No acetyl-dead and no acetyl-mimetic KMT2A was ever put in a methyltransferase assay - the link from K1130/K1133 to catalysis runs entirely through SIRT1, and that is a weaker claim than the abstract makes it sound
also called what PMID 25751424 actually tested · K1130R evidence audit
The paper everyone cites says a chemical tag on MLL1 controls how well the enzyme works. Reading the actual experiments, the enzyme test was done by adding or removing the enzyme that strips the tag - never by making a version of MLL1 that simply cannot carry the tag. So the tag and the enzyme activity are linked by association, not by a clean direct test of the two spots.
The same thing, for a scientist
A full read of PMC4732879 separates what PMID 25751424 demonstrated at K1130/K1133 from what it demonstrated about SIRT1. The residues themselves carry exactly two experiments: antibody-specificity validation on a K→R fragment (Fig 3e) and a K→R peptide competition on a circadian reporter (Supplementary Fig 5b). The in vitro histone methyltransferase assay (Fig 4c) was run on full-length Flag-MLL1-myc with wild-type or catalytically dead SIRT1 - not with any site mutant. There is no acetyl-mimetic (K→Q) construct anywhere in the paper. The abstract's 'impacts the methyltransferase activity of MLL1' is therefore an inference across a SIRT1 intervention, not a site-attributed result.
The complete inventory of experiments that touch the two residues
| figure | construct / material | what it shows | what it does not show |
|---|---|---|---|
| Fig 3e | Flag-M2 fragment, WT vs K1130R, K1133R, K1130/3R, in 293 cells | the custom anti-acetyl-MLL1 antibody is specific to these lysines; the M2 fragment IS acetylated at them in cells | nothing functional |
| Fig 3f | full-length, ± SIRT1, ± SIRT1(H363Y), ± SIRT2, ± EX527 | SIRT1 catalytic activity lowers ac-K1130/K1133; SIRT2 does not | nothing about consequence |
| Fig 3g,h | MEF nuclear extracts, circadian times; FK866, EX527, NAD⁺, β-NMN | the acetyl mark is rhythmic (peaks CT18) and NAD⁺-responsive | nothing about consequence |
| Suppl Fig 5b | peptide, WT vs K1130R/K1133R, competing against MLL1 on a CLOCK–BMAL1 circadian promoter | the WT peptide competes, the K→R peptide does not | this is a competition readout on a reporter, not enzymology, and a peptide is not a protein |
| Suppl Fig 5c | mouse liver at ZT timepoints, WT vs Sirt1-null | the mark is rhythmic in vivo (peaks ZT9) and SIRT1-dependent | nothing about consequence |
| Suppl Fig 6a | truncation removing an N-terminal region containing the two residues | the truncation loses transactivation | a multi-hundred-residue deletion cannot attribute anything to two lysines |
| Fig 4c | full-length Flag-MLL1-myc ± SIRT1 / SIRT1(H363Y) — the HKMT assay on recombinant H3 | SIRT1 deacetylase activity lowers MLL1 methyltransferase output | no site mutant is in this experiment |
🔴 The gap, named precisely
Searched the full text: K1130Q, K1133Q, "glutamine", "mimetic" — all absent. There is no acetyl-mimetic construct in this paper 1.
So the causal chain as published is:
SIRT1 activity ──(Fig 3f)──> ac-K1130/K1133 down
SIRT1 activity ──(Fig 4c)──> MLL1 HKMT output down
and the step joining them — ac-K1130/K1133 → HKMT output — is inferred, not measured. SIRT1 has many substrates, including histones themselves (H3K9, H4K16) and the transcription factors in the same assay. A full-length K1130R/K1133R MLL1 in Fig 4c, showing no further effect of SIRT1, would have closed it. That experiment is not in the paper.
This is a correction to how this graph has been stating the finding. ana-the-clock-sirt1-acetylation-site-is-at-kmt2a-1130-1133-... quotes the claim as "it affects the methyltransferase activity of MLL1" in bold, sourced to the abstract. The abstract does say that. The experiments do not support it at the residue level, and the difference matters here more than usual, because the beat's central objection is precisely about whether the methyltransferase consequence transfers to a SET-less fusion. It turns out the methyltransferase consequence is not firmly attached to the site even in wild-type KMT2A.
