The LEDGF PWWP–H3K36me nucleosome contact: how the fusion learns its address
also called LEDGF PWWP domain · PWWP-H3K36me2/me3 · LEDGF nucleosome anchor · chromatin address of the menin-KMT2A module
The far end of LEDGF is a small reader module that grips nucleosomes carrying a specific methyl tag, H3K36 methylation. That grip is what tells the whole menin-KMT2A machine which stretches of DNA to sit on, and there are six good cryo-EM pictures of it.
The same thing, for a scientist
The N-terminal PWWP domain of LEDGF/p75 (PSIP1 residues ~1-93) binds nucleosomes bearing H3K36me2 or H3K36me3, making a bipartite contact with the methylated H3 tail and with nucleosomal DNA. Six independent cryo-EM structures resolve it on mono- and dinucleosomes at 2.69-4.0 Å. Because LEDGF is bridged to KMT2A by menin, this is the DNA-side anchor of the menin-KMT2A-LEDGF module, and it is the arm of LEDGF that determines genomic targeting rather than complex assembly. It is a compact methyl-lysine reader pocket of exactly the class the field already drugs in other proteins, and there is no chemical matter against it at all.
Tell me more about this picture
PDB 8PEO, cryo-EM at 2.69 A - the highest-resolution structure of this contact.
Tell me more about this picture
The contact is bipartite: the methylated H3 tail into an aromatic cage, plus a separate grip on nucleosomal DNA. Both halves would have to be considered by any inhibitor.
Tell me more about this picture
H3K36me3 versions of the same complex are 8PC5 and 8PC6; a dinucleosome version is 8CBN.
Tell me more about this picture
Alternative poses exist at comparable resolution for the same substrate, so the target conformation is not single-valued - a real problem for structure-based design.
The structures, and why there are so many of them2 deeper
Every identifier below was resolved against the RCSB API rather than recalled 1:
| PDB | Nucleosome substrate | Method | Resolution | Primary citation |
|---|---|---|---|---|
| 6S01 | H3K36-methylated mononucleosome (165-mer DNA) | cryo-EM | 3.2 Å | PMID 31819277 |
| 8PEO | H3K36me2 mononucleosome | cryo-EM | 2.69 Å | PMID 37615563 |
| 8PEP | H3K36me2, alternative pose | cryo-EM | 3.33 Å | PMID 37615563 |
| 8PC5 | H3K36me3 mononucleosome | cryo-EM | 3.02 Å | PMID 37615563 |
| 8PC6 | H3K36me3, alternative pose | cryo-EM | 3.04 Å | PMID 37615563 |
| 8CBN | H3K36me3 dinucleosome | cryo-EM | 3.34 Å | PMID 37615563 |
| 8CBQ | H3K36me3 dinucleosome | cryo-EM | 4.0 Å | PMID 37615563 |
Three things are worth noticing about that table.
Every entry models PSIP1 residues 1-530, i.e. the full-length protein was used even where only the PWWP is ordered on the nucleosome. The rest of LEDGF is disordered in these maps, which is consistent with its AlphaFold model (global pLDDT 62.6, 47.4% of residues in the very-low band) and is a warning against modelling the linker.
Both me2 and me3 are captured, separately and at good resolution. That matters because H3K36me2 and H3K36me3 have different genomic distributions — me2 broad and intergenic, me3 concentrated over transcribed gene bodies — so which mark LEDGF prefers is a statement about where the fusion goes, not a technical detail.
There are alternative poses and a dinucleosome. Two poses at comparable resolution for the same substrate (8PC5/8PC6, 8PEO/8PEP) say the domain is not locked in one orientation, and the dinucleosome structures say the relevant unit may be more than one nucleosome. For a structure-based drug-design campaign this is inconvenient: the target conformation is not single-valued.
There is a very close precedent for drugging exactly this fold
BRPF1 carries an analogous PWWP domain that reads H3K36me3, solved at 1.50-1.85 Å with an H3K36me3 peptide (2X4W, 2X4X, 2X4Y, and 3MO8 at 1.69 Å) 2. Unlike LEDGF, BRPF1 has been drugged: ChEMBL target CHEMBL3132741 carries 173 activity rows at pChEMBL ≥ 6 with a best Kd of 31 nM from a 1,3-dimethylbenzimidazolone series 3.
