Oncofusion

start / graph / mechanism-arginine-methylation-gates-kmt2-access-independently-of-menin-and-the-win-motif

Mechanismmoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q103-arginine-marks

Arginine methylation gates KMT2-complex access to chromatin at two residues, in both directions, independently of the menin tether AND of the WIN motif - a parallel control point, not a rival to menin dependence

also called the arginine switch on WRAD access · H3R2/H4R3 methylation switch · arginine-mark gating of KMT2 complexes

Two arginines on the histone tails act as switches: depending on which enzyme methylates them and how, the KMT2 machine can either grip the chromatin or not. Neither switch touches the menin handle or the WDR5 handle this project is built around, so it is a second, separate lock on the same door.

The same thing, for a scientist

Two histone arginines each carry two isobaric dimethyl states with **opposite** consequences for KMT2-complex engagement, and the mechanism in every case is on the **histone substrate**, not on the KMT2A polypeptide. **H3R2**: the asymmetric state H3R2me2a prevents WDR5 binding to the H3 tail and prevents binding of an ASH2/WDR5/MLL-family complex [E1]; quantitative binding plus crystallography makes this a **binary** switch in which me2a is the only non-permissive state and me0, me1 and me2s are equally permissive [E2]. **H4R3**: the symmetric state H4R3me2s inhibits the tandem PHD(4-6) of KMT2D - a module **inside the catalytic subunit** that KMT2D requires for nucleosomal H3K4 methylation - while H4R3me0 and H4R3me2a are permissive, and PRMT7 but not PRMT5 raises the inhibitory state at KMT2D targets including many HOXA/HOXB genes [E3]. Independently, loss of H3R2me2a at enhancers **reduces** KMT2D binding [E4]. **Neither switch requires menin, the menin-binding motifs, or the WIN motif**, so this is a control point that operates on chromatin the fusion has already reached. **The correct inference is parallel control, not menin-independence of the disease**: the fusion's own recruitment to chromatin is menin- and CXXC-dependent on this graph's evidence, and an arginine switch acting on WRAD's or KMT2D's grip on the tail is downstream of and orthogonal to that. **What would refute the model is already partly published**: PMID 30232013 reports the same mark activating at enhancers and its loss *increasing* KMT2A binding at promoters [E4], and UniProt flatly calls H3R2me2a "a specific tag for epigenetic transcriptional repression" [E5] - which that paper contradicts. And the whole model is **untestable by antibody**, because the two states at each residue are isobaric and predict opposite things [E6].

Go deeper
The four states, and which of them is actually established
residuestatewritereffect on KMT2-family accessstrength
H3R2me2aPRMT6BLOCKS WDR5 and an ASH2/WDR5/MLL-family complex on the H3 tail 1; the only non-permissive H3R2 state 2established - three groups, two organisms, plus a structural re-examination
H3R2me2sPRMT5 (PRMT7 contested)PERMISSIVE, and indistinguishable from me0 and me1 in WDR5 affinity 2 - originally reported as enhancing 7the permissive call is strong; the enhancing call is revised and should not be leaned on
H4R3me2aPRMT1 (major), PRMT6, PRMT3PERMISSIVE for KMT2D PHD(4-6) - but so is H4R3me0, so the mark is not required 3strong for permissiveness, weak for instruction
H4R3me2sPRMT5, PRMT7BLOCKS KMT2D PHD(4-6) 3; separately recruits DNMT3A and couples to DNA methylation 8strong

Read down the 'effect' column and the honest summary is: the switch is real and it is one-sided at each residue. Both established effects are BLOCKS - H3R2me2a blocking WDR5, H4R3me2s blocking KMT2D's PHD. Neither permissive state is established as a gain over unmodified chromatin. So this is a pair of OFF switches with default-on positions, not two-way dials, and that distinction changes the drug logic: relieving a block means inhibiting the writer of the blocking state (PRMT6 for H3R2, PRMT7 for H4R3), and there is no writer to activate for a gain.

This is a correction to the two-way framing that reached this graph through gene-prmt5 and gene-prmt6 and into this beat's own dispatch brief, and it comes from two papers those entries do not cite: PMID 32207970 2 and PMID 23249737 3.

What this means for the menin-interface hypothesis - stated carefully, because the careless version is wrong

The careless version: "WRAD access is controlled by arginine marks without menin, so menin is not the control point." That does not follow and this entry does not assert it.

