Oncofusion

start / graph / mark-h4r3me2a

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

H4R3me2a - the activating arginine mark that a KMT2-family enzyme's OWN tandem PHD module reads, where the Kabuki-syndrome mutations sit, and which no KMT2A-fusion study has ever measured

also called H4R3me2a · H4R3me2 · asymmetric dimethylation of histone H4 arginine 3 · H4 Arg-3 asymmetric dimethylation · H4R3me2a/H2AR3me2a

H4R3me2a is an activating tag on histone H4, written mainly by PRMT1, and it is the tag a reading module inside one of the KMT2 enzymes grabs in order to work on chromatin at all. The mutations that cause Kabuki syndrome sit in that reading module and weaken exactly this grip.

The same thing, for a scientist

H4R3me2a is the asymmetric dimethyl mark on histone H4 arginine 3. **PRMT1 is the main enzyme**, established by mass spectrometry and a dedicated antibody in human 293T cells and called "the major, if not exclusive, H4 Arg-3 methyltransferase" there [E1]; the same year, purification of a H4-specific methyltransferase identified PRMT1 and showed that **H4R3 methylation facilitates subsequent acetylation of H4 tails by p300**, while H4 acetylation inhibits its methylation by PRMT1 - the founding statement of the mark as activating [E2]. UniProt records PRMT1 as "the main enzyme" for H4R3me1 and H4R3me2a and calls the mark "a specific tag for epigenetic transcriptional activation" [E3]. PRMT6 methylates H4/H2A R3 in vitro [E4] and PRMT3 deposits H4R3me2a at a promoter in human mesenchymal stem cells [E5]. It has a genuine **reader**: the Tudor domain of TDRD3, identified by protein-domain microarray as an effector of H4R3me2a and H3R17me2a, and itself a transcriptional coactivator whose activity requires an intact Tudor domain [E6]. **The result that makes this mark load-bearing for this project is the trans-tail one.** The tandem plant homeodomain PHD(4-6) of KMT2D (the paper's "MLL4", also called MLL2 and ALR) **recognises unmethylated OR asymmetrically dimethylated H4R3 and is required for KMT2D's nucleosomal methyltransferase activity**, with Kabuki-syndrome mutations in PHD(4-6) reducing both binding and catalysis - while the symmetric state H4R3me2s inhibits that binding [E7]. Independently, loss of H3R2me2a at enhancers reduces KMT2D binding [E8]. So **an asymmetric arginine mark is permissive for a KMT2-family enzyme through that enzyme's own reader module, with no involvement of menin, the WIN motif or WDR5.** **No eraser is established** [E9][E10]. **In a KMT2A-rearranged context there is nothing**: `"H4R3me2a"[tiab] AND KMT2A[tiab]` = 0 and `AND MLL[tiab]` = 0, both re-run, against `KMT2A[tiab]` = 1652 [E11].

Go deeper
The trans-tail result, and why it is the most important thing this beat found

Everything else in this graph's WRAD arm reasons about recruitment of the complex to the gene. PMID 23249737 is about something different: a reading module inside a KMT2-family enzyme, which must engage the H4 tail for the enzyme to methylate a nucleosome at all.

The findings, in the authors' order 7:

  1. The tandem PHD(4-6) of KMT2D recognises unmethylated or asymmetrically dimethylated H4 Arg-3 (H4R3me0 or H4R3me2a).
  2. That module is required for KMT2D's nucleosomal methyltransferase activity and for KMT2D-mediated differentiation of NT2/D1 stem cells.
  3. Kabuki-syndrome mutations in PHD(4-6) reduce its binding ability and KMT2D's catalytic activity - a human genetic disease mapping onto this exact contact.
  4. H4R3me2s inhibits PHD(4-6) binding, and PRMT7 - but not PRMT5 - represses KMT2D target genes by raising H4R3me2s, antagonising KMT2D-mediated differentiation; PRMT7 knockdown increases KMT2D-catalysed H3K4me3.
  5. During differentiation, falling H4R3me2s tracks rising H3K4me3 at a cohort of genes including many HOXA and HOXB genes.

Three consequences for this project.

(a) The switch is not confined to H3R2, and the H4R3 version acts on the enzyme, not on the scaffold. The H3R2 switch gates WDR5's grip on the tail. The H4R3 switch gates a PHD cassette inside a KMT2 catalytic subunit. These are different machines and both are menin-independent.

(b) The genes are the right genes. HOXA and HOXB. This is the only arginine-mark result in the literature that lands on the gene family this disease lives on, and it does so through a KMT2-family enzyme.

(c) It is about KMT2D, not KMT2A, and that limit is load-bearing. KMT2A has PHD1-3 and a bromodomain in a cassette this graph already treats as a unit (subtraction-kmt2a-phd3-bromo-cassette, interface-kmt2a-phd3-h3k4me3, mechanism-phd3-cyp33-repressor-switch-lost), and that whole cassette is LOST from every 5'-KMT2A fusion. Whether KMT2A's PHD1-3 reads H4R3 at all has not been tested in any source read for this beat. So the honest statement is: a homologous module in a paralogue does this, KMT2A's counterpart module is deleted in the disease, and nobody has looked. That is a specific, cheap experiment, not a gap that needs a programme.

