Oncofusion

start / graph / gene-psip1

Proteinstrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q5b-interactome

PSIP1 / LEDGF p75 — the tether with the best therapeutic index in the whole interactome

also called LEDGF · LEDGF/p75 · p75 · PSIP2 · DFS70 · lens epithelium-derived growth factor · PC4 and SFRS1-interacting protein

LEDGF is the protein that reads a chemical mark on chromatin and uses it to park the menin-KMT2A complex on the right genes. Cells in general do not seem to need it at all, but KMT2A-rearranged leukemia cells do, which makes it the most promising second target after menin.

The same thing, for a scientist

PSIP1 (O75475, 530 aa) encodes LEDGF/p75, a two-module chromatin tether: an N-terminal PWWP domain that reads H3K36me2/me3 on the nucleosome, and a C-terminal integrase-binding domain (IBD, ~aa 347-435) that clamps a short peptide on partner proteins. In KMT2A biology the IBD binds a KMT2A segment around residues 140-160 and forms a ternary complex with menin, converting the menin-KMT2A dimer into a chromatin-bound module. Its dependency profile is the most favourable of any partner examined: mean DepMap Chronos across 1,258 lines is +0.028, i.e. neutral, yet it falls to -0.444 in KMT2A-rearranged lines and -0.346 in KMT2A::AFF1 lines. There is essentially no oncology chemical matter against it.

loading 3U88…
PDB 3U88menin + KMT2A + LEDGF IBD ternary complex
Three proteins in one crystal: menin gripping the start of KMT2A, with LEDGF holding a nearby piece of the same KMT2A chain.
Tell me more about this picture

PDB 3U88, 3.0 A. Menin residues 2-550, KMT2A 103-153, LEDGF/PSIP1 347-435.

Tell me more about this picture

Menin and LEDGF are not competitors for KMT2A; they form a composite clamp on one KMT2A segment, menin supplying affinity and LEDGF supplying the chromatin address.

Tell me more about this picture

Menin-inhibitor resistance mutations preserve KMT2A binding, which means they should leave the LEDGF arm of this clamp entirely intact.

Tell me more about this picture

At 3.0 A the side-chain detail is not reliable enough to enumerate hot-spot residues from this structure alone.

loading 6S01…
PDB 6S01LEDGF PWWP on an H3K36-methylated nucleosome
The other end of LEDGF, docked onto a nucleosome that carries the H3K36 methyl mark.
Tell me more about this picture

PDB 6S01, cryo-EM at 3.2 A; the same interaction is also captured in 8PC5, 8PC6 (H3K36me3), 8PEO, 8PEP (H3K36me2) and 8CBN, 8CBQ (dinucleosome).

Tell me more about this picture

This is the DNA-side anchor of the menin-KMT2A-LEDGF module and a compact methyl-lysine reader pocket, which is a chemotype the field knows how to drug.

Tell me more about this picture

H3K36me2/me3 is an abundant, broadly distributed mark, so blocking this arm is not KMT2A-selective either.

What it looks like
PC4 and SFRS1-interacting protein (LEDGF/p75)530 aa
PWWP (H3K36me2/me3 reader)PWWP (H3K36me2/me3 reader) · residues 1–93 · structurally resolved on nucleosomes in 6S01, 8PC5, 8PEO, 8CBNAT-hook-like / charged regionAT-hook-like / charged region · residues 146–197 · DNA-binding, not resolved in complexIntegrase-binding domain (IBD)Integrase-binding domain (IBD) · residues 347–435 · modelled as 344-442 across 2MSR, 3U88, 2B4J0200400
loading AF-O75475-F1…
AlphaFold AF-O75475-F1 ↗
colourshape

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.

very high (90+)confident (70–90)low (50–70)very low (<50)
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.

