Daily digest — 2026-10-28

Four-for-four Gram-negative ESKAPE coverage suggests transporter-dependent uptake, not the RNAP pharmacophore, was the limiting factor. A cationic CPP is a blunt fix, though: expect serum binding, protease exposure and mammalian membrane liability, so selectivity indices and stability data will decide whether it beats engineering the loop for better transporter recognition. Grafting onto the threaded scaffold looks tractable.

A Pt(II) warhead plus a vinylsulfonamide is a dual-electrophile design, so selectivity worries me first: PKM2 has several reactive cysteines, and Pt itself hits glutathione and everything soft. Chemoproteomics naming one target is encouraging, but the cross-linking story needs the cysteine mutant and a Pt-free vinylsulfonamide control. Synthetically easy, mechanistically messy.

Binding retained, analgesia gone: deleting the N-terminal amine decouples affinity from function, so Ki alone will mislead your SAR here. That amine likely anchors an extracellular-loop contact needed for the active state, so treat it as untouchable in the pharmacophore. The non-additive combinations also warn that stacking 'good' substitutions in a macrocycle needs re-validation, not assumption.

S7's cleaner mitochondrial profile versus niclosamide is the data to chase: salicylanilides are classic protonophore frequent hitters, so RdRp inhibition decoupled from uncoupling is what makes the scaffold usable. N-pocket binding rests on docking plus mutagenesis, though, and without a co-crystal I wouldn't trust the SAR yet. A conserved pocket and easy amide chemistry make pan-flavivirus follow-up cheap.

Swapping aripiprazole's hydroxy-dihydroquinolinone for a naphthyridinol adds a ring nitrogen, which should lower logD and open fresh IP space on a crowded core. The catch is that reversed cocaine hyperactivity without catalepsy is only a phenotypic readout; with no D2 partial agonism or 5-HT1A/2A data, it's unclear whether the hop preserved the pharmacology.

A nitrile reaching the catalytic lysine is a smart way around gatekeeper and activation-loop mutations, yet the new ATP-site pose probably costs avapritinib's selectivity, so a kinome panel comes first. Squaric esters are the weak link: hydrolytically fragile and prone to hitting surface lysines. Does ATP-competitive labeling hold in cells, not just buffer?

Two cycles and a few thousand members is a toy library, and recovering carbonic anhydrase proves little; it's the easiest control in DEL. The numbers that matter are yield and DNA integrity per on-resin step, since organic-solvent chemistry only pays off if it opens reactions aqueous conditions can't, like Pd couplings or SNAr. Automation helps reproducibility, not diversity.

Three pairings from one cheap linker is a useful toolkit; tyrosine has been the missing handle in a field dominated by all-hydrocarbon and Cys-based staples. The ESF-derived sulfonyl aryl ether needs a stability check against lactam staples, both hydrolytic and conformational. One RGD example says nothing about αvβ3 versus α5β1 selectivity, so the SAR stays unproven until that panel exists.