Degrade one CXXC reader, ChIP the fusion: the one experiment that settles whether the address is contested
also called CXXC competition experiment · reader-degron fusion-occupancy experiment
Take a leukaemia cell whose KMT2A fusion is doing the damage, rapidly remove one of the other proteins that reads the same DNA addresses, and see whether the fusion spreads into the vacated spots. Nobody has done this, and it would answer in one go a question the whole field has only guessed at.
The same thing, for a scientist
**Design.** In a KMT2A-rearranged human line with an endogenously tagged fusion (SEM or RS4;11 for KMT2A::AFF1, MV4-11 or MOLM-13 for KMT2A::MLLT3), introduce a degron on one CXXC reader at a time - **CXXC1, KDM2A and KDM2B are the three to run first**, being the family members with the closest reader chemistry [E1], an existing entry in this graph, and measurable expression in leukaemia lines. Deplete acutely (dTAG or auxin, 2 h and 6 h, so that transcriptional secondary effects are minimised), then measure, in the same cells: **(a)** fusion occupancy by CUT&RUN or ChIP-seq against the endogenous tag, **(b)** occupancy of the degraded reader as a positive control for the depletion, **(c)** H3K4me3 and H3K79me2 at fusion targets, **(d)** nascent transcription at HOXA9/MEIS1 by PRO-seq or 4sU-seq. **Mapping must use an rDNA-containing assembly**, because the standard human reference omits rDNA repeats and a second CXXC convergence sits there [E2][E3]. **Controls.** A degron on a non-CXXC chromatin protein of comparable abundance, to show that removing any abundant chromatin factor does not move the fusion; and a menin-inhibitor arm, because this project already knows the fusion is evicted by that and it calibrates what a real occupancy change looks like. **Why it has never been done.** The field has run the reciprocal - manipulating a reader and measuring the fusion's *neighbours* - once, and found that KDM2A overexpression moved LEDGF and wild-type MLL and left the fusion in place [E4]. The direct version requires a degron line per reader, which is three lines' construction, and the question has not been asked in print. **Cost tier: medium.** Three degron lines, four CUT&RUN conditions each, one nascent-transcription readout. **Readout that decides it:** fusion occupancy at shared CpG islands, acutely, with the depletion verified in the same sample.
Which three readers, and why those
| reader | why first | what this graph already holds |
|---|---|---|
| CXXC1/CFP1 | the SET1 complex's CpG-reading targeting subunit, so a depletion tests SET1-versus-KMT2A at CpG islands directly - a question the field has posed and never answered ((CXXC1 OR CFP1 OR SETD1A...) AND (MLL1 OR KMT2A) AND CpG returns 2 records, both reviews); and it is rank 414 of 17,728 on this project's DepMap scan | gene-cxxc1, including the negative CXXC-swap result |
| KDM2A | the closest reader chemistry to KMT2A's, and the only family member with a published antagonism result in MLL leukaemia; rank 442 of 17,728 | gene-kdm2a, written this cycle |
| KDM2B | the PRC1.1 DNA-targeting subunit, so a depletion tests the activating-versus-repressive CpG hand-off, and it carries the nucleolar rDNA address | gene-kdm2b, protein-prc1-1-complex |
Deliberately not first: TET1 and TET3 (their effect would run through DNA chemistry over hours to days, which an acute degron is the wrong instrument for), DNMT1 (broadly required, so acute depletion has confounding proliferative effects), CXXC4 and FBXL19 (may not be predominantly nuclear), MBD1 (its CXXC did not bind DNA in the one assay that tested it), KMT2B (a paralogue of the index protein, so a positive result would be ambiguous).
What each outcome would mean
- Fusion occupancy rises at islands the degraded reader occupied. The address is contested, at named loci, in the index disease. This would be the first such demonstration in the family and would make the competitor set a legitimate part of the fusion's mechanism.
hyp-the-cxxc-address-is-a-contested-resourcemoves from unsupported to supported. - Fusion occupancy does not move. The null holds, and the fusion's address is set by its own multivalent anchoring - AT-hooks, CXXC and the menin-LEDGF tether - rather than by who else is there. This is the outcome current evidence predicts, and it would be a positive result for the menin-interface hypothesis, because it locates the fusion's stability in protein-protein contacts.
