Oncofusion

start / graph / concept-rbbp5-what-is-established-for-wild-type-kmt2a-and-what-for-the-fusion

Explainerstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q23-wrad-adjudication

RBBP5: the thinnest fusion literature of the four subunits, and one chromatin datum

also called RBBP5 versus the fusion · RBQ3 and KMT2A fusion · RBBP5 evidence split

RBBP5 clamps onto the enzyme end of KMT2A together with ASH2L, and that end is deleted from the usual leukemia fusion. Searching every name RBBP5 has ever had turns up just two papers pairing it with any KMT2A fusion, and neither tests whether the two proteins touch.

The same thing, for a scientist

**(a) Wild-type.** Established. RBBP5 and ASH2L form a heterodimer that engages a non-active-site surface of the MLL1 SET domain [E1]; RBBP5 directly binds H2B-conjugated ubiquitin on the nucleosome and orients MLL1 or MLL3 on it [E2]; and WDR5 plus RBBP5 plus ASH2L is the minimal module that confers sequential nucleosomal H3K4 methylation on MLL1 [E3]. Endogenous RBBP5 coprecipitates with transfected full-length MLL and is lost on SET-domain deletion [E4]. **(b) Fusion.** Effectively nothing. The alias-expanded PubMed query pairing RBBP5 with the named KMT2A fusions returns **2 records**, the smallest count of the four subunits, and neither is an association experiment [E6]. The retired symbol RBQ3 returns 1 record with MLL, KMT2A or leukaemia and it is not on topic. The only substantive datum is indirect and comes from a t(4;11) system: in SEM and RS4;11, RUNX1 knockdown reduces chromatin occupancy of AF4-N, MLL-C and 'the MLL-C complex component RBBP5' at shared target genes, in a paper whose model places RBBP5 in a complex with the reciprocal fusion AF4-MLL [E5]. That is chromatin co-occupancy in fusion-bearing cells, not a measured contact with a fusion protein. **(c) Domain.** A non-active-site surface of the SET domain, plus the H2B-ubiquitin contact on the nucleosome - LOST from every direct 5'-KMT2A fusion, RETAINED in reciprocal der products and in NUP98::KMT2A.

Go deeper
Why the count is so low, and why the low count is informative

WDR5 has an inhibitor industry and 611 PubMed records; RBBP5 has 165 and no chemical matter. The 2-record fusion count is therefore partly a measure of attention rather than of biology. But it is the right measure for this beat's question, because an association between a WRAD subunit and a fusion protein is the kind of result that would be reported in a title or abstract if anyone had found it. What the count establishes is that nobody has looked, in either direction, and the founding negative that the field cites for RBBP5 was run on a deletion mutant rather than a fusion.

The alias trap was checked in both directions. RBQ3 is the retired symbol; it returns 10 PubMed records in total and 1 when paired with MLL, KMT2A or leukaemia, and that record is not about this biology. There is no large hidden older literature under the old symbol, unlike the septin case that produced this project's tenfold-undercount rule.

The one datum in a KMT2A::AFF1 cell, and exactly what it can support

Two independent reports place RBBP5 on chromatin in t(4;11) lines. Wilkinson et al. show RUNX1-dependent RBBP5 occupancy at SPI1, MEF2D, JUND and JUNB in SEM and RS4;11 5. Wu et al. show that ASH2L knockdown in RS4;11 reduces RBBP5, WDR5 and MLL occupancy at the HOXC8 promoter 7.

Neither result can distinguish three sources of the chromatin-bound RBBP5 signal: the wild-type KMT2A allele, which is present and required in KMT2A-rearranged cells 8; SETD1A/SETD1B/KMT2B/KMT2C/KMT2D, which share the same WRAD module; and the reciprocal AF4-MLL fusion, which is present in SEM and RS4;11 and retains the full RBBP5 interface. Reading either result as 'RBBP5 works with the fusion' requires an assumption the experiments do not supply.