The nomenclature trap, stated because this project has been bitten by it

PMID 23249737 calls its enzyme "MLL4", and "MLL4" is the single worst symbol in this gene family. This graph already records the trap: MLL4 is an alias of both KMT2D (with MLL2 and ALR) and KMT2B (with WBP7), and searching the bare symbol pulls skeletal-muscle KMT2D papers in one direction and movement-disorder KMT2B papers in the other.

The disambiguation here is unambiguous and rests on two independent tells in the abstract itself:

  1. The paper writes "MLL4; also called MLL2 and ALR". gene-kmt2d in this graph carries exactly the aliases MLL2, MLL4, ALR. MLL2 and ALR are not KMT2B synonyms.
  2. The paper attributes Kabuki syndrome mutations to the PHD(4-6) it studies. Kabuki syndrome is a KMT2D disease.

So "MLL4" in PMID 23249737 means KMT2D (O14686, 5537 aa), and the edges from this entry go to gene-kmt2d. Recorded explicitly so that no later reader re-derives it, and so that nobody attaches this result to KMT2B on the strength of the bare alias.

Writers, and the honest ranking among them
enzymeevidence for H4R3me2astrength
PRMT1H4 Arg-3 methylation shown in vivo by mass spectrometry in asynchronously growing human 293T cells plus a dedicated anti-H4R3me antibody, conserved throughout eukaryotes; PRMT1 is "the major, if not exclusive" H4 Arg-3 methyltransferase in those cells 1. Independently purified as the H4-specific methyltransferase, with the p300 hand-off 2. UniProt: "the main enzyme" 3.established
PRMT6methylates histones H4/H2A at R3 in vitro; UniProt records H4R3me depositionmoderate - in vitro, and PRMT6's established in vivo target is H3R2 4
PRMT3deposits H4R3me2a at the miR-3648 promoter to promote transcription and osteogenesis in human mesenchymal stem cells 5moderate - one locus, one cell type; PRMT3 is mostly cytoplasmic (see gene-prmt3)

Note what the two UniProt annotations do together. PRMT1's record calls H4R3me1/me2a "a specific tag for epigenetic transcriptional activation" 3; PRMT6's record calls H3R2me2a "a specific tag for epigenetic transcriptional repression". Two type I enzymes, two asymmetric dimethylarginines two residues apart on two different histone tails, annotated with opposite signs in the same database. One of those two annotations has already been contradicted by genome-wide work (H3R2me2a, PMID 30232013; see mark-h3r2me2a). Nothing found in this beat contradicts the activation call for H4R3me2a, but the symmetry is a reason to hold it as an annotation rather than a law.

Distribution, and the one developmental map that exists

There is no genome-wide H4R3me2a map in a haematopoietic system in anything read for this beat. What exists is an immunohistochemical developmental series 12: in the proliferative neuroepithelium, during the stage of mainly symmetric expansive divisions, H4R3me2s predominates and H4R3me2a is almost undetectable; later, once post-mitotic neuronal layers begin forming, both states are detected in post-mitotic neurons and in oligodendrocyte precursors. The author's proposal is that H4R3me2s is part of the histone code of undifferentiated precursors and the appearance of H4R3me2a marks the onset of differentiation.

That is the same directional statement as the trans-tail result, arrived at independently and by a different method: me2a with differentiation, me2s with the undifferentiated state. For a disease defined by a differentiation block, the arrow points the interesting way - and it is a mouse cortex, by antibody immunohistochemistry, with no locus resolution and no quantification of specificity between the two isomers. The claim this graph can carry is the direction, in a non-haematopoietic tissue, at low resolution.

Bounded absence in this disease, with translations, warninglists and re-runs

Run 2026-09-22, translation read back, every zero re-run identical.

querynids / warninglist
"H4R3me2a"[tiab]63-
H4R3[tiab]109-
"H4R3me2a"[tiab] AND KMT2A[tiab]0outputmessages: ["No items found."]; re-run 0
"H4R3me2a"[tiab] AND MLL[tiab]0same; re-run 0
"H4R3me2a"[tiab] AND menin[tiab]0same; re-run 0
H4R3[tiab] AND KMT2A[tiab]0same; re-run 0
H4R3[tiab] AND MLL[tiab]317891136, 22689669, 26766589
"H4R3me2a"[tiab] AND leukemia[tiab]323249737, 37144835, 39668478
PRMT1[tiab] AND "H4R3me2a"[tiab] (control)41the writer-mark pair is well populated
operand control KMT2A[tiab]1652

Read the pattern, not the zeros. H4R3[tiab] AND KMT2A[tiab] is zero while H4R3[tiab] AND MLL[tiab] is three - which is this project's rule-71/77 shape: the modern symbol is the one that fails, because the arginine-methylation literature that touches this family is old enough to say MLL. And the three MLL records are not H4R3me2a chromatin measurements in a fusion: 17891136 is PRMT1 as an MLL-EEN cofactor via Sam68, 22689669 is WDR5/PRMT5 at the gamma-globin promoter, 26766589 is PRMT1/KDM4C in AML. "H4R3me2a"[tiab] AND MLL[tiab] is itself zero, so even the old symbol does not reach an H4R3me2a measurement in this family.