That is the strongest single argument for the LEDGF PWWP as a target. It is not a speculative pocket class. A structurally homologous reader in the same beat has a hundred-fold more chemical matter and nanomolar probes. The LEDGF PWWP has nothing — PSIP1's entire ChEMBL footprint is three rows of antiviral integrase-interface chemistry aimed at the opposite end of the protein 3.
Also in the same beat: a 2026 structure putting LEDGF inside an elongating polymerase complex
9XYC (cryo-EM, 3.5 Å) models PSIP1 2-530 within a Pol II-DSIF-SPT6-PAF1c-TFIIS-IWS1-ELOF1-LEDGF-nucleosome assembly 1. We have not verified the primary paper behind that entry, so no mechanistic claim is made from it here and it is listed only as an identifier. If it holds up, it places LEDGF not merely on chromatin but inside the elongation machinery the fusion hijacks, which would make the PWWP arm a much more central target than a simple tether model implies. That structure is the highest-value single thing for the next cycle to verify.
Which arm of LEDGF should we drug? An honest comparison
| PWWP arm (this node) | IBD arm (interface-psip1-ibd-kmt2a) | |
|---|---|---|
| What blocking it does | removes the chromatin address; the KMT2A contact survives | uncouples KMT2A; LEDGF stays on chromatin |
| Structural basis | 6 cryo-EM structures, 2.69-4.0 Å, on real nucleosomes | 1 NMR complex plus a 3.0 Å ternary crystal |
| Pocket quality | compact methyl-lysine aromatic cage; classic reader pocket | extended peptide clamp; harder PPI class |
| Precedent chemistry | none for LEDGF, but 173 potent rows for the homologous BRPF1 PWWP | LEDGIN antiviral small molecules against the same domain, different footprint |
| Published affinity | none verified | none verified |
| Selectivity liability | H3K36me2/me3 is abundant and genome-wide; HDGF-family paralogues share the fold | mutually exclusive with integrase, POGZ, JPO2, MED19 |
| Cellular proof of concept | none | yes, cyclic peptide CP65 in murine MLL-AF9 blasts |
On structure quality and pocket class the PWWP arm is the better target. On demonstrated cellular effect the IBD arm is ahead by a mile, because it is the only one anybody has actually tried. The two also have different theories of failure, which is an argument for making both probes rather than choosing.
The uncomfortable part: this arm is the least KMT2A-specific thing in the dossier
H3K36me2 and H3K36me3 are among the most abundant histone marks in the genome, deposited by NSD1/NSD2/NSD3 and by SETD2 respectively and present across essentially all transcribed chromatin. LEDGF's PWWP domain does not read a KMT2A-specific signal; it reads a general one.
So the selectivity of a PWWP-directed drug would come entirely from the dependency of the cell, not from the chemistry — the same situation as the menin pocket, which is shared with JUND. The project should stop expecting any of these interfaces to be intrinsically fusion-selective and start treating dependency as the only source of selectivity available. The one place where genuine fusion-specificity could exist is the junction epitope itself (target-fusion-junction-epitope), and that is a different modality problem.
A second worry specific to this arm: KMT2A fusions in which the partner is NSD1 (NUP98::NSD1 is the famous case, but KMT2A-partner biology intersects here) change the H3K36 methylation landscape directly. If the mark that LEDGF reads is itself reshaped by fusion biology, a PWWP inhibitor's effect becomes context-dependent in a way that is hard to predict from a mononucleosome structure.
Counter-argument sweep (2026-09-19): the bipartite problem is measured, and the answer is unfavourable
This entry's druggability case rests on an analogy - BRPF1's PWWP is a tractable methyl-lysine pocket with 173 potent ChEMBL rows, so LEDGF's should be too - and on an open question about whether the bipartite contact is fatal for a small molecule. That question has a published answer and it is discouraging.
van Nuland et al. measured the PSIP1 PWWP directly: it binds an H3K36me3 peptide and DNA with low affinity, through distinct adjacent surfaces, and its affinity for an H3K36me nucleosome is about 10,000-fold higher than for the methylated peptide [counter-evidence to E2]. The specificity and the strength both come from concerted engagement of the tail and of nucleosomal DNA, which the later cryo-EM work describes as multivalent cross-gyre binding to both DNA gyres (PMID 31819277, the source of 6S01 in E1).