Why it does not follow. The two mechanisms answer different questions.

questionanswer in this graphdoes arginine methylation bear on it?
How does the fusion get to a locus?menin tether plus the CXXC CpG address; this is the basis of the drug classNo. Both arginine switches act on the H3 or H4 tail, not on the fusion's targeting modules
Once there, can a WRAD-containing KMT2 complex grip the H3 tail?WIN motif and the WDR5 contact - and the WIN motif is lost from every 5'-KMT2A fusionYes. H3R2me2a blocks WDR5's grip on the substrate regardless of the WIN motif's state
Can KMT2D methylate a nucleosome it has reached?its own tandem PHD(4-6) must engage the H4 tailYes, and this is independent of menin, of WDR5 and of the WIN motif

So the right statement is a PARALLEL CONTROL POINT. Menin inhibition removes the fusion from chromatin; an arginine switch changes whether a KMT2-family enzyme can act on chromatin it is already on. They are composable, not competing, and the interesting predictions are about the combination rather than about which one is 'the' mechanism.

And there is a specific reason this is worth writing down rather than filing as an aside. This graph's WRAD adjudication (concept-wdr5-what-is-established-for-wild-type-kmt2a-and-what-for-the-fusion, concept-ash2l-..., concept-rbbp5-...) is stuck on one question - what brings WRAD to a fusion that has thrown away the WIN motif? Those entries consider protein-protein contacts only. A substrate-side gate is a different kind of answer to the same question, and it is the kind that does not require the fusion to have kept anything. Whether it is the right answer is unknown, because nobody has measured an arginine mark at a fusion target in any fusion, in any lineage.

The one asymmetry that makes this testable within a single cell. If the arginine gate acts on the H3/H4 substrate, it should restrain the wild-type KMT2A allele - which is SET-competent and WIN-competent - while leaving the fusion's menin/CXXC-driven occupancy untouched. PMID 30232013 reports exactly the wild-type half: loss of promoter H3R2me2a enhances KMT2A binding 4. Measuring both alleles in one KMT2A-rearranged cell under a PRMT6 perturbation is one experiment with a clear two-outcome readout.

What would refute this model, with the published counter-evidence rather than a balanced-sounding sentence

Refutation 1 - the mark's effect is not uniform, and this is published. PMID 30232013 is the genome-wide test of the H3R2me2a arm and it splits the mark by compartment: at promoters, losing it enhances KMT2A binding and H3K4me3 (consistent with a block); at enhancers, losing it reduces KMT2D binding and H3K4me1/H3K27ac (the opposite sign) 4. A model of the form 'H3R2me2a blocks the KMT2 machine' is therefore false at enhancers in the one system where it was tested genome-wide. The compartment-split version survives; the simple version does not.

Refutation 2 - the database's own verdict is the one the data contradicts. UniProt Q96LA8 states that H3R2me2a "represents a specific tag for epigenetic transcriptional repression" 5. PMID 30232013 finds it at ACTIVE genes and activating at enhancers. A curated one-line functional verdict on an arginine mark has already been wrong once in this exact family, which is a reason to hold the parallel annotation for H4R3me2a - "a specific tag for epigenetic transcriptional activation" on PRMT1's record - as an annotation rather than a law.

Refutation 3 - the permissive states may do nothing. If H3R2me2s is merely permissive rather than enhancing 2, and if KMT2D's PHD(4-6) binds H4R3me0 as well as H4R3me2a 3, then neither 'ON' state is an instruction, and any experiment that raises a permissive mark expecting a gain in KMT2 occupancy will see nothing. That is a prediction of no effect, and it is the cheapest way to kill the two-way framing.

Refutation 4 - it could be absent from this disease entirely. The measurement that would decide it does not exist. "H3R2me2a" AND KMT2A = 1 record and it is murine neural differentiation; "H3R2me2s" AND KMT2A = 0; "H4R3me2a" AND KMT2A = 0 and AND MLL = 0; "H4R3me2s" AND KMT2A = 1 and it is a bulk writer-abundance result in murine MLL-AF9 9. If isomer-resolved mass spectrometry in a t(4;11) line found H3R2 and H4R3 essentially unmethylated, or methylated uniformly regardless of fusion occupancy, the model would have no purchase in this disease - and that is a two-week experiment.

