Daily digest — 2026-09-16

A canonical G-loop-mediated ternary complex — reassuring, not exotic: this is the same validated CRBN surface, not a new liability profile to characterize from scratch. The compact ~41-residue degron is the real tell for tractability — that's a defined element, not a diffuse IDR, so SAR here should be rational rather than lucky. Nanomolar potency with reported high proteome-wide selectivity is the headline number, though for a spliceosome-associated protein I'd want orthogonal splicing/transcriptomic data before fully trusting "minimal downstream effects." The piece I'd actually chase into my own work is wTAG: a monovalent, hook-effect-free degron tag beats dTAG-style bifunctional systems for clean dose-response in target-validation studies, regardless of whether WBP4 itself ever becomes a drug target.

Smart move folding the VHL ligand and linker into a bifunctional amino acid for direct SPPS incorporation — that sidesteps the PEGylation/conjugation bottleneck that makes cyclic peptide degraders such a pain to iterate on, and it turns linker SAR into a synthesis-cycle variable instead of a separate coupling step. The SICLOPPS-derived warhead is a nice proof that a cyclic peptide can grip a transcription-factor surface a small molecule can't easily touch, so on chemotype comparison this favors the peptide side. But the real flag for my own work is the hypoxia dropout: ternary complex still forms, HIF-1α still gets marked, degradation just can't keep up. That's not a binding problem I'd fix by optimizing the VHL warhead — it's a turnover-rate problem, and it says something uncomfortable about any PROTAC campaign against a target with fast resynthesis kinetics.

Five million virtual candidates funneled down to a handful that actually work is the real story — that pipeline (docking → ML force fields → QM → nonadiabatic dynamics → FEP) is doing what a med chem team's SAR intuition used to do by hand, just faster and across chemical space no bench chemist would ever hand-explore. The binding-mode hypothesis is straightforward: a photoswitch appended to (presumably) a nicotinamide-mimetic PARP1 pharmacophore toggles pocket occupancy between light and dark isomers, so "selectivity" here means spatiotemporal control over tumor tissue, not isoform selectivity against PARP2/tankyrases — a different axis entirely from what olaparib or niraparib were ever optimized for. Red-shifted absorption for tissue penetration plus a therapeutically workable thermal half-life are the tractability bottlenecks that usually kill photopharmacology leads, so getting synthetically accessible hits past both, and into validated wet-lab data, is the genuinely notable part.

The SAM-pocket cysteine hit on PRMT1 is the part I trust most here — that's a genuinely druggable site, and confirming occupancy by intact-protein LC-MS rather than an activity readout alone is the right bar. Where I'd slow down is selectivity: tuning reactivity through metal identity and arene substituents is a real lever, but it's one more variable than iodoacetamide/chloroacetamide fragment libraries already juggle, and a lysate screen doesn't tell me the off-target cysteine load in cells. Chemically, this is genuinely unfamiliar ground for most of us — ligand exchange, redox lability, and batch-to-batch reproducibility are liabilities organic warheads simply don't have. Against established cysteine-directed chemoproteomic platforms, r-mS adds an interesting third tuning axis, not yet a clear win. Metallodrugs as a class still carry that same tractability tax relative to organic small molecules — worth watching, not yet worth reaching for.

Sub-picomolar affinity out of a 125,000-member DCC library gets my attention immediately — that's an avidity effect, not a single tight pocket, and it's essentially bivalirudin's active-site/exosite-1 logic pushed one contact further to include exosite II as well. Exosite engagement should also buy real selectivity over trypsin-like serine proteases that don't share thrombin's exosite topology, though I'd want that confirmed head-to-head against fXa and trypsin. My hesitation is tractability: a DNA/PNA-templated trivalent assembly is a fantastic proof-of-concept and chemical tool, not yet a lead — that's a long way from oral. Against DOACs' one-pocket, small-molecule simplicity, this wins on affinity and reversibility but loses badly on drug-likeness; the real test is whether target-templated DCC can be steered toward a druggable combination, not just the tightest binder.

The crosslinker sits inside the acetyllysine pocket itself, at L922 — that's not a fortuitous surface hit, it's evidence TRIM24 reads acetylated lysine on NPM1 and CDK9 through the same pocket it uses on histones, not a peripheral contact. That reframes TRIM24 as an acetyl-lysine reader with a broader substrate scope, and it means any inhibitor competing for that pocket is also competing off a nonhistone binding event, not just a histone mark. Practically, TRIM24 probe programs need NPM1/CDK9 engagement counter-screens now, and the TSS co-occupancy by ChIP-seq suggests that off-target isn't background noise but transcriptionally active. Worth flagging: this is GCE-based interactome mapping, not medicinal chemistry — no chemotype, no SAR, just a map of what a pocket-directed probe will also touch. BET programs learned this same lesson with nonhistone acetyl-lysine substrates years ago; TRIM24 is only now catching up to that complexity.