What that does to the target case. A cage-occupying ligand is competing with the weak half of the interaction. To evict LEDGF it would have to beat an avidity effect that the methyl-lysine site contributes little to. That is precisely the failure mode listed in bite_back, and it is now a measurement rather than a worry. The BRPF1 precedent does not rescue it, because the BRPF1 chemistry and structures are peptide-complex work; nobody has shown a PWWP ligand displacing any PWWP protein from a nucleosome.
Two things that would change the verdict. A DNA-competitive or interface-spanning ligand, or a degrader, would not face this problem - the entry should say so, because it reframes the modality question from 'occupy the cage' to 'remove the protein', exactly as this graph's ENL entry concluded for a different reader. And a direct measurement of nucleosome-displacement potency for any PWWP chemotype, positive or negative, would be the most informative single experiment for this interface. Neither exists.
Selectivity is worse than the DepMap number implies. HRP-2/HDGFL2 carries the same PWWP+IBD architecture, and its depletion reduces colony formation in leukaemic cells regardless of MLL rearrangement [counter-evidence to E4]. PSIP1's Chronos score was measured with HRP-2 intact; a PWWP-directed ligand would not be so discriminating.
| partner a | gene-psip1 |
|---|---|
| partner b | nucleosome-h3k36me2-me3 |
| motif a | LEDGF/PSIP1 PWWP domain, approximately residues 1-93; deposited constructs are full-length PSIP1 1-530 |
| motif b | Histone H3 tail around Lys36 bearing me2 or me3, plus a direct contact with nucleosomal DNA |
| interaction type | methyl-lysine reader plus DNA contact (bipartite) |
| kd status | No dissociation constant for the LEDGF PWWP-H3K36me nucleosome interaction was verified in this cycle. None of the cryo-EM primary citations we checked reports a numeric affinity in its abstract. Any Kd quoted for this contact elsewhere should be traced to a full text before use. |
| druggability | Downgraded 2026-09-19. Structurally the pocket is attractive - six cryo-EM structures on real nucleosomes, one at 2.69 A, and a homologous BRPF1 PWWP with 173 potent ChEMBL rows - but the binding mode is the problem. Peptide and DNA binding are each LOW affinity and the nucleosome is bound about 10,000-fold more tightly than the peptide (PMID 23656834), so occupancy of the aromatic cage competes with the minor component of an avidity-driven interaction, and no PWWP ligand has ever been shown to displace a PWWP protein from a nucleosome. Selectivity is also worse than the PSIP1 CRISPR score implies because HRP-2/HDGFL2 shares the fold and matters for leukaemic cell survival (PMID 33477970). The modality question should be reopened around degradation or interface-spanning ligands rather than reader-pocket occupancy. |
kd measurements
[]
structures
[ "8PEO", "8PC5", "8PC6", "8PEP", "6S01", "8CBN", "8CBQ", "9XYC" ]
structure details — the numbers behind the chart
[
{
"pdb": "8PEO",
"substrate": "H3K36me2 mononucleosome",
"method": "cryo-EM",
"resolution_A": 2.69,
"pmid": "37615563",
"title": "Multivalency of nucleosome recognition by LEDGF.",
"authors": "Koutná E, Lux V, Kouba T, et al.",
"venue": "Nucleic Acids Res",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "8PC5",
"substrate": "H3K36me3 mononucleosome",
"method": "cryo-EM",
"resolution_A": 3.02,
"pmid": "37615563",
"title": "Multivalency of nucleosome recognition by LEDGF.",
"authors": "Koutná E, Lux V, Kouba T, et al.",
"venue": "Nucleic Acids Res",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "8PC6",
"substrate": "H3K36me3 mononucleosome, pose 2",
"method": "cryo-EM",
"resolution_A": 3.04,
"pmid": "37615563",
"title": "Multivalency of nucleosome recognition by LEDGF.",
"authors": "Koutná E, Lux V, Kouba T, et al.",
"venue": "Nucleic Acids Res",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "8PEP",
"substrate": "H3K36me2 mononucleosome, pose 2",
"method": "cryo-EM",
"resolution_A": 3.33,
"pmid": "37615563",
"title": "Multivalency of nucleosome recognition by LEDGF.",