And the identifier trap runs the other way here from the residue spelling. ID-set comparison, retmax=500 so nothing truncated: H4R3[tiab] = 109 records; the union of "H4R3me2a", "H4R3me2s", "H4R3me1"[tiab] = 152; the intersection is 24 (Jaccard 10.1%). 128 records are reachable only through a compact spelling and 85 only through the residue spelling. The brief's premise - that the residue form returns fewer records than the compact form - is right in count for H4R3 (109 vs 152) and wrong as a containment claim in both directions: neither form is a superset, and each misses roughly five-sixths of the other's records.

Facts and figures
markH4R3me2a
histoneH4 (and H2A R3, the same N-terminal sequence)
residueR3 (Arg-3), N-terminal tail
chemistryasymmetric omega-N,N-dimethylarginine (aDMA) - both methyls on one terminal guanidino nitrogen. Isobaric with H4R3me2s, which has the OPPOSITE consequence for the KMT2D PHD(4-6) module.
eraser statusNO ESTABLISHED ERASER. JMJD6 was claimed for H3R2 and H4R3 (PMID 17947579) and reassigned as a lysyl-5-hydroxylase (PMID 19574390, review PMID 25997831). PADI4 citrullination CANNOT remove a dimethyl state (PMID 15339660) although it does target the PRMT1 H4 Arg-3 site at lower methylation degrees (PMID 15345777). Purified JmjC KDMs show in vitro arginine-demethylase capability with bona fide RDMs described as controversial (PMID 27337104). Removal is by histone turnover or nucleosome eviction.
nomenclature notePMID 23249737's enzyme is called 'MLL4' and IS KMT2D (O14686, 5537 aa). Two independent tells inside its own abstract settle it: the aliases given are MLL2 and ALR, both KMT2D-only, and the disease attributed to PHD(4-6) mutations is Kabuki syndrome, a KMT2D disease. The competing reading - KMT2B, whose MLL4 alias travels with WBP7 - is excluded. This graph already carries the MLL4 alias collision; recorded here so nobody re-derives it or misattributes the result.
crosstalkMethylation of H4 Arg-3 by PRMT1 FACILITATES subsequent acetylation of the H4 tail by p300, and conversely H4 acetylation inhibits PRMT1 methylation - a directional, mutually inhibitory pair (PMID 11387442). This is the basis of the mark's activating reputation, independent of any KMT2 enzyme.
antibody constraintLOAD-BEARING. H4R3me2a and H4R3me2s are isobaric and have OPPOSITE effects on the KMT2D PHD(4-6) module, so an assay that cannot separate them cannot test either direction. The one developmental distribution map for this mark (PMID 21072194) is an antibody comparison between the two isomers and its entire result would be reproduced by a reagent affinity difference. Base rates: >25% of 200+ histone-modification antibodies fail specificity by dot/western and >20% of the specific ones fail in ChIP (PMID 21131980); 52 antibodies against H3K4 methylforms often fail to distinguish states, changing the biological conclusion (PMID 30244833). The Histone Antibody Specificity Database (PMID 26212453) exists and was NOT queried in this beat. Isomer-resolving mass spectrometry is the arbiter.
relationship to menin and the win motifINDEPENDENT OF BOTH, and in the KMT2D case independent of WDR5 as well - PHD(4-6) is a module inside the catalytic subunit, not a WRAD component. So this is a third, separate control point on KMT2-family catalysis: menin tether, WIN/WDR5 contact, and the enzyme's own PHD reading of the H4 tail.
cell type and stateMeasured in: human 293T (PMID 11448779, 11387442); NT2/D1 embryonal carcinoma differentiation (PMID 23249737); human mesenchymal stem cells during osteogenesis (PMID 31378783); murine developing cortex, neural precursors, post-mitotic neurons and oligodendrocyte precursors (PMID 21072194); breast/oestrogen-responsive systems (PMID 21172665). The recurring cell STATE association is DIFFERENTIATION ONSET - the mark appears as precursors commit. NOT measured in fetal liver, LMPP, pro-B, HSC, or any KMT2A-rearranged line or patient sample.
shown for fusionNO - nobody has looked in any fusion. Not KMT2A::AFF1, not KMT2A::MLLT3, not mouse, not human. The KMT2-family result that exists (PMID 23249737) is wild-type KMT2D in a non-haematopoietic stem-cell differentiation model.
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, a writer-pair control and an ID-set comparison'
writers
[
  "PRMT1 - main enzyme; 'the major, if not exclusive, H4 Arg-3 methyltransferase' in human 293T cells (PMID 11448779, 11387442); UniProt Q99873 'the main enzyme'",
  "PRMT6 - methylates H4/H2A R3 in vitro (PMID 18079182); degree not stated; in vivo target is H3R2",
  "PRMT3 - H4R3me2a at the miR-3648 promoter in human MSCs (PMID 31378783); one locus, one cell type"
]
erasers
[]
readers
[
  "TDRD3 - Tudor domain; identified by protein-domain microarray as an effector of H4R3me2a and H3R17me2a, and is itself a transcriptional coactivator requiring an intact Tudor domain (PMID 21172665). NO ENTRY IN THIS GRAPH as of 2026-09-22.",
  "KMT2D PHD(4-6) - tandem plant homeodomain; recognises H4R3me0 OR H4R3me2a, required for KMT2D nucleosomal methyltransferase activity; Kabuki-syndrome mutations here reduce binding and catalysis (PMID 23249737). PERMISSIVE, not instructive: unmodified H4R3 binds equally."
]
effect on kmt2 family
{
  "kmt2d_direct": "PHD(4-6) of KMT2D must engage the H4 tail for KMT2D to methylate a nucleosome; H4R3me0 and H4R3me2a both permit it and H4R3me2s inhibits it (PMID 23249737). The genes whose H3K4me3 tracks this during differentiation include many HOXA and HOXB genes.",
  "kmt2d_enhancers_indirect": "loss of the analogous asymmetric mark on the other tail, H3R2me2a, reduces KMT2D binding at enhancers (PMID 30232013) - a second, independent observation that asymmetric arginine dimethylation is permissive for KMT2D",
  "kmt2a": "UNTESTED. KMT2A carries its own PHD1-3 plus bromodomain cassette, which this graph treats as a unit and which is LOST from every 5'-KMT2A fusion. Whether KMT2A's PHD1-3 reads H4R3 at all was not established in any source read for this beat.",
  "kmt2a_fusion": "NEVER MEASURED in any fusion, in any lineage"
}
literature bounds
{
  "date": "2026-09-22",
  "operand_control": {
    "query": "KMT2A[tiab]",
    "n": 1652
  },
  "writer_pair_control": {
    "query": "PRMT1[tiab] AND \"H4R3me2a\"[tiab]",
    "n": 41
  },
  "counts": [
