Separate positioning from catalysis: an ASH2L mutant that keeps RBBP5 and loses DPY30
also called ASH2L SDI separation-of-function · WRAD division-of-labour test
Build a version of ASH2L that still assembles the H3K4 enzyme but cannot hold DPY30, put it into a KMT2A::AFF1 leukemia line, and see whether the cells behave as if DPY30 were gone. If they do, DPY30's job really is aiming the enzyme.
The same thing, for a scientist
Structure-guided separation-of-function test of the WRAD division-of-labour hypothesis. Replace endogenous ASH2L in SEM (KMT2A::AFF1) with either wild-type ASH2L or an SDI-interface mutant that abolishes DPY30 binding while preserving the RBBP5 heterodimer and the MLL1 SET-domain contact, then compare proliferation, differentiation and genome-wide H3K4me2/3 placement against DPY30 knockout and full ASH2L knockout arms. The interface coordinates needed to design the mutant are available at 2.23 A (4RIQ) and 2.24 A (6E2H).
Design
Cells. SEM (KMT2A::AFF1, infant-derived B-ALL) as the primary line, with SEMK2 as a second KMT2A::AFF1 line and NALM-6 (non-KMT2A B-ALL) as a specificity control. MOLM-13 (KMT2A::MLLT3) extends to myeloid if the primary result is positive.
Arms. Endogenous ASH2L knockout with doxycycline-inducible re-expression of (a) wild-type ASH2L, (b) ASH2L SDI-interface mutant, (c) empty vector; plus separate DPY30 knockout and non-targeting arms. Five arms, each in two KMT2A::AFF1 lines and one control line.
Mutant design. The DPY30 homodimer clamps the ASH2L SDI motif at 2:1 stoichiometry, resolved at 2.23 A in 4RIQ and with full-length DPY30 at 2.24 A in 6E2H. Mutate the SDI residues that contact the DPY30 Dpy-30 motif (PF05186, DPY30 residues 52-92) while leaving the SPRY domain and the RBBP5-binding surface intact. Confirm separation biochemically before any cell work: co-immunoprecipitation must show loss of DPY30 and retention of RBBP5, and the reconstituted complex must retain measurable H3K4 methyltransferase activity on nucleosomes.
Readouts.
- Proliferation and competitive growth over 21 days.
- Differentiation markers by flow cytometry (CD19, CD33, CD11b) — the fusion's differentiation block is the phenotype of interest.
- CUT&RUN for H3K4me2 and H3K4me3, and the key analysis: is the loss preferential at KMT2A-fusion-bound loci (defined by menin or KMT2A CUT&RUN in the same cells) or uniform across the genome?
- RNA-seq, scored against the HOXA9/MEIS1/MEF2C output signature.
Controls. The specificity claim rests on the non-KMT2A line showing no differential between wild-type and mutant ASH2L. Include a catalytically dead MLL1 arm as a positive control for global H3K4 loss.
What each outcome means
Mutant phenocopies DPY30 knockout, and H3K4 loss is preferential at fusion-bound loci. The hypothesis stands. DPY30 is the aiming mechanism, and the graph gains a general rule for reading dependency data in mis-targeting oncoproteins.
Mutant phenocopies full ASH2L knockout. The subunits are not functionally separable in cells, the in vitro division of labour does not transfer, and the DPY30 co-dependency needs a different explanation.
Mutant has no phenotype at all. The DPY30-ASH2L interface is dispensable in this context, which would be surprising given the genome-wide placement data and would point at a DPY30 function outside WRAD.
H3K4 loss is uniform rather than preferential. The positioning claim fails even if the growth phenotype reproduces, which would be the most informative negative because it separates the two halves of the hypothesis.
Why this is worth doing before anything else in this beat
DPY30 is currently the only entry in this domain that is simultaneously a corrected-significant co-dependency, non-pan-essential, and mechanistically unexplained. Every downstream decision — whether to pursue a DPY30 degrader, whether to trust the ten-survivor list, whether the 'targeting versus catalysis' rule generalises to other fusion families — depends on whether the DPY30/ASH2L asymmetry is real biology.
The reagents are ordinary, the structures needed to design the mutant already exist, and the whole thing is a single graduate-student project rather than a programme.
| design | Endogenous ASH2L knockout in SEM and SEMK2 (KMT2A::AFF1) with doxycycline-inducible re-expression of wild-type ASH2L, an SDI-interface mutant that abolishes DPY30 binding while retaining RBBP5 binding, or empty vector; parallel DPY30-knockout and non-targeting arms; NALM-6 as a non-KMT2A specificity control and MOLM-13 as a myeloid extension. |
|---|---|
| readout | 21-day competitive proliferation; differentiation markers (CD19, CD33, CD11b) by flow cytometry; CUT&RUN for H3K4me2/me3 scored for preferential versus uniform loss against KMT2A/menin-bound loci in the same cells; RNA-seq scored against the HOXA9/MEIS1/MEF2C output signature. |
| controls | Biochemical confirmation of separation before cell work (co-IP showing DPY30 loss and RBBP5 retention; retained nucleosomal H3K4 methyltransferase activity). Non-KMT2A line for specificity. Catalytically dead MLL1 arm as a positive control for global H3K4 loss. Non-targeting guides throughout. |
| cost tier | med |
| timeline | 9-12 months including mutant design and biochemical validation |
| decision rule | Declared before running: the hypothesis is supported only if the SDI mutant phenocopies DPY30 knockout on proliferation AND H3K4 loss is preferential at fusion-bound loci. A growth phenotype without preferential H3K4 loss supports a DPY30 requirement but refutes the positioning mechanism. |
| tests hypothesis | hyp-dpy30-positions-ash2l-catalyses |
- A negative result kills the most promising new entry in this beat, and a positive result still leads to a 99-residue protein with no pocket and a pancytopenic mouse knockout.
