Every partner this project imports from the only MLL-fusion AP-MS interactome was pulled from a Strep-HA transgene in Jurkat T-ALL cells — so the host's own lineage factors are in the prey pool, and one of them is LEF1
also called Jurkat host confound · PXD009338 system · Skucha 2018 experimental system
The one experiment that has ever mapped what MLL fusion proteins bind was done by putting a tagged copy of each fusion into Jurkat cells — a T-cell leukaemia line that does not have the fusion and is the wrong lineage for the infant B-cell disease we study. Anything the host cell is full of will show up in the pulldown, and for a T-cell line that includes T-cell transcription factors.
The same thing, for a scientist
PXD009338 / Skucha 2018 is the only MLL-fusion AP-MS interactome and 72 entries in this corpus cite it. Its system: **isogenic Jurkat (human T-ALL) lines carrying a Doxycycline-inducible, single-copy, Strep-HA-tagged MLL-fusion transgene installed by FRT/Flp cassette exchange**, nuclear extract, benzonase/RNase, single-step StrepTactin, LTQ Orbitrap Velos, Mascot+Phenyx against SwissProt 2012/2013 at <1% protein FDR [E1][E2]. Three consequences the corpus mostly states in caveats and should state in claims (rule 73): the host is **T-lineage**, not the infant B-lineage disease; the bait is an **ectopic tagged transgene**, not the endogenous fusion at its own locus; and the prey pool is the **Jurkat nuclear proteome**, so any abundant T-cell nuclear factor is a candidate false positive. That last point is not hypothetical — **LEF1, UniProt's 'T cell-specific transcription factor 1-alpha', is one of the six p<0.05 MLL-AF4-specific preys**, and its independent leukaemia literature is T-ALL.
What the system is, precisely
| element | value | source |
|---|---|---|
| host cell | Jurkat, human T-cell acute lymphoblastic leukaemia | paper body 1 |
| bait installation | FRT/Flp-mediated locus-specific cassette exchange, single copy | paper body 1 |
| bait induction | Doxycycline-inducible | paper body 1 |
| tag | Strep-HA (SH) | paper body 1 |
| baits | MLL-AF1p, MLL-AF4, MLL-AF9, MLL-CBP, MLL-EEN, MLL-ENL, MLL-GAS7 | paper body 1 |
| input | nuclear extract from 1e9 cells, 420 mM NaCl extraction adjusted to 150 mM | PRIDE PXD009338 2 |
| nucleic acid | benzonase 20 U/ml + RNase A 50 ng/ml, 15 min 4 C | PRIDE PXD009338 2 |
| purification | single-step StrepTactin sepharose, 2 h, 3 washes, biotin elution | PRIDE PXD009338 2 |
| instrument | LTQ Orbitrap Velos, top-15 CID | PRIDE PXD009338 2 |
| search | Mascot 2.3.02 + Phenyx 2.6 vs human SwissProt, <1% protein FDR | PRIDE PXD009338 2 |
| contaminant removal | trypsin and keratin only | PRIDE PXD009338 2 |
Two of these deserve emphasis. Single-step affinity purification carries substantially more background than the tandem purification the same group's earlier protocol supports, and the authors compensate statistically rather than by a second purification step. And benzonase plus RNase means no interaction in this dataset is DNA- or RNA-bridged: a genuine chromatin co-occupancy that requires the template would have been destroyed, which cuts against the dataset in the opposite direction to everything else here.
Why the host lineage is not a caveat but a claim-level fact
This corpus's standing finding is that its load-bearing evidence is partner-imported and system-mismatched: 0 of 24 sampled load-bearing experiments rest on a dedicated human KMT2A::AFF1 experiment. PXD009338 is a particularly clean instance — it does study MLL-AF4, and it does so in a cell that is the wrong lineage, with a bait that is not the endogenous protein.
The operational test is whether the host's identity changes which preys are plausible. It does:
- LEF1 (Q9UJU2, UniProt protein name "T cell-specific transcription factor 1-alpha") is MLL-AF4's fourth-ranked specific prey at p = 0.0261 on 18 spectra, and is the only one of the six with non-zero negative-control signal (1|3). Its independent leukaemia literature is T-ALL:
LEF1[tiab] AND "T-cell acute lymphoblastic"[tiab]returns 20 records,LEF1[tiab] AND "B-cell precursor"[tiab]returns 2, andLEF1[tiab] AND KMT2A[tiab]returns 1 off-topic record 3. In a T-ALL host, LEF1 is an abundant resident nuclear factor. Seeinterface-kmt2a-lef1-declined-a-t-lineage-factor-in-a-t-lineage-host. - Conversely, a genuine B-lineage-restricted KMT2A::AFF1 partner — anything requiring EBF1, PAX5 or an early-B chromatin state — cannot appear in this dataset at all, because the protein is not expressed in the host. The AF4-specific list is therefore biased against exactly the partners the PI's standing priority is about.