Go deeper
Why this is the standout number in the dossier1 deeper

Put the DepMap numbers side by side 1:

mean ChronosKMT2A-r linesselectivity vs leukemiacommon-essential
MEN1-0.363-1.205-0.631no
PSIP1+0.028-0.444-0.346no
DOT1L-0.148-0.565-0.166no
WDR5-1.159-1.051+0.191yes
CDK9-1.536-1.687-0.077yes

PSIP1 is the only gene in the interactome whose average effect of knockout across 1,258 cancer cell lines is positive while still showing a real selective requirement in KMT2A-rearranged lines. Only 20.5% of lines are depleted at the -0.5 threshold. Menin has the larger selective effect; LEDGF has the cleaner baseline. For a project whose stated worry is essential-function toxicity, that combination is the point.

The caution is that gnomAD says something different: pLI is effectively 1.0 with LOEUF 0.373 and 19 observed loss-of-function variants against 74.7 expected 1. As with MEN1, this is a constraint signature, not a viability signature, and the two questions are not the same — but it does mean human heterozygous LoF is selected against, and a tissue-wide inhibitor is not obviously free.

Normal-tissue expression argues both ways

The Human Protein Atlas classifies PSIP1 as low tissue specificity, detected in all tissues, low immune cell specificity 1. There is no tissue in which it is absent, so an inhibitor would engage it everywhere.

That is a weaker objection than it sounds, for the same reason it was weak for menin: ubiquitous expression predicts exposure, not toxicity. Toxicity depends on whether any normal cell's survival runs through the specific contact being blocked. The relevant unknown is whether human haematopoietic stem cells need the LEDGF IBD, and nobody has asked that with an interface-specific mutant.

Two modules, two independent interfaces, and only one of them is fusion-relevant2 deeper

LEDGF is unusual among these partners in having two structurally separate, independently druggable surfaces, and there are verified structures of both.

The PWWP domain (N-terminal) reads methylated H3K36 on the nucleosome. This is the DNA-side anchor. Cryo-EM structures exist at real resolution: 6S01 (LEDGF PWWP on an H3K36-methylated nucleosome, 3.2 Å), 8PC5 and 8PC6 (H3K36me3 mononucleosome, 3.02 and 3.04 Å), 8PEO and 8PEP (H3K36me2, 2.69 and 3.33 Å), and 8CBN/8CBQ (H3K36me3 dinucleosome, 3.34 and 4.0 Å) 2. All of them model PSIP1 residues 1-530, so the construct is full-length even where only the PWWP is resolved on the nucleosome.

The IBD (aa ~347-435) is a peptide clamp with promiscuous partners. Verified complexes: 2MSR, the solution structure of the LEDGF IBD with a KMT2A 140-160 peptide (NMR); 3U88, the menin-KMT2A-LEDGF ternary crystal structure at 3.0 Å, which models PSIP1 347-435, KMT2A 103-153 and menin 2-550 together; 2N3A, the IBD with a POGZ peptide; and 2B4J, 3F9K, 3HPG and several cryo-EM intasome structures with lentiviral integrases 2.

That list is the warning as much as the opportunity. The same IBD surface that holds KMT2A also holds HIV-1 integrase, POGZ, JPO2/CDCA7L and MED19. A drug that blocks the clamp blocks all of them.

The 3U88 geometry is the reason LEDGF matters at all

3U88 is the structure that makes the argument concrete: in one crystal, menin holds KMT2A 103-153 while the LEDGF IBD holds an adjacent stretch of the same KMT2A chain. Menin and LEDGF are not alternative binders competing for KMT2A; they are a composite clamp on a single KMT2A segment, with menin supplying the affinity and LEDGF supplying the chromatin address 2.

This predicts something testable and, as far as we can find, untested: blocking menin and blocking the LEDGF IBD should not be additive but synergistic below additivity on the target and additive on the tumour, because each alone leaves a partially competent clamp. It also predicts that menin-inhibitor resistance mutations, which preserve KMT2A binding, should leave the LEDGF arm fully intact — so LEDGF is a candidate second-line mechanism, not just a parallel one.

The PWWP domain is the better drug target of the two, and nobody is working on it for oncology

Reasons to prefer the PWWP arm: it is a small, well-defined, compact reader module with a methyl-lysine pocket, which is a chemotype the field already knows how to attack (see BRPF1's 1,3-dimethylbenzimidazolone series, 173 potent activity rows in ChEMBL against an analogous reader 3); it has six independent high-resolution nucleosome-bound structures to design against; and blocking it should evict the whole menin-KMT2A-LEDGF module from chromatin rather than merely loosening it.

