Taspase1 and its two KMT2A cleavage sites: an enzyme-substrate interface with four apo structures, one co-crystal with a small molecule, and no structure with MLL at all
also called taspase1-MLL · CS1 CS2 · D2666 D2718 · TASP1
A protease called taspase1 cuts normal KMT2A in two places, and that cut controls how long the protein lasts. The fusion protein is truncated long before either cut site, so taspase1 never touches it - and no structure of the enzyme holding its KMT2A substrate has ever been solved.
The same thing, for a scientist
Taspase1 (TASP1, Q9H6P5) is a threonine aspartase that cleaves MLL at two conserved D/GADD and D/GVDD sites, generating the MLL-N320 and MLL-C180 fragments that reassociate into a stable complex. UniProt Q03164 places the cleavages after residues 2666 and 2718. Ten PDB entries contain Q9H6P5 - the 2.0 A apo enzyme 2A8I and companions 2A8J/2A8L/2A8M, the 3.04 A circularly permuted construct 6VIN, and a series of co-crystals with synthetic ligands - and an RCSB search for Q9H6P5 together with Q03164 returns an empty result set. The enzyme-substrate interface is therefore defined entirely by the cleavage-site sequence and by mutagenesis, not by any structure. Both sites sit roughly 1,300 residues C-terminal to the KMT2A breakpoint cluster region, so no der(11) KMT2A fusion protein is a taspase1 substrate.
What is structurally known, and what is not
Fetched from RCSB on 2026-09-20 34:
| entry | what it is | method | resolution |
|---|---|---|---|
| 2A8I | human taspase1, apo | X-ray | 2.0 A |
| 6VIN | circularly permuted human taspase1 | X-ray | 3.04 A |
| 9D02 | circularly permuted taspase1 with a piperazine ligand | X-ray | 2.15 A |
The last row is worth noticing on its own: taspase1 has small-molecule co-crystals. Whatever else is true of this interface, the enzyme is not chemically untouched.
What does not exist: any deposit containing taspase1 and MLL together, and therefore any experimental picture of how the D/GADD and D/GVDD sites are recognised.
An interface that the fusion is defined by not having
The graph's subtraction-kmt2a-taspase1-cleavage-sites records the loss and its consequence, and that entry should be read first for the functional argument. What this entry adds is the interface-layer statement: this is a substrate-recognition interface with measured sites, a catalogued enzyme, tractable chemistry, and zero structural information about the recognition event itself.
It is also the cleanest case in this beat of an interface where 'lost in the fusion' is not a loss of function but a gain of stability: the uncleaved fusion escapes a cleavage-coupled degradation route that the wild-type protein cannot.
| partner a | gene-tasp1 |
|---|---|
| partner b | gene-kmt2a |
| motif b | KMT2A taspase1 cleavage site 1 (D/GADD) after residue 2666 and site 2 (D/GVDD) after residue 2718, per UniProt Q03164 |
| motif a | taspase1 type-2 asparaginase fold active site; substrate register unresolved |
| kd status | none published; this is an enzyme-substrate interface characterised by cleavage-site mutagenesis rather than by equilibrium binding |
| fate in fusion | lost - both cleavage sites lie roughly 1,300 residues C-terminal to the breakpoint cluster region, so no der(11) KMT2A fusion is a taspase1 substrate |
| consequence of loss | the fusion escapes a cleavage-coupled route rather than losing a function; this is a loss that plausibly benefits the oncoprotein |
| druggability | the enzyme has apo structures, a catalytically active permuted construct and at least one small-molecule co-crystal, so it is chemically tractable; what is missing is the substrate-recognition geometry that would make a KMT2A-selective inhibitor conceivable |
| resolution note | PDB metadata fetched from data.rcsb.org 2026-09-20; cleavage-site residues taken from the graph's existing UniProt Q03164 annotation rather than re-derived, and flagged as such |
interface residues a
[]
interface residues b
[ "KMT2A D2666", "KMT2A D2718" ]
kd measurements
[]
structures
[ "2A8I", "2A8J", "2A8L", "2A8M", "6VIN", "9D02", "9D03", "9D04", "9E6U", "6UGK" ]
structure details — the numbers behind the chart
[
{
"pdb": "2A8I",
"title_fetched": "Crystal Structure of human Taspase1",
"method": "X-ray",
"resolution_A": 2,
"contents": "one 420-residue TASP1 entity, no ligand",
"fetched": "2026-09-20"
},
{
"pdb": "6VIN",
"title_fetched": "Crystallographic structure of the circularly permuted human Taspase1 protein",
"method": "X-ray",
"resolution_A": 3.04,
"fetched": "2026-09-20"
},
{
"pdb": "9D02",
"title_fetched": "Co-crystal structure of circularly permuted human taspase-1 bound to ligand SMDC1014883",
"method": "X-ray",
"resolution_A": 2.15,
"fetched": "2026-09-20"
}
]structures absent
[ "no PDB entry contains taspase1 together with KMT2A", "no structure of any taspase1 substrate peptide in the active site" ]
- Wrong direction: inhibiting taspase1 stabilises uncleaved wild-type MLL1, which in a KMT2A-rearranged cell may compete with the fusion or may simply add another player; the sign of the effect is not obvious from the interface.