"authors": "Koutná E, Lux V, Kouba T, et al.",
"venue": "Nucleic Acids Res",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "6S01",
"substrate": "H3K36-methylated mononucleosome, 165-mer DNA",
"method": "cryo-EM",
"resolution_A": 3.2,
"pmid": "31819277",
"title": "Structure of H3K36-methylated nucleosome-PWWP complex reveals multivalent cross-gyre binding.",
"authors": "Wang H, Farnung L, Dienemann C, et al.",
"venue": "Nat Struct Mol Biol",
"year": 2020,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "8CBN",
"substrate": "H3K36me3 dinucleosome",
"method": "cryo-EM",
"resolution_A": 3.34,
"pmid": "37615563",
"title": "Multivalency of nucleosome recognition by LEDGF.",
"authors": "Koutná E, Lux V, Kouba T, et al.",
"venue": "Nucleic Acids Res",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "8CBQ",
"substrate": "H3K36me3 dinucleosome",
"method": "cryo-EM",
"resolution_A": 4,
"pmid": "37615563",
"title": "Multivalency of nucleosome recognition by LEDGF.",
"authors": "Koutná E, Lux V, Kouba T, et al.",
"venue": "Nucleic Acids Res",
"year": 2023,
"title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
},
{
"pdb": "9XYC",
"substrate": "Pol II-DSIF-SPT6-PAF1c-TFIIS-IWS1-ELOF1-LEDGF-nucleosome",
"method": "cryo-EM",
"resolution_A": 3.5,
"pdb_verified": "2026-09-20 against data.rcsb.org: 9XYC released 2026-03-18, ELECTRON MICROSCOPY, 3.5 A, title 'Pol II-DSIF-SPT6-PAF1c-TFIIS-IWS1-ELOF1-LEDGF-nucleosome LEDGF+nucleosome map Q' - matches the substrate, method and resolution stored here.",
"primary_citation": "Structure and function of IWS1 in transcription elongation.",
"pmid_absent_because": "The PDB primary citation carries NO PubMed id and NO DOI in the RCSB record, i.e. the structure paper is not yet indexed. The resolving identifier for this row is the PDB accession 9XYC, not a PMID. The previous value of this field was the free-text string 'unverified this cycle', which is not a PMID and which every citation checker in this project silently skipped.",
"url": "https://www.rcsb.org/structure/9XYC"
}
]homologous druggable precedent
{
"protein": "BRPF1",
"uniprot": "P55201",
"pwwp_structures": [
"2X4W",
"2X4X",
"2X4Y",
"3MO8"
],
"pwwp_resolution_range_A": [
1.5,
1.85
],
"chembl_target": "CHEMBL3132741",
"n_activity_rows_pchembl_ge_6": 173,
"best_activity": {
"molecule": "CHEMBL3752151",
"type": "Kd",
"value": 31,
"unit": "nM"
}
}ligands
[]
ID corrections
[ "The LEDGF nucleosome structures are 6S01, 8PC5, 8PC6, 8PEO, 8PEP, 8CBN and 8CBQ; 8PEP and 8PC6 are alternative poses of the same complexes rather than different interactions", "The KAT7/HBO1 structural partner is BRPF2/BRD1 (UniProt O95696), not BRPF1 (P55201) - BRPF1 appears in this node only as a homologous PWWP precedent, not as a KAT7 subunit", "No AlphaFold model exists for full-length KMT2A (AF-Q03164-F1 returns HTTP 404); the AlphaFold model of PSIP1 itself (AF-O75475-F1) has global pLDDT 62.6 with 47.4% very-low confidence and must not be used for the linker between the PWWP and the IBD" ]
binding mode
{
"summary": "bipartite and avidity-driven: low-affinity H3K36me3 peptide binding and low-affinity DNA binding through distinct adjacent surfaces on the PWWP, with approximately 10,000-fold enhancement on an H3K36me nucleosome relative to the methylated peptide",
"source_pmid": "23656834",
"structural_corroboration": "cryo-EM describes multivalent cross-gyre binding to both gyres of nucleosomal DNA (PMID 31819277)",
"drug_design_consequence": "a methyl-lysine-cage-only inhibitor competes with the weak component; occupancy is the wrong modality unless a ligand spans the DNA contact, and degradation is the obvious alternative"
}- The contact is bipartite and the numbers are now known: the PWWP binds an H3K36me3 peptide and DNA each with LOW affinity through distinct surfaces, and nucleosome binding is about 10,000-fold tighter than peptide binding (PMID 23656834). A ligand occupying the methyl-lysine cage competes with the weak half of an avidity effect and should not be expected to evict the module. This is the strongest single argument against the target as framed.