    {
      "query": "\"H4R3me2a\"[tiab]",
      "translation": "\"H4R3me2a\"[Title/Abstract]",
      "n": 63
    },
    {
      "query": "H4R3[tiab]",
      "translation": "\"H4R3\"[Title/Abstract]",
      "n": 109
    },
    {
      "query": "\"H4R3me2a\"[tiab] AND KMT2A[tiab]",
      "translation": "\"H4R3me2a\"[Title/Abstract] AND \"KMT2A\"[Title/Abstract]",
      "n": 0,
      "warninglist": "phrasesignored [], quotedphrasesnotfound [], outputmessages [\"No items found.\"]",
      "rerun_n": 0
    },
    {
      "query": "\"H4R3me2a\"[tiab] AND MLL[tiab]",
      "n": 0,
      "warninglist": "outputmessages [\"No items found.\"]",
      "rerun_n": 0
    },
    {
      "query": "\"H4R3me2a\"[tiab] AND menin[tiab]",
      "n": 0,
      "warninglist": "outputmessages [\"No items found.\"]",
      "rerun_n": 0
    },
    {
      "query": "H4R3[tiab] AND KMT2A[tiab]",
      "n": 0,
      "warninglist": "outputmessages [\"No items found.\"]",
      "rerun_n": 0
    },
    {
      "query": "H4R3[tiab] AND MLL[tiab]",
      "n": 3,
      "ids": [
        "17891136",
        "22689669",
        "26766589"
      ]
    },
    {
      "query": "\"H4R3me2a\"[tiab] AND leukemia[tiab]",
      "n": 3,
      "ids": [
        "23249737",
        "37144835",
        "39668478"
      ]
    },
    {
      "query": "PRMT3[tiab] AND H4R3[tiab]",
      "n": 0,
      "warninglist": "outputmessages [\"No items found.\"]",
      "rerun_n": 0
    },
    {
      "query": "PRMT3[tiab] AND \"H4R3me2a\"[tiab]",
      "n": 6,
      "ids": [
        "31378783",
        "36637351",
        "39513743",
        "39892499",
        "40344412",
        "40374607"
      ]
    }
  ],
  "identifier_trap": "ID-SET comparison, retmax=500: H4R3[tiab] = 109; union of \"H4R3me2a\"/\"H4R3me2s\"/\"H4R3me1\"[tiab] = 152; INTERSECTION = 24 (Jaccard 10.1%). 128 records reachable only by a compact spelling, 85 only by the residue spelling. NEITHER FORM IS A SUPERSET OF THE OTHER. Also note PRMT3[tiab] AND H4R3[tiab] = 0 while PRMT3[tiab] AND \"H4R3me2a\"[tiab] = 6 - the dispatch brief for this beat recorded that pair as 1 and 6; measured with [tiab] on both operands it is 0 and 6, so the contrast is sharper than the brief stated, not weaker.",
  "bound_type": "titles and abstracts only",
  "symbol_scoping_warning": "H4R3[tiab] AND KMT2A[tiab] = 0 while H4R3[tiab] AND MLL[tiab] = 3. The arginine-methylation literature touching this family is old enough to say MLL. Any bounded absence for an arginine mark in this disease MUST run both symbols or it is a nomenclature artefact."
}
How this could bite us back
  • PRMT1 is the main writer of this mark and is DepMap common-essential (mean Chronos -1.399 across 1,258 lines, 98.7% of lines depleted, per gene-prmt1), and the first-in-human PRMT1 inhibitor GSK3368715 (NCT03666988) was TERMINATED after 31 participants because the overall benefit-risk profile did not support continuation, with lack of efficacy, extensive treatment-emergent effects and dose-limiting toxicities. Lowering H4R3me2a pharmacologically has been attempted and the essential-function toxicity is the documented outcome, not a hypothetical one.
  • The direction is ambiguous and could be the wrong one. If H4R3me2a is permissive for KMT2D catalysis and its symmetric counterpart is inhibitory, then a type I PRMT inhibitor lowers a permissive mark - but PRMT1 loss in mouse haematopoiesis already causes anaemia, leukopenia and reduced terminal erythroid and lymphoid differentiation (gene-prmt1). An agent that blocks differentiation in a disease defined by a differentiation block is the worst available direction.
  • The PRMT7 arm has the opposite sign from the PRMT5 arm on the SAME mark's symmetric counterpart - PRMT7 but not PRMT5 raises H4R3me2s and represses KMT2D targets (PMID 23249737). A drug developed against 'symmetric arginine methylation' as a class would hit the wrong enzyme for this axis, and a PRMT5 inhibitor would not be expected to move it at all. Class-level reasoning about PRMTs is unsafe here and the graph should not permit it.
  • Every distribution claim for this mark rests on antibodies that must discriminate isobaric isomers, and the one developmental map (PMID 21072194) is literally a comparison between two such antibodies. A pharmacodynamic endpoint built on that reagent class could report a clean change caused entirely by differential affinity, and a trial would have no way to tell.
Still open
  1. Does KMT2A's own PHD1-3 cassette read H4R3, as KMT2D's PHD(4-6) does? Nothing read for this beat tests it. If it does, then the fusion - which loses PHD1-3, the bromodomain, FYRN, the TAD and SET - has lost an H4-tail-reading requirement that the wild-type allele still has, which is a clean, measurable asymmetry between the two alleles in one patient-derived cell. If it does not, the H4R3 switch is a KMT2D story and should be filed as such.
  2. Is the H4R3me2a/me2s balance different at KMT2A-fusion target loci from elsewhere in the same genome? This is the single measurement that would open or close this whole axis, it needs no new reagent if done by isomer-resolving mass spectrometry on fractionated chromatin, and the answer is currently unknown for every fusion.
  3. Does PRMT7 loss raise KMT2D-catalysed H3K4me3 at HOXA in a haematopoietic cell as it does in NT2/D1? PMID 23249737 reports exactly that in a stem-cell differentiation model and names HOXA and HOXB. PRMT7 had no entry in this graph before this beat and no leukaemia dependency data had been joined to it.
  4. Is TDRD3 present and doing anything at KMT2A-fusion targets? It is the only established reader of this mark, it is a transcriptional coactivator localised to transcription start sites, and it has no entry in this graph - so the graph currently records a mark with a reader it cannot name.
Every source behind this entry (12)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Histone H4 isolated from asynchronously growing human 293T cells is methylated at arginine 3 in vivo, shown by mass spectrometry and independently by a novel antibody directed against H4 methylated at Arg-3, and the modification is highly conserved throughout eukaryotes; PRMT1 is the major, if not exclusive, H4 Arg-3 methyltransferase in human 293T cells.

