STRC Research

STRC h01 Phase 8 Wet-Lab Triage SOP

12 min read

STRC h01 Phase 8 Wet-Lab Triage SOP

Concrete protocol for wet-lab validation of pharmacochaperone hits from Phase 3c v3b / Phase 5e. Intended hand-off to Misha’s collaborator molecular biology lab (TBD). Must be executable from a GREEN compute hit within 6–8 weeks.

design philosophy

Three-gate triage, each stage kills most candidates so the next stage’s cost stays bounded. Gate thresholds are calibrated against iododiflunisal (known TTR pharmacochaperone at related ΔG range) as an internal positive control and DMSO as negative.

GATE 1 (1 wk, $3k, 96-well)  — ThermoFluor ΔTm on mini-STRC apo + mutant
  ├── ≥ 1.5 °C stabilization on mutant → pass
  └── fail → killed, feedback to medchem
GATE 2 (2 wk, $8k, n=3 replicates) — MST Kd + stoichiometry
  ├── Kd ≤ 50 µM AND Hill 0.8–1.2 → pass
  └── fail → killed
GATE 3 (4 wk, $25k, cryoEM co-complex, cell rescue) — structural + functional
  ├── co-complex density at K1141 pocket AND E1659A HEK293 rescue ≥ 30% → validated lead
  └── fail → de-prioritized

construct design (Gate 0 — pre-req)

Already in progress with AF3-validated Ultra-Mini boundary. Cloning target:

ConstructResidues (full-length)PurposeTag
STRC-UM-WT594–1294WT Ultra-Mini, ThermoFluor + MST + cryoEMN-His6-TEV
STRC-UM-E1659A594–1775extended to include mutation site; probe fold stabilityN-His6-TEV
STRC-FL-WT1–1775full length for cell rescueN-GFP (rescue readout)
STRC-FL-E1659A1–1775E1659A full length; Misha alleleN-GFP

Cloning: pcDNA3.4 mammalian expression for HEK293 rescue; pFastBac1 baculovirus/Sf9 for milligram-scale protein. SignalP-6 predicts N-terminal signal peptide at residues 1–26 — cleave for mature protein. C-term propeptide (1775–1809) not included in any construct (removed post-ER in vivo per Verpy 2001).

Protein production:

  • Sf9 baculovirus: 1 L → 0.8–2 mg soluble mini-STRC after Ni-NTA + SEC (reference yield from STRC-related stereociliary proteins, see Grillet 2019 Nature Commun for stereocilin-CDH23 complex protocol)
  • HEK293 GnTI⁻ for homogeneous glycosylation (stereocilin has 7 N-linked sites in UM boundary per NetNGlyc)
  • pI ≈ 5.8 → buffer: 25 mM HEPES pH 7.4, 150 mM NaCl, 5% glycerol, 1 mM TCEP
  • Stability check: SEC trace monodispersity + DLS PDI < 0.15 before any compound assay

GATE 1 — ThermoFluor ΔTm (differential scanning fluorimetry)

Purpose: first-pass stabilization readout. Chaperone binding raises Tm.

Reagents:

  • Protein: 5 µM mini-STRC-UM-WT and mini-STRC-UM-E1659A
  • SYPRO Orange 5× in assay buffer
  • Compounds: 1 mM DMSO stock → final 100 µM (2% DMSO), also 30 µM and 10 µM
  • Buffer: 25 mM HEPES pH 7.4, 150 mM NaCl, 1 mM TCEP
  • Positive control: iododiflunisal 100 µM
  • Negative control: 2% DMSO vehicle

Plate layout: 96-well PCR plate, n=3 per compound per concentration per construct. 20 compounds × 3 conc × 2 constructs × 3 reps = 360 wells → 4 plates.

Instrument: QuantStudio 3 or Applied Biosystems StepOnePlus qPCR, melt 25 → 95 °C at 1 °C/min. Filter: TAMRA (SYPRO Orange).

Analysis: fit Boltzmann sigmoid to -dF/dT peak. ΔTm = Tm(compound) - Tm(DMSO). Calibrate:

  • iododiflunisal positive control should give ≥ 2 °C shift on WT.
  • Mutant baseline Tm expected 2–4 °C lower than WT (E1659A destabilizes).
  • GATE 1 PASS: ΔTm ≥ 1.5 °C on E1659A at 100 µM, ≥ 0.5 °C at 30 µM, monotonic with dose.

Expected timing: 3 days (prep + plates + analysis), cost ≈ $3k compound + consumables.

Probes to include first pass — orderable NOW (do not wait for v4):

Per STRC h01 Phase 4h Tafamidis Playbook Library 2026-04-23 Tier-1 commercial set + Phase 5e mutant-validated benchmarks. Phase 5e (2026-04-24) confirmed WT-ensemble docking valid for E1659A (max kd_ratio 2.09×) → ThermoFluor on E1659A is the right validation gate for WT-derived predictions.