Both directions matter and the second is the more important one: this dataset can generate T-lineage false positives and is structurally blind to B-lineage true positives.
| status | run |
|---|---|
| result summary | The only MLL-fusion AP-MS interactome is a Dox-inducible single-copy Strep-HA transgene in Jurkat T-ALL, purified in one StrepTactin step from benzonase/RNase-treated nuclear extract. The prey pool is the Jurkat nuclear proteome; LEF1, a T-lineage factor, is among the six p<0.05 MLL-AF4-specific preys. |
| shown for fusion | MLL-AF4 specifically, as one of seven baits - but as a Dox-inducible single-copy Strep-HA transgene in a Jurkat T-ALL host, not the endogenous KMT2A::AFF1 in B-lineage disease, and no B-lineage KMT2A-rearranged line appears in the AP-MS arm. |
| epigenetic readout | none in the AP-MS arm - no histone mark or occupancy measurement; the benzonase/RNase treatment removes the chromatin template before purification, so no interaction in this dataset can be DNA- or RNA-bridged. |
| absence class | never-measured |
| absence class note | There is no AP-MS interactome of the endogenous KMT2A::AFF1 in a B-lineage host. Closing it needs an endogenous-tagging or antibody-based pulldown in SEM or RS4;11, i.e. real wet work, not a re-analysis. |
inputs
[ "PMC5959866 full-text XML, fetched 2026-09-24", "PRIDE archive REST record for PXD009338, fetched 2026-09-24", "PubMed esearch via scripts/pubmed_safe.py, 2026-09-24" ]
measurements
[
{
"quantity": "count",
"subject": "corpus entries citing PMID 29777171",
"value": 72,
"units": null,
"n": 3032,
"universe": "all node files under data/graph/nodes as of 2026-09-24",
"null": null,
"note": "grep -rl on the PMID string; the dataset is load-bearing for this project"
},
{
"quantity": "pubmed_records",
"subject": "LEF1[tiab] AND \"T-cell acute lymphoblastic\"[tiab]",
"value": 20,
"units": "records",
"n": null,
"universe": "PubMed titles and abstracts, 2026-09-24",
"null": 2,
"note": "null column is LEF1[tiab] AND \"B-cell precursor\"[tiab]; LEF1 AND KMT2A returns 1, off-topic"
}
]host cell lines
[ "Jurkat (human T-ALL)" ]
- Any therapeutic argument built on a partner from this dataset inherits a T-lineage host. A partner that is real in Jurkat and absent in infant B-ALL blasts would fail in exactly the patients the programme is for, and nothing in the source can distinguish the two cases.
- The blindness runs the other way too: a B-lineage-restricted vulnerability cannot be discovered from this dataset at all, so an absence here is not evidence of absence in the index disease.
- Which of the 291 MLL-AF4 preys are genes Jurkat expresses unusually highly relative to a B-lineage KMT2A-rearranged line such as SEM? A CCLE/DepMap expression join would turn 'the host lineage is a confound' from an argument into a ranked list, and would say which of the 110 are simply abundant in the host. No such join has been run here or anywhere in this corpus.
- Does any of the seven baits recover a partner that the Jurkat proteome makes structurally impossible to see - i.e. is there a published KMT2A::AFF1 partner from B-lineage work that is absent from all seven columns because the host does not express it? EBF1 and PAX5 are the obvious candidates to check.
- Given benzonase and RNase pre-treatment, is any reported chromatin co-occupancy of KMT2A::AFF1 expected to be MISSING from this dataset? If so, the pan-fusion core is also incomplete in a specific, predictable way.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Skucha 2018 generated isogenic Jurkat T-ALL lines by FRT/Flp-mediated locus-specific cassette exchange allowing Doxycycline-inducible, single-copy expression of Strep-HA-tagged MLL-AF1p, MLL-AF4, MLL-AF9, MLL-CBP, MLL-EEN, MLL-ENL and MLL-GAS7, and purified protein complexes from nuclear lysates of those lines.
MLL-fusion-driven leukemia requires SETD2 to safeguard genomic integrity. · Skucha A, Ebner J, Schmöllerl J, Roth M, et al. (20 authors; Grebien F senior author) · Nature Communications 2018What this evidence cannot show
Read from the Europe PMC full-text XML of PMC5959866 on 2026-09-24. The expected_absent_terms record that no B-lineage KMT2A-rearranged line (SEM, RS4;11) and no primary infant sample appears anywhere in the AP-MS arm - that absence is part of the claim. The paper's LATER functional work does use MV4-11 and primary AML cells, so a reader searching the whole paper will find leukaemia-relevant systems; they belong to the SETD2 validation, not to the interactome.