    PMID 11448779DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. Human 293T embryonic kidney cells - not haematopoietic, not leukaemic. The abstract says 'methylated at arginine 3' without always specifying the degree; the asymmetric assignment for PRMT1 comes from the enzyme's type I chemistry and from UniProt (E3), not from a degree-resolved measurement in this abstract. 'Major, if not exclusive' is scoped to 293T and does not exclude other writers in other cell types - PRMT6 and PRMT3 both deposit H4R3 marks elsewhere (E4, E5).

    Evidence that argues against this (1)
    • The same residue carries a symmetric dimethyl state written by PRMT5 and PRMT7 with the opposite transcriptional consequence, so 'H4R3 is methylated by PRMT1' is degree-ambiguous in a way that matters: an experiment measuring total H4R3 methylation in a PRMT1 knockdown would see the asymmetric pool fall while the symmetric pool is untouched, and a non-discriminating reagent would report a partial change of unclear meaning. Zhao Q et al. 2009
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Purification and molecular identification of a histone H4-specific methyltransferase identified PRMT1, which specifically methylates arginine 3 of H4 in vitro and in vivo; methylation of Arg-3 by PRMT1 facilitates subsequent acetylation of H4 tails by p300, whereas acetylation of H4 inhibits its methylation by PRMT1, and a mutation in the S-adenosyl-methionine-binding site of PRMT1 substantially crippled its nuclear-receptor coactivator activity.