CompoundSourceRoleCat# / vendor
tafamidis (Vyndaqel API)Phase 4h #5Tier-1 positive, FDA pediatric-cleared scaffoldSelleck S2839
iododiflunisalPhase 4h 12Tier-1 positive, known TTR pharmacochaperoneSigma SML0345
diflunisalPhase 5b/5e benchPositive control, h01 reference KdSigma D3281
benzoxazole-3CN-COOH (4h #1 / 4h #6 thio)Phase 4h Tier-2Lead bioisostere; in-house 2-step synthTCI / synth
benzoxazole-3OMe-COOH (4h #2 / 4h #7 thio)Phase 4h Tier-2Polar variant, Cx50-killingsynth
niflumic acidPhase 3c v2/v3b/5eMD probe only per Tox Audit, NOT Gate 2/3Sigma N0630
flufenamic acidPhase 3c v2/v3b/5eMD probe only per Tox AuditSigma F9005
sulfasalazinePhase 3c v2parent probe (Phase 5e flagged unstable, std 4.7)Sigma S0883
meclofenamic acidPhase 3c v2/v3b/5eparent probeSigma M4531
Phase 3c v3b top-15 YELLOWsv3b (delivered 04-24)Lead compounds — synth Tier-2/3, ETA 4-6 wkCRO quote (Enamine REAL or WuXi)
Phase 3c v4 top-N (TBD)v4 (running, ETA 17:57 local 04-24)Fragment-grow leads — feed in within 24 hTBD post-verdict
DMSO 2% vehicle—Negative—
aspirin—Negative (unrelated NSAID; COX+ but no chaperone activity)Sigma A5376

Action this week: order tafamidis, iododiflunisal, diflunisal + 4 fenamic parents (Tier-1, ~$500 total) → run G1 ThermoFluor on existing mini-STRC-UM-WT/E1659A constructs as soon as protein production wraps. Establishes baseline Tm shifts + iododiflunisal positive-control calibration before v3b/v4 medchem leads need to be tested.

Kill criteria: ΔTm < 0.5 °C at 100 µM on E1659A OR destabilizing (ΔTm < -1 °C) on either construct. Destabilizers sent back to medchem as counter-probes (useful geometry info).

GATE 2 — MicroScale Thermophoresis (MST) Kd + stoichiometry

Purpose: precise Kd, cross-check ΔTm rank order.

Instrument: NanoTemper Monolith NT.115 (RED channel; use RED-tris-NTA dye on His6 tag, no covalent labeling needed).

Reagents:

  • Labeled mini-STRC-UM-E1659A: 50 nM in assay buffer + 0.05 % Tween-20 (prevent aggregation)
  • Compound titration: 16-point 1:1 serial from 500 µM → 15 nM (2% DMSO constant).
  • Capillaries: standard treated.
  • Temperature: 25 °C and 37 °C (human cochlear temp).

Data analysis: fit normalized F_norm vs [compound] with Hill equation. Require Kd ≤ 50 µM AND Hill slope 0.8–1.2 (no cooperativity / no aggregation artifacts). Compound-induced aggregation presents as U-shaped curve and is a kill.

Replicates: n ≥ 3 per compound × 2 temperatures. Report Kd ± SEM.

GATE 2 PASS criteria:

  • Kd ≤ 50 µM (calibrated to phase5b ensemble expectations, 3× loosened for first wet-lab pass to catch soft hits).
  • Hill 0.8–1.2.
  • Kd on WT within 2-fold of mutant (selectivity is not required but extreme mutant > WT discrimination flags allostery).
  • Reproducible 25 °C vs 37 °C (factor ≤ 2×).

Expected timing: 2 weeks post-G1, cost ≈ $8k (capillaries, dye, compound, instrument time).

Expected winnowing: 20 → 3–6 compounds pass G1; 3–6 → 1–3 pass G2.

GATE 2.5 — COX + TMEM16A activity kill gate (post-MST, pre-cryoEM)

Why before cryoEM: Phase 6c full 6-target panel (STRC h01 Phase 6c Full 6-Target Panel 2026-04-24) showed every Phase 4h Tier-1/Tier-2 candidate (tafamidis, iododiflunisal, all benzoxazole-COOH bioisosteres) hits both COX-1/-2 (Kd 1-3 µM) and TMEM16A (Kd 2-5 µM) — sub-therapeutic-dose off-target liability. cryoEM at $25k+ for a compound that fails COX/TMEM16A activity is wasted budget. G2.5 kills these before structural commitment.

2.5a COX-1 / COX-2 activity

Assay: PGE₂ ELISA (Cayman Chemical kit 514010) on isolated ovine COX-1 and human COX-2 enzymes. 16-point dose-response 100 µM → 30 nM at 37 °C, 5 min reaction, indomethacin positive control.

Reagents:

  • ovine COX-1 (Cayman 60100, 50 U/well)
  • human COX-2 (Cayman 60122, 50 U/well)
  • arachidonic acid substrate 30 µM final
  • compound stocks 10 mM DMSO → 100 µM final (1% DMSO)
  • indomethacin positive 1 µM (IC₅₀ ~10 nM COX-1, ~30 nM COX-2)

Pass criteria: IC₅₀ ≥ 50 µM on both COX-1 and COX-2 at 37 °C. Sets a 50× selectivity floor over predicted h01 K1141 Kd ~1 µM (compound) — below this floor, COX activity at therapeutic 10 µM dose would exceed 17% inhibition (PGE₂ pathway disturbance). Tighter than 50 µM threshold = 30%+ inhibition, unacceptable for pediatric chronic dosing.