- Occupancy unchanged but transcription changes. The readers interact through the histone marks rather than through the DNA, which is a more tractable mechanism and points at the mark-writing and mark-erasing enzymes rather than at the CXXC domains.
- Occupancy changes only at rDNA. The convergence bounded in
absence-kdm2b-and-the-kmt2a-fusion-are-both-nucleolar-and-nobody-has-put-them-togetheris real, and the competition is compartment-specific rather than genome-wide.
The cheap computational precursor, which should be run first
Before building three degron lines, two analyses cost nothing but compute:
- Re-map existing KMT2A-fusion ChIP-seq against an rDNA-containing assembly. Every published dataset in this field was processed against a reference that omits the rDNA repeats, so any fusion signal there has been discarded. Custom rDNA assemblies were what made the KMT2A-and-SETD1A-on-rDNA result visible at all 2.
- Intersect published CXXC-reader occupancy with published KMT2A-fusion occupancy across cell types, with a matched-promoter null. This cannot establish competition - overlap of a promoter class is a base rate - but it would say whether the islands even coincide, and if they do not, the degron experiment is unnecessary.
Neither exists. Both are within this project's demonstrated capability.
| design | Degron-mediated acute depletion of one CXXC-family reader at a time (CXXC1, KDM2A, KDM2B) in a KMT2A-rearranged human line carrying an endogenous tag on the fusion; 2 h and 6 h depletion; CUT&RUN or ChIP-seq for the fusion and for the degraded reader in the same samples. |
|---|---|
| readout | Fusion occupancy at shared CpG islands (primary); reader occupancy as depletion control; H3K4me3 and H3K79me2 at fusion targets; nascent transcription at HOXA9 and MEIS1. |
| controls | A degron on a non-CXXC chromatin protein of comparable abundance, to show that removing any abundant chromatin factor does not move the fusion; a menin-inhibitor arm to calibrate what a real occupancy change looks like in the same cells; spike-in normalisation, because a genome-wide occupancy change is invisible to a ratio-normalised experiment. |
| mapping requirement | Use an rDNA-containing genome assembly. Standard human references omit rDNA repeats and discard reads from a locus where a second CXXC convergence sits. |
| cost tier | med |
| timeline | 9-12 months including degron line construction; the two computational precursors are weeks |
| decides | hyp-the-cxxc-address-is-a-contested-resource |
| predicted outcome on current evidence | No change in fusion occupancy - the null - which would be a positive result for the menin-interface hypothesis because it locates the fusion's stability in protein-protein contacts rather than in DNA affinity. |
cell systems
[ "SEM (KMT2A::AFF1)", "RS4;11 (KMT2A::AFF1)", "MV4-11 or MOLM-13 (KMT2A::MLLT3)" ]
computational precursors
[ "Re-map existing KMT2A-fusion ChIP-seq against an rDNA-containing assembly - no new wet work.", "Intersect published CXXC-reader occupancy with published KMT2A-fusion occupancy, with a matched-promoter null; if the islands do not coincide, the degron experiment is unnecessary." ]
cxxc dna kd conditions
{
"_why": "Added by king sweep 2026-09-21 (cycle 46). This entry quotes the KMT2A CXXC-DNA 44 uM. That number is a property of five choices, not of the molecule, and quoting it bare is how this graph manufactured a 1,300-fold 'discrepancy' with the menin contact.",
"value": "44 uM",
"construct": "isolated CXXC domain, KMT2A 1147-1203, untagged, 15N-labelled",
"ligand": "DNA pal oligonucleotide, one CpG element",
"method": "NMR titration",
"buffer": "50 mM phosphate, pH 7.1, 300 mM NaCl, 2 mM DTT, 25 C",
"nacl_mM": 300,
"source_pmid": "20010842",
"source_quote": "\"Determination of the dissociation constants by NMR for the CXXC domain (MLL 1147-1203) and mutants was performed by titration of 15N-labeled protein with concentrated solutions of DNA pal oligonucleotide. Experiments were carried out in 50 mM phosphate buffer (pH 7.1), 300 mM NaCl, 2 mM DTT at 25 C.\" -- Cierpicki 2010 NSMB, Methods, fetched and read from PMC2908503 on 2026-09-21. The result sentence reads \"The wild type protein yielded a Kd of 44 uM\".",