Refutation 5 - the writers point the wrong way for therapy. PRMT6, whose mark is the H3R2 block, is not a DepMap dependency and PRMT6 gain LOWERS HOX transcription. PRMT7, whose mark is the H4R3 block, is not a DepMap dependency either. If the blocks are brakes on a HOXA program, their writers are tumour suppressors in this disease and inhibiting them is the wrong direction - which does not refute the mechanism but does refute the therapeutic reading of it, and is the more likely way this line of work fails.

The measurement constraint, which is not a caveat but a gate on the whole model

At each residue the two states are isobaric - the same residue, the same two methyl groups, differing only in whether both sit on one terminal guanidino nitrogen or one on each - and they predict opposite consequences for KMT2-complex access 23.

Therefore an assay that cannot separate me2a from me2s cannot test this model in either direction, and its result is uninterpretable rather than merely noisy. A ChIP with a non-discriminating reagent reports the sum of an inhibitory and a permissive mark. A drug that converts one state into the other reads as no change. A real inversion can be scored as a null result.

The field's base rate makes this a live risk: more than 25% of over 200 antibodies against 57 histone modifications failed specificity by dot or western blot and more than 20% of the specific ones failed in ChIP 6; and for the residue two positions away, 52 commercial antibodies against the H3K4 methylforms often fail to distinguish states, with reagent choice changing the biological conclusion and producing "substantial divergence from the literature" 10.

Bounded honestly: no published head-to-head me2a-versus-me2s panel was found for either residue in this beat. "H3R2me2a"[tiab] AND antibod*[tiab] = 1; "H4R3me2s"[tiab] AND antibod*[tiab] = 3, none of them a specificity study of this reagent class 9. The Histone Antibody Specificity Database exists and was not queried in this beat - that is a named next step, not a finding.

Consequence, and it belongs in every experiment this graph proposes on this axis: the arbiter is isomer-resolving mass spectrometry, and an antibody may be used only after its discrimination is demonstrated on the specific clone and lot. This is recorded as its own entry - concept-an-assay-that-cannot-separate-me2a-from-me2s-cannot-test-the-arginine-switch - so that it can be cited from any future experiment node rather than restated.

