Oncofusion

start / graph / gene-suz12

Genemoderate confidence 3 of 5: One strong study, or several smaller consistent ones.Q41-prc2-arm

SUZ12 - the scaffold that holds PRC2 together and decides which PRC2 it is, required in this disease and studied in it exactly once

also called SUZ12 · JJAZ1 · CHET9 · KIAA0160 · suppressor of zeste 12 protein homolog

SUZ12 is the backbone that the Polycomb machine is built on: without it the machine falls apart. It is also the piece that decides which of two different versions of the machine gets assembled, and therefore where in the genome it goes.

The same thing, for a scientist

SUZ12 (UniProt Q15022, 739 aa) is the scaffolding subunit of PRC2. Its VEFS-Box binds EZH2 and is required for catalytic activity, and its N-terminal region is the platform that binds the mutually exclusive accessory modules that define **PRC2.1** (the PCL proteins MTF2/PHF19/PHF1, plus EPOP or PALI1) and **PRC2.2** (JARID2 and AEBP2) [E1][E5]. So SUZ12 is not merely structural: it is the subunit through which PRC2's targeting is specified. In KMT2A-rearranged leukaemia the direct evidence is thin but consistent: shRNA suppression of *Suz12*, like *Eed* or *Ezh1/Ezh2*, arrests proliferation and forces differentiation in MLL-AF9;Nras-G12D AML with minimal impact on non-transformed haematopoietic lines [E2], and SUZ12 is described as essential for HSC maintenance [E3]. In this project's DepMap panel SUZ12 is **required but not selective**: index mean **-0.482** against **-0.452** in comparator leukaemia lines (difference -0.030, rank 5,332 of 17,728, pan-essentiality -0.405) [E4]. **The literature gap is stark.** `SUZ12 AND (KMT2A OR MLL) AND leukemia` returns **5** records and title-checking leaves **one** on topic - the 2013 MLL-AF9 shRNA screen already cited here [E6]. The other four are an HTLV-1 ATL sequencing study, a murine ESC aryl-hydrocarbon-receptor paper, a prostate paper and a review of epithelial translocations.

Go deeper
PRC2.1 versus PRC2.2, and why this graph cannot yet write those entries

SUZ12 is the subunit that makes the PRC2.1/PRC2.2 distinction a real thing rather than a naming convention: the two accessory modules bind the same SUZ12 surface and are mutually exclusive 15. PRC2.1 (PCL proteins MTF2/PHF19/PHF1 with EPOP or PALI1) is associated with CpG-island targeting; PRC2.2 (JARID2/AEBP2) reads H2AK119ub1 and couples to PRC1. The 2021 cryo-EM structure of PRC2:EZH1-AEBP2-JARID2 (PDB 7KSO) is a PRC2.2 assembly 5.

The brief that dispatched this beat asked whether the evidence distinguishes PRC2.1 from PRC2.2. It does - but only outside leukaemia, and this beat nearly recorded a false absence finding out.

The first query run was keyed to subunit symbols and the word leukaemia:

queryrecordson topic after title check
(JARID2 OR MTF2 OR PHF19 OR AEBP2 OR EPOP) AND (KMT2A OR MLL) AND leukemia10 - a paediatric AML SNP-array landscape (PMID 27022003) with an incidental mention
"PRC2.1" OR "PRC2.2"5none in a KMT2A context

Dropping the word leukemia changes the answer. (JARID2 OR MTF2 OR PHF19 OR AEBP2 OR EPOP OR "PRC2.1" OR "PRC2.2") AND (KMT2A OR MLL OR menin) returns 4, and the first is PMID 36635503, Targeting Menin disrupts the KMT2A/B and polycomb balance to paradoxically activate bivalent genes 8. A whole-genome CRISPR screen in that paper identifies MTF2-PRC2.1, PCGF1-PRC1.1 and Menin-KMT2A/B as the complexes maintaining bivalency, and finds that menin loss or inhibition phenocopies polycomb disruption. The subtype implicated is PRC2.1, not PRC2.2, and it is implicated in the same axis this graph cares most about.

Two lessons, both already in this project's rulebook and both nearly repeated here. The paper does not say leukaemia in its abstract - it says cancer cells and pluripotent stem cells - so a query that conjoins a disease word to a subunit symbol misses it; that is the same shape as the FACT and Mediator absences. And this graph already cited PMID 36635503 in two entries (gene-pcgf1, protein-prc1-1-complex) before this beat existed, so the repo grep prescribed by the subtract-before-you-report rule was what caught it.