🟢 What IS solidly established, and it is not nothing
- The residues are acetylated in cells — antibody validated against three K→R constructs 1.
- SIRT1, specifically, is the eraser. Catalytically dead SIRT1(H363Y) fails; SIRT2 fails; SIRT3, SIRT6 and SIRT7 fail to deacetylate MLL1 in vitro; TSA (class I/II HDAC inhibitor) has no effect 1.
- CBP and p300 are the writers for MLL1 acetylation overall. PCAF does not acetylate MLL1; neither do CLOCK, MOF or HAT1 1. ⚠️ Caveat the graph must carry: the CBP/p300 experiment (Fig 3b,c) used the pan-acetyl-lysine antibody, which the paper itself says "preferentially detects C-terminal acetylation of MLL1". CBP/p300 were never shown with the site-specific antibody to write K1130/K1133. The writer for these two residues specifically is, strictly, unassigned in this paper.
- The mark is rhythmic and metabolically coupled — CT18 peak in MEFs, ZT9 peak in mouse liver, up under FK866 and EX527, down under NAD⁺ and β-NMN 1.
- The SIRT1-interaction region of MLL1 is aa 650–1327, mapped with a six-fragment truncation panel and co-IP 1. See the separate entry on what that coordinate implies for the fusion.
And a note on the systems used
Every cellular experiment is HEK 293, mouse embryonic fibroblasts (including Mll1-null MEFs) or mouse liver. No haematopoietic cell of any kind appears in this paper, let alone a leukaemia line. That is worth stating in the same breath as the retention geometry, because the geometry makes the site look immediately relevant to leukaemia and the experimental base is a liver clock study.
| script | computed inline (Europe PMC fullTextXML text mining), 2026-09-24 |
|---|---|
| status | run |
| method | Stripped PMC4732879 XML to plain text, sentence-split, and enumerated every sentence and figure legend naming K1130/K1133, a mutant, an assay or an acetyltransferase. Token search for K1130Q, K1133Q, 'glutamine', 'mimetic'. |
| result summary | No acetyl-mimetic anywhere. Site mutants appear only in antibody validation and a peptide competition. The HKMT assay carries no site mutant, so 'acetylation impacts MLL1 methyltransferase activity' is an inference across a SIRT1 intervention. |
| acetyl mimetic tested | false |
| acetyl dead in hkmt assay | false |
| eraser | SIRT1 |
| writer for k1130 k1133 specifically | unassigned - the CBP/p300 result used a pan-acetyl antibody the authors describe as preferentially detecting C-terminal acetylation |
| haematopoietic system used | false |
| corrects | ana-the-clock-sirt1-acetylation-site-is-at-kmt2a-1130-1133-and-lies-208-residues-inside-the-retained-portion, which quotes the abstract's 'affects the methyltransferase activity of MLL1' in bold without noting that no site mutant was assayed for catalysis |
inputs
[ "PMC4732879 fullTextXML, 113,427 bytes, HTTP 200, fetched 2026-09-24" ]
site mutants used
[ "K1130R", "K1133R", "K1130/3R" ]
eraser negatives
[ "SIRT2", "SIRT3", "SIRT6", "SIRT7", "class I/II HDACs (TSA-insensitive)" ]
writer for mll1 overall
[ "CREBBP", "EP300" ]
writer negatives
[ "KAT2B/PCAF", "CLOCK", "KAT8/MOF", "HAT1" ]
sirt1 interaction region AA
[ 650, 1327 ]
systems used
[ "HEK 293", "mouse embryonic fibroblasts", "Mll1-null MEFs", "mouse liver" ]
measurements — the numbers behind the chart
[
{
"quantity": "point mutants of K1130/K1133 in a catalytic assay",
"subject": "PMID 25751424 Fig 4c HKMT assay",
"value": 0,
"units": "constructs",
"n": 3,
"universe": "the 3 site mutants the paper made (K1130R, K1133R, K1130/3R)",
"null": null,
"note": "all three appear only in Fig 3e antibody validation; none reaches the enzymology"
},
{
"quantity": "haematopoietic systems used",
"subject": "PMID 25751424 all figures",
"value": 0,
"units": "systems",
"n": 4,
"universe": "the 4 systems in the paper: HEK293, MEFs, Mll1-null MEFs, mouse liver",
"null": null,
"note": "no leukaemia line, no primary blood cell, no haematopoietic tissue"
},
{
"quantity": "C-terminal boundary of the mapped SIRT1-interaction region",
"subject": "MLL1 truncation panel M1-M6, co-IP with SIRT1 in 293 cells",
"value": 1327,
"units": "residue (Q03164 numbering)",
"n": 6,
"universe": "six MLL1 truncation fragments spanning the protein",
"null": null,
"note": "region is aa 650-1327; ends 11 residues before the first breakpoint-cluster exon boundary at 1338"
}
]- If an intervention were built on the premise 'block acetylation at K1130/K1133 to reduce KMT2A methyltransferase activity', the premise has never been tested with a site mutant and the fusion has no SET domain to inhibit. Two independent reasons the rationale fails before any chemistry starts.