Facts and figures
net shape of the switchTWO OFF SWITCHES WITH DEFAULT-ON POSITIONS, not two two-way dials. Both established effects are blocks (H3R2me2a on WDR5, H4R3me2s on KMT2D PHD4-6); neither permissive state is established as a gain over unmodified chromatin. DRUG CONSEQUENCE: relieving a block means inhibiting the writer of the BLOCKING state - PRMT6 for H3R2, PRMT7 for H4R3 - and there is no writer to activate for a gain. This corrects the two-way framing carried in gene-prmt5, gene-prmt6 and this beat's own dispatch brief.
independence from meninCOMPLETE, and this is a claim about the MECHANISM's inputs, not about the disease. Both switches act on the H3 or H4 N-terminal tail. Neither touches the KMT2A menin-binding motifs, the CXXC CpG-binding module, or the WIN motif. The H4R3 switch does not even involve WDR5 - PHD(4-6) is a module inside the KMT2D catalytic subunit.
relationship to the menin interface hypothesisPARALLEL CONTROL POINT, NOT COUNTER-EVIDENCE. Menin plus CXXC answers 'how does the fusion reach a locus'; the arginine switches answer 'can a KMT2-family complex grip and act on the tail once there'. The two are composable. Specifically: this entry does NOT support a menin-independent model of the disease, and any future entry citing it for that purpose is misreading it. What it DOES do is offer a substrate-side answer to the standing WRAD-adjudication question - what brings WRAD to a fusion that has thrown away the WIN motif - which is the only kind of answer that does not require the fusion to have retained anything.
testable allele asymmetryIf the gate acts on the substrate, it should restrain the SET- and WIN-competent WILD-TYPE KMT2A allele while leaving the fusion's menin/CXXC-driven occupancy untouched. PMID 30232013 reports the wild-type half: loss of promoter H3R2me2a ENHANCES KMT2A binding. Measuring both alleles in one KMT2A-rearranged cell under a PRMT6 perturbation gives a clean two-outcome readout and has never been done.
measurement gateEach residue's two states are ISOBARIC and predict OPPOSITE consequences, so an assay that cannot separate them cannot test this model in either direction and a state inversion reads as no change. Isomer-resolving mass spectrometry is the arbiter; an antibody may be used only after per-clone, per-lot discrimination is demonstrated. See concept-an-assay-that-cannot-separate-me2a-from-me2s-cannot-test-the-arginine-switch.
kmt2a fusion evidenceTWO records across all four marks, and neither is locus-resolved at a fusion target. (1) PMID 30232013: wild-type KMT2A in murine neural differentiation. (2) PMID 28945229: bulk H4R3me2s reduction in murine MLL-AF9 downstream of Cdc73 excision and of Prmt5 transcription - a writer-abundance result. Zero for menin, zero for revumenib, zero for MLL-AF4.
what this graph had beforeNothing. 14 mark-* nodes on 2026-09-22 and every one a lysine or ubiquitin mark; H3R2/H4R3 named in 11 entries with no node behind them; and concept-wdr5-what-is-established-for-wild-type-kmt2a-and-what-for-the-fusion contains the word 'arginine' three times, all three referring to the WDR5 arginine pocket and KMT2A Arg3765 - with ZERO occurrences of PRMT, H3R2 or H4R3. The brief's premise that the WDR5 adjudication does not consider this layer is CONFIRMED by grep.
absence classliterature-absence-bounded
absence class noteMIGRATED 2026-09-23, mechanically and not by judgement: the class string itself begins 'literature'. The field previously held the whole verdict as free text and that string is preserved here verbatim because it says more than the token can: 'literature - bounded with counts, querytranslations, warninglists, re-runs, an operand control and two measured symbol artefacts'
switch table
[
  {
    "residue": "H3R2",
    "state": "me2a",
    "writer": "PRMT6",
    "target_module": "WDR5 / ASH2-WDR5-MLL-family complex on the H3 tail",
    "effect": "BLOCKS",
    "strength": "established"
  },
  {
    "residue": "H3R2",
    "state": "me2s",
    "writer": "PRMT5; PRMT7 contested",
    "target_module": "WDR5",
    "effect": "PERMISSIVE - equal to me0 and me1; the earlier 'enhances' call is revised",
    "strength": "permissive call strong; enhancing call revised"
  },
  {
    "residue": "H4R3",
    "state": "me2a",
    "writer": "PRMT1 (major), PRMT6, PRMT3",
    "target_module": "KMT2D tandem PHD(4-6)",
    "effect": "PERMISSIVE - but H4R3me0 binds equally, so not instructive",
    "strength": "strong for permissiveness, weak for instruction"
  },
  {
    "residue": "H4R3",
    "state": "me2s",
    "writer": "PRMT5, PRMT7",
    "target_module": "KMT2D tandem PHD(4-6); separately recruits DNMT3A",
    "effect": "BLOCKS KMT2D; recruits DNMT3A",
    "strength": "strong"
  }
]
falsifiers
[
  "Genome-wide, the H3R2me2a effect is NOT uniform - already published: at enhancers its loss REDUCES KMT2D binding, the opposite sign from promoters (PMID 30232013). The simple 'arginine methylation blocks the KMT2 machine' model is false at enhancers in the one system tested genome-wide; only the compartment-split version survives.",
  "Raising a permissive state produces NO change in KMT2 occupancy. Predicted by the binary switch (PMID 32207970) and by PHD(4-6)'s equal binding to H4R3me0 (PMID 23249737). This is the cheapest way to kill the two-way framing and it predicts a null.",