Reasons for caution: H3K36me2/me3 is a broad, abundant mark, so the PWWP arm is not KMT2A-specific either; and LEDGF's own transcriptional and DNA-repair roles run through the same domain.

Chemical matter: effectively zero, and what exists is from the wrong disease

ChEMBL v34 has a human PSIP1 single-protein target (CHEMBL3988590) with exactly 3 activity rows at pChEMBL >= 6, all from one 2015 document, best IC50 470 nM (CHEMBL3981070). Every one of them is an AlphaScreen/HTRF assay of the HIV-1 integrase-LEDGF/p75 protein-protein interaction, developed as an antiviral (the LEDGINs programme), not as an oncology agent. There is no curated mechanism of action against PSIP1 3.

That is a genuinely good position to be in. The antiviral field has already shown that the IBD clamp is a druggable protein-protein interface with small molecules, and has published chemotypes. Nobody has repurposed that chemistry against the KMT2A arm of the same clamp. Whether LEDGIN-class compounds displace the KMT2A 140-160 peptide as well as they displace integrase is an experiment that could be run in a month, and it is the single highest-value cheap experiment in this dossier.

What is missing from this node and should worry us

There is no published Kd for the LEDGF IBD-KMT2A interaction that we could verify. 2MSR establishes the geometry of the IBD with KMT2A 140-160 by NMR, and 3U88 establishes the ternary arrangement, but neither is a calorimetry or fluorescence-polarisation number, so the menin-style affinity table (53 nM for MBM1) has no LEDGF counterpart here. Until that is measured, statements about which arm of the clamp dominates are qualitative.

There is no AlphaFold-model gap to worry about: AF-O75475-F1 exists and resolves (HTTP 200), but its global pLDDT is 62.6 with 47.4% of residues in the very-low confidence band, which is what one expects for a protein that is largely intrinsically disordered between two folded modules. Do not use the AlphaFold model for the linker.

Two open questions retired, and a premise about the HIV field that was wrong

The Kd question is answered and was answered in 2014 - see interface-psip1-ibd-kmt2a. This entry carried "no verified number exists" for several cycles after the number was in the graph.

The LEDGIN question was malformed. It assumed LEDGINs/ALLINIs are LEDGF ligands inherited from the HIV field. They are integrase ligands. The decisive experiment is PMID 22396646: in cells with the PSIP1 IBD exons deleted, LEDGINs remained active, by acting through HRP-2's IBD; co-crystals place them in the integrase CCD dimer interface as molecular glue. No tool compound is inherited from HIV for LEDGF itself. What does transfer is assay format and the 2B4J surface map.

The chemical-matter position also needs updating in one direction: the PWWP arm is not empty - 68 crystallographic fragment hits with double-digit micromolar leads (PMID 39461037) - though they were raised on HRP-2 crystals, bind both paralogues, and come with the authors' own stated affinity ceiling.