- H3K36me2/me3 is an abundant, genome-wide mark and HDGF-family proteins share the PWWP fold, so a PWWP-directed drug is neither KMT2A-selective nor obviously LEDGF-selective; the paralogue HRP-2/HDGFL2 carries the same PWWP+IBD architecture and its depletion reduces leukaemic colony formation independently of MLL rearrangement (PMID 33477970).
- Compensation: blocking the chromatin address leaves menin still gripping the KMT2A MBM1 motif and CXXC still binding unmethylated CpG, so the fusion retains two independent routes to chromatin.
- Alternative poses at equal resolution mean the target conformation is not single-valued, which degrades the reliability of structure-based design and of any docking or co-folding campaign built on one deposited model.
- LEDGF's role in DNA double-strand break repair runs partly through chromatin engagement; removing the anchor during genotoxic therapy could increase mutagenesis in surviving cells.
- The upstream efficacy question is unsettled independently of chemistry: LEDGF/p75 is essential for the INITIATION of MLL-mediated leukaemia and dispensable for steady-state haematopoiesis, with no published test of LEDGF removal from an established human KMT2A::AFF1 leukaemia (PMID 29084774).
- Does the LEDGF PWWP prefer H3K36me2 or H3K36me3, and by how much? The two marks have different genomic distributions, so the answer determines where the menin-KMT2A-LEDGF module is directed. Both structures exist; the comparative affinity does not.
- Do the 1,3-dimethylbenzimidazolone chemotypes developed against BRPF1 show any measurable binding to the LEDGF PWWP domain? A cross-screen of existing PWWP and bromodomain probes against LEDGF is a small, cheap, purely in vitro experiment that nobody appears to have published.
- What does 9XYC actually show? If LEDGF is genuinely a component of an elongating Pol II-PAF1c-nucleosome assembly rather than a passive tether, the PWWP arm becomes a far more central target. Verifying that primary paper is the highest-value single task for the next cycle.
- Does the H3K36 methylation landscape of KMT2A::AFF1 leukemia differ from normal progenitors in a way that would make a PWWP inhibitor more or less selective than the structure suggests?
- Given that nucleosome binding is about 10,000-fold tighter than peptide binding and is driven by nucleosomal DNA (PMID 23656834), can ANY methyl-lysine-cage ligand displace a PWWP protein from chromatin? No published example exists for any PWWP protein, and finding one - or failing to - would settle the modality question for this interface in one experiment.
- Would a LEDGF degrader be the right modality instead of a PWWP occupancy inhibitor? The avidity argument that defeats cage occupancy does not apply to degradation, and this graph already reached the same conclusion for the ENL YEATS reader.
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Six cryo-EM structures resolve the LEDGF/PSIP1 PWWP domain on H3K36-methylated nucleosomes: 6S01 (3.2 A, methylated mononucleosome), 8PEO (2.69 A) and 8PEP (3.33 A) with H3K36me2, 8PC5 (3.02 A) and 8PC6 (3.04 A) with H3K36me3, and 8CBN (3.34 A) and 8CBQ (4.0 A) with an H3K36me3 dinucleosome; all model PSIP1 residues 1-530. A further entry, 9XYC (3.5 A), models PSIP1 2-530 inside a Pol II-DSIF-SPT6-PAF1c-TFIIS-IWS1-ELOF1-LEDGF-nucleosome assembly.