  "Isomer-resolved mass spectrometry in a t(4;11) line finds H3R2 and H4R3 essentially unmethylated, or methylated uniformly regardless of fusion occupancy. Then the model has no purchase in this disease. Nobody has run it.",
  "The writers point the wrong way therapeutically: PRMT6 (H3R2 block) and PRMT7 (H4R3 block) are both non-dependencies in DepMap, and PRMT6 gain LOWERS HOX transcription. If the blocks are brakes on a HOXA program, their writers behave as tumour suppressors here and inhibiting them is pro-leukaemic. This refutes the therapeutic reading without refuting the mechanism, and is the likeliest way the line of work fails.",
  "A demonstrated me2a/me2s antibody cross-reactivity in the reagents used by PMID 30232013, 23249737, 19234465 or 21072194 would make those distribution results uninterpretable rather than merely uncertain - and no published discrimination panel exists for either residue."
]
How this could bite us back
  • The therapeutic direction may be inverted. Both established arms of this switch are BLOCKS on KMT2-family activity, and removing them raises H3K4me3 at HOXA/HOXB genes (PMID 30232013, 23249737). In a HOXA-driven leukaemia that is the wrong way. An inhibitor of PRMT6 or PRMT7 - the writers of the two blocking states - would be inhibiting a brake, and neither enzyme is a DepMap dependency, so no screen would have flagged the risk.
  • The available drugs hit the wrong enzymes. The type I PRMT class inhibitor that reached humans (GSK3368715, NCT03666988, terminated after 31 participants, benefit-risk did not support continuation) hits PRMT1 and PRMT6 together - one a putative fusion cofactor, the other a brake - so the net direction on this axis is unpredictable from either arm. The PRMT5 inhibitor that reached myeloid patients (NCT03614728, terminated after 30 participants on internal review of clinical data) targets the enzyme whose H3R2 product is merely permissive under the current best evidence, and whose dominant biology is snRNP assembly.
  • An intervention cannot ADD a mark. There is no established eraser for any of these four states, so the only pharmacological handle is writer inhibition - which can only relieve a block, never impose one. If the therapeutically useful direction is to impose a block on wild-type KMT2A or KMT2D activity, this mechanism offers no route to it.
  • Every pharmacodynamic on this axis is antibody-dependent and the states are isobaric. A trial declaring target engagement on an H3R2 or H4R3 readout with a non-discriminating reagent has measured a sum, and a complete state inversion would report as no change. This is not a caveat about precision - it is a scenario in which a dose-escalation decision is made on a number containing no information.
Still open
  1. Do arginine marks exist at KMT2A-fusion target loci at all? Two records in the entire literature co-occur an arginine mark with KMT2A and neither is locus-resolved at a fusion target. Isomer-resolving mass spectrometry on fractionated chromatin from SEM or RS4;11 versus a non-rearranged B-ALL line needs no new reagent and would open or close this axis in one experiment.
  2. Does the gate discriminate the wild-type allele from the fusion within one cell? The mechanism predicts it should, because the fusion lacks the SET domain, the WIN motif and the PHD1-3/bromodomain cassette, so the substrate-side gate has nothing on the fusion to act through. A PRMT6 perturbation with allele-resolved KMT2A occupancy is the discriminating experiment.
  3. Does KMT2A's own PHD1-3 cassette read H4R3 as KMT2D's PHD(4-6) does? Untested anywhere found in this beat. If it does, the fusion has lost an H4-tail-reading requirement the wild-type allele retains - which converts the arginine gate from a KMT2D story into a direct statement about what the fusion no longer needs.
  4. Is the compartment split of PMID 30232013 - repressive at promoters, activating at enhancers - reproducible outside neural differentiation? It is the most consequential result on this axis and rests on one lineage. If it holds in a haematopoietic system, 'H3R2me2a is repressive' has to be retired corpus-wide including from UniProt's annotation.
  5. Would relieving a block help or harm? The mechanism says removing H3R2me2a raises wild-type KMT2A occupancy and H3K4me3, and removing H4R3me2s raises KMT2D-catalysed H3K4me3 at HOXA/HOXB. In a HOXA-driven leukaemia both are plausibly pro-leukaemic, which means the therapeutically interesting direction may be to ADD the blocks - and there is no way to do that pharmacologically.
Every source behind this entry (10)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    PRMT6 catalyses H3R2 dimethylation in vitro and controls global H3R2me2a in vivo; H3K4me3 on the H3 tail prevents PRMT6 from acting, and conversely the H3R2me2a mark prevented methylation of H3K4 as well as binding to the H3 tail by an ASH2/WDR5/MLL-family methyltransferase complex; independently, H3R2 methylation was shown to inhibit H3K4 trimethylation and the recruitment of WDR5 to histone H3 in vitro, with PRMT6 overexpression reducing MLL1/WDR5 occupancy, H3K4me3 and transcription at HoxA2.