- Substrate breadth: taspase1 cleaves substrates other than MLL, so an inhibitor is not an MLL-axis agent.
- No selectivity handle: without a substrate-bound structure there is no basis for designing an inhibitor that blocks MLL cleavage and spares other substrates.
- Target absent from the fusion: the fusion protein contains neither site, so this interface cannot be a fusion-directed target under any modality.
- How does taspase1 recognise the D/GADD and D/GVDD register? No substrate-bound structure exists, so no selective inhibitor can be designed against the MLL sites specifically.
- Do the published taspase1 small-molecule co-crystals represent active-site binders or allosteric ones, and does either inhibit MLL cleavage in cells?
- Since the fusion is not a substrate, does taspase1 inhibition in a KMT2A-rearranged cell act purely through the wild-type allele - and is that the mechanism of the published pharmacological result?
- E1established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Taspase1 is the threonine aspartase that cleaves MLL and is required for proper HOX gene expression.
Taspase1: a threonine aspartase required for cleavage of MLL and proper HOX gene expression. · Hsieh JJ, Cheng EH, Korsmeyer SJ · Cell 2003What this evidence cannot show
The founding identification; the HOX phenotype is from taspase1-deficient cells and mice, in which MLL is only one of several substrates, so the transcriptional phenotype is not attributable to MLL cleavage alone.
Evidence that argues against this (1)
- Taspase1 has substrates besides MLL, so a taspase1 loss-of-function phenotype cannot be read as an MLL-cleavage phenotype without a non-cleavable MLL allele as the control. Hsieh JJ et al. 2003
- E2established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
Proteolytic cleavage of MLL generates N- and C-terminal fragments that reassociate into a complex conferring protein stability and subnuclear localisation.
Proteolytic cleavage of MLL generates a complex of N- and C-terminal fragments that confers protein stability and subnuclear localization. · Hsieh JJ, Ernst P, Erdjument-Bromage H, et al. · Mol Cell Biol 2003What this evidence cannot show
Biochemical fractionation and imaging in cell lines; the stability claim is from the reassociated complex rather than from a kinetic measurement of turnover.
Evidence that argues against this (1)
- Later work reframed cleavage as priming MLL1 for degradation rather than simply stabilising it, so the direction of the effect on protein lifetime is not as originally described - the graph's subtraction-kmt2a-taspase1-cleavage-sites carries that correction. Hsieh JJ et al. 2003
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB 2A8I is a 2.0 A crystal structure of human taspase1 containing a single polymer entity and no MLL peptide.
PDB 2A8I: Crystal Structure of human Taspase1 · Khan JA, Dunn BM, Tong L · RCSB Protein Data Bank 2005What this evidence cannot show
Title, method, resolution and the single 420-residue entity fetched from data.rcsb.org 2026-09-20. An apo enzyme structure constrains the active site geometry but says nothing about substrate register.
Evidence that argues against this (1)
- A later study solved a circularly permuted, catalytically active construct and argued that the apo structure does not represent the active enzyme, so the 2005 structure's relevance to substrate recognition is itself contested. Nagaratnam N et al. 2021
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.database
A circularly permuted, catalytically active human taspase1 has been crystallised (6VIN, 3.04 A), and ligand co-crystals of that construct exist; no PDB entry contains taspase1 together with KMT2A.
PDB 6VIN: Crystallographic structure of the circularly permuted human Taspase1 protein · Nagaratnam N, Delker SL, Jernigan R, et al. · RCSB Protein Data Bank 2021What this evidence cannot show
Entry metadata fetched from data.rcsb.org 2026-09-20; the absence established by RCSB search API, exact_match on reference_sequence_identifiers.database_accession for Q9H6P5 paired with Q03164, which returned an empty result set. Bounded by UniProt cross-annotation. Neither Khan et al. 2005 nor Nagaratnam et al. 2021 was retrievable in full text this cycle - both lack a PMCID.
Evidence that argues against this (1)
- The original apo structure and the circularly permuted construct disagree about which arrangement is the active enzyme, so the structural literature on taspase1 is internally unsettled even before substrate recognition is considered. Khan JA et al. 2005
enzyme-substrate interface, structurally unsolved
cleavage sites after D2666 and D2718
the interface layer of that subtraction entry; adds the structural absence and the enzyme's chemical tractability
the phospho-licensing step for the same cleavage
- related to →KMT2A::AFF1 (MLL-AF4), t(4;11)(q21;q23)
the index fusion contains neither site and is not a substrate
the epitope lies within the taspase1 N320 product, which the fusion is never cut to generate
zero taspase1-KMT2A entries
- ← depends onKMT2A / MLL1 — the gene that breaks
the two-fragment state that carries the subnuclear address is produced by taspase1 cleavage