RCSB PDB entries 6S01, 8PC5, 8PC6, 8PEO, 8PEP, 8CBN, 8CBQ and 9XYC, with per-entity UniProt accessions, methods, resolutions and modelled residue ranges resolved against the RCSB search and data GraphQL APIs · RCSB Protein Data Bank · RCSB PDB 2026What this evidence cannot show
Identifiers, methods, resolutions and residue ranges were read from the RCSB API in this session, but the residue ranges are the deposited construct ranges as RCSB annotates them against UniProt, not the ordered region in the map, so 'models PSIP1 1-530' does not mean 530 residues are resolved. Resolutions of 3.0-4.0 A do not support side-chain-level design; only 8PEO at 2.69 A approaches that. The primary paper behind 9XYC has not been verified in this cycle and no mechanistic claim is made from it.
Evidence that argues against this (1)
- Counting PDB entries measures how tractable a protein was to CRYSTALLISE, which correlates with being small, ordered and well-behaved rather than with being druggable. What decides design value is whether the deposited density covers the site that matters, at a resolution that supports modelling, in the relevant state — and this graph has repeatedly found structures described in the direction that overstates what is resolved. Six cryo-EM structures of one domain on reconstituted nucleosomes are six views of the same in vitro system, and H3K36 methylation marks most transcribed gene bodies — so the interface explains binding and cannot explain locus selection. Cancer Research 2014
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The PWWP domain of BRPF1 is a structurally characterised H3K36me3 reader, solved with an H3K36me3 peptide at 1.50-1.85 A (2X4W, 2X4X, 2X4Y) and at 1.69 A (3MO8), establishing that this domain class forms a tractable, well-ordered methyl-lysine binding site.
Molecular basis of histone H3K36me3 recognition by the PWWP domain of Brpf1 · Vezzoli A, Bonadies N, Allen MD, et al. · Nature Structural and Molecular Biology 2010What this evidence cannot show
BRPF1 and LEDGF PWWP domains are homologous but not identical, and the BRPF1 structures are peptide complexes rather than nucleosome complexes, so transferability of chemotypes is an assumption rather than a result. The abstract reports no numeric affinity.
Evidence that argues against this (1)
- The reader-pocket framing that carries this entry's druggability argument does not survive the thermodynamics. van Nuland et al. measured the PSIP1 PWWP domain directly and report that it binds an H3K36me3 PEPTIDE and DNA with low affinity, through distinct adjacent binding surfaces, and that binding to H3K36me nucleosomes is enhanced approximately 10,000-fold compared with the methylated peptide. The high-affinity, specific recognition of the nucleosome is therefore produced by CONCERTED binding to the methylated tail and to nucleosomal DNA, not by the aromatic cage. The later cryo-EM structure of the same complex describes the same thing structurally as multivalent cross-gyre binding to both gyres of nucleosomal DNA (PMID 31819277, already cited here as the source of 6S01). The practical consequence is that this entry's open question about whether the bipartite contact is fatal for a small molecule already has a discouraging published answer: a ligand occupying only the methyl-lysine cage competes against the weak half of a 10,000-fold avidity effect and should not be expected to evict LEDGF from chromatin. The BRPF1 comparison in E2 and E3 is a comparison of peptide-complex tractability, which is not the binding mode that matters here. van Nuland R et al. 2013
- E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
The homologous BRPF1 PWWP-containing reader has substantial chemical matter in ChEMBL v34 (target CHEMBL3132741, 173 activity rows at pChEMBL >= 6, best Kd 31 nM from a 1,3-dimethylbenzimidazolone series) whereas PSIP1 has three rows in total (CHEMBL3988590, best IC50 470 nM), all of them assaying the HIV-1 integrase-LEDGF interface at the opposite end of the protein, and no curated mechanism of action.
ChEMBL database version 34, targets CHEMBL3132741 (Peregrin/BRPF1) and CHEMBL3988590 (PC4 and SFRS1-interacting protein): target_search, get_bioactivity and get_mechanism queries · EMBL-EBI ChEMBL · ChEMBL 2026What this evidence cannot show
The BRPF1 chemical series is directed at its bromodomain as well as its PWWP domain and the returned page does not separate them, so '173 potent rows' overstates PWWP-specific matter. Neither target has a curated clinical candidate. Distinct-compound counts are page-local because of the 25-row API limit.