The functional argument that removes RBBP5 from the target list, and its limit

RBBP5's entire mechanistic role in this complex is to stimulate SET-domain catalysis. That catalysis is genetically dispensable for MLL-AF9 leukaemogenesis in mouse 9. So even a confirmed RBBP5-fusion contact would not by itself make RBBP5 a target in this disease. The limit of the argument is that it only covers the catalytic role: RBBP5 is DepMap common-essential (-1.105, 97.6% of lines) and a pan-essential gene may well have functions nobody has separated, as the BOD1L result showed for SETD1A, where a non-catalytic requirement was invisible to genome-wide screening and appeared only in a tiling screen 10.

Facts and figures
subunitRBBP5
uniprotQ15291
wild type contactnon-active-site surface of the MLL1 SET domain, via the RBBP5/ASH2L heterodimer; plus a direct contact with nucleosomal H2B-ubiquitin
contact retained in direct fusionfalse
contact retained in reciprocal fusiontrue
contact retained in nup98 kmt2atrue
fusion association evidencenone published; two indirect chromatin-occupancy observations in t(4;11) lines
pubmed records subunit x fusion2
pubmed queryRBBP5 AND (MLL-AF4 OR MLL-AF9 OR MLL-ENL OR MLL-AF10 OR MLL-AF6 OR MLL-ELL OR "MLL fusion" OR "MLL fusion protein" OR "KMT2A fusion" OR "MLL-rearranged" OR "KMT2A-rearranged")
verdictwild-type KMT2A SET-domain interactor; no measured association with any KMT2A fusion protein; present and functional on chromatin in KMT2A::AFF1 cells from an unattributed source
alias record counts
RBQ3 with MLL/KMT2A/leuke…1.000RBQ3 with MLL/KMT2A/leukemia: 1RBQ310.0RBQ3: 10RBBP5165.0RBBP5: 165
aliases searched
[
  "RBBP5",
  "RBQ3",
  "RBQ-3"
]
alias record counts — the numbers behind the chart
{
  "RBBP5": 165,
  "RBQ3": 10,
  "RBQ3 with MLL/KMT2A/leukemia": 1
}
How this could bite us back
  • RBBP5 is DepMap common-essential in 97.6% of lines with gnomAD LOEUF 0.146, so any RBBP5-directed agent starts from a very narrow tolerability assumption.
  • The catalytic role RBBP5 supports is genetically dispensable in this disease, so a successful RBBP5 agent would most likely be working through something nobody has characterised - which is the worst position from which to design a clinical trial.
  • RBBP5 is shared by all six SET1/MLL complexes; removing it removes H3K4 methylation capacity globally, and the developmental phenotypes of that loss are severe.
Still open
  1. Is the chromatin-bound RBBP5 in SEM and RS4;11 loaded on wild-type KMT2A, on SETD1A/B, or on the reciprocal AF4-MLL fusion? A degron on the wild-type allele plus RBBP5 CUT&RUN would separate them, and nothing published does.
  2. Does RBBP5 have any non-catalytic function in a KMT2A-rearranged cell? Its pan-essentiality hides the question and the BOD1L precedent shows genome-wide screening cannot answer it.
  3. Why has no group ever blotted RBBP5 on a KMT2A fusion immunoprecipitate in twenty-two years? The reagents are commercial and the experiment is a day's work, which suggests the community regards the answer as settled - on the basis of an inference from a deletion mutant.
Every source behind this entry (10)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    A non-active-site surface of the MLL1 SET domain is crucial for the interaction of MLL1 with the RBBP5/ASH2L heterodimer within MLL-family core complexes.

    PMID 24680668DOI
    What this evidence cannot show

    In vitro reconstitution with recombinant fragments; the mapped surface is a requirement for the interaction in that system, not necessarily the only contact in a cell.

    Evidence that argues against this (1)
    • A reconstituted MLL1 core complex assembles hierarchically and its subunit occupancy depends on temperature and component concentration, so a mapped contact surface describes one state of a conditional assembly rather than a fixed architecture. Namitz KEW et al. 2023
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    RBBP5 directly binds the H2B-conjugated ubiquitin on the nucleosome and orients the association between MLL1 or MLL3 and the nucleosome.

