Ebselen: the only small molecule reported to bind LEDGF itself - and it does it by making selenium-sulphur bonds with whatever thiols it meets
also called 2-phenyl-1,2-benzisoselenazol-3-one · PZ 51 · SPI-1005
Only one small molecule has ever been reported to stick to LEDGF directly. It works by forming bonds with sulphur atoms in the protein, which it will also do to many other proteins, so it is a starting clue rather than a drug.
The same thing, for a scientist
Screening for inhibitors of the LEDGF/p75-integrase interaction identified ebselen as a moderate micromolar inhibitor that, unusually for that screen, acts **by binding LEDGF/p75 rather than integrase**. Biolayer interferometry indicated formation of selenium-sulphur bonds with reduced thiols in LEDGF, and the inhibition was reversed by dithiothreitol. The authors state it as the first report of a small molecule inhibiting the LEDGF/p75-integrase interaction by directly binding LEDGF/p75, and explicitly position it as a probe compound rather than a lead. For this graph it occupies a specific slot: it is the counterexample that keeps the statement 'there is no chemical matter against LEDGF' from being literally true, while being exactly the kind of compound that a target dossier should not count. Ebselen is a known promiscuous cysteine-reactive agent with activity reported against many unrelated targets, and a DTT-reversible covalent adduct at unspecified cysteines is not a selective ligand for a defined pocket.
What was measured, and what was not
Measured 1: ebselen inhibits the LEDGF/p75-integrase interaction at moderate micromolar potency; the binding partner is LEDGF/p75, not integrase; BLI is consistent with selenium-sulphur bond formation to reduced thiols; DTT reverses the effect.
Not measured: which cysteines; whether the adduct is in or near the IBD; whether the KMT2A IBM contact is disrupted; anything in a leukaemia cell; selectivity against any other thiol-containing protein.
Why it still matters
Two reasons, both narrow.
First, the graph's chemical-matter statement needs a footnote. target-phase2-psip1-ibd-dossier and gene-psip1 both say ChEMBL holds three potent rows against PSIP1, all from HIV integrase chemistry, and both frame that as near-zero chemistry. That framing is right in substance, and ebselen is the one published exception to the letter of it. Note the database caveat this project already records: a ChEMBL activity-row count measures how much medicinal chemistry has been done, and the compounds in those rows are frequently probes for something else.
Second, a covalent route is a real route. If LEDGF carries a ligandable cysteine, a covalent fragment screen is a legitimate and cheap way to find a starting point for a protein with no conventional pocket - and LEDGF's IBD is a shallow, extended peptide-binding surface of exactly the class that resists conventional ligands. Ebselen says there are reactive cysteines. It does not say where.
What would make this useful
One experiment: intact-mass or peptide-level mass spectrometry of LEDGF after ebselen treatment, to localise the adduct. If it maps into the IBD, this becomes the first covalent starting point for the leukaemic interface. If it maps to a cysteine in the disordered linker or the PWWP, ebselen is a nonspecific thiol reagent and should be recorded as such and set aside.
| target | gene-psip1 |
|---|---|
| modality | small molecule, thiol-reactive organoselenium |
| status | research probe only; not developed for this target |
| mechanism | selenium-sulphur bond formation with reduced thiols in LEDGF/p75; DTT-reversible |
| potency | moderate micromolar inhibition of the LEDGF/p75-integrase interaction |
| adduct site | unmapped |
trials
[]
key ae
[ "promiscuous cysteine reactivity" ]
resistance
[]
- Counting ebselen as chemical matter for LEDGF would flatter the dossier: a DTT-reversible promiscuous thiol reagent is not a selective ligand and would not survive a counter-screen.
- A covalent LEDGF agent would modify the protein in every tissue that expresses it, which is all of them, and LEDGF's DNA-repair and R-loop functions are exactly the ones a covalent agent would take out systemically.
- Ebselen has been in clinical development for other indications, which makes it tempting to repurpose; the temptation should be resisted until the adduct site is known.
- Which LEDGF cysteines does ebselen modify, and is any of them in the IBD? One mass-spectrometry experiment answers it.
- Does ebselen disrupt the LEDGF IBD-KMT2A interaction, as opposed to the integrase one? Nobody has run the obvious counter-assay, even though this project's structural beat established the two interactions share a surface.
- Is there a ligandable cysteine anywhere on the IBD by covalent-fragment screening? The IBD is a shallow peptide-binding surface and a covalent approach may be the only realistic modality.
- E1moderate confidence 3 of 5: One strong study, or several smaller consistent ones.paper
Ebselen was identified as a moderate micromolar inhibitor of the LEDGF/p75-HIV-1 integrase interaction that acts by binding LEDGF/p75 directly, probably through selenium-sulphur bonds with reduced thiols, with the inhibition reversible by dithiothreitol; the authors describe it as the first small molecule shown to inhibit this interaction by binding LEDGF/p75.
The selenium-containing drug ebselen potently disrupts LEDGF/p75-HIV-1 integrase interaction by targeting LEDGF/p75. · Zhang DW, Yan HL, et al. · J Enzyme Inhib Med Chem 2020What this evidence cannot show
A single report, in an HIV assay, with no cellular leukaemia data and no mapping of the modified residues. Ebselen is a well-known promiscuous thiol-reactive compound and DTT reversibility indicates a labile covalent adduct rather than a defined binding site. Open access in Europe PMC (PMC7170385).
Evidence that argues against this (1)
- The same paper positions ebselen as a probe compound for designing new disrupters rather than as a lead, and thiol-reactive hits from protein-protein-interaction screens are the classic false-positive class in that assay format. Hou Y et al. 2008
- E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper
The LEDGF IBD surface that ebselen would need to reach is the same surface used by HIV-1 integrase and by the KMT2A IBM2 motif, with integrase Trp131 occupying the position of KMT2A Phe148.
The same site on the integrase-binding domain of lens epithelium-derived growth factor is a therapeutic target for MLL leukemia and HIV. · Murai MJ, Pollock J, et al. · Blood 2014What this evidence cannot show
Establishes that an agent disrupting the IN-IBD interaction is acting at a site relevant to KMT2A. It does not establish that ebselen binds at that site, which is unmapped.
Evidence that argues against this (1)
- A compound that disrupts the IN-IBD interaction by modifying a distant cysteine and destabilising the domain would produce the same assay result without touching the KMT2A-relevant surface. Zhang DW et al. 2020
- E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper
The structural basis for the recognition between HIV-1 integrase and the LEDGF/p75 IBD is defined at 2.02 A in PDB 2B4J.
Structural basis for the recognition between HIV-1 integrase and transcriptional coactivator p75. · Cherepanov P, Ambrosio AL, et al. · Proc Natl Acad Sci U S A 2005What this evidence cannot show
Title taken from PubMed, not from the RCSB primary-citation string, which differs for this record - the trap logged in cycle 67. Cited for the structural definition of the surface.
Evidence that argues against this (1)
- A solved interface is not a ligandable pocket: the IBD-integrase contact is an extended helix-bundle surface rather than a cavity, which is why no conventional small molecule against it has been reported. Blokken J et al. 2017
binds LEDGF/p75 through reduced thiols; DTT-reversible
the single exception that proves the class distinction
would need to reach this surface to be relevant; adduct site unmapped