    PMID 11387442DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. The p300 hand-off is an in vitro histone-tail result; the coactivator arm is a nuclear-receptor reporter system. Neither is haematopoietic and neither involves a KMT2 enzyme. The ordering claim (methylation then acetylation) establishes a directional crosstalk, not a genome-wide rule, and the mutual inhibition means the two marks cannot be read as simply additive.

    Evidence that argues against this (1)
    • A later study of the same enzyme-mark pair reports that PRMT1-deposited H4R3 methylation is a prerequisite for a wider set of H4 modifications, which broadens the mechanism beyond p300 and means an H4R3me2a perturbation cannot be attributed to any single downstream acetylation event without measuring the others. Barrero MJ et al. 2006
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    UniProt Q99873 (ANM1_HUMAN, PRMT1, 371 aa) records that PRMT1 'constitutes the main enzyme that mediates monomethylation and asymmetric dimethylation of histone H4 Arg-3 (H4R3me1 and H4R3me2a, respectively), a specific tag for epigenetic transcriptional activation', and separately that it methylates H4R3 in genes involved in glioblastomagenesis in a CHTOP- and/or TET1-dependent manner (PubMed:25284789).

    What this evidence cannot show

    A curated annotation, not a measurement; accession fetched live rather than recalled. 'A specific tag for epigenetic transcriptional activation' is the curator's editorial summary and is the exact mirror of the sentence the same database applies to H3R2me2a for repression - and that one has been contradicted by genome-wide work. IDENTIFIER NOTE: PRMT1 is 371 aa and PRMT6 is 375 aa, four residues apart, which is inside this project's accession-discrimination limit, so a PRMT1/PRMT6 swap cannot be caught by length - the accessions are asserted explicitly here.

    Evidence that argues against this (1)
    • The database's parallel annotation for the other asymmetric arginine mark - H3R2me2a as 'a specific tag for epigenetic transcriptional repression' - is directly contradicted by genome-wide PRMT6-knockout data showing activating behaviour at enhancers. Nothing found in this beat contradicts the activation call for H4R3me2a, but a curator's one-line functional verdict on an arginine mark has already been wrong once in this exact family. Bouchard C et al. 2018
  4. E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    PRMT6 methylates histone H3 at R2 and histones H4/H2A at R3 in vitro, and overexpression and knockdown identify PRMT6 as the major in vivo H3 R2 methyltransferase.

    PRMT6-mediated methylation of R2 in histone H3 antagonizes H3 K4 trimethylation. · Hyllus D, Stein C, Schnabel K, Schiltz E, et al. · Genes Dev 2007
    PMID 18079182DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. The H4/H2A R3 activity is IN VITRO only in this paper; the in vivo claim established here is for H3R2, not H4R3. The degree (me1 versus me2a) is not stated for the H4/H2A activity in the abstract, and PRMT6's type I chemistry is the basis for reading it as asymmetric. So PRMT6 should be recorded as a possible but unquantified H4R3me2a writer, not a major one.

    Evidence that argues against this (1)
    • PRMT1 is independently established as the major, if not exclusive, H4 Arg-3 methyltransferase in human cells, so an in vitro PRMT6 activity on H4/H2A R3 does not make PRMT6 a physiological writer of this mark. Attributing an H4R3me2a change to PRMT6 without a knockdown-with-rescue would be unsupported. Strahl BD et al. 2001
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    PRMT3 mediates asymmetric arginine dimethylation of histone H4 Arg-3 (H4R3me2a) in the promoter region of microRNA miR-3648 to promote its transcription and osteogenesis, in human mesenchymal stem cells.

    Asymmetrical methyltransferase PRMT3 regulates human mesenchymal stem cell osteogenesis via miR-3648. · Min Z, Xiaomeng L, Zheng L, Yangge D, et al. · Cell Death Dis 2019
    PMID 31378783DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. A single locus in a single primary cell type; not haematopoietic and not leukaemic. PRMT3's principal characterised substrate is the small ribosomal subunit protein RPS2 and the enzyme is largely cytoplasmic (see gene-prmt3), so a chromatin activity is the minority of its biology and one promoter is not a genome-wide writer claim.

    Evidence that argues against this (1)
    • PRMT3 is not a DepMap dependency (mean gene effect +0.0419 across 1,258 lines, per gene-prmt3), and its nuclear pool is conditional on ZNF200. A chromatin-writing role that requires a permissive partner and confers no fitness cost anywhere is unlikely to be a major source of this mark in a leukaemic cell. Gupta S et al. 2024
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CARM1 and PRMT1 deposit H3R17me2a and H4R3me2a respectively, and a protein-domain microarray approach identified the Tudor-domain-containing protein TDRD3 as a reader of these marks; TDRD3 is itself a transcriptional coactivator, its coactivator activity requires an intact Tudor domain, it is recruited to an oestrogen-responsive element in a CARM1-dependent manner, and ChIP-seq shows it predominantly localised to transcriptional start sites.