"this_is_the_high_salt_end": "300 mM NaCl is roughly twice physiological ionic strength and SIX TIMES the 50 mM the menin-MBM1 53 nM was measured in. The same domain measured tagless at 150 mM NaCl gives 5.7 uM -- 7.7-fold tighter from salt alone.",
"do_not_compare_naively": {
"against": "menin-MBM1 53 nM (PMID 20961854), measured at 50 mM NaCl by FP",
"best_matched_pair": "108-fold, not ~1,000-fold",
"see": "ana-the-menin-affinity-comparison-was-made-at-one-sixth-the-salt"
},
"added": "2026-09-21"
}- If the answer is the null, the project should stop treating the CXXC family as a mechanism question and treat it purely as a selectivity problem for chemistry - which is cheaper and is what the phase-2 dossier already assumes.
- A positive result would be awkward for the menin-interface hypothesis only if the occupancy change were large; a small change would show the address is jointly determined and would not displace menin from the centre of the mechanism.
- Are CXXC1, KDM2A and KDM2B all expressed at measurable protein level in SEM and RS4;11? If one is not, it drops out of the design and the family question narrows further.
- Would a 2-hour depletion be long enough for an occupancy change at a site whose vacancy has to be found by diffusion, and short enough to exclude transcriptional secondary effects?
- Does any existing KMT2A-fusion ChIP-seq dataset have raw reads available for re-mapping against an rDNA-containing assembly?
- E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The human CXXC domains fall into distinct measured DNA-selectivity classes, with KDM2A's and KMT2A's among the closest in reader chemistry - both preferring non-methylated CpG with binding blocked by methylation.
DNA Sequence Recognition of Human CXXC Domains and Their Structural Determinants. · Xu C, Liu K, et al. · Structure 2018What this evidence cannot show
Isolated domains on short duplexes in vitro; the closeness of KDM2A and KMT2A is a reading of the selectivity classification plus the KDM2A recruitment paper, not a single head-to-head measurement.
Evidence that argues against this (1)
- A head-to-head that was done - MLL1 versus TET3 - found the earlier family measurements had not controlled for interference from flanking cytosines, so pairwise closeness in this classification carries an uncontrolled variable. Stroynowska-Czerwinska A et al. 2018
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
KMT2A and KMT2F/SETD1A bind entire rDNA loci, a result obtained only by using customised rDNA genome assemblies because standard assemblies omit the rDNA repeats.
The KMT2F histone methyltransferase interacts with the RNA polymerase I machinery to promote ribosomal RNA transcription. · Lone KA, Karole AM, et al. · PLoS Biol 2026What this evidence cannot show
Mammalian cells, wild-type KMT2A rather than a fusion; abstract read this session, full text not read. The methodological point - that standard assemblies discard these reads - is what makes it relevant to the design here.
Evidence that argues against this (1)
- KDM2B was shown to demethylate H3K4me3 at rDNA and repress it, so the H3K4me3 signal at rDNA is under active opposition and a single-factor ChIP there may be hard to interpret. Frescas D et al. 2007
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
KDM2B is a nucleolar protein that preferentially binds the transcribed region of rDNA and represses rRNA transcription by demethylating H3K4me3 in the nucleolus.
JHDM1B/FBXL10 is a nucleolar protein that represses transcription of ribosomal RNA genes. · Frescas D, Guardavaccaro D, et al. · Nature 2007What this evidence cannot show
Human somatic cell lines, 2007; nothing haematopoietic or leukaemic, and the paper does not test whether the ZF-CxxC is what targets KDM2B to rDNA.
