APC tumour suppressor counteracts both β-catenin occupancy and H3K4 methylation there, which supplies the only published mechanistic route from a TCF/LEF factor to KMT2A — an indirect one, through β-catenin.
The APC tumor suppressor counteracts beta-catenin activation and H3K4 methylation at Wnt target genes. · Sierra J, Yoshida T, Joazeiro CA, Jones KA · Genes & Development 2006What this evidence cannot show
Cited from the PubMed record and title; the full text was not read for this entry, so no figure, cell line or antibody is claimed. The system is colon-cancer cells, not leukaemia, and the recruitment is to beta-catenin, not to LEF1 - LEF1's involvement is inferred from its being beta-catenin's DNA anchor, which is an inference this entry makes and not something the paper states about LEF1. The expected_absent_terms record that the paper says nothing about any fusion or about leukaemia.
Evidence that argues against this (1)
- A 2006 paper predating the KMT2A/KMT2B/KMT2D nomenclature says 'MLL' in a sense that may not be KMT2A specifically - rule 66's trap, where MLL2 and MLL4 each name two genes and the protein names are swapped. Which KMT2 family member carries the H3K4 activity at Wnt targets is therefore not settled by this citation, and the entry should not be read as saying KMT2A itself was assayed. UniProt Consortium 2026
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
In Skucha 2018's 960-protein seven-bait network, LEF1 appears under MLL-AF4 only, and CTNNB1, TCF7, TCF7L1, TCF7L2, APC, GSK3B, TLE1 and TLE3 are absent from the network entirely — so no Wnt co-factor was co-recovered with LEF1 under any of the seven MLL fusions.
Wnt-effector sweep of Skucha 2018 Supplementary Table 5, run 2026-09-24 by Q180. Script analyses/q180-is-af4-specificity-a-depth-artefact/compartment_of_the_tail.py, output compartment_of_the_tail.json, key wnt_effectors_in_network. · This repository (script and JSON committed; deterministic) 2026What this evidence cannot show
'Absent from the network' means absent from the top-300-by-p list of every one of the seven baits. It does NOT mean absent from the raw data: the paper recovered 4,600 proteins in total and the table carries only 960, so beta-catenin may well have been detected below the rank cut in one or more baits. That is a measured-not-searched absence and it is answerable from PXD009338. CREBBP is in the table as a singleton of MLL-CBP, i.e. as the bait's own moiety, not as a Wnt co-factor.
Evidence that argues against this (1)
- Beta-catenin is a notoriously difficult AP-MS prey: the free nuclear pool is small, the bulk of cellular beta-catenin is at adherens junctions and therefore not in a nuclear extract at all, and the beta-catenin-TCF interaction is salt-labile. Its absence from a 150 mM StrepTactin pulldown is therefore weak evidence that no bridge exists, and a reader should not treat 'the bridge protein was not co-purified' as 'there is no bridge'. The honest statement is that the dataset neither supports nor excludes the beta-catenin route. ProteomeXchange / PRIDE, EMBL-EBI 2018
the one AF4-specific prey at p<0.05 that already has its own entry, found only by alias-resolving coverage
LEF1 is beta-catenin's DNA-binding partner at Wnt-responsive elements; CTNNB1 is absent from all seven baits in the Skucha network
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
named as an MLL-AF4-specific prey at p 0.0261 on 18 spectra in the only MLL-fusion AP-MS interactome; adjudicated and declined
the adjudication of LEF1's single AP-MS row; DECLINED on host-lineage confound plus the only non-zero control of the six
the host-lineage confound has a named instance: LEF1, UniProt's 'T cell-specific transcription factor 1-alpha', in a T-ALL host