Oncofusion

start / graph / interface-kmt2a-acadm-declined-a-mitochondrial-matrix-enzyme-in-a-nuclear-pulldown

Interfaceweak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.Q180-the-af4-specific-interactors

ACADM is a mitochondrial matrix beta-oxidation enzyme with an experimentally coded matrix-only localisation, recovered on 8 spectra from a nuclear-extract pulldown whose specific tail is 14.5% mitochondrial — declined

also called MLL-AF4-ACADM · KMT2A-ACADM · MCAD · Medium-chain specific acyl-CoA dehydrogenase

ACADM is an enzyme that burns fat inside mitochondria. The experiment pulled proteins out of cell nuclei, so ACADM should not be there at all — and one in seven of the proteins said to be uniquely bound by MLL-AF4 is a mitochondrial enzyme like it. This is what contamination looks like, and we are declining it.

The same thing, for a scientist

**Verdict: DECLINED on compartment.** ACADM (P11310, 421 aa, MCAD) carries the single UniProt subcellular-location comment *"Mitochondrion matrix"* with an experimental code, ECO:0000269|PubMed:16020546 — **no nuclear annotation at any evidence level** [E1]. Its evidence as an MLL-AF4 partner is one row in Skucha 2018: p = 0.0208, spectral counts 5|3 = 8, neg_ctrl 0|0, rank 53 of 291, seen in no other bait [E2]. The pulldown is from **nuclear extract** [E3], and the class ACADM belongs to — the 110 MLL-AF4-only preys — is **14.5% mitochondrial against 2.4% for the 42 pan-fusion preys of the same bait, Fisher two-sided p = 0.042** [E4]. ACADM is one of sixteen mitochondrial proteins in that class, four of which (ACADM, ECHS1, GLUD1, PDHA1) are matrix enzymes of beta-oxidation, the TCA cycle and pyruvate entry. **A matrix enzyme in a nuclear pulldown, in a class enriched for matrix enzymes, is background until something else says otherwise, and nothing else does:** `ACADM[tiab] AND leukemia[tiab]` returns 1 record [E5].

Go deeper
Why compartment is the deciding argument here and not merely a caveat

This corpus has declined partners on compartment before — interface-kmt2a-tmco1-declined-an-er-calcium-channel-crosslinked-in-nuclei, interface-psip1-fastkd2-declined-a-mitochondrial-matrix-bait-for-a-nuclear-prey — and in each case the caveat was that the compartments touch: the ER membrane is continuous with the outer nuclear membrane, and mitochondria contact the nuclear envelope. That escape hatch is narrower for a matrix protein than for a membrane protein. ACADM is inside the inner membrane, behind two bilayers, in a compartment with no continuity with the nucleoplasm and no described transport route for a folded 421-residue flavoenzyme.

What makes this decisive rather than merely suggestive is that the rate was measured rather than asserted: 16 of 110 in the specific tail against 1 of 42 in the shared core of the same bait, in the same preparation, scored by the same annotation source 4. The single pan-fusion "mitochondrial" protein is PML, dual-annotated to nucleus and PML body. ACADM is not an unlucky individual; it is a member of a class that the experiment's own internal comparison identifies as contamination-enriched.

The preparation makes this expected. Nuclei were pelleted at 500 x g after hypotonic 0.1% NP-40 lysis and extracted at 420 mM NaCl 3 — a protocol that routinely carries mitochondria into the nuclear pellet, after which a high-salt extraction solubilises some of what came along. Known contaminants were removed after the search, but the removal list was trypsin and keratin 3; no organelle filter was applied.

What would reopen it

A moonlighting nuclear pool. Several metabolic enzymes do have documented non-canonical nuclear functions, and this is the only route by which ACADM becomes interesting. Nothing supports one for ACADM specifically: UniProt records no nuclear pool, and the entirety of ACADM's clinical and biochemical literature is medium-chain acyl-CoA dehydrogenase deficiency, the commonest inherited disorder of fatty-acid oxidation. A single paper reporting nuclear ACADM in any cell type would change this verdict; none was found 5.

The cheap check, if anyone wants one, is the same as for the rest of the tail: the peptide-level evidence in PXD009338. Eight spectra is not nothing, and if they are clean unique peptides then ACADM really was in the sample — which would still be an argument about the preparation rather than about KMT2A::AFF1.