Facts and figures
uniprotO75475
gene symbolPSIP1
length AA530
protein namePC4 and SFRS1-interacting protein (LEDGF/p75)
alphafold IDAF-O75475-F1
alphafold verifiedtrue
alphafold global plddt62.62
alphafold fraction plddt very low0.474
PDB verified against rcsbtrue
subnuclear localizationnuclear interior with H3K36me2/me3, not the H3K9me3 nuclear rim (ExEpi, HeLaP4, human); isoform-unresolved between LEDGF/p75 and LEDGF/p52
shown for fusionNO - no KMT2A fusion system appears anywhere in the localisation literature for PSIP1
domains
PWWP (H3K36me2/me3 reader)1.000PWWP (H3K36me2/me3 reader): 1AT-hook-like / charged re…146.0AT-hook-like / charged region: 146Integrase-binding domain …347.0Integrase-binding domain (IBD): 347
key structures
2B4J2.0202B4J: 2.028PEO2.6908PEO: 2.693U883.0003U88: 38PC53.0208PC5: 3.026S013.2006S01: 3.28CBN3.3408CBN: 3.34
domains — the numbers behind the chart
[
  {
    "name": "PWWP (H3K36me2/me3 reader)",
    "start": 1,
    "end": 93,
    "note": "structurally resolved on nucleosomes in 6S01, 8PC5, 8PEO, 8CBN"
  },
  {
    "name": "AT-hook-like / charged region",
    "start": 146,
    "end": 197,
    "note": "DNA-binding, not resolved in complex"
  },
  {
    "name": "Integrase-binding domain (IBD)",
    "start": 347,
    "end": 435,
    "note": "modelled as 344-442 across 2MSR, 3U88, 2B4J"
  }
]
PDB IDS
[
  "3U88",
  "2MSR",
  "6S01",
  "8PC5",
  "8PC6",
  "8PEO",
  "8PEP",
  "8CBN",
  "8CBQ",
  "2N3A",
  "2B4J",
  "3F9K",
  "3HPG",
  "9XYC"
]
key structures — the numbers behind the chart
[
  {
    "pdb": "3U88",
    "what": "menin + KMT2A 103-153 + LEDGF IBD 347-435 ternary",
    "method": "X-ray",
    "resolution_A": 3,
    "pmid": "22327296",
    "title": "The same pocket in menin binds both MLL and JUND but has opposite effects on transcription.",
    "authors": "Huang J, Gurung B, Wan B, et al.",
    "venue": "Nature",
    "year": 2012,
    "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": "2MSR",
    "what": "LEDGF IBD 344-426 + KMT2A 140-160 peptide",
    "method": "NMR",
    "resolution_A": null,
    "pmid": "25082813",
    "title": "Validation and structural characterization of the LEDGF/p75-MLL interface as a new target for the treatment of MLL-dependent leukemia.",
    "authors": "Cermáková K, Tesina P, Demeulemeester J, et al.",
    "venue": "Cancer Res",
    "year": 2014,
    "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",
    "what": "LEDGF PWWP on H3K36-methylated nucleosome",
    "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": "8PC5",
    "what": "LEDGF PWWP on H3K36me3 nucleosome",
    "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": "8PEO",
    "what": "LEDGF PWWP on H3K36me2 nucleosome",
    "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": "8CBN",
    "what": "LEDGF PWWP on 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": "2B4J",
    "what": "LEDGF IBD + HIV-1 integrase dimerisation domain",
    "method": "X-ray",
    "resolution_A": 2.02,
    "pmid": "16260736",
    "title": "Structural basis for the recognition between HIV-1 integrase and transcriptional coactivator p75.",
    "authors": "Cherepanov P, Ambrosio AL, Rahman S, et al.",
    "venue": "Proc Natl Acad Sci U S A",
    "year": 2005,
    "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."
  }
]
essentiality
{
  "depmap_common_essential": false,
  "chronos_mean": 0.028,
  "n_lines": 1258,
  "fraction_lines_depleted": 0.205,
  "chronos_mean_leukemia": -0.098,
  "chronos_mean_kmt2ar": -0.444,
  "chronos_mean_kmt2a_aff1": -0.346,
  "selectivity_vs_leukemia": -0.346,
  "selectivity_vs_all_lines": -0.472,
  "rank_in_interactome": 3,
  "gnomad_pLI": 1,
  "gnomad_loeuf": 0.373,
  "gnomad_obs_lof": 19,
  "gnomad_exp_lof": 74.7,
  "gnomad_mis_z": 0.23,
  "mouse_ko": "not established in this node; no verified citation was obtained this cycle",
  "hpa_rna_tissue_specificity": "Low tissue specificity",
  "hpa_rna_tissue_distribution": "Detected in all",
  "hpa_rna_blood_cell_specificity": "Low immune cell specificity"
}
most depleted kmt2ar lines
[
  {
    "line": "MOLM-13",
    "fusion": "KMT2A::MLLT3"
  },
  {
    "line": "SEMK2",
    "fusion": "KMT2A::AFF1"
  },
  {
    "line": "MV4-11",
    "fusion": "KMT2A::AFF1"
  }
]
chemical matter
{
  "chembl_target": "CHEMBL3988590",
  "chembl_pref_name": "PC4 and SFRS1-interacting protein",