Expected timing: 1 week, ~$2k for 6 compounds in duplicate.

2.5b TMEM16A activity

Assay: Patch-clamp (whole-cell) in HEK293-TMEM16A stably expressing line. 8-point dose-response 100 µM → 100 nM at +60 mV with 1 µM intracellular Ca²⁺. Niclosamide positive control (IC₅₀ ~0.1 µM).

Reagents:

  • HEK293-TMEM16A stable line (commercial: ChanTest CT3017 or in-house transfection)
  • niclosamide 1 µM positive control
  • pipette buffer: 130 mM CsCl, 1 mM CaCl₂ (free Ca buffered with EGTA), 10 mM HEPES pH 7.3
  • bath: 140 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 10 mM HEPES, 10 mM glucose pH 7.4

Pass criteria: IC₅₀ ≥ 30 µM. At therapeutic 10 µM dose this corresponds to ≤ 25% TMEM16A current inhibition — borderline acceptable given stria vascularis K⁺ secretion compensation buffer. Tighter than 30 µM = ≥30% inhibition = endocochlear potential risk → kill.

Why TMEM16A and not BK or Cx50: Phase 6c said TMEM16A at sub-5 µM is the second-largest selectivity wall after COX. BK has no NSAID-bound PDB so wasn’t dock-screenable; Cx50 was via Cx36 paralog (cross-paralog Kd uncertain by 5-10×). Both BK and Cx50 stay deferred to G3 cochlear-channel panel where wet-lab patch-clamp is the ground truth.

Expected timing: 2 weeks (CT3017 + patch-clamp instrument time), ~$5k.

2.5c TRPM4 + KCNQ4 + BK + Cx50 panel (concurrent, single-pass)

Per Phase 6c: KCNQ4 cleanly de-selected for benzoxazole-COOH bioisosteres (Kd 442-559 µM); TRPM4 mildly de-selected (3-4× vs flufenamic, borderline at 10 µM). Single-dose patch-clamp at 10 µM compound on each of TRPM4 (HEK293-TRPM4), KCNQ4 (HEK293-KCNQ4), BK (HEK293-BK), Cx50 (dye-transfer assay HEK293-Cx50).

Pass criteria (all 4): ≤ 25 % current inhibition at 10 µM. KCNQ4 expected easy pass per Phase 6c; TRPM4 is the nervous one.

Expected timing: 2 weeks parallel with 2.5b, ~$8k.

Combined 2.5 winnowing: 1-3 compounds out of G2 → 0-2 compounds carrying through to G3 cryoEM. If all 1-3 fail 2.5a (COX) or 2.5b (TMEM16A), feed back to medchem with the structure-activity data and trigger v4b carboxylate-deselect synth ladder. If 0/3 pass, hypothesis chemistry is wrong — not pocket-wrong.

GATE 3 — cryoEM co-complex + HEK293 E1659A functional rescue

3a cryoEM

Purpose: structural confirmation that compound binds K1141 pocket, not a decoy cryptic site.

Reagents: mini-STRC-UM-E1659A 5 mg/mL + compound 500 µM (2× saturation). Grids: Quantifoil R1.2/1.3 or gold-coated for anti-aggregation. Vitrify on Vitrobot Mark IV.

Collection: K3 or Falcon 4, 300 kV, 0.8 Å/px, 40 frames/movie, 40 e⁻/Ų. Target 4000 movies for 3–3.5 Å map.

Analysis: RELION-5 or cryoSPARC standard single-particle pipeline. Check ligand density in refined map at K1141 pocket. Bonus: unbiased difference map (apo − holo) at ≥ 4σ isolated at pocket coordinates.

GATE 3a PASS criteria:

  • Ligand density at Phase 4a pocket coordinates (centered near full-length K1141 / residue 548 in chain-A numbering) at ≥ 4σ in difference map.
  • Modeled pose within 1.5 Å RMSD of Phase 5b/5e best Vina pose.
  • Protein conformation within 2 Å Cα RMSD of apo (confirms pocket doesn’t collapse on compound dissociation).

Expected timing: 6–8 weeks (grid optimization + 3 days collection + 2 weeks processing), cost ≈ 3k labor.

3b HEK293 E1659A rescue assay

Purpose: functional endpoint — does the compound rescue surface trafficking of E1659A stereocilin in a cell?

Reagents:

  • HEK293-T stably transfected with STRC-FL-E1659A-GFP (or transient via Lipofectamine 3000).
  • Compound pre-treatment: 10 µM (therapeutic intracochlear target conc), 100 µM (saturating), in complete medium for 24 h.
  • Positive control: WT construct (full surface trafficking).
  • Negative: E1659A + DMSO (baseline impaired trafficking).

Readouts:

  1. Surface GFP vs total GFP (live-cell flow cytometry, non-permeabilized vