UBE2A (RAD6A) — the E2 for H2B ubiquitination in one primary account, contradicted by another, and with no KMT2A literature at all
also called UBE2A · RAD6A · HR6A · hRAD6 · Ubiquitin-conjugating enzyme E2 A · UBC2
Before histone H2B can be tagged, a small carrier enzyme has to hand the ubiquitin over; UBE2A is the carrier one major paper names, and a different major paper names a different enzyme instead. Nobody has ever tested either of them in KMT2A-rearranged leukaemia, and the cell-line dependency data say neither is needed for cells to survive.
The same thing, for a scientist
UBE2A (P49459, 152 aa; human RAD6A/HR6A) is the ubiquitin-conjugating E2 that Kim et al. identify as the cognate partner for human H2B K120 ubiquitination, recruited to transcribed genes through direct interaction with hPAF-bound hBRE1 (RNF20/RNF40) [E1]. This directly contradicts Zhu et al., who assign the same reaction to UbcH6/UBE2E1 [E2]. Both are primary papers in strong journals from groups that work on this pathway, and **this beat does not consider the question settled.** Three facts make the disagreement more than a curiosity. First, in the project's genome-wide DepMap scan none of the three candidate E2s — UBE2A, UBE2B, UBE2E1 — is required by cancer cell lines (mean gene effects -0.162, +0.049 and -0.007 respectively) and none shows KMT2A selectivity, while the E3 subunits are broadly essential [E3]; the simplest reading is E2 redundancy. Second, a controlled PubMed search finds **no paper at all** intersecting UBE2A/UBE2B/RAD6 with MLL or KMT2A: the intersection returns 7 records and none is about this biology, against 722 for the E2 terms alone and 1,686 for the KMT2A-rearrangement terms alone [E4]. Third, UBE2A has a well-characterised human loss-of-function phenotype — X-linked UBE2A deficiency (Nascimento-type intellectual disability) with seizures, dysmorphism, urogenital anomalies, and recurrent pineal cysts/tumours, congenital heart defects and hypogammaglobulinaemia [E5] — which is a better guide to the cost of inhibiting it than anything in the cancer literature.
This is a prediction, not a measurement. No one has seen this protein in this shape. The colours say how much the software trusts itself, and that is worth reading before anything else — the orange stretches are places the model is effectively saying this part has no fixed shape.
I have never read one of these pictures before
A protein is one long chain of amino acids that folds up. The ribbon traces that chain from one end to the other; it is not a wire, it is a simplification of where the backbone runs. Drag to rotate, scroll to zoom.
Domains are the stretches of chain that do a particular job, and they are what the coloured blocks above the picture mark out. Click one and the view flies to it. In this project the important question about any domain is usually whether the fusion still has it — the labels say lost where it does not.
Anything in yellow sticks is a bound small molecule, which is usually the drug.
The disagreement about which E2 this is
| Paper | E2 named | Key support |
|---|---|---|
| Zhu et al. 2005, Mol Cell 2 | UbcH6 (UBE2E1) | physically interacts with RNF20/40 and with hPAF; trimeric complex stimulates H2Bub in vitro; all three recruited to active genes |
| Kim et al. 2009, Cell 1 | hRAD6 (UBE2A/UBE2B) | serves as the cognate E2; recruited via direct interaction with hPAF-bound hBRE1; ubiquitylates chromatinised H2B at K120 |
The yeast precedent favours RAD6 — Bre1 works with Rad6 in S. cerevisiae, and every yeast paper in this beat's crosstalk chain uses Rad6 [see mechanism-h2bub-dot1l-trans-histone-crosstalk E1, E6]. Conservation is an argument, not a result.
A reconciliation that neither paper makes, but that the DepMap numbers are consistent with, is that more than one E2 can charge RNF20/RNF40, and which one dominates depends on context and on what is available. If that is right, the E2 step is the wrong place to intervene in this pathway, because there is no single node to hit. This is recorded as a hypothesis-level reading in hyp-h2bub-required-but-not-selective, not as a finding.
The negative search, reported with both halves
This project's standing rule is that a claim of absence must report the counts for each half of the intersection, because a zero from an over-constrained query is uninterpretable. Run against PubMed eutils on 2026-09-19 with [Title/Abstract] tags 4:
| Query | Count |
|---|---|
UBE2A OR UBE2B OR RAD6 OR RAD6A OR RAD6B | 722 |
MLL-rearranged OR KMT2A-rearranged OR MLL-AF9 OR MLL-AF4 | 1686 |
(UBE2A OR UBE2B OR RAD6 OR HR6A) AND (MLL OR KMT2A) | 7, none of which concerns the E2 step in KMT2A-rearranged leukaemia (the hits are a paediatric short-stature genotype study, a Cornelia de Lange review, a cognitive-genetics review, an embryogenesis paper, and three yeast/Dot1 crosstalk papers that mention MLL only in passing) |
RNF20 AND infant | 0 |
So the absence is real and it is not a search artefact. The E2 step of the H2Bub axis has never been studied in this disease.
| uniprot | P49459 |
|---|---|
| gene symbol | UBE2A |
| length AA | 152 |
| protein name | Ubiquitin-conjugating enzyme E2 A |
| alphafold ID | AF-P49459-F1 |
| accession resolution | UniProt query gene_exact:UBE2A AND organism_id:9606 AND reviewed:true; primary gene name read back as UBE2A and length 152 confirmed. The near-identical paralogue UBE2B is P63146, also 152 aa. |
| enzyme class | E2 ubiquitin-conjugating enzyme |
| pathway role | charges the RNF20/RNF40 E3 for H2B K120 monoubiquitination - contested; see E1 vs E2 |
| mark written | mark-h2bk120ub |
| competing assignment | UBE2E1 (UbcH6) |
| paralogue | UBE2B (RAD6B, P63146) |
| kmt2a literature | none - zero indexed primary papers intersect this gene family with MLL/KMT2A leukaemia biology |
| human lof phenotype | X-linked UBE2A deficiency / intellectual disability Nascimento type, MIM 300860 |
| druggability | E2 enzymes are a difficult target class and no selective UBE2A inhibitor is known to this beat. Redundancy among candidate E2s would defeat a single-agent E2 inhibitor regardless. |
essentiality
{
"depmap_common_essential": false,
"chronos_mean": -0.1617,
"n_lines": 1208,
"chronos_mean_kmt2a_rearranged": -0.1727,
"chronos_mean_other_leukemia": -0.1527,
"kmt2a_selectivity_difference": -0.0201,
"genome_wide_rank": "6504/17728",
"paralogue_ube2b_chronos_mean": 0.0492,
"alternative_ube2e1_chronos_mean": -0.0072,
"interpretation": "Not required, no selectivity, and the same is true of both alternative E2s - the pattern of redundancy."
}- If three different E2 enzymes can charge the same E3, inhibiting any one of them will not remove the mark - the single most likely failure mode for any intervention at this step, and it is visible before the first experiment.
- Germline UBE2A loss in humans causes severe neurodevelopmental disease. In infants, that is the wrong organ to put at risk.
- UBE2A's best-known role outside chromatin is in DNA damage tolerance and translesion synthesis; inhibiting it would perturb mutagenesis control in a population already receiving genotoxic chemotherapy.
- This entry rests on zero KMT2A-specific experiments. Any therapeutic reasoning built on it is reasoning from pathway position alone, which is exactly the error this beat's DepMap results caught for RNF20.
- Which E2 actually charges RNF20/RNF40 on chromatin in a KMT2A-rearranged leukaemia cell? A combinatorial degron of UBE2A, UBE2B and UBE2E1 with an H2Bub western is a week of work and would close a twenty-year-old disagreement in the context that matters here.