    TDRD3 is an effector molecule for arginine-methylated histone marks. · Yang Y, Lu Y, Espejo A, Wu J, et al. · Mol Cell 2010
    PMID 21172665DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. The recruitment arm demonstrated in the abstract is CARM1-dependent, i.e. via H3R17me2a rather than H4R3me2a specifically, so the H4R3me2a half of the reader claim rests on the domain-microarray binding rather than on a locus experiment. Breast-cancer and oestrogen-receptor biology, not haematopoietic. TDRD3 has NO entry in this graph as of 2026-09-22 and this beat did not write one - it is named here as an unwritten reader.

    Evidence that argues against this (1)
    • TDRD3's better-characterised chromatin function in later work is R-loop resolution through DHX9 recruitment, which is a genome-stability role rather than a mark-reading coactivator role. A TDRD3 perturbation therefore cannot be interpreted as a clean readout of H4R3me2a signalling. Yuan W et al. 2021
  7. E7strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The tandem plant homeodomain PHD(4-6) of MLL4 - the paper's name for the enzyme it also calls MLL2 and ALR, i.e. KMT2D - recognises unmethylated or asymmetrically dimethylated histone H4 Arg 3 (H4R3me0 or H4R3me2a) and is required for MLL4's nucleosomal methyltransferase activity and for MLL4-mediated differentiation of NT2/D1 stem cells; Kabuki-syndrome mutations in PHD(4-6) reduce its binding ability and MLL4's catalytic activity; PHD(4-6) binding is inhibited by H4R3 symmetric dimethylation (H4R3me2s); PRMT7, but not PRMT5, represses MLL4 target genes by upregulating H4R3me2s and antagonises MLL4-mediated differentiation, and PRMT7 knockdown increases MLL4-catalysed H3K4me3; during differentiation, decreased H4R3me2s levels are 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, and NO residue numbers for PHD(4-6) are asserted here because none were fetched from a sequence. NOMENCLATURE: the paper's 'MLL4' is KMT2D, established by two independent tells in its own abstract - the aliases MLL2 and ALR, and the Kabuki-syndrome attribution - and NOT KMT2B, whose MLL4 alias travels with WBP7. The system is NT2/D1 embryonal carcinoma stem-cell differentiation, not haematopoiesis and not leukaemia, and the enzyme is KMT2D rather than KMT2A. Critically, PHD(4-6) binds H4R3me0 AS WELL AS H4R3me2a, so the mark is PERMISSIVE and not instructive - the module does not require the methyl mark, it only requires the absence of the symmetric one.

    Evidence that argues against this (1)
    • Because PHD(4-6) binds unmodified H4R3 equally, this result does not establish H4R3me2a as an activating instruction for KMT2D - only that H4R3me2s is inhibitory. The activating reading of H4R3me2a in this graph therefore rests on PRMT1/p300 crosstalk and the TDRD3 reader, not on this paper, and an experiment that raised H4R3me2a expecting a KMT2D gain could see nothing. Migliori V et al. 2012
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    In a murine neural-differentiation model, loss of PRMT6-dependent H3R2me2a from enhancers - where its peaks co-localise with H3K4me1 and H3K27ac - results in reduced KMT2D binding and reduced H3K4me1/H3K27ac deposition with decreased transcription of associated genes, while loss from promoters enhances KMT2A binding and H3K4me3.

    PMID 30232013DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. THIS IS H3R2me2a, NOT H4R3me2a - it is cited on this entry solely because it is the second independent observation that an ASYMMETRIC arginine dimethyl mark is permissive or supportive for KMT2D, converging with E7 from a different residue, a different histone tail and a different method. It is not evidence about H4R3. Murine neural differentiation; PRMT6 knockout removes the writer rather than the mark.

    Evidence that argues against this (1)
    • Convergence across two different residues is suggestive, not additive. The mechanisms are distinct - PHD(4-6) reads the H4 tail directly, while the enhancer effect on KMT2D in this paper has no stated reader - and a common explanation could be indirect, since a PRMT6 knockout changes many non-histone substrates. Treating the two as one 'asymmetric marks help KMT2D' rule would overreach. Dhar SS et al. 2012
  9. E9weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    JMJD6 was reported to demethylate histone H3 at arginine 2 and histone H4 at arginine 3 in biochemical and cell-based assays - the only eraser ever claimed for H4R3me2a - and the assignment has not held: the same enzyme was shown to catalyse lysyl-5-hydroxylation of the splicing factor U2AF65, and a dedicated review states the arginine-demethylase role is subject to conflicting reports.

    JMJD6 is a histone arginine demethylase. · Chang B, Chen Y, Zhao Y, Bruick RK · Science 2007
    PMID 17947579DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. Confidence WEAK on purpose despite the venue: cited to record what was claimed, not to assert an eraser exists. Note that this PMID appears in BOTH the H3R2[tiab] and H4R3[tiab] result sets, which is why the same paper anchors the eraser section of all four arginine-mark entries in this beat.

