GenemoderateQ92-kmt2a-sarcoma

LDB1 (Q86U70) - and the sarcoma fusion keeps the part that self-associates while losing the LIM interaction domain entirely

also called LDB1 · CLIM2 · CLIM-2 · NLI · Nuclear LIM interactor · hLdb1 · Q86U70 · LDB1_HUMAN

LDB1 is a small adaptor protein that holds transcription factors together on DNA, most famously in blood cells. In the one sarcoma reported with an LDB1-KMT2A fusion, the fusion keeps LDB1's self-sticking front end and cuts off the very end that grips its usual partners.

The same thing, for a scientist

LDB1 (Q86U70, 411 aa; primary gene name and length read back from the fetched record) is the scaffolding adaptor of the LDB1/LMO2/TAL1/GATA1 complex and the mediator of long-range enhancer-promoter contacts through its N-terminal self-association region. UniProt annotates one domain: the LIM interaction domain (LID) at aa 336-375. It enters this graph as the 5' partner of the single reported LDB1::KMT2A sarcoma. Mapping the reported exon range onto the canonical 11-exon transcript ENST00000673968 shows something the case report does not state: LDB1 exons 1-10 encode aa 1-335, and the LID (336-375) is encoded entirely within exon 11. So the fusion retains LDB1's self-association and dimerisation region and DISCARDS the LIM interaction domain - it cannot recruit LMO2 or any other LIM-domain partner. What it can do is dimerise, which would make the chimera a self-associating, near-full-length KMT2A with an intact SET domain.

What it looks like

LDB1 (Q86U70) - and the sarcoma fusion keeps the part that self-associates while losing the LIM interaction domain entirely411 aa

No three-dimensional model is available for this protein, so the map above is the whole picture. That is itself informative: AlphaFold does not model chains this long, and the parts that have been crystallised are held under the entries for the individual domains.

Go deeper

What the fusion keeps and loses, computed


LDB1 canonical transcript	ENST00000673968, 11 exons, 411 aa 2
exons 1-10 (retained in the fusion)	aa 1-335 - the N-terminal self-association / dimerisation region 2
exon 11 (lost)	aa 336-412, which contains the whole LIM interaction domain (LID, 336-375) 12

The case report states the fusion as LDB1 exons 1-10 to KMT2A exons 4-36 3; the consequence for the LID is derived here and is not stated in that paper.

Why the lost domain is the interesting half

LDB1's best-characterised function in this project's neighbouring literature is as the adaptor that assembles LMO2 with TAL1 and GATA1 - and it does that through the LIM interaction domain. A fusion that keeps everything except the LID therefore cannot be a vehicle for LIM-domain recruitment. What survives is the self-association region, which is how LDB1 bridges distal enhancers to promoters.

So the plausible reading is that LDB1 contributes dimerisation, not partner recruitment - and dimerisation of a near-full-length, catalytically intact KMT2A is a different kind of molecule from anything in the leukaemia literature. That is a hypothesis from one case and a coordinate derivation, and it is recorded as such.

Facts and figures

uniprot	Q86U70
length AA	411
ensembl gene ID	ENSG00000198728
ensembl transcript ID	ENST00000673968
cytoband	10q24.32 (chr10, minus strand)
resolution note	gene_exact:LDB1 AND organism_id:9606 AND reviewed:true returned exactly one row; primary gene name (LDB1) and length (411) read back from the fetched record, 2026-09-21
partner class	LIM domain binding transcriptional adaptor; core component of the LDB1/LMO2/TAL1/GATA1 complex and a mediator of long-range enhancer-promoter contact through N-terminal self-association
shown for fusion	LDB1::KMT2A in one human spindle-cell sarcoma; nothing in KMT2A::AFF1, in any leukaemia, or in B-lineage
retained in the sarcoma fusion	aa 1-335 (self-association region)
lost in the sarcoma fusion	aa 336-411 including the entire LIM interaction domain - derived here, not stated in the case report

domains

[
  {
    "name": "LID (LIM interaction domain)",
    "start": 336,
    "end": 375
  }
]

exon AA ranges

{
  "1": [
    1,
    9
  ],
  "2": [
    9,
    43
  ],
  "3": [
    43,
    58
  ],
  "4": [
    58,
    83
  ],
  "5": [
    84,
    118
  ],
  "6": [
    118,
    175
  ],
  "7": [
    176,
    216
  ],
  "8": [
    217,
    244
  ],
  "9": [
    245,
    286
  ],
  "10": [
    286,
    335
  ],
  "11": [
    336,
    412
  ]
}

fusion partner of

[
  "gene-kmt2a (spindle-cell sarcoma, 1 published case)"
]

How this could bite us back

This entry's central claim - that the LID is lost - is a coordinate derivation from one abstract's exon range on an assumed transcript. If the case report numbered exons differently, the claim inverts, and nothing in the graph would flag it.
LDB1 is essential for erythropoiesis and for long-range enhancer contacts generally, so any intervention aimed at an LDB1 function would have a broad essential-function toxicity profile - a reason not to treat this partner as a target.