- Is the E2 step redundant, as the DepMap pattern suggests? If a double or triple knockout is needed to lose H2Bub, the E2 step is not a target and the field should stop listing it as one.
- Do individuals with UBE2A deficiency syndrome have measurable H2Bub reduction in accessible tissue? If not, the syndrome is telling us about a non-histone UBE2A function and is not informative about this axis at all.
- Why do UBE2A loss-of-function individuals recurrently develop pineal cysts and tumours? Unexplained, and a loss-of-function-to-tumour link in an H2Bub enzyme is the sort of observation that occasionally turns out to matter.
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
hRAD6 serves as the cognate E2-conjugating enzyme for human H2B ubiquitylation and, through direct interaction with hPAF-bound hBRE1, is recruited to transcribed genes and ubiquitylates chromatinised H2B at lysine 120.
RAD6-Mediated transcription-coupled H2B ubiquitylation directly stimulates H3K4 methylation in human cells. · Kim J, Guermah M, McGinty RK, Lee JS, Tang Z, Milne TA, Shilatifard A, Muir TW, Roeder RG · Cell 2009What this evidence cannot show
Abstract read in full; full text not read. 'hRAD6' in the abstract does not distinguish UBE2A from its near-identical paralogue UBE2B (both 152 aa), so this entry's assignment to UBE2A specifically is an inference from the gene name RAD6A and is not what the abstract states. Reconstituted systems plus cells; no leukaemia.
Evidence that argues against this (1)
- Zhu et al. assign the E2 role to UbcH6 (UBE2E1), not RAD6, in the same reaction in human cells, on the basis of physical interaction with RNF20/40 and hPAF and in vitro stimulation of H2Bub activity. Zhu B et al. 2005
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
UbcH6 is the ubiquitin E2-conjugating enzyme for human H2B-Lys120 monoubiquitination, physically interacts with RNF20/40 and with the hPAF complex, and formation of the trimeric complex with hPAF stimulates H2B monoubiquitination activity in vitro.
Monoubiquitination of human histone H2B: the factors involved and their roles in HOX gene regulation. · Zhu B, Zheng Y, Pham AD, Mandal SS, Erdjument-Bromage H, Tempst P, Reinberg D · Mol Cell 2005What this evidence cannot show
Abstract read in full; full text not read. In vitro reconstitution plus interaction data; an E2 that can perform a reaction in vitro is not necessarily the one that performs it on chromatin in a given cell type. UbcH6 is UBE2E1 (this beat resolved the alias but did not verify a UniProt accession for it).
Evidence that argues against this (1)
- UbcH6 being the E2 in a reconstituted system with a trimeric hPAF complex establishes sufficiency in vitro. Cells express several E2 enzymes able to charge this pathway, so in vitro identification of one does not establish that it is the operative E2 in a leukaemic cell. Molecular Cell 2005
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.dataset
In the project's genome-wide DepMap KMT2A-selectivity scan, none of the three candidate E2 enzymes is required by cancer lines or shows KMT2A selectivity: UBE2A mean gene effect -0.162 across 1,208 lines with a KMT2A difference of -0.020 (rank 6,504/17,728); UBE2B +0.049 with a difference of +0.007 (rank 10,016); UBE2E1 -0.007 with a difference of +0.108 (rank 17,038) - while the E3 subunits RNF20 and RNF40 are broadly essential (-0.734 and -0.490).
DepMap CRISPR (Chronos) genome-wide KMT2A-selectivity scan of 17,728 protein-coding genes, 11 Cellosaurus-annotated KMT2A-rearranged leukaemia lines vs 43 other leukaemia lines · Broad Institute DepMap; Cellosaurus; Open Targets Platform; analysis run by this project (analyses/depmap-kmt2a-selective-scan-full) · Database and derived analysis 2026What this evidence cannot show
A non-essential single-gene knockout is exactly what redundancy looks like, but it is also what a poorly targeted guide set looks like; DepMap cannot distinguish them. Eleven KMT2A-rearranged lines, AML-weighted, none infant-derived. The redundancy interpretation offered in this entry is an inference from the pattern, not a tested result.
