PHD3 and the CyP33 RRM: Kd 7-15 uM, and a 2023 ternary structure that settles the 2010 disagreement - H3K4me3 and the RRM CAN be held at once, and the real switch is RNA
also called PHD3-CyP33 · PPIE RRM · 2KU7 · mutually inhibitory
A prolyl isomerase called CyP33 grabs the same small finger of KMT2A that reads the histone mark, and for thirteen years two papers disagreed about whether both can be held at once. A 2023 structure settles it: they can, and what actually throws the switch is RNA binding to CyP33.
The same thing, for a scientist
Park et al. measured the MLL PHD3-CYP33 RRM interaction at Kd 14.7 uM by ITC, identified V1617 as the RRM hot spot, mapped the H3K4me3 and RRM ligands to surfaces 180 degrees apart, and concluded the two bindings were mutually inhibitory. Wang et al., the same year, placed the ligands on distinct surfaces and concluded the opposite - that a ternary complex forms. Blatter et al. arbitrated it in 2023 by determining solution structures of the Cyp33 RRM free, bound to RNA, bound to MLL1 PHD3 (Kd about 7 uM by ITC), and bound to BOTH MLL1 PHD3 and H3K4me3 (PDB 7ZEZ). The ternary complex exists. What is mutually exclusive is not H3K4me3 and the RRM on PHD3, but RNA and PHD3 on the RRM beta-sheet: Cyp33 binds UAAUGU RNA at Kd 300 uM through the same surface, and an alpha3 helix N-terminal to the RRM adopts three positions that sequence the binding events. Adding RNA to the ternary complex releases the histone tail from PHD3, which the released RNA then traps. So the switch is real, the repression model survives, and the mechanism is RNA-gated rather than steric competition on PHD3. The whole module is absent from every leukaemogenic KMT2A fusion.
The disagreement, and how it was settled
Both 2010 papers agree that PHD3 binds H3K4me3 preferentially, that PHD3 binds the CyP33 RRM, that the two ligands occupy distinct surfaces, and that CyP33's isomerase acts on a proline in the PHD3-bromodomain linker. They disagreed on the consequence:
- Park et al. 1: the two bindings are mutually inhibitory, making PHD3 a molecular switch.
- Wang et al. 2: the two ligands can bind simultaneously to form a ternary complex.
Blatter et al. 2023 determined that ternary complex - Cyp33 RRM-delta-alpha with MLL1 PHD3 and an H3K4me3 peptide, deposited as 7ZEZ - and state in terms that their results explain 'the previously unclear and controversial role of RNA binding by Cyp33 as well as the existence and role of a ternary complex between Cyp33, MLL1, and H3K4me3' 4. Wang et al.'s reading was right about coexistence.
Park et al.'s intuition that something is exclusive was also right, but about a different pair: RNA and PHD3 compete for the same beta-sheet surface of the RRM, and titrating RNA into the ternary complex releases the histone peptide from PHD3 4.
What this graph should now say
mechanism-phd3-cyp33-repressor-switch-lost carries "mutual_exclusivity": "H3K4me3 binding and CyP33-RRM binding to PHD3 are mutually inhibitory" as a flat assertion. That should be revised, citing PMID 37075125, to: the two ligands form a ternary complex; the mutually exclusive pair is RNA versus PHD3 on the Cyp33 RRM; and RNA binding is the trigger that releases the histone tail. The entry's larger claim - that PHD3 is a switch between activation and repression and that the fusion loses it - is unaffected and if anything strengthened, because the switch now has a physiological input.
The structures are chimeras, except the 2023 ones
2KU7 is a single chain annotated to both Q03164 and Q9UNP9 with a GGSGGSGGSG linker - PHD3 portion Q03164 1586-1628 3. The 2023 deposits are genuinely multi-chain: 7ZEY has separate Cyp33 RRM (Q9UNP9) and MLL1 PHD3 (Q03164, 64-mer) entities, and 7ZEZ adds a 13-residue H3 peptide as a third chain 5. The bimolecular and trimolecular versions of this interface were only solved in 2023, thirteen years after the chimera.
| partner a | gene-kmt2a |
|---|---|
| partner b | gene-ppie |
| motif a | KMT2A PHD3, Q03164 1586-1628 in the 2KU7 chimera and 1564-1628 in 2KYU |
| motif b | CyP33/PPIE RRM domain, using its canonical RNA-binding surface |
| kd NM | 14700 |
| fate in fusion | lost - PHD3 lies C-terminal to the breakpoint cluster region and is absent from every leukaemogenic KMT2A fusion |
| chimera note | 2KU7 joins the third of three chimeric single-chain deposits underpinning this disease's interface literature, alongside 2MTN (MLL-LEDGF) and 2LM0 (AFF1-AF9) |
| druggability | a micromolar contact between a zinc finger and an RRM, with a single identified hot spot; the interesting modality is not inhibition but induced proximity - recruiting CyP33/HDAC activity to a fusion that lacks PHD3 |
| resolution note | PDB metadata and the 2KU7 entity sequence fetched from data.rcsb.org 2026-09-20; PHD3 span aligned to UniProt Q03164 fetched the same day |
| switch input | RNA. Cyp33 binds UAAUGU RNA (Kd 300 uM) through the same RRM beta-sheet that binds PHD3, and an alpha3 helix N-terminal to the RRM adopts three positions that order the binding events; adding RNA to the ternary complex releases the H3K4me3 tail from PHD3. |
interface residues a
[ "KMT2A V1617 (RRM hot spot; V1617A costs ~6.2-fold)", "KMT2A W1594 (H3K4me3 cage; dispensable for RRM binding)" ]
kd measurements — the numbers behind the chart
[
{
"ligand": "MLL PHD3 + CYP33 RRM",
"kd_nM": 14700,
"method": "ITC",
"source_pmid": "20677832"
},
{
"ligand": "MLL PHD3 W1594A + CYP33 RRM",
"kd_nM": 19600,
"method": "ITC",
"source_pmid": "20677832"
},
{
"ligand": "MLL PHD3 V1617A + CYP33 RRM",
"fold_loss": "~6.2",
"method": "ITC",
"source_pmid": "20677832"
},
{
"ligand": "Cyp33 RRM + MLL1 PHD3",
"kd_nM": 7000,
"method": "ITC",
"source_pmid": "37075125"
},
{
"ligand": "Cyp33 wild type + UAAUGU RNA",
"kd_nM": 300000,
"method": "NMR titration",
"source_pmid": "37075125"
},
{
"ligand": "Cyp33 alpha3-helix mutant (WLF) + UAAUGU RNA",
"kd_nM": 70000,
"method": "ITC and NMR",
"source_pmid": "37075125"
}
]structures
[ "2KU7", "7ZEY", "7ZEZ", "7ZEV", "7ZEW", "7ZEX" ]
structure details
[
{
"pdb": "2KU7",
"title_fetched": "Solution structure of MLL1 PHD3-Cyp33 RRM chimeric protein",
"method": "NMR",
"resolution_A": null,
"contents": "ONE 140-residue entity annotated to Q03164 and Q9UNP9; PHD3 portion maps to Q03164 1586-1628",
"fetched": "2026-09-20"
},
{
"pdb": "7ZEY",
"title_fetched": "Complex Cyp33-RRM : MLL1-PHD3",
"method": "NMR",
"resolution_A": null,
"contents": "two chains: Q9UNP9 117-mer + Q03164 64-mer",
"fetched": "2026-09-20"
},
{
"pdb": "7ZEZ",
"title_fetched": "Trimolecular complex Cyp33-RRMdelta alpha : MLL1-PHD3 : H3K4me3",
"method": "NMR",
"resolution_A": null,
"contents": "three chains: Q9UNP9 isoform 3 93-mer + Q03164 64-mer + 13-residue histone H3 peptide",
"fetched": "2026-09-20"
}
]open dispute
{
"question": "Are H3K4me3 and CyP33 RRM binding to PHD3 mutually exclusive?",
"position_a": "mutually inhibitory (PMID 20677832, 2010)",
"position_b": "distinct surfaces, simultaneous ternary complex (PMID 20541251, 2010)",
"resolution": "SETTLED in favour of position B by PMID 37075125 (2023), which determined the trimolecular structure 7ZEZ; the genuinely exclusive pair is RNA versus PHD3 on the Cyp33 RRM beta-sheet",
"graph_status": "mechanism-phd3-cyp33-repressor-switch-lost still asserts position A with no counter-note; correction owed"
}- Model correction: an induced-proximity strategy designed on the 2010 mutual-exclusivity model is designed against a mechanism the 2023 structures contradict.
- RNA as a competitor: cellular RNA competes with PHD3 for the Cyp33 RRM at micromolar-to-millimolar affinity, so any engineered recruitment operates against a large endogenous competitor pool.
- Wrong allele: PHD3 is absent from the fusion, so a direct ligand acts on wild-type MLL1.
- Weak contact: 7-15 uM is a modest affinity to build chemistry against, on an RRM surface that is also an RNA-binding site.
- Which cellular RNA supplies the switch input in a haematopoietic cell, and is it present at concentrations comparable to the 300 uM in vitro Kd?
- Does the alpha3 helix occlude the ternary arrangement in the wild-type RRM? No wild-type ternary structure exists - 7ZEZ uses a helix-deleted isoform-3 construct.
- If PHD3 re-insertion into a fusion blocks transformation, does it require the RNA-gated step, or only the CyP33 tether?
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
MLL PHD3 binds the CYP33 RRM with Kd 14.7 uM by ITC; V1617A weakens RRM binding about 6.2-fold and W1594A leaves it nearly unchanged (19.6 uM), the two ligands occupy surfaces 180 degrees apart, and the authors conclude the two bindings are mutually inhibitory.
The PHD3 domain of MLL acts as a CYP33-regulated switch between MLL-mediated activation and repression . · Park S, Osmers U, Raman G, et al. · Biochemistry 2010What this evidence cannot show
ITC on isolated domains; 'mutually inhibitory' is the authors' interpretation of competition data on separated surfaces rather than a directly observed exclusion, and the isolated PHD3 lacks the bromodomain. The mutual-exclusivity conclusion is SUPERSEDED by the 2023 ternary structure; the affinity and hot-spot mapping in this paper stand.
Evidence that argues against this (1)
- Wang et al. state explicitly that the two ligands can bind simultaneously to form a ternary complex, which is the opposite conclusion from the same starting observation that the surfaces are distinct. Wang Z et al. 2010
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
H3K4me3/2 and the CyP33 RRM target distinct surfaces of MLL1 PHD3 and can bind simultaneously to form a ternary complex; CyP33's PPIase activity isomerises a proline in the PHD3-bromo linker, disrupting the interdomain interface and exposing an occluded PHD3 surface for RRM binding.
Pro isomerization in MLL1 PHD3-bromo cassette connects H3K4me readout to CyP33 and HDAC-mediated repression. · Wang Z, Song J, Milne TA, et al. · Cell 2010What this evidence cannot show
The ternary complex claim rests on NMR titration of the cassette and a chimeric construct; the proline isomerisation model is inferred from chemical-shift changes and PPIase activity assays rather than from a structure of both states. Full text reached by efetch db=pmc after fullTextXML returned HTTP 500.
Evidence that argues against this (1)
- Park et al., working on the isolated PHD3 in the same year, conclude the two bindings are mutually inhibitory, and the graph's own mechanism entry has asserted that reading for several cycles. Park S et al. 2010
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB 2KU7 is a solution NMR structure of a single 140-residue chimeric chain annotated to Q03164 and Q9UNP9, joining MLL1 PHD3 (Q03164 1586-1628) to the CyP33 RRM through a glycine-serine linker.
PDB 2KU7: Solution structure of MLL1 PHD3-Cyp33 RRM chimeric protein · Wang Z, Song J, Milne TA, et al. · RCSB Protein Data Bank 2010What this evidence cannot show
Title, method and the single entity sequence fetched from data.rcsb.org 2026-09-20; the PHD3 span is this cycle's alignment to the UniProt Q03164 sequence fetched the same day. A tethered chimera can stabilise a 14.7 uM interaction to the point of structural determination and cannot report on the association's entropic cost.
Evidence that argues against this (1)
- The isolated PHD3 solution structure 2KYU, from the other group, is a genuine single-domain structure and maps to a longer span (1564-1628), so the two deposits describe different constructs of the same finger. Park S et al. 2010
- E4strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
Solution structures of the Cyp33 RRM free, RNA-bound, PHD3-bound (Kd about 7 uM by ITC) and bound to both MLL1 PHD3 and H3K4me3 show that a ternary complex forms; RNA and PHD3 compete for the same RRM beta-sheet, Cyp33 binds UAAUGU RNA at Kd 300 uM, and adding RNA releases the histone tail from PHD3.
RNA binding induces an allosteric switch in Cyp33 to repress MLL1-mediated transcription. · Blatter M, Meylan C, Clery A, et al. · Sci Adv 2023What this evidence cannot show
Solution NMR on isolated domains and short RNA and peptide ligands; the ternary structure 7ZEZ uses a Cyp33 RRM construct lacking the alpha helix (RRM-delta-alpha, annotated to isoform 3), not the full-length wild-type RRM, so the trimolecular arrangement is demonstrated for a truncated form. The 300 uM RNA affinity is weak in absolute terms and was measured by NMR titration for wild type.
Evidence that argues against this (1)
- Because 7ZEZ uses an RRM lacking the alpha3 helix, a reader could argue the helix would occlude the ternary arrangement in the wild-type protein - which is precisely the allosteric mechanism the paper proposes, so the truncation is the model rather than an artefact, but it does mean no wild-type ternary structure exists. Park S et al. 2010
- E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database
PDB 7ZEZ is a solution NMR trimolecular complex with three separate polymer entities: Cyp33 RRM-delta-alpha (Q9UNP9 isoform 3), MLL1 PHD3 (Q03164, 64 residues) and a 13-residue histone H3 peptide; 7ZEY is the corresponding binary Cyp33 RRM-PHD3 complex.
PDB 7ZEZ: Trimolecular complex Cyp33-RRMdelta alpha : MLL1-PHD3 : H3K4me3 · Blatter M, Meylan C, Clery A, et al. · RCSB Protein Data Bank 2023What this evidence cannot show
Titles, methods and entity annotations fetched from data.rcsb.org 2026-09-20. NMR, so no resolution applies. The H3 entity in 7ZEZ is annotated to Q6NXT2 (histone H3.Y-family accession) rather than to P68431, which is a cross-reference quirk of the deposit rather than a different histone.
Evidence that argues against this (1)
- The Cyp33 entity in 7ZEZ is an isoform-3, helix-deleted construct, so the deposit is not a structure of the wild-type full-length RRM in a ternary complex. Blatter M et al. 2023
PHD3 1586-1628 binds the CyP33 RRM, Kd 14.7 uM
CyP33 RRM canonical RNA-binding surface
that entry asserts mutual exclusivity; PMID 20541251 concludes a simultaneous ternary complex
the competing or coexisting ligand on the same finger
the cellular experiment that would bear on which model holds
the 2023 arbitration of the 2010 disagreement