Still open

Does the LDB1::KMT2A chimera dimerise? The retained region is LDB1's self-association region and the answer decides whether this fusion is a dimerising KMT2A, which would be unlike anything in the leukaemia literature.
Has an LDB1 fusion with KMT2A ever been reported in leukaemia? LDB1 is a haematopoietic adaptor and its absence from the KMT2A recombinome would be notable, but no systematic check against that census was run for this entry.
Is the LID loss functionally important, or is exon 11 simply where the intron the break fell into happens to be?

Every source behind this entry (3)

E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Fetched from live UniProt (2026-09-21): gene_exact:LDB1 with organism_id:9606 and reviewed:true returns exactly one row, Q86U70 / LDB1_HUMAN, primary gene name LDB1, 411 aa, LIM domain-binding protein 1, with a single annotated domain - the LIM interaction domain (LID) at aa 336-375.
UniProtKB Q86U70 (LDB1_HUMAN), LIM domain-binding protein 1, feature table and sequence · UniProt Consortium · UniProtKB 2026
What this evidence cannot show
One reviewed human row, so this graph's usual several-rows hazard does not arise; the primary gene name and the length were nonetheless read back from the fetched record. UniProt annotates only the LID as a domain - the N-terminal self-association region that most of the functional literature is about carries no domain feature, so its boundaries are not asserted here.
Evidence that argues against this (1)
- LDB1's synonym CLIM2 and its description as 'Nuclear LIM interactor' mean a literature search on the current symbol alone will split the older literature by era, the same retired-symbol hazard this project has measured elsewhere. No alias-expanded search was run for this entry, so its literature coverage is not bounded. Ensembl (EMBL-EBI) 2026
E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
Computed here from live Ensembl REST (2026-09-21): LDB1 canonical transcript ENST00000673968 has 11 CDS exons encoding 411 aa, with exons 1-10 encoding aa 1-335 and exon 11 encoding aa 336-412 - so the LIM interaction domain at aa 336-375 lies entirely within exon 11 and is absent from an LDB1(ex1-10) fusion.
Ensembl REST API lookup, LDB1 (ENSG00000198728), canonical transcript ENST00000673968 with exon and translation coordinates · Ensembl (EMBL-EBI) · Ensembl REST 2026
What this evidence cannot show
Codon ranges computed by cumulative CDS length on the minus-strand gene, with boundary codons appearing in both flanking exons. The case report does not name a transcript accession, so the mapping of its 'exons 1-10' to this transcript is an inference; LDB1 has other transcripts and a different one would renumber the exons.
Evidence that argues against this (1)
- If the case report numbered LDB1 exons on a transcript with a different 5' exon structure, the boundary could shift and part of the LID could be retained. The qualitative point - that the LID sits in the last exon and a fusion stopping at exon 10 loses it - is robust to small shifts but was derived from one transcript only. UniProt Consortium 2026
E3moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
In the single reported LDB1::KMT2A sarcoma, next-generation sequencing identified the fusion as LDB1 exons 1-10 joined to KMT2A exons 4-36, verified by FISH and RT-PCR, in a 19-year-old man's spindle-cell sarcoma with a concurrent pathogenic BCOR frameshift.
LDB1::KMT2A Fusion in a Spindle-Cell Sarcoma: A Case Report. · Wang Z, Zhang Y, Zhu M, et al. · Genes Chromosomes Cancer 2025
PMID 40944346 DOI
What this evidence cannot show
One case, abstract-level reading, with a second pathogenic driver present and only 5 months of follow-up. The abstract does not discuss the LID or any LDB1 domain, so all domain reasoning here is this graph's derivation and not the authors'.
Evidence that argues against this (1)
- With a pathogenic BCOR frameshift in the same tumour, the LDB1::KMT2A fusion may be incidental, in which case its domain content is a curiosity rather than a mechanism. A second case would change that and none exists. Nakamura H et al. 2023

Connected to

part of →LDB1::KMT2A - the clearest statement in this literature that a sarcoma fusion keeps KMT2A exons 4 to 36 and discards exons 1 to 3
supplies LDB1 aa 1-335 (self-association region); the LIM interaction domain in exon 11 is lost
fuses with →KMT2A / MLL1 — the gene that breaks
LDB1 exons 1-10 to KMT2A exons 4-36, one published sarcoma case