Facts and figures
partner agene-kmt2a
partner a uniprotQ03164
partner a length AA3969
partner bACADM
partner b uniprotP11310
partner b length AA421
verdictDECLINED
verdict scopeDeclined on compartment plus class-level contamination rate. Not a claim that fatty-acid oxidation is irrelevant to KMT2A-rearranged leukaemia - that is a separate question this row is simply not evidence about.
adjudicated byQ180-the-af4-specific-interactors, cycle 159, 2026-09-24
source publication pmid29777171
source datasetPXD009338, Skucha 2018 seven-bait MLL-fusion AP-MS
detection methodsingle-step StrepTactin affinity purification from Jurkat nuclear extract, LC-MS/MS spectral counting with a GLM background model
experimental role partnerprey
experimental role kmt2abait (as the MLL-AF4 fusion)
bait exists in datasettrue
n publications1
n methods1
orthogonal method existsfalse
intact curatedfalse
af4 p0.0208
af4 spectral5|3
af4 spectral total8
neg ctrl0|0
rank in af4 by p53
n fusions1
druggabilityNot assessable and not pursued. No interface, no affinity, no second method, and a compartment mismatch that is not bridgeable.
absence classmeasured-not-searched
absence class noteThe peptide-level evidence for these eight spectra is in PXD009338 and has never been read. It would establish whether ACADM was genuinely in the sample - which would indict the nuclear preparation rather than support a contact.
shown for fusionMLL-AF4 only, as one prey row in one AP-MS experiment in a Jurkat T-ALL host with a tagged transgene.
epigenetic readoutnone - no histone mark, occupancy or chromatin measurement exists for ACADM in any context.
host cell lines
[
  "Jurkat (human T-ALL)"
]
compartment mismatch
{
  "partner_compartments": [
    "Mitochondrion matrix [ECO:0000269|PubMed:16020546]"
  ],
  "partner_nuclear_annotation": "none at any evidence level",
  "sample_compartment": "nuclear extract",
  "escape_hatch": "none of the usual kind - a matrix protein sits behind two membranes with no continuity to the nucleoplasm, unlike the ER-membrane and outer-mitochondrial-membrane cases this corpus has previously declined",
  "class_rate": "16 of 110 MLL-AF4-only preys are mitochondrial against 1 of 42 pan-fusion; Fisher two-sided p 0.0417"
}
measurements
[
  {
    "quantity": "spectral_counts",
    "subject": "ACADM|MLL-AF4",
    "value": 8,
    "units": "spectra (5|3 over two replicates)",
    "n": 2,
    "universe": "Skucha 2018 Suppl Table 5, MLL-AF4 column",
    "null": 0,
    "note": "null column is the neg_ctrl (0|0); p = 0.0208, rank 53 of 291"
  }
]
interface residues a
[]
structures
[]
ligands
[]
How this could bite us back
  • This corpus has a live metabolism thread. If a beat later argues for an FAO or CPT1 intervention in KMT2A-rearranged disease, it must not cite this row as physical evidence - the argument would have to stand on dependency or metabolomic data, and citing a contaminated pulldown alongside real data would make the case look stronger than it is.
Still open
  1. Has a nuclear pool of ACADM ever been reported in any cell type? A single such report would reopen this. The query was run under the official symbol only; MCAD was not searched.
  2. Are the sixteen mitochondrial members of the AF4-specific tail present in the OTHER six baits' specific tails as well? If the same sixteen recur across baits, they are a preparation signature and every one of them can be declined in a single pass rather than individually.
  3. Does any published KMT2A-rearranged leukaemia study report a fatty-acid-oxidation dependency? That question is independent of this row and is not answered by declining it.
Every source behind this entry (5)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    ACADM is UniProt P11310 (ACADM_HUMAN), primary gene symbol ACADM, 421 aa, 'Medium-chain specific acyl-CoA dehydrogenase, mitochondrial (MCAD) (EC 1.3.8.7)', with a single subcellular-location comment 'Mitochondrion matrix' coded ECO:0000269|PubMed:16020546 and no nuclear annotation at any evidence level.

    What this evidence cannot show

    Resolved by accession with the primary gene name read back (rules 62, 65); ACADM has no gene synonyms listed and no row in data/symbol-collisions.json. The expected_absent_terms entry records that the ABSENCE of any nuclear comment is part of the claim. Rule 91 applies to the positive annotation too: ECO:0000269|PubMed:16020546 means a curator saw an experiment in that paper, not that the paper is the founding or best localisation study for MCAD; the cited paper was not opened.

    Evidence that argues against this (1)
    • UniProt's silence on a nuclear pool is not evidence against one. This corpus has measured that roughly 10.7% of its entries rest on uncoded UniProt localisation and rule 91 records that even coded annotations often point at the wrong paper; an under-curated moonlighting pool would look exactly like this. The decline rests on the class-level rate in E4, where a single protein's annotation gaps average out, rather than on this record alone. This repository, entry ana-uniprot-localization-is-not-a-measurement-in-a-quarter-of-this-graph 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    ACADM's entire evidence as an MLL-AF4 partner is one prey row in Skucha 2018 Supplementary Table 5: p = 0.0208, spectral counts 5|3 (8 summed), neg_ctrl 0|0, rank 53 of 291, recovered under MLL-AF4 and none of the other six baits.