Evidence that argues against this (1)
- KDM2B's better-characterised address is CpG islands genome-wide as the PRC1.1 targeting subunit, a different compartment; which dominates in a leukaemic blast is unknown. Blackledge NP et al. 2010
- E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
The reciprocal experiment has been run once: KDM2A overexpression reduced occupancy of LEDGF and wild-type MLL and not of the MLL fusion, while antagonising MLL-associated leukaemogenesis.
ASH1L Links Histone H3 Lysine 36 Dimethylation to MLL Leukemia. · Zhu L, Li Q, et al. · Cancer Discov 2016What this evidence cannot show
Abstract read this session, with the occupancy detail taken from this graph's existing mark-h3k36me2-3 and ledgf-axis entries. Overexpression not depletion, murine MLL leukaemia, H3K36me2 as the manipulated variable. It is the closest existing result and it motivates the design rather than replacing it.
Evidence that argues against this (1)
- Because the manipulation was overexpression of an H3K36me2 demethylase, a negative for fusion displacement could reflect that the fusion does not depend on H3K36me2 rather than that it resists CXXC competition. Zhu L et al. 2016
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Structure-guided point mutants that disrupt the KMT2A CXXC-DNA contact abolish immortalisation and leukaemia in mice; the contact's affinity is NOT a single number - the NMR titration behind the widely quoted 44 uM used the isolated 1147-1203 domain, a single-CpG 12-mer and 300 mM NaCl, and ten published measurements of the same contact span 44 uM to 33 nM.
Structure of the MLL CXXC domain-DNA complex and its functional role in MLL-AF9 leukemia. · Cierpicki T, Risner LE, et al. · Nat Struct Mol Biol 2010What this evidence cannot show
Isolated-domain biophysics plus murine MLL-AF9 transduction and transplantation. CORRECTED 2026-09-21 by Q82: this claim previously read that the contact is 'individually weak - NMR titration gives Kd approximately 44 uM'. That figure is the extreme low-affinity point of a ten-measurement series (44 uM to 33 nM) and was taken at twice nuclear ionic strength on the shortest construct and the poorest DNA ligand. The nearest-to-physiological tagless measurement is 5.7 uM at 150 mM NaCl (PMID 34280370). The mutant arm of the claim is unaffected: C1188D abolishes binding and abolishes leukaemogenesis. See ana-cxxc-dna-affinity-ladder.
Evidence that argues against this (2)
- Single-molecule work on long DNA with a real CpG island shows spatially dependent collaboration among CpGs, so the effective affinity on a genomic island may be considerably higher than 44 uM. Liang L et al. 2021
- The single-molecule paper already cited in this array turns out to carry a dissociation constant too - Kd 5.7 +/- 0.2 uM by EMSA at 150 mM NaCl for a tagless CXXC domain - which is the most physiologically apt measurement of this contact in the literature and is nearly eight-fold tighter than the 44 uM this claim used to quote. It was invisible to this project's PubMed sweeps because MLL[tiab] does not match the token MLL1. Liang L et al. 2021
The one experiment that decides the hypothesis, by measuring fusion occupancy under acute reader depletion.
The experiment the counted absence says has never been run.
First of the three readers to deplete - the SET1 complex's CpG-reading subunit, and the SET1-versus-KMT2A question the field has posed and never answered.
Second reader: the closest reader chemistry to KMT2A's and the only family member with a published antagonism result.
Third reader: the PRC1.1 targeting subunit, and the one carrying the nucleolar rDNA address.
Tests whether this weak (~44 uM) interface is displaceable in a cell by a rival reader.
Motivates the experiment better than the absence did: KDM2B co-resides on nearly every uCpG site the fusion uses, so a KDM2B degron is now a targeted test rather than a fishing trip.
The experiment entry states the CXXC-DNA contact is individually weak at about 44 uM; the ladder shows that number is the extreme point of a nine-measurement series.