What remains true: no study distinguishes PRC2.1 from PRC2.2 in a KMT2A-REARRANGED leukaemia. DepMap does not separate them either - MTF2 -0.072, PHF19 -0.027, JARID2 -0.022, AEBP2 +0.014, EPOP +0.050, all mid-genome 4, and single-gene knockouts cannot see redundancy among the three PCL paralogues. So no PRC2.1 or PRC2.2 sub-entry is written, but protein-prc2-complex now records MTF2-PRC2.1 as the subtype with the only supporting evidence, and mechanism-menin-inhibition-shifts-the-kmt2a-polycomb-balance carries the finding.

Accession resolution

gene_exact:SUZ12 AND organism_id:9606 AND reviewed:true returns exactly one reviewed human record, Q15022, primary gene name SUZ12, synonyms CHET9, JJAZ1, KIAA0160, 739 aa, protein name Polycomb protein SUZ12 7. Resolved 2026-09-20 with gene_primary and length in fields. No paralogue, so no Rule 63/65 hazard - but JJAZ1 is the symbol under which SUZ12 appears in the endometrial-stromal-sarcoma fusion literature (JAZF1-SUZ12), and a literature count keyed only to 'SUZ12' will miss that body of work entirely.

Facts and figures
symbolSUZ12
uniprotQ15022
length AA739
uniprot primary gene nameSUZ12
uniprot resolution notegene_exact:SUZ12 AND organism_id:9606 AND reviewed:true returned exactly one row; primary gene name SUZ12 matched; synonyms CHET9, JJAZ1, KIAA0160; length 739 asserted. Resolved 2026-09-20.
paraloguenone
complexprotein-prc2-complex
functionscaffold; VEFS-Box binds EZH2 and is required for catalysis; the N-terminal region is the mutually exclusive binding platform for the PRC2.1 (PCL/EPOP/PALI1) and PRC2.2 (JARID2/AEBP2) accessory modules
expression hpa nTPM median kmt2a r panel57.4
essentiality
{
  "depmap_kmt2a_r_mean": -0.4818,
  "depmap_other_leukaemia_mean": -0.4515,
  "depmap_difference": -0.0303,
  "depmap_rank_of_17728": 5332,
  "depmap_pan_essentiality_all_lines": -0.4053,
  "depmap_index_myeloid_mean": -0.5255,
  "depmap_index_lymphoid_mean": -0.4054,
  "source": "analyses/prc2-dependency-lineage-split/results.json",
  "interpretation": "required in both arms, not selective"
}
prc2 subtype accessory depmap
{
  "MTF2": -0.072,
  "PHF19": -0.0269,
  "JARID2": -0.0221,
  "AEBP2": 0.0142,
  "EPOP": 0.0503,
  "note": "all mid-genome; single-gene knockouts cannot see within-module paralogue redundancy (MTF2/PHF19/PHF1). NOTE: MTF2-PRC2.1 is implicated in bivalency maintenance alongside Menin-KMT2A/B by PMID 36635503, outside leukaemia."
}
literature absence
{
  "query": "SUZ12 AND (KMT2A OR MLL) AND leukemia",
  "records": 5,
  "on_topic_after_title_check": 1,
  "aliases_not_searched": [
    "JJAZ1",
    "CHET9",
    "KIAA0160"
  ],
  "date": "2026-09-20",
  "correction": "The original absence was keyed to subunit symbols AND the word 'leukemia' and missed PMID 36635503, which names MTF2 and Menin-KMT2A/B but not leukaemia. Dropping the disease word recovers it. The surviving absence is narrower: no study distinguishes PRC2.1 from PRC2.2 in a KMT2A-REARRANGED leukaemia."
}
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.4818,
  "chronos_mean_other_leukaemia": -0.4515,
  "difference": -0.0303,
  "pan_essentiality_all_lines": -0.4053,
  "selectivity_rank": 5332,
  "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."
}
How this could bite us back
  • SUZ12 is described as essential for haematopoietic stem cell maintenance and is required at similar depth in every leukaemia line tested, so it is a complex-integrity target with no selectivity - removing it removes PRC2 everywhere, including in the stem-cell compartment that has to reconstitute the marrow.
  • Because SUZ12 is the platform for both accessory modules, an agent that disrupted the scaffold would remove PRC2.1 and PRC2.2 together and could not be titrated toward whichever subtype the leukaemia actually uses - a specificity that an EED- or EZH-directed agent also lacks but that a PCL- or JARID2-directed one might have.
Still open
  1. Which PRC2 subtype operates at fusion target loci in KMT2A-rearranged leukaemia - PRC2.1 or PRC2.2? Nothing in the literature or in DepMap distinguishes them here, and the distinction matters because PRC2.2 couples to PRC1 through H2AK119ub1 and this graph already holds a PRC1 arm (protein-prc1-1-complex, gene-cbx8).
  2. Does the JAZF1-SUZ12 fusion literature contain anything transferable? SUZ12 is itself a fusion partner in endometrial stromal sarcoma under the alias JJAZ1, which is a fusion-oncoprotein system this graph's comparative arm has never looked at.
  3. Is there a SUZ12 separation-of-function allele that keeps the VEFS-Box catalytic contribution and removes accessory-module binding? That is the only clean way to ask whether PRC2 targeting rather than PRC2 catalysis is what this disease needs.
Every source behind this entry (7)
  1. E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    PRC2 consists of three core subunits - EZH2, EED and SUZ12 - with the catalytic subunit EZH2 methylating H3K27 and its activity further enhanced by EED binding to H3K27me3.