    Methylation of histone H3R2 by PRMT6 and H3K4 by an MLL complex are mutually exclusive. · Guccione E, Bassi C, Casadio F, Martinato F, et al. · Nature 2007
    PMID 17898714DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. The complex-binding block is in vitro on H3 tails, not on nucleosomes and not in a leukaemic cell; the HoxA2 arm is from the companion Genes Dev paper (PMID 18079182) in a neuronal-differentiation model. 'An ASH2/WDR5/MLL-family complex' is the paper's wording and identifies no specific KMT2 paralogue. Nothing here is a measurement on a KMT2A FUSION, which has no SET domain and no WIN motif.

    Evidence that argues against this (1)
    • Genome-wide, the same mark is found at active genes and its loss at enhancers REDUCES KMT2D binding - the opposite sign from the promoter block. So the in vitro occlusion result does not generalise into a uniform statement about KMT2-family access across the genome. Bouchard C et al. 2018
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Quantitative binding analysis with high-resolution crystal structures of WDR5 in complex with unmodified and monomethyl/symmetric-dimethyl L-arginine and with the H3R2me1 peptide supports an interaction regulated by a binary arginine methylation switch: H3R2me2a prevents interaction with WDR5, whereas H3R2me0, me1 and me2s are equally permissive; H3R2 methylarginine isoforms comprise PRMT5-catalysed me1 and me2s and PRMT6-catalysed me1 and me2a.

    A Binary Arginine Methylation Switch on Histone H3 Arginine 2 Regulates Its Interaction with WDR5. · Lorton BM, Harijan RK, Burgos ES, Bonanno JB, et al. · Biochemistry 2020
    PMID 32207970DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read, and no PDB accession is asserted because none was fetched. Isolated recombinant WDR5 with free amino acids and short peptides - not WDR5 inside an assembled WRAD module, not on a nucleosome, not in a cell. It addresses WDR5 affinity only and does not test the RBBP7 exclusion, which is the unrevised other half of the original model.

    Evidence that argues against this (1)
    • The earlier study reported, from SILAC MS profiling of H3-tail interactors plus a WDR5/H3R2me2s co-crystal, that H3R2me2s ENHANCES WDR5 binding - a graded rather than binary model. A free-peptide affinity measurement need not predict chromatin occupancy by a six-subunit complex, so the me2s question is genuinely unresolved; only the me2a block is agreed. Migliori V et al. 2012
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A tandem plant homeodomain PHD(4-6) of MLL4 - the enzyme the paper also calls MLL2 and ALR, i.e. KMT2D - recognises unmethylated or asymmetrically dimethylated histone H4 Arg 3 and is required for MLL4's nucleosomal methyltransferase activity and MLL4-mediated differentiation; Kabuki-syndrome mutations in PHD(4-6) reduce binding and catalytic activity; PHD(4-6) binding is inhibited by H4R3 symmetric dimethylation; PRMT7 but not PRMT5 represses MLL4 target genes by upregulating H4R3me2s and antagonises MLL4-mediated differentiation, with PRMT7 knockdown increasing MLL4-catalysed H3K4me3; and decreased H4R3me2s during differentiation is associated with increased H3K4me3 at a cohort of genes including many HOXA and HOXB genes.

    PMID 23249737DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read; no residue numbers asserted for PHD(4-6). NOMENCLATURE: 'MLL4' here is KMT2D, settled by the MLL2/ALR aliases and the Kabuki-syndrome attribution in the abstract, and not KMT2B whose MLL4 alias travels with WBP7. System is NT2/D1 embryonal carcinoma differentiation - not haematopoietic, not leukaemic, and the enzyme is KMT2D not KMT2A. CRITICALLY, PHD(4-6) binds H4R3me0 as well as H4R3me2a, so the asymmetric mark is PERMISSIVE and not instructive.

    Evidence that argues against this (1)
    • The result concerns KMT2D. KMT2A's homologous PHD1-3 and bromodomain cassette sits at 1431-1765 of Q03164, well 3' of the breakpoint cluster region, and is therefore LOST from every direct 5'-KMT2A fusion; whether KMT2A's PHD1-3 reads H4R3 at all has not been tested anywhere found in this beat. Reading this as a statement about a KMT2A fusion would be a paralogue substitution. UniProt 2026
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a murine neural-differentiation model with PRMT6 knockout, PRMT6-dependent H3R2me2a is detected at ACTIVE genes at both promoter and enhancer sites; loss from promoters leads to enhanced KMT2A binding and increased H3K4me3 with increased target-gene transcription, while at enhancers H3R2me2a peaks co-localise with H3K4me1 and H3K27ac and loss reduces KMT2D binding and H3K4me1/H3K27ac with decreased transcription - so the mark exerts opposing transcriptional effects by genomic location.