Evidence that argues against this (1)
- The authors' choice of Jurkat is defensible and deliberate: an isogenic single-copy system in a cell lacking any endogenous MLL fusion is what makes a seven-bait comparison interpretable at all, and they demonstrate the baits are nuclear and induce HOXA5/HOXA9/HOXA10/MEIS1, i.e. the fusions are functional in that host. A B-lineage KMT2A-rearranged host would have carried its own endogenous fusion and destroyed the comparison. The criticism is therefore not of the design but of importing its prey lists into B-lineage claims without the lineage in the claim. Nature Communications 2018
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset
The PXD009338 deposit records single-step StrepTactin purification from 15 mg of benzonase- and RNase-A-treated Jurkat nuclear extract with three washes and biotin elution, analysed on an LTQ Orbitrap Velos and searched with Mascot 2.3.02 plus Phenyx 2.6 against human SwissProt at <1% protein FDR, with trypsin and keratin the only contaminants removed.
PXD009338 - Protein Interactions of 7 MLL-fusion Proteins (sample-processing and data-processing protocols) · Skucha A, et al., deposited via PRIDE 2018-04-23, public 2018-05-25 · ProteomeXchange / PRIDE, EMBL-EBI 2018What this evidence cannot show
The PRIDE record gives the SwissProt version as v2012.05 and the paper's Methods give v2013.01; the discrepancy is unresolved and matters only for which paralogues were in the search space (see interface-kmt2a-rgpd8-declined-ninety-six-percent-shared-tryptic-peptides). The expected_absent_terms record that this is explicitly NOT a tandem affinity purification despite the Strep-HA tag naming two epitopes - only the Strep step was used - which is part of the claim about background. Number and composition of the control purifications is not stated in either source.
Evidence that argues against this (1)
- A single-step purification with a statistical background model is a legitimate and widely used design, and the model here is more careful than a simple fold-change: it fits a GLM with a per-prey baseline estimated from control purifications and a per-replicate normalisation anchored on DDX5/DDX17. Whether that compensates for the missing second purification step is not testable from outside the deposit, and the correct statement is that the format has higher background exposure than TAP and the authors addressed it statistically rather than experimentally. Nature Communications 2018
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
LEF1's independent leukaemia literature is T-lineage, matching the host cell rather than the index disease: on 2026-09-24 PubMed returns 20 records for LEF1[tiab] AND "T-cell acute lymphoblastic"[tiab], 2 for LEF1[tiab] AND "B-cell precursor"[tiab], 1 for LEF1[tiab] AND KMT2A[tiab] (a bladder-cancer PRMT5-WDR5 paper), 0 for LEF1[tiab] AND AFF1[tiab] and 0 for 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); querytranslations read back and non-empty for every query, so no zero here is an instrument failure (rule 97). · NCBI PubMed E-utilities 2026What this evidence cannot show
Bounded by titles and abstracts only; a LEF1 measurement inside a supplementary figure of a KMT2A paper is invisible to this (rule 80). Names tried: LEF1, LEF-1 (which returns 1 record with MLL, PMID 16510874). The retired/alternative symbols TCF1-alpha and TCF1A were NOT tried and LEF1 is old enough that some 1990s literature uses them; the count is therefore a floor. Querytranslations were non-empty in every case, e.g. LEF1[tiab] AND KMT2A[tiab] -> "LEF1"[Title/Abstract] AND "KMT2A"[Title/Abstract].
Evidence that argues against this (1)
- The one substantive MLL-LEF1 connection in the literature is real and points the other way: PMID 16510874 (Sierra et al., Genes Dev 2006) shows beta-catenin recruits MLL to Wnt target genes with H3K4 methylation, and LEF1 is beta-catenin's DNA-binding partner, so a KMT2A-LEF1 proximity has a mechanism. Against that: CTNNB1 and every other TCF/LEF family member (TCF7, TCF7L1, TCF7L2) are ABSENT from all seven baits' lists in Skucha 2018, so the bridge protein the mechanism requires was not co-recovered. Sierra J et al. 2006
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
bounds what the only MLL-fusion AP-MS interactome can say about the index fusion: T-ALL host, tagged single-copy transgene, nuclear extract, benzonase-treated
the same dataset's absence from IntAct is why none of this is visible to the project's interactome coverage figures
the decline turns on the host lineage established there
the host-lineage confound has a named instance: LEF1, UniProt's 'T cell-specific transcription factor 1-alpha', in a T-ALL host
the dataset's absence from IntAct, established here, is the third of the four mechanisms that produced the coverage inversion