    An allosteric PRC2 inhibitor targeting the H3K27me3 binding pocket of EED. · Qi W, Zhao K, Gu J, Huang Y, Wang Y, Zhang H, et al. · Nat Chem Biol 2017
    PMID 28135235DOI
    What this evidence cannot show

    Abstract read from the eutils record; full text not read. Cited for the core-subunit composition only. The abstract names three core subunits and omits RBBP4/RBBP7, which are also core by most definitions - so 'three' is the minimal catalytic core rather than the complete complex, and this entry does not use the number.

    Evidence that argues against this (1)
    • Describing PRC2 as three subunits understates the targeting problem this entry is about: the accessory modules that bind SUZ12 are what determine where PRC2 acts, and a cryo-EM structure of PRC2:EZH1-AEBP2-JARID2 exists as a distinct assembly. A model with three subunits cannot represent the PRC2.1/PRC2.2 distinction at all. Grau D et al. 2021
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    shRNA-mediated suppression of the PRC2 component Suz12 caused proliferation arrest and differentiation of MLL-AF9;Nras-G12D acute myeloid leukaemia cells with minimal impact on the growth of several non-transformed haematopoietic cell lines, partly through direct PRC2 repression of self-renewal-limiting genes including Cdkn2a.

    The Polycomb complex PRC2 supports aberrant self-renewal in a mouse model of MLL-AF9;Nras(G12D) acute myeloid leukemia. · Shi J, Wang E, Zuber J, Rappaport A, Taylor M, Johns C, et al. · Oncogene 2013
    PMID 22469984DOI
    What this evidence cannot show

    Abstract read from the eutils record; full text not read. This is the ONLY on-topic record in the SUZ12 literature for this disease, so every SUZ12-specific statement in this entry rests on one 2013 mouse study. shRNA knockdown, not knockout. Mouse MLL-AF9 with a co-operating Nras-G12D, myeloid, no B-lineage or infant arm. The therapeutic-window claim is against immortalised non-transformed lines, not primary HSPCs.

    Evidence that argues against this (1)
    • The same result set includes Eed and Ezh1/Ezh2 arms, so nothing in it isolates SUZ12 - the finding is that PRC2 is required and SUZ12 is one of three ways the paper removed it. A SUZ12-specific claim cannot be made from a screen whose conclusion is about the complex, and this entry treats SUZ12 as part of that conclusion rather than as an independently validated node. Neff T et al. 2012
  3. E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.review

    PRC2 components, particularly EZH1/2, SUZ12 and EED, are essential for the maintenance of haematopoietic stem cells; variant PRC1 and PRC2 subtypes use distinct recruitment mechanisms enabling both PRC2-dependent and PRC2-independent silencing; and therapeutic targeting of H3K27me3 with EZH2-selective inhibitors such as tazemetostat is limited by functional redundancy with EZH1, which the dual EZH1/2 inhibitor valemetostat overcomes, with resistance emerging through PRC2 gatekeeper mutations and compensatory DNA methylation.

    Polycomb repressive complexes as therapeutic targets in hematologic malignancies. · Yamagishi M, Iwama A, Kitabayashi I · Exp Hematol 2026
    PMID 41443386DOI
    What this evidence cannot show

    A 2026 review, cited as a map and for the field's own statement of the EZH1-redundancy problem and the resistance routes - not for any primary result. Reviews compress; the HSC-essentiality statement in particular aggregates several mouse studies not individually checked in this beat. The review does not discuss KMT2A-rearranged leukaemia specifically.

    Evidence that argues against this (1)
    • A review's summary of a therapeutic problem is not evidence that the problem applies in a given disease. The EZH1-redundancy claim is independently supported here by mouse KMT2A genetics (PMID 22396593) and by a head-to-head in B-ALL lines (PMID 36280659); the HSC-essentiality claim is NOT independently checked in this beat and should be treated as a pointer rather than a finding. Neff T et al. 2012
  4. E4moderate confidence 3 of 5: One strong study, or several smaller consistent ones.dataset

    SUZ12 gene effect in 11 KMT2A-rearranged DepMap leukaemia lines is -0.482 against -0.452 in 43 other leukaemia lines (difference -0.030, rank 5,332 of 17,728, pan-essentiality -0.405); the PRC2 accessory subunits are all mid-genome and none is a dependency - MTF2 -0.072, PHF19 -0.027, JARID2 -0.022, AEBP2 +0.014, EPOP +0.050.