    PMID 30232013DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. Murine neural differentiation, not haematopoiesis; this is the ONLY record returned by '"H3R2me2a"[tiab] AND KMT2A[tiab]'. The KMT2A and KMT2D ChIP are on wild-type alleles with no fusion protein present. A PRMT6 knockout removes the writer, not the mark, so indirect effects through PRMT6's non-histone substrates are not excluded by the design as described in the abstract.

    Evidence that argues against this (1)
    • This contradicts the 2007 human and yeast distribution maps and the UniProt annotation, all of which place the mark on inactive chromatin and call it repressive. It is a single study in one lineage against two independent 2007 reports plus a yeast genome-wide map, so the graph holds both and presents neither distribution as settled. Kirmizis A et al. 2007
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt Q96LA8 (PRMT6) states that 'H3R2me2a represents a specific tag for epigenetic transcriptional repression and is mutually exclusive with methylation on histone H3 Lys-4 (H3K4me2 and H3K4me3)', and UniProt Q99873 (PRMT1) states in parallel that PRMT1 is the main enzyme for H4R3me1 and H4R3me2a, 'a specific tag for epigenetic transcriptional activation' - two curated one-line functional verdicts on two asymmetric arginine dimethyl marks, with opposite signs.

    What this evidence cannot show

    Curated annotations, not measurements; both accessions fetched live rather than recalled. IDENTIFIER NOTE: PRMT6 is 375 aa and PRMT1 is 371 aa, four residues apart, inside this project's accession-discrimination limit - the accessions are asserted explicitly and were fetched separately. These sentences are editorial summaries of primary papers examining promoters, and neither is a genome-wide statement.

    Evidence that argues against this (1)
    • The PRMT6 sentence is directly contradicted by genome-wide PRMT6-knockout data showing H3R2me2a at active genes and activating at enhancers. Since the two annotations are parallel in form, the activation verdict for H4R3me2a should be treated with the same suspicion - and independently, the KMT2D PHD(4-6) result shows H4R3me2a is merely permissive rather than activating. Bouchard C et al. 2018
  6. E6established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    More than 200 antibodies raised against 57 different histone modifications were tested in Drosophila, C. elegans and human cells; more than 25% failed specificity tests by dot blot or western blot, and among the specific antibodies more than 20% failed in chromatin immunoprecipitation.

    An assessment of histone-modification antibody quality. · Egelhofer TA, Minoda A, Klugman S, Lee K, et al. · Nat Struct Mol Biol 2011
    PMID 21131980DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. A base rate across 57 modifications; it does NOT report the failure rate for H3R2 or H4R3 methylarginine antibodies, nor whether any tested reagent was assayed for symmetric-versus-asymmetric discrimination. Cited to establish that assuming specificity is unsafe, not to assert that a particular reagent fails.

    Evidence that argues against this (1)
    • Most antibodies in that panel performed well, so a blanket presumption of failure would also be wrong. The actionable constraint is per-clone and per-lot, and the Histone Antibody Specificity Database is where it can be checked - a query this beat did not run. Rothbart SB et al. 2015
  7. E7moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    H3R2 is symmetrically dimethylated by PRMT5 and PRMT7 to give H3R2me2s, present in euchromatic regions; SILAC MS profiling of H3-tail interactors showed the mark excludes RBBP7, a central component of the Sin3a, NuRD and PRC2 co-repressor complexes, and enhances binding of WDR5, a common component of the MLL, SET1A, SET1B, NSL1 and ATAC coactivator complexes, with a crystal structure of WDR5 and the H3R2me2s peptide.

    Symmetric dimethylation of H3R2 is a newly identified histone mark that supports euchromatin maintenance. · Migliori V, Müller J, Phalke S, Low D, et al. · Nat Struct Mol Biol 2012
    PMID 22231400DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. HeLa and NT2 differentiation, not haematopoietic. Confidence MODERATE because the 'enhances WDR5' half is revised by E2; the RBBP7 exclusion is unrevised and, as far as this beat found, never independently re-tested. The PRMT7 attribution is weaker still and is recorded as weak on gene-prmt7.

    Evidence that argues against this (1)
    • The binary-switch study finds me0, me1 and me2s equally permissive for WDR5, removing the gain-of-function reading and with it the symmetric half of the 'two-way switch'. This is the claim this beat's dispatch brief and this graph's PRMT5 entry both lean on. Lorton BM et al. 2020
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Symmetric methylation of histone H4 arginine 3 by PRMT5 is required for subsequent DNA methylation at the human beta-globin locus; H4R3me2s is a direct binding target for DNMT3A through the ADD domain containing the PHD motif, and PRMT5 knockdown loses the mark, loses DNMT3A binding, loses DNA methylation and activates the gene, with the mark present on inactive methylated globin genes in primary erythroid progenitors from adult human bone marrow.