- SIRT1 activators and inhibitors both exist in the clinic. Reaching for one on the strength of this paper would be acting on a SIRT1 phenotype, not on a KMT2A site, and SIRT1 has hundreds of substrates including p53 - this graph already records a SIRT1 direction conflict in leukaemia that this result does not resolve.
- Does a full-length KMT2A K1130R/K1133R still lose methyltransferase output when SIRT1 is added? This is the experiment that would attach the catalytic consequence to the site, and it has never been run. It is a one-plasmid extension of the paper's own Fig 4c.
- Do CBP or p300 write K1130/K1133 specifically? The paper's writer result used a pan-acetyl antibody it describes as biased toward the C-terminus, and the site-specific antibody it made was never used in the CBP/p300 experiment. An orthogonal answer is already available in PXD005252 (the CBP/p300 acetylome), which contains K1133 - see the PTM-database entry.
- What is the Supplementary Fig 5b construct actually - a peptide or a protein? The main text says both in one sentence. The supplementary is not in the Europe PMC XML and this ambiguity is load-bearing for how strong the only functional site result is.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
In PMID 25751424 the only K1130/K1133 point mutants are arginine substitutions used for antibody validation (Fig 3e) and a peptide competition (Suppl Fig 5b); the in vitro HKMT assay (Fig 4c) uses full-length MLL1 with wild-type or H363Y SIRT1 and no site mutant; no lysine-to-glutamine acetyl-mimetic appears anywhere; CBP and p300 acetylate MLL1 while PCAF, CLOCK, MOF and HAT1 do not; and the SIRT1-interacting region maps to aa 650-1327.
NAD(+)-SIRT1 control of H3K4 trimethylation through circadian deacetylation of MLL1 · Aguilar-Arnal L, Katada S, Orozco-Solis R, Sassone-Corsi P · Nature Structural & Molecular Biology 22(4):312-8 2015What this evidence cannot show
This audit reads the main text, the Methods and the figure legends of the Europe PMC author-manuscript XML. SUPPLEMENTARY FIGURES ARE NOT IN THAT XML - Supplementary Fig 5b and 6a are described only through the main-text sentences that cite them, so the peptide-competition design is known from the authors' description and not from the figure. If Supplementary Fig 5b in fact contains a full-length K->R protein rather than a peptide, the central criticism here softens. The main text's own wording - 'a protein that is functionally unable to compete ... whereas the wild type form of the PEPTIDE is able to do so' - is internally inconsistent about protein versus peptide, and that inconsistency is unresolved.
Evidence that argues against this (1)
- The absence of an acetyl-mimetic is itself weak evidence about acetylation biology: K->Q is a poor mimetic for many sites, and referees in 2015 frequently accepted K->R alone. The criticism here is not that the authors were careless; it is that this graph has been quoting the abstract's causal phrasing as though the site-level experiment existed.
the abstract's catalytic claim is not supported at the residue level; no site mutant reached the HKMT assay
SIRT1 is the specific eraser; SIRT2/3/6/7 and class I/II HDACs are negatives
supplies the site-resolved CBP/p300 evidence the source paper could not