Evidence that argues against this (1)
- SCAN CONFOUND: this number comes from the project's genome-wide KMT2A-selectivity scan, whose index arm is MYELOID-ENRICHED (8 AML lines against 5 B-ALL). A gene whose requirement differs by lineage is mis-scored in BOTH directions by that design — CEBPA ranks 25 genome-wide with no dependency at all in KMT2A-rearranged B-ALL, and conversely a gene required in one lineage and not the other can average to apparent non-selectivity. See `ana-lineage-confound-myeloid-tf-dependencies`. Here the confound cuts toward a false negative: three candidate E2 enzymes each showing no requirement is exactly what redundancy produces, and a scan over single-gene knockouts cannot detect a dependency shared among paralogues whatever its lineage balance. Oncofusion knowledge graph (internal analysis) 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
A controlled PubMed eutils search on 2026-09-19 finds no literature intersecting the H2Bub E2 enzymes with KMT2A-rearranged leukaemia: the E2 terms alone return 722 records, the KMT2A-rearrangement terms alone return 1,686, and the intersection with MLL or KMT2A returns 7 records none of which concerns this biology; RNF20 intersected with 'infant' returns 0.
NCBI Entrez Programming Utilities (esearch/efetch), PubMed, searches run 2026-09-19 with [Title/Abstract] field tags; both halves of every intersection counted · US National Library of Medicine; searches run by this project · Database 2026What this evidence cannot show
Title/abstract searching misses work described only in a full text, and misses preprints not indexed in PubMed. A gene symbol can be absent from a title and abstract while being central to a paper. The claim is therefore 'no indexed primary literature that names both', which is weaker than 'nobody has looked'.
Evidence that argues against this (1)
- An absence across three E2 enzymes intersected with KMT2A leukaemia is a conjunction of three counted absences, so it inherits three chances to have missed an alias. UBE2A in particular publishes as RAD6A and HR6A, and RAD6 is how the older H2Bub literature names it — which is precisely where a functional link would sit. Cell 2009
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
UBE2A deficiency (intellectual disability, Nascimento type; MIM 300860) is an X-linked syndrome of developmental delay, moderate to severe intellectual disability, seizures, dysmorphism, skin anomalies and urogenital malformations; with forty affected subjects reported by 2020 and pineal gland cysts/tumours, congenital heart defects and hypogammaglobulinaemia emerging as recurrent features, several missense variants destabilising the E3-binding domain.
Refinement of the clinical and mutational spectrum of UBE2A deficiency syndrome. · Cordeddu V, Macke EL, Radio FC, Lo Cicero S, Pantaleoni F, Tatti M, Bellacchio E, Ciolfi A, Agolini E, Bruselles A, Brunetti-Pierri N, Suri M, Josephs KS, McEntagart M, Lanpher B, et al. · Clin Genet 2020What this evidence cannot show
Abstract read in full; full text not read. Germline hemizygous loss from conception is not a model for pharmacological inhibition. The syndrome phenotype is dominated by neurodevelopment; no leukaemia predisposition is reported, but a cohort of roughly 48 individuals cannot exclude a modest one. The recurrence of pineal tumours in a loss-of-function syndrome is an interesting and unexplained signal.
Evidence that argues against this (1)
- UBE2A deficiency is an X-linked germline neurodevelopmental syndrome. It establishes that the gene matters in human development and says nothing about a leukaemic requirement; importing a constitutional syndrome into a dependency argument crosses both tissue and mechanism. Clinical Genetics 2020
E2 conjugating enzyme for H2B K120 ubiquitination in one primary account; contested
hRAD6 recruited through direct interaction with hPAF-bound hBRE1 (RNF20/RNF40)