    PMID 19234465DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. One locus (the beta-globin cluster) in K562 plus primary adult bone-marrow erythroid progenitors - human and haematopoietic, but erythroid rather than lymphoid and with no leukaemic fusion present. Antibody-based chromatin readout with no isomer-specificity panel named. The epistatic ordering rests on a knockdown, so indirect routes through PRMT5's non-histone substrates are not excluded.

    Evidence that argues against this (1)
    • The writer of H4R3me2s appears to be locus- and context-dependent: a KMT2D trans-tail study reports PRMT7 and explicitly NOT PRMT5 raising the mark at KMT2D targets, and UniProt names PRMT7 for it. So this globin result should not be generalised into 'PRMT5 writes H4R3me2s' genome-wide, and the DNMT3A arm has not been shown at a KMT2-controlled locus. Dhar SS et al. 2012
  9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    PubMed esearch 2026-09-22, translations read back and every zero re-run identical: '"H3R2me2a"[tiab] AND KMT2A[tiab]' = 1 (30232013, murine neural differentiation); '"H3R2me2s"[tiab] AND KMT2A[tiab]' = 0; '"H4R3me2a"[tiab] AND KMT2A[tiab]' = 0 and AND MLL[tiab] = 0; '"H4R3me2s"[tiab] AND KMT2A[tiab]' = 1 (28945229, murine MLL-AF9, bulk reduction downstream of Cdc73 excision and Prmt5 transcription); all four marks AND menin[tiab] = 0; '"H3R2me2a"[tiab] AND revumenib[tiab]' = 0; '"H3R2me2a"[tiab] AND antibod*[tiab]' = 1; '"H4R3me2s"[tiab] AND antibod*[tiab]' = 3. Operand control KMT2A[tiab] = 1652. So across all four arginine marks there are exactly TWO records co-occurring with KMT2A in a title or abstract, neither of them a locus-resolved measurement at a fusion target.

    What this evidence cannot show

    Bounds TITLES AND ABSTRACTS only; no Europe PMC full-text sweep was run. Two symbol artefacts were measured and must be carried with these numbers: (i) H4R3[tiab] AND KMT2A[tiab] = 0 while H4R3[tiab] AND MLL[tiab] = 3, so the legacy symbol reaches records the modern one does not; (ii) '"H4R3me2s" AND "MLL-AF9"' = 0 while the single KMT2A record IS an MLL-AF9 study whose abstract never writes 'MLL-AF9'. An absence written against one symbol alone would have been false in both directions.

    Evidence that argues against this (1)
    • This project has a measured case of a four-spelling bounded zero on a histone mark being false because the measurement was in a supporting figure. These zeros bound announcements, not literature, and the founding KMT2A-fusion chromatin papers' supplementary tables have not been searched for arginine marks. Bernt KM et al. 2011
  10. E10established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Fifty-two commercial antibodies purported to distinguish the H3K4 methylforms were characterised by histone peptide array and internally calibrated ChIP; many widely used antibodies poorly distinguish the methylforms, and high- versus low-specificity reagents yield dramatically different biological interpretations, producing substantial divergence from the literature for numerous H3K4 methylform paradigms.

    Examining the Roles of H3K4 Methylation States with Systematically Characterized Antibodies. · Shah RN, Grzybowski AT, Cornett EM, Johnstone AL, et al. · Mol Cell 2018
    PMID 30244833DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. About H3K4 methylforms, NOT about arginine methylarginine isomers - it is the residue two positions from H3R2, cited to establish that discrimination failures among closely related methyl states of one residue are common and consequential. It reports nothing about H3R2 or H4R3 antibodies, and no claim here says it does.

    Evidence that argues against this (1)
    • Because it concerns a lysine and me1/me2/me3 rather than symmetric versus asymmetric dimethylarginine, the transfer is by analogy. It would be refuted as a constraint if a published panel showed the arginine reagents discriminate cleanly; none was found ('"H3R2me2a"[tiab] AND antibod*[tiab]' = 1), so the constraint stands as unresolved rather than demonstrated. Rothbart SB et al. 2015
Connected to