- Separation-of-function mutants routinely fail to separate cleanly, and a mutant that partially destabilises WRAD would produce an uninterpretable intermediate phenotype.
- Cell lines are not infants, and SEM is a decades-old line whose chromatin state may not represent a primary infant blast.
- Can an ASH2L SDI mutant be designed that abolishes DPY30 binding without destabilising the SPRY fold? If not, the experiment needs a DPY30-side mutant instead, which is harder because DPY30's only structured element is the interface itself.
- What is the right definition of a fusion-bound locus for the preferential-loss analysis? Using menin CUT&RUN and KMT2A CUT&RUN will give overlapping but non-identical sets, and the conclusion may depend on which is chosen. This should be pre-registered.
- Should the experiment include a menin-inhibitor arm to test the predicted epistasis directly? It would roughly double the scale but would answer the combination question in the same study.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
The structural information needed to design an ASH2L mutant that loses DPY30 while retaining the rest of WRAD exists: the DPY-30 dimerisation/docking domain bound to the Ash2L SDI motif is solved at 2.23 A (4RIQ) and the Ash2L SPRY plus SDI region with full-length DPY-30 at 2.24 A (6E2H).
RCSB PDB entries 4RIQ and 6E2H, verified against the RCSB data API for title, experimental method and resolution · RCSB Protein Data Bank · RCSB PDB 2026What this evidence cannot show
These are isolated-domain crystal structures, not the assembled MLL1 core complex on a nucleosome, so an interface mutation designed from them may have unintended effects on complex assembly. The biochemical separation check before cell work is not optional.
Evidence that argues against this (1)
- A design caution that bears on whether the experiment can return an interpretable answer. The premise that WRAD composition is a fixed four-subunit module is not correct across the SET1 family: in the absence of WDR5, MLL3 still interacts with RAD and shows enhanced activity, and MLL2/MLL4 and SETD1B complexes tolerate Win-motif disruption with only modest loss of assembly and little loss of in vitro activity. An ASH2L separation-of-function mutant will therefore be read out against a background in which other SET1-family complexes can redistribute, and the proposed DPY30-phenocopy readout may be confounded by paralogue compensation rather than reporting on positioning. The experiment should include a readout that discriminates which SET1-family complex is responsible for any residual H3K4 methylation. Alicea-Velazquez NL et al. 2016
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The functional premise of the experiment is that DPY30 governs genome-wide placement of H3K4me2/3 through its ASH2L interface while contributing only modestly to catalytic stimulation.
Structural Analysis of the Ash2L/Dpy-30 Complex Reveals a Heterogeneity in H3K4 Methylation. · Haddad JF, Yang Y, Takahashi YH, Joshi M, Chaudhary N, Woodfin AR, Benyoucef A, Yeung S, Brunzelle JS, Skiniotis G, Brand M, Shilatifard A, Couture JF · Structure 2018What this evidence cannot show
Not performed in leukemia models. The present experiment is precisely the test of whether that finding transfers to a KMT2A-fusion context.
Evidence that argues against this (1)
- This is the decisive objection to the whole WRAD thread and it should be read before any of it. The SET domain of MLL1 -- the catalytic activity that WRAD exists to stimulate -- is dispensable for maintaining haematopoietic stem cells AND for supporting leukemogenesis driven by the MLL-AF9 fusion oncoprotein in mouse. On Mll1 deletion, what is lost at MLL1 target genes is H4K16 acetylation, and inhibiting SIRT1 is sufficient to restore both the acetylation and target-gene expression. The conclusion the authors draw is that recruited MOF activity, not MLL1's intrinsic methyltransferase activity, is what maintains these genes. If H3K4 methylation by MLL1 is not required for MLL-fusion leukemia, then the subunit whose job is to enable that methylation is not a target in this disease, whatever its biochemistry. Mishra BP et al. 2014
- E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
DPY30 is required for the growth of several MLL1-fusion-mediated leukemia cell lines, which is the prior that makes a DPY30-phenocopy readout meaningful.
The DPY30 subunit in SET1/MLL complexes regulates the proliferation and differentiation of hematopoietic progenitor cells. · Yang Z, Augustin J, Chang C, Hu J, Shah K, Chang CW, Townes T, Jiang H · Blood 2014What this evidence cannot show
A short shRNA result reported at the end of a paper about normal progenitors, in a small number of lines, none of them a KMT2A::AFF1 infant model. The DPY30 knockout arm of this experiment should be treated as establishing the phenotype, not confirming it.
Evidence that argues against this (1)
- This project's own deeper re-analysis withdraws the DPY30 result that this evidence item rests on. Against variance-matched nulls drawn from 2,743 usable random protein-coding genes across the same 54 leukemia lines, the Holm-corrected survivors are KMT2A (r 0.783), PSIP1 (0.686), DOT1L (0.681), FLT3 (0.645), MEF2C (0.526) and KAT7 (0.517). DPY30 is explicitly among those that do NOT clear correction, alongside MEIS1 (Holm 0.067), MLLT1 and BCL2. The earlier 16-gene panel gave DPY30 Holm p = 0.015 using an 'exceeds the maximum of the null' criterion that the deep-pool run shows to be unstable by construction, because the null maximum rises mechanically as the pool grows. The DPY30-versus-other-WRAD-subunits contrast is therefore a contrast between one uncorrected near-signal and three nulls, not between a survivor and three failures. Broad Institute DepMap; Open Targets Platform; analysis run by this project 2026
structure-guided separation-of-function test with a declared decision rule
run in SEM and SEMK2, both KMT2A::AFF1