  "n_activity_rows_pchembl_ge_6": 3,
  "best_activity": {
    "molecule": "CHEMBL3981070",
    "type": "IC50",
    "value": 470,
    "unit": "nM",
    "pchembl": 6.33,
    "assay_type": "B"
  },
  "curated_mechanism_drugs": [],
  "note": "all rows assay the HIV-1 integrase-LEDGF interface (LEDGIN antiviral chemistry), not an oncology programme"
}
interfaces
[
  "interface-psip1-ibd-kmt2a",
  "interface-psip1-pwwp-h3k36me-nucleosome"
]
localization
{
  "subcellular": "nucleus; chromatin-associated throughout the cell cycle, including on condensed mitotic chromosomes (live-cell EGFP-fusion imaging, human HeLa)",
  "subnuclear": "nuclear interior co-distributed with H3K36me2/me3, and explicitly NOT the H3K9me3 nuclear rim (expansion microscopy, human HeLaP4). Isoform-unresolved: the imaging antibody detects p75 and p52 together.",
  "how_shown": "live-cell EGFP/HcRed fusion imaging + RNAi (PMID 12796494); EGFP deletion-mutant imaging in interphase and mitosis (PMID 16549878); FRAP/FCS scan-and-lock kinetics (PMID 20974633); expansion microscopy ExEpi (PMID 35716125)",
  "systems": "human HeLa and HeLaP4 only; no haematopoietic and no KMT2A-rearranged cell has ever been imaged for this protein",
  "node": "localization-psip1-ledgf-chromatin-tether-subnuclear"
}
cell type
{
  "evidence_class_key": "concept-expressed-required-maintains-three-evidence-classes",
  "expressed_in": {
    "class": 1,
    "summary": "broadly expressed; expression carries no cell-type discrimination for this gene and is not used here as evidence of requirement"
  },
  "required_in": {
    "class": 2,
    "headline": "the cleanest normal-versus-leukaemic dissociation in the whole core set: LEDGF/p75 is DISPENSABLE for murine haematopoiesis and ESSENTIAL for MLL-rearranged leukemogenesis",
    "pmid": "29084774",
    "kmt2a_rearranged_lines_mean_gene_effect": -0.553,
    "all_1258_lines_mean_gene_effect": 0.029,
    "pct_all_lines_depleted": 1.1,
    "floor_compressed": false,
    "myeloid_kmt2ar_mean": -0.637,
    "blineage_kmt2ar_mean": -0.406,
    "same_fusion_myeloid_minus_blineage": -0.201,
    "verdict": "required in KMT2A-rearranged leukaemia cells and in essentially no other screened cell line (1.1% of 1,258). The largest normal-to-leukaemic contrast available for any gene in this beat.",
    "source": "PMID 29084774 for the mouse; DepMap via Open Targets fetched 2026-09-20 for the lines",
    "title": "LEDGF/p75 is dispensable for hematopoiesis but essential for MLL-rearranged leukemogenesis.",
    "authors": "El Ashkar S, Schwaller J, Pieters T, et al.",
    "venue": "Blood",
    "year": 2018,
    "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."
  },
  "context": "BOTH, and the contrast between them is the finding rather than a caveat. Murine system for the normal-haematopoiesis arm (brief rule 73).",
  "graph_links": [
    "cell-type-hsc-bone-marrow"
  ]
}
cell state
{
  "maintains": {
    "class": 3,
    "states": [],
    "claim": "none established"
  },
  "bounded_query": {
    "date": "2026-09-20",
    "tool": "NCBI esearch db=pubmed",
    "query": "(PSIP1[tiab] OR LEDGF[tiab] OR \"LEDGF/p75\"[tiab] OR \"p75\"[tiab] OR \"DFS70\"[tiab]) AND <cell-state clause> AND <leukaemia clause>",
    "count": 5,
    "on_topic_after_title_check": 0,
    "controls": {
      "gene_clause_alone": 7404,
      "state_clause_alone": 96743
    },
    "alias_precision_warning": "the alias p75 inflates the bare count enormously (p75 neurotrophin receptor, p75 NTR, p75 in Grb2 signalling); two of the five hits are 1990s p75 tyrosine-phosphorylation papers with no relation to PSIP1, and one is the unrelated NUP98::PSIP1 fusion in MDS",
    "reading": "PSIP1/LEDGF has never been placed in a defined cell STATE in leukaemia. Zero in titles and abstracts; supporting figures not searched."
  },
  "full_texts_opened_before_recording_the_absence": [
    "the beat did not open a full text for PSIP1; the absence is recorded at abstract level only and is labelled accordingly"
  ]
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.5529,
  "chronos_mean_other_leukaemia": 0.0116,
  "difference": -0.5645,
  "pan_essentiality_all_lines": 0.0278,
  "selectivity_rank": 7,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
menin mbm1 kd conditions
{