    What this evidence cannot show

    SUZ12 is required in both arms and the difference is negligible, so this is a dependency and explicitly NOT a selectivity claim. 11 index lines, none infant. The accessory-subunit nulls are single-gene knockouts and PRC2.1 and PRC2.2 members are partially redundant within each module (MTF2/PHF19/PHF1 are three PCL paralogues), so the nulls are uninformative about the modules as units. Raw effects, no z, no permutation null. DepMap release version not stated by the Open Targets API.

    Evidence that argues against this (1)
    • Within-module paralogue redundancy is the same trap as EZH1/EZH2, one level down: MTF2, PHF19 and PHF1 are three PCL paralogues binding the same SUZ12 surface, so a single-gene screen cannot see a PRC2.1 requirement even if one exists. The accessory-subunit nulls should not be read as evidence that PRC2.1 and PRC2.2 do not matter here. Neff T et al. 2012
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    A cryo-EM structure of PRC2:EZH1 in complex with AEBP2 and JARID2 has been determined (RCSB entry 7KSO, titled 'Cryo-EM structure of PRC2:EZH1-AEBP2-JARID2'), alongside a dimeric PRC2:EZH1 bound to a nucleosome (7KSR), establishing the PRC2.2 accessory module as a structurally defined assembly.

    What this evidence cannot show

    Entry titles were fetched and read back on 2026-09-20; polymer entities were NOT fetched, so no chain-to-UniProt mapping, construct composition or residue span is asserted. Per the QUEEN_BRIEF structural rules, several deposits central to this field are single-chain engineered chimeras and residue indices read straight off a deposit are often wrong - neither was checked here. RCSB returned no primary-citation PubMed identifier for either entry, so the association with PMID 33514705 is by subject and not by a stored identifier. A structure existing is not a structure suitable for design.

    Evidence that argues against this (1)
    • Both structures are of PRC2:EZH1, not PRC2:EZH2, and this graph's own position is that the two complexes are not interchangeable - PRC2:EZH1 dimerises and compacts chromatin through a loop that diverges between the paralogues. A PRC2.2 architecture solved on EZH1 should not be assumed to describe the EZH2 complex. Grau D et al. 2021
  6. E6strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database

    PubMed `SUZ12 AND (KMT2A OR MLL) AND leukemia` returns 5 records, of which title-checking leaves one on topic (PMID 22469984, the MLL-AF9;Nras shRNA screen); the other four are an HTLV-1 adult T-cell leukaemia sequencing study (26880370), a murine embryonic stem cell aryl-hydrocarbon-receptor paper (24316986), a prostate chromatin-regulator paper (23038103) and a review of translocations in epithelial cancers (19406209).

    What this evidence cannot show

    The alias JJAZ1 - under which SUZ12 appears in the JAZF1-SUZ12 endometrial stromal sarcoma literature - was NOT included, nor were CHET9 or KIAA0160. Per Rule 64 the complex form was run and is larger ('Polycomb repressive complex 2' AND (KMT2A OR MLL) AND leukemia returns 47, 'PRC2' AND (KMT2A OR MLL-rearranged) AND leukemia returns 24) and recovers the PRC2 leukaemia genetics but no additional SUZ12-specific study. This is a bounded absence under these symbols.

    Evidence that argues against this (1)
    • The complex form reaching papers no subunit symbol can is the documented failure mode of exactly this kind of count, and in this beat it reached 47 records against the subunit's 5. Nothing SUZ12-specific was recovered, but the ratio is a reminder that the subunit number is not the size of the relevant literature - it is the size of the literature that names the subunit. NCBI PubMed 2026
  7. E7established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    SUZ12 resolves to a single reviewed human UniProt record, Q15022 (SUZ12_HUMAN), primary gene name SUZ12, synonyms CHET9, JJAZ1 and KIAA0160, 739 amino acids, protein name Polycomb protein SUZ12.

    What this evidence cannot show

    Resolved 2026-09-20. SUZ12 has no paralogue so the Rule 63/65 family hazards do not apply, but JJAZ1 is a live alias in the sarcoma fusion literature and a symbol-keyed retrieval will miss it.

    Evidence that argues against this (1)
    • This graph has measured a roughly tenfold under-count from searching a current symbol when the literature uses a retired or alternative one (SEPT9 20 records against SEPTIN9 1). The SUZ12 absence in this entry is therefore bounded by three unsearched aliases and should be reopened by anyone who needs it to be tight. NCBI PubMed 2026
Connected to