    PMID 31485071DOI
    What this evidence cannot show

    Cryo-EM of reconstituted complexes on designer nucleosomes; the in-cell relevance of the H2Bub contact in a leukaemia cell has not been tested.

    Evidence that argues against this (1)
    • Multistate cryo-EM of the MLL1-WRAD complex on H2B-ubiquitinated nucleosomes shows variable docking and ambiguity in the positions of ASH2L and DPY30, so a single-state contact map overstates the rigidity of the arrangement. Rahman S et al. 2022
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    WRAD forms a non-SET-domain multi-subunit methyltransferase required for sequential nucleosomal H3 methylation by the MLL1 core complex, with WDR5 plus RBBP5 plus ASH2L as the minimal active module.

    PMID 21106533DOI
    What this evidence cannot show

    In vitro biochemistry on reconstituted complexes and nucleosomal substrates; it establishes what the module does chemically, not what it does in a leukaemia cell.

    Evidence that argues against this (1)
    • The activity this module exists to support is dispensable in vivo for MLL-AF9 leukaemogenesis, so the biochemical necessity does not transfer to a therapeutic rationale in this disease. Mishra BP et al. 2014
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    Endogenous RBBP5 coprecipitates with transfected full-length MLL in 293 cells and is lost when the SET domain is deleted; RBBP5 was never blotted on any fusion protein in the paper that established this.

    PMID 15199122DOI
    What this evidence cannot show

    Transient overexpression in an epithelial line. The paper's fusion experiments blot menin, HCF-1, MLL-N, MLL-C and ENL only.

    Evidence that argues against this (1)
    • The same paper reports that an earlier 'MLL super complex' purification contained WDR5 and RBBP5 but lacked ASH2L and shared no other components with the complex described there, so even in 2004 the composition of the MLL complex depended on how it was purified. Yokoyama A et al. 2004
  5. E5moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper

    In the t(4;11) lines SEM and RS4;11, RUNX1 knockdown reduces chromatin binding of AF4-N, MLL-C and the MLL-C complex component RBBP5 at SPI1, MEF2D, JUND and JUNB, in a study whose model places these proteins in a complex with the reciprocal fusion AF4-MLL.

    PMID 23352661DOI
    What this evidence cannot show

    ChIP occupancy, not association with a fusion protein. The anti-MLL-C antibody cannot separate wild-type MLL-C from the MLL-C fragment released from AF4-MLL by taspase1 cleavage. The authors' AF4-MLL-specific siRNAs failed, so the causal role of the reciprocal fusion was not tested.

    Evidence that argues against this (1)
    • Chromatin-bound RBBP5 in these cells could equally come from the wild-type KMT2A allele, which is present and required in KMT2A-rearranged cells, and no experiment in this paper separates the two sources. Thiel AT et al. 2010
  6. E6moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    An alias-expanded PubMed query pairing RBBP5 with the named common KMT2A fusions returns 2 records, neither of which tests association between RBBP5 and a fusion protein; the retired symbol RBQ3 paired with MLL, KMT2A or leukaemia returns 1 record, off topic.

    What this evidence cannot show

    Counted absence, bounded by the symbol set RBBP5 and RBQ3/RBQ-3. A result reported at complex level ('the MLL1 core complex') rather than at subunit level would be invisible to this query, a trap this project has previously been caught by. Supplementary-only data are also invisible.

    Evidence that argues against this (1)
    • Complex-level framing does hide subunit results in this literature: a study of MLL1-complex targeting in CEBPA-mutated AML is framed entirely at complex level and would not be retrieved by any RBBP5 query. Schmidt L et al. 2019
  7. E7weak confidence 2 of 5: Preliminary, small, indirect, or not yet peer reviewed.paper

    ASH2L siRNA knockdown in the t(4;11) KMT2A::AFF1 line RS4;11 and in THP-1 reduces occupancy of RBBP5, WDR5, MLL and BPTF at regions of the HOXC8 promoter by ChIP, and lowers HOXC8 transcript and protein.