  "_why": "Added by king sweep 2026-09-21 (Q82 corrections-owed #3). This entry quotes the menin-MBM1 53 nM figure. That number is NOT comparable to the KMT2A CXXC-DNA affinities without stating its buffer, and the omission is how a 1,300-fold 'discrepancy' was manufactured.",
  "value": "53 nM",
  "motif": "KMT2A MBM1, residues 4-15, FITC-labelled peptide at 15 nM",
  "method": "fluorescence polarisation",
  "buffer": "50 mM Tris, pH 7.5, 50 mM NaCl, 1 mM DTT",
  "nacl_mM": 50,
  "companion_itc": "72 nM by ITC, 50 mM phosphate pH 7.5, 50 mM NaCl, 1 mM beta-mercaptoethanol, 27 C",
  "source_pmid": "20961854",
  "source_quote": "\"in the FP buffer (50 mM Tris, pH 7.5, 50 mM NaCl, 1 mM DTT) used for binding experiments (Kd measurements)\" -- Grembecka 2010 JBC, Methods, fetched and read from PMC3003368 on 2026-09-21",
  "do_not_compare_naively": {
    "against": "KMT2A CXXC-DNA Kd 44 uM (PMID 20010842, Cierpicki 2010)",
    "because": "that measurement was made at 300 mM NaCl -- verbatim from its Methods, \"50 mM phosphate buffer (pH 7.1), 300 mM NaCl, 2 mM DTT at 25 C\" (PMC2908503, read 2026-09-21) -- SIX-FOLD more salt, on a salt-sensitive protein-DNA contact, against a hydrophobic peptide pocket that is not.",
    "best_matched_pair": "44 uM at 300 mM NaCl vs the 5.7 uM tagless CXXC measurement at 150 mM NaCl gives 108-fold against menin, not ~1,000-fold",
    "see": "ana-the-menin-affinity-comparison-was-made-at-one-sixth-the-salt"
  },
  "added": "2026-09-21"
}
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"
}
How this could bite us back
  • The IBD is a promiscuous clamp shared with HIV-1 integrase, POGZ, JPO2/CDCA7L and MED19. Blocking it displaces all of them, and the consequences of losing the non-KMT2A partners are unmeasured.
  • H3K36me2/me3 is an abundant, genome-wide mark, so the PWWP arm is no more KMT2A-selective than the menin pocket is; selectivity again comes from the cell's dependency, not the chemistry.
  • gnomAD pLI ~1.0 and LOEUF 0.373 say human heterozygous loss is selected against. The reassuring DepMap number is from immortalised cancer lines and does not speak to stem-cell or developmental function.
  • LEDGF has independent roles in transcriptional coupling and in DNA double-strand break repair; inhibiting it during genotoxic chemotherapy could plausibly increase, not decrease, mutagenesis in surviving cells.
  • Redundancy risk: the shorter p52 isoform of PSIP1 shares the PWWP domain but lacks the IBD, and HDGF-family PWWP proteins read the same mark. A PWWP-directed drug may be buffered by paralogues that a CRISPR knockout of PSIP1 does not reveal.
Still open
  1. SUPERSEDED 2026-09-20: the Kd is published. 14.7 +/- 0.7 uM for MLL 1-160 binding the IBD without menin (NMR) and 1.4 uM against the preformed menin-MLL complex (ITC), a ~10-fold cooperative enhancement, with a bipartite interface - IBM1 124-133, IBM2 147-152 (PMID 25305204). What remains open is whether 14.7 uM is enough to hold the fusion at cellular concentrations.
  2. REWRITTEN 2026-09-20: this asked whether LEDGIN-class compounds displace the KMT2A peptide from the IBD, which presupposes LEDGINs bind LEDGF. **They bind HIV-1 integrase** - they remain active in cells with the PSIP1 IBD exons deleted (PMID 22396646) and co-crystallise in the integrase CCD dimer pocket. The real question: can ANY small molecule occupy the IBD's IBM2 site? The only reported direct LEDGF binder is ebselen, a promiscuous thiol reagent.
  3. Is the PWWP arm or the IBD arm the better target? A PWWP-selective probe and an IBD-selective probe would answer it, and only the IBD side has precedent chemistry.
  4. Do human haematopoietic stem cells require the LEDGF IBD specifically, as opposed to the whole protein? A knock-in IBD point mutant rather than a knockout is the experiment, and it has not been done.
  5. Because menin-inhibitor resistance mutations preserve KMT2A binding, does the menin-KMT2A-LEDGF clamp stay chromatin-bound in resistant cells, making LEDGF a rational second-line rather than merely a parallel target?
Every source behind this entry (6)
  1. E1strong 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 with 20.5% of lines depleted, yet falls to -0.444 across 10 KMT2A-rearranged lines and -0.346 across 4 KMT2A::AFF1 lines; gnomAD gives pLI ~1.0, LOEUF 0.373, 19 observed versus 74.7 expected loss-of-function variants; HPA classifies it as low tissue specificity, detected in all tissues.

