Hydroxamic Acid Divalent Cation Liability
Distilled from DR3 — CRO Wet-Lab Vendor Menu §I.6. Direct hit on H01 lead v5.2__aq3__adamantyl__CONHOH__-Cl.
The principle
The hydroxamic acid (R–CONHOH) warhead is a strong bidentate chelator of divalent cations — Zn²⁺ first, then Ca²⁺, Mg²⁺, Fe²⁺ at lower affinity. This is the entire reason hydroxamates are the dominant zinc-metalloenzyme inhibitor pharmacophore in marketed drugs (vorinostat, romidepsin precursors, ilomastat MMP inhibitors).
For an ear-targeted small molecule the same chemistry creates a population-wide off-target footprint:
- Cochlear ion channels with intracellular divalent regulatory sites — KCNQ4 (calmodulin-Ca²⁺ gated, DFNA2 paralog, OHC potassium recycling), TRPM4 (Ca²⁺-activated nonselective), TMEM16A (Ca²⁺-activated chloride), BK (Ca²⁺/voltage), Cx50/Cx26 connexins (Ca²⁺-modulated gap junctions).
- Stria vascularis Zn²⁺ enzymes — carbonic anhydrase II/IV maintain endocochlear potential; CA inhibition by hydroxamates is documented.
- HDAC family — most relevant HDACs are Zn²⁺-active-site; hydroxamates are the standard pan-HDAC inhibitor scaffold. Note this is two-faced: SAHA is also a documented otoprotectant via HSP32 induction (DR5 §C), so the off-target signal here is bidirectional and not a clean liability.
Why the existing H01 selectivity work does not yet close this
STRC h01 Phase 8g v5.2 Off-Target Panel 2026-04-25 — Vina docking against the 6-channel cochlear panel failed for all 6 ligands including v3b reference. The conclusion landed on “Vina limitation; wet-lab required.” That is methodologically correct for Vina but it leaves the hydroxamate-specific divalent question unanswered: Vina has no force-field term for Zn²⁺ chelation, so even a perfect docking score against a Zn-loaded channel pocket would say nothing about whether the warhead pulls the metal out.
Phase 8h-lite #5 Tanimoto-against-ototoxin-panel returned max 0.127, no class motif — this falsifies inheritance from quinoline-class retinopathy precedents (DR4) but says nothing about the warhead itself, because Tanimoto on Morgan FP r=2 weights the scaffold heavily and the CONHOH group is a small structural fraction of the molecule.
Computational paths to actually close it
- APBS on each ion-channel pocket loaded with its native divalent — directly computes whether the hydroxamate’s electronegative O,O bidentate face is energetically pulled toward the channel’s Zn²⁺ or Ca²⁺ binding site relative to bulk solvent. This is the APBS-on-enclosed-pockets task already queued as Phase 8h-lite heavy follow-up; framing it specifically around metal-coordination electrostatics (not just generic charge complementarity) is the right scope.
- Hydroxamate-vs-Zn DFT/QM-MM on a model active-site cluster — for a final paper claim, a single QM-level energy on hydroxamate displacing the channel’s native ligating water vs. its native substrate gives a defensible per-channel selectivity number. Single-cluster DFT is feasible locally.
- Cochlear melanin / stria vascularis Zn²⁺-protein side-search — DR4 review (log entry 01:1x) flagged melanin accumulation risk specific to intratympanic confinement. Same logic applies to Zn²⁺-rich proteins concentrated in stria vascularis. UniProt query for “Zn-binding” + cochlear-expression filter, screen lead against shortlist.
Bioisosteric escape routes (not yet tried)
If the metal-chelation off-target case turns red, the warhead has known weaker-chelating or non-chelating bioisosteres:
- Acyl sulfonamide (R–CONHSO₂R’) — already in the Phase 4h library; preserves H-bond donor + anion at physiological pH, removes O,O bidentate face.
- Tetrazole (R–C(=N–N=N–NH)) — anion bioisostere, monodentate at best, no Zn² chelation.
- N-hydroxy formamide / α-hydroxy amide — partial chelation, often used to soften MMP-inhibitor zinc affinity by 1-2 log.
These are already in the v5.2 design space (STRC h01 Phase 8d 8e 8f v5.2 Library Design 2026-04-25). The relevant question is whether the combined-score ranking after a proper APBS-Zn off-target screen still places __CONHOH__ at #1, or whether __CONHSO₂Me__ / __tetrazole__ analogs of the same aq3__adamantyl__-Cl body now climb. That is a re-rank, not a new screen.