    Evidence that argues against this (1)
    • The hydroxylase reassignment plus the review of conflicting assignments means the graph's position must be that H4R3me2a has no established eraser; removal is by histone turnover or nucleosome eviction. Purified JmjC KDMs do show in vitro arginine-demethylase capability, with the authors themselves calling bona fide arginine demethylases controversial. Webby CJ et al. 2009
  10. E10established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    PAD4 converts methyl-arginine to citrulline, targeting multiple sites in histones H3 and H4 including H4 Arg-3 as methylated by PRMT1, but the companion study establishes that dimethylation of arginines prevents deimination by PADI4 while monomethylation still allows it - so citrullination can deplete H4R3me1 and unmodified H4R3 but cannot remove H4R3me2a.

    Human PAD4 regulates histone arginine methylation levels via demethylimination. · Wang Y, Wysocka J, Sayegh J, Lee YH, et al. · Science 2004
    PMID 15345777DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. This paper names PRMT1's H4 Arg-3 site as a PAD4 target but does not itself state the degree restriction; that comes from PMID 15339660, cited in the counter-evidence and in full on mark-h3r2me2a. Demethylimination is not demethylation in any case - it destroys the arginine and produces citrulline, a third species with its own consequences. HL-60 granulocytes and MCF-7, so there IS a myeloid system here, but the readout is bulk arginine methylation, not H4R3me2a specifically.

    Evidence that argues against this (1)
    • The degree restriction is decisive and comes from the companion paper: dimethylated arginines are not deiminated, monomethylated ones are. Citing PAD4 as an eraser of H4R3me2a would therefore be wrong, and an experiment seeing H4R3me2a fall after raising PAD4 is most likely seeing the monomethyl substrate pool removed upstream of PRMT1. Cuthbert GL et al. 2004
  11. E11strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    PubMed esearch, run 2026-09-22 with querytranslation and warninglist read back and every zero re-run: '"H4R3me2a"[tiab]' = 63; H4R3[tiab] = 109; '"H4R3me2a"[tiab] AND KMT2A[tiab]' = 0; AND MLL[tiab] = 0; AND menin[tiab] = 0; H4R3[tiab] AND KMT2A[tiab] = 0; H4R3[tiab] AND MLL[tiab] = 3 (17891136, 22689669, 26766589); '"H4R3me2a"[tiab] AND leukemia[tiab]' = 3 (23249737, 37144835, 39668478); writer-pair control PRMT1[tiab] AND '"H4R3me2a"[tiab]' = 41; operand control KMT2A[tiab] = 1652. ID-SET comparison at retmax=500: H4R3[tiab] = 109, union of the three compact H4R3 spellings = 152, intersection = 24 (Jaccard 10.1%) - 128 records reachable only by a compact spelling and 85 only by the residue spelling.

    What this evidence cannot show

    Bounds TITLES AND ABSTRACTS only; no Europe PMC full-text sweep was run. The three H4R3[tiab] AND MLL[tiab] records were title-checked and none is an H4R3me2a chromatin measurement in a fusion. The zeros carried non-empty translations with warninglist.outputmessages = ['No items found.'] and repeated on re-run. Note that 23249737 appears in the '"H4R3me2a" AND leukemia' set only because MLL4's name contains 'leukemia' - it is a stem-cell differentiation study, and counting it as leukaemia literature would be a symbol artefact.

    Evidence that argues against this (1)
    • That H4R3[tiab] AND KMT2A[tiab] is zero while H4R3[tiab] AND MLL[tiab] is three is a warning about the modern-symbol scoping this project has hit before: the arginine-methylation literature touching this family predates the KMT2A symbol. A bounded absence written only against KMT2A would be an artefact of nomenclature rather than a fact about the literature. Cheung N et al. 2007
  12. E12moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    Immunohistochemistry of the developing mouse cortex with lineage-specific and histone-modification-specific antibodies shows that the proliferative neuroepithelium, during the stage of mainly symmetric expansive divisions, is characterised by the prevalence of H4R3me2s with almost no detectable H4R3me2a, whereas at a later stage when cortical layers of post-mitotic neurons have begun forming both H4R3me2a and H4R3me2s are detected in post-mitotic neurons and in developing oligodendrocyte precursors - so the appearance of H4R3me2a tracks the onset of neurogenesis and gliogenesis and commitment to differentiate.

    PMID 21072194DOI
    What this evidence cannot show

    Abstract read from the eutils record 2026-09-22; full text not read. Single-author study, mouse cortex, immunohistochemistry only - no locus resolution, no quantification, and the whole result rests on the two antibodies discriminating isobaric H4R3 dimethyl isomers, which is exactly the constraint this beat records as load-bearing and which the abstract does not address with a competition panel. Not haematopoietic. The inference that the mark is 'part of the histone code of undifferentiated precursors' is the author's proposal, not a measurement.

    Evidence that argues against this (1)
    • Because the entire result is a comparison between two antibodies against isobaric isomers, a cross-reactivity or affinity difference between the two reagents would produce exactly this apparent developmental switch with no change in the underlying chromatin. The direction agrees with an independent mechanism paper, which is reassuring, but the two are not independent of the antibody problem - both use antibodies. Egelhofer TA et al. 2011
Connected to