Evidence that argues against this (1)
- A ChEMBL activity-row count measures how much medicinal chemistry has been DONE, not how tractable the target is, and the best reported affinity belongs to the most optimised compound in whichever series someone chose to pursue. Zero rows means nobody tried; thousands of rows means a target had a decade of industrial attention. The BRPF1 comparison is the useful part and it cuts the other way: BRPF1's 173 rows exist because a bromodomain programme was already running next to its PWWP, so the contrast measures adjacency to a funded campaign rather than intrinsic ligandability. ChEMBL, EMBL-EBI 2026
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
PSIP1 is not DepMap common-essential and has a mean Chronos gene effect of +0.028 across 1,258 cell lines, yet falls to -0.444 across 10 KMT2A-rearranged lines, giving the best combination of general dispensability and KMT2A-rearrangement-selective requirement of any non-menin partner examined.
DepMap CRISPR (Chronos) gene effect via the Open Targets Platform GraphQL API; loss-of-function constraint via the gnomAD GraphQL API; normal-tissue expression via the Human Protein Atlas API · Broad Institute DepMap; gnomAD; Human Protein Atlas; Open Targets Platform · Databases 2026What this evidence cannot show
Means were computed by us from per-line API values rather than read from a DepMap release table, and the release version is not stated by the API. Only 10 KMT2A-rearranged lines (4 KMT2A::AFF1) contribute. Critically for this node, the CRISPR score reflects loss of the whole protein and therefore of both the PWWP and the IBD arms at once; it cannot be attributed to this interface alone.
Evidence that argues against this (1)
- A single-gene CRISPR score overstates the selectivity of this module because the PWWP+IBD architecture is shared with a paralogue that leukaemic cells also need. Van Belle et al. show HRP-2/HDGFL2 carries both a PWWP and an integrase-binding domain, interacts with MLL through a conserved interface less menin-dependently than LEDGF does, and - although HRP-2-null primary cells were efficiently transformed by MLL-ENL, so HRP-2 is dispensable for transformation - depletion of HRP-2 in leukaemic cell lines and transformed primary murine cells reduced colony formation independently of MLL rearrangement. PSIP1's Chronos score of -0.444 in KMT2A-rearranged lines is therefore measured in cells that still have an intact HRP-2 PWWP, and a PWWP-directed small molecule would hit both. That cuts both ways: it removes the ligand-level selectivity this entry hoped for, and it means the family-level dependency may be deeper than the PSIP1 number suggests. Van Belle S et al. 2021
STORED 'unverified this cycle' IN A PMID FIELD inside a status:verified entry; PDB 9XYC verified live and recorded as the resolving identifier instead.
how the tether reads the chromatin it sits on
constrains what the PWWP arm can be doing for the fusion
supplies the chemical matter that entry recorded as absent, and its ceiling
both functions are exercised on the same H3K36-methylated chromatin
whether the fusion requires this contact at all
Counter-evidence (van Nuland 2013, PMID 23656834): the PWWP binds H3K36me3 peptide and DNA each with LOW affinity through distinct surfaces, and nucleosome binding is ~10,000-fold tighter than peptide binding. Methyl-lysine-cage occupancy competes with the weak half of an avidity effect; druggability downgraded and the modality question reopened around degradation.
Shared counter-evidence (Van Belle 2021, PMID 33477970): HRP-2/HDGFL2 carries the same PWWP+IBD architecture and matters for leukaemic cell survival independent of MLL rearrangement, so the PSIP1 CRISPR selectivity that both interfaces rely on was measured with the paralogue intact.
bite-back: multivalent PWWP/linker anchoring may defeat IBD-only blockade
the mark that interface binds - previously a forward reference with no node behind it
- ← analogous toHDGF - the third PWWP paralogue, which pulls down LEDGF but has the reader domain WITHOUT the domain that grips MLL
homologous PWWP fold (12-69); affinity for H3K36me nucleosomes not measured for HDGF
- ← related toU2AF1 - the 3' splice-site recognition factor, a speckle protein whose BioID neighbourhood contains LEDGF
H3K36me2/3, the mark PSIP1 reads, is an established chromatin input to co-transcriptional splicing - the mechanistic reason this row is plausible and also the reason it is uninformative