    PMID 32021283DOI
    What this evidence cannot show

    siRNA, a single non-HOXA target gene, no growth or survival readout, no comparison of fusion-bearing with fusion-free cells, and a low-impact venue. It establishes that the WRAD module is present and functional on chromatin in a KMT2A::AFF1 cell; it establishes nothing about the fusion protein.

    Evidence that argues against this (1)
    • The MLL antibody used in such ChIP experiments detects the wild-type allele, which is itself required for KMT2A-fusion leukaemia maintenance, so the occupancy being measured is most simply read as wild-type KMT2A complex occupancy. Thiel AT et al. 2010
  8. E8strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Menin recruits both wild-type MLL and the MLL-AF9 fusion to HOX loci, and the wild-type Mll allele is required for MLL-AF9-induced leukemogenesis and for maintenance of MLL-AF9-transformed cells.

    MLL-AF9-induced leukemogenesis requires coexpression of the wild-type Mll allele. · Thiel AT, Blessington P, Zou T, et al. · Cancer Cell 2010
    PMID 20159607DOI
    What this evidence cannot show

    Mouse genetics with MLL-AF9; whether the wild-type-allele requirement generalises to KMT2A::AFF1 and to human infant disease is assumed rather than shown.

    Evidence that argues against this (1)
    • A later genome-wide study reports that wild-type MLL1 and MLL-AF9 bind distinct chromatin regions with divergent transcriptional consequences, which complicates a model in which the wild-type allele licenses the fusion at the same loci. Xu J et al. 2016
  9. E9strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    The histone methyltransferase activity of MLL1 - the activity RBBP5 exists to stimulate - is dispensable for haematopoiesis and for MLL-AF9-driven leukemogenesis in mouse; the functionally critical output at MLL1 target genes is recruited MOF/H4K16 acetylation opposed by SIRT1.

    PMID 24813891DOI
    What this evidence cannot show

    A SET-domain deletion in mouse, in the MLL-AF9 model only; no t(4;11) or lymphoid model, and a mouse SET deletion may be compensated by paralogues in ways a human cell would not be.

    Evidence that argues against this (1)
    • Pharmacological blockade of the same module with MM-401 does kill MLL-rearranged cells without general toxicity to normal marrow, so 'dispensable' has been established genetically and contradicted pharmacologically and the discrepancy is unresolved. Cao F et al. 2014
  10. E10strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    BOD1L is classified as non-essential in all leukaemia cell lines in DepMap, yet a CRISPR-tiling screen in doxycycline-inducible Cas9 MOLM-13 and MV4-11 human leukaemia lines identifies a strong dependency on the N-terminal Shg1 domain and not on the C-terminal disordered regions or AT-hook; BOD1L loss reduces SETD1A occupancy at transcription start sites and causes an elongation defect without reducing H3K4me3 - a non-catalytic SET1-family requirement that genome-wide screening cannot see.

    BOD1L mediates chromatin binding and non-canonical function of H3K4 methyltransferase SETD1A. · Hoshii T, Kikuchi S, Kujirai T, et al. · Nucleic Acids Res 2024
    PMID 38989615DOI
    What this evidence cannot show

    Full text read via PMC11381347. The example is SETD1A-adjacent biology, not RBBP5, and it argues for caution rather than for a specific RBBP5 function. The screen lines are MOLM-13 (KMT2A::MLLT3) and MV4-11 (KMT2A::AFF1), both myeloid; no lymphoid t(4;11) line was screened. RBBP5 is pan-essential, which is the opposite problem from BOD1L's and limits how far the precedent transfers.

    Evidence that argues against this (1)
    • RBBP5's near-universal essentiality (97.6% of lines) means a tiling screen would have almost no dynamic range in which a KMT2A-selective sub-domain requirement could be resolved, so the BOD1L strategy may not transfer to this subunit. Broad Institute DepMap; Open Targets Platform 2026
Connected to