    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 2018
    PMID 29777171DOI
    What this evidence cannot show

    Read from the cached Supplementary Table 5 workbook with the column map asserted against the sheet's header row; the value appears nowhere in the paper's abstract or body, and ACADM is not discussed anywhere in the paper. The p-value is from the study's own GLM and is a rank-selected top-300 call, not a threshold pass - the largest p retained in the MLL-AF4 column is 0.175.

    Evidence that argues against this (1)
    • Eight spectra across two replicates (5 and 3) is reproducible detection, not a single lucky PSM, and the zero negative control means ACADM was not detected in the control purification at all. On the study's own statistic ACADM is the second-best-supported of the 110. If this beat is wrong about any of the six, ACADM is a candidate, because the reproducibility across replicates is better than for RGPD8, RPA2 or CEP112. Nature Communications 2018
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.dataset

    The purification was from nuclear extract — nuclei pelleted at 500 x g after hypotonic 0.1% NP-40 lysis, extracted at 420 mM NaCl and adjusted to 150 mM — with a single StrepTactin step and three washes, and the only contaminants removed after the database search were trypsin and keratin.

    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 2018
    What this evidence cannot show

    The expected_absent_terms entry records that no organelle or mitochondrial filter is named in either protocol; that absence is part of the claim. The protocols do not state whether a sucrose cushion or any further nuclear purification step was used beyond the single buffer-N wash, so the degree of mitochondrial carry-over is not quantified anywhere by the authors.

    Evidence that argues against this (1)
    • The study's GLM does address background statistically: it estimates a per-prey baseline from control AP-MS experiments and normalises per replicate on DDX5/DDX17. A mitochondrial protein present in the nuclear preparation should therefore also be present in the control preparation and be subtracted - and ACADM's control is 0|0, which means the model saw no baseline for it. That is the strongest single argument against this decline, and it is unresolved: a zero control on a low-abundance protein is also what insufficient control depth produces. Nature Communications 2018
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    The class ACADM belongs to is contamination-enriched: 16 of the 110 MLL-AF4-only preys (14.5%) carry a mitochondrial UniProt annotation against 1 of 42 (2.4%) of the pan-fusion preys of the same bait, Fisher exact two-sided p = 0.0417, and the single pan-fusion case is PML, which is dual-annotated to nucleus and PML body.

    What this evidence cannot show

    Membership in the pan-fusion class requires clearing the top-300 cut in all seven baits, which selects for abundant well-behaved nuclear proteins, so part of the contrast is definitional rather than evidence of contamination. At n = 42 in the smaller arm the Fisher p is fragile - two more mitochondrial calls in the pan-fusion class would put it above 0.05. Annotations were counted at any ECO level (rule 91).

    Evidence that argues against this (1)
    • A class-level enrichment does not convict an individual member, and using it to do so is the ecological fallacy. 94 of the 110 are NOT mitochondrial; if ACADM were the one real metabolic partner in the list, this statistic would look identical. The class rate raises ACADM's prior of being background from unknown to high; it does not establish it, which is why this entry's verdict is a decline on the balance of evidence rather than a refutation. This repository 2026
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.database

    There is no independent literature connecting ACADM to KMT2A, to any MLL fusion, or to leukaemia biology: ACADM[tiab] AND leukemia[tiab] returns 1 record on 2026-09-24, and no report of a nuclear ACADM pool was found.

    What this evidence cannot show

    Bounded by titles and abstracts (rule 80) - a nuclear ACADM result in a supplementary figure is invisible to this. Name list tried: ACADM only. NOT tried: MCAD, MCADH, 'medium-chain acyl-CoA dehydrogenase', or the British spelling leukaemia - and MCAD is the name almost all of the clinical literature uses, so this count badly under-samples ACADM's literature as a whole. It is adequate only for the narrow question of whether ACADM has been linked to leukaemia under its official symbol, and is a floor even for that.

    Evidence that argues against this (1)
    • The name list here is weak in exactly the way rule 71 warns about, and the caveat says so: MCAD was not searched and it is the dominant name. A reader should treat this evidence item as the weakest in the entry. It carries no weight in the verdict, which rests on E1 and E4; it is recorded so that a later beat can see what was and was not searched rather than repeating it. UniProt Consortium 2026
Connected to