    What this evidence cannot show

    Means were computed by us from per-line API values, not 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 to the selective score, so the confidence interval is wide and has not been bootstrapped. Whole-gene knockout removes the PWWP domain and the IBD simultaneously and therefore overstates what an interface-specific inhibitor would do. A positive mean Chronos across immortalised cancer lines says nothing about human haematopoietic stem cells under stress.

    Evidence that argues against this (1)
    • gnomAD constraint points the other way: pLI ~1.0 and LOEUF 0.373 mean human heterozygous loss of PSIP1 is selected against, so a tissue-wide inhibitor is not obviously benign. The same tension exists for MEN1 and is discussed there. gnomAD Consortium 2026
  2. E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    Verified experimental structures exist for both LEDGF modules: the PWWP domain on H3K36-methylated mono- and dinucleosomes (6S01 3.2 A, 8PC5 3.02 A, 8PC6 3.04 A, 8PEO 2.69 A, 8PEP 3.33 A, 8CBN 3.34 A, 8CBQ 4.0 A), the IBD with a KMT2A 140-160 peptide (2MSR, NMR), the menin-KMT2A-LEDGF ternary complex (3U88, 3.0 A, modelling PSIP1 347-435 with KMT2A 103-153 and menin 2-550), the IBD with a POGZ peptide (2N3A) and the IBD with lentiviral integrases (2B4J 2.02 A, 3F9K 3.2 A, 3HPG 3.28 A).

    What this evidence cannot show

    All identifiers, resolutions and residue ranges here were read from the RCSB API in this session rather than recalled, but the residue ranges are the crystallographically or NMR-modelled segments as RCSB annotates them against UniProt, not independently curated domain boundaries. 2MSR is an NMR structure of an isolated peptide with an isolated domain and carries no affinity measurement. 3U88 at 3.0 A does not resolve side-chain detail reliably enough to enumerate hot-spot residues.

    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. The PWWP coverage is on reconstituted nucleosomes with purified domains, and H3K36 methylation marks most transcribed gene bodies, so the structures explain binding and not selection of the loci the fusion occupies. Cancer Research 2014
  3. E3strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    ChEMBL v34 contains essentially no chemical matter against PSIP1: the human single-protein target CHEMBL3988590 has 3 activity rows at pChEMBL >= 6 (best IC50 470 nM, CHEMBL3981070), all from one document and all assaying the HIV-1 integrase-LEDGF protein-protein interaction, with no curated mechanism of action; by contrast the analogous reader domain protein BRPF1 (CHEMBL3132741) has 173 potent rows with a best Kd of 31 nM.

    What this evidence cannot show

    Absence from ChEMBL is not absence from the world: the LEDGIN antiviral series is larger than three rows in the primary literature and ChEMBL curation lags. Conversely, activity against the integrase-LEDGF interface does not imply activity against the KMT2A-LEDGF interface even though both use the IBD, because the two ligands occupy overlapping but non-identical surfaces.

    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. Three rows from one document, all assaying the HIV-1 integrase-LEDGF interaction, means the chemistry that exists was built for a different interface on the same protein. ChEMBL, EMBL-EBI 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    PSIP1 is 530 amino acids (UniProt O75475) and an AlphaFold model exists (AF-O75475-F1) but has a global pLDDT of 62.6 with 47.4% of residues in the very-low confidence band, consistent with long disordered regions between two folded modules; by contrast no AlphaFold model exists for full-length KMT2A (AF-Q03164-F1 returns HTTP 404).

    UniProtKB entry O75475 (PSIP1_HUMAN) and AlphaFold Protein Structure Database entry AF-O75475-F1, both queried directly by API · UniProt Consortium; AlphaFold Protein Structure Database (EMBL-EBI / Google DeepMind) · Databases 2026
    What this evidence cannot show

    The AlphaFold entry is a monomer prediction and says nothing about the IBD-bound conformation; use 2MSR or 3U88 for the complex. The pLDDT distribution was read from the AlphaFold API summary and not recomputed.

    Evidence that argues against this (1)
    • A global pLDDT of 62.6 with 47.4% very-low-confidence residues is a statement about DISORDER, not about how well the protein is characterised. Predictor confidence tracks foldedness, and a chromatin adaptor is expected to be largely disordered between its two ordered modules. UniProt Consortium 2026
  5. E5strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Menin is the adaptor that physically links MLL/KMT2A to LEDGF, and LEDGF is required both for MLL-dependent transcription and for leukemic transformation; MEN1 patient mutations that abolish LEDGF binding while sparing MLL binding still compromise function.

    PMID 18598942DOI
    What this evidence cannot show

    This establishes requirement by knockdown and by mutation, not by pharmacology, and the transformation assays are mouse haematopoietic progenitor colony and transplant models rather than human infant leukemia. The abstract reports no affinity measurement, so no Kd should be attributed to this work.

    Evidence that argues against this (1)
    • The separation-of-function MEN1 alleles are patient mutations from endocrine tumours, so their behaviour in a leukaemia assay crosses a disease boundary. LEDGF is also broadly required as a chromatin adaptor and as the HIV integration cofactor, so 'required for transformation' is compatible with 'required for a great deal else'. Nature 2012
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    NMR defines a second LEDGF/p75 IBD interface on MLL/KMT2A in which MLL Phe148 and Phe151 are the key residues; this interface is mutually exclusive with HIV-1 integrase, POGZ and JPO2 binding to the same IBD; LEDGF mutants defective for MLL binding impair colony formation by MLL-AF9 cells, and a cyclic peptide (CP65) selectively impairs clonogenic growth of primary murine MLL-AF9 blasts.

    PMID 25082813DOI
    What this evidence cannot show

    The structural work is NMR on an isolated IBD with an isolated MLL peptide; no calorimetric affinity is reported in the abstract, so the quantitative comparison with the 53 nM menin-MBM1 interaction cannot be made from this source. CP65 is a cyclic peptide tool compound, not a drug-like molecule, and the cellular experiments are murine MLL-AF9 rather than human KMT2A::AFF1. The mutual exclusivity with integrase, POGZ and JPO2 is exactly the selectivity problem noted in bite_back.

    Evidence that argues against this (1)
    • Mutual exclusivity with HIV-1 integrase, POGZ and JPO2 on one IBD surface is the tractability argument and the selectivity problem in one sentence: a ligand for that surface would displace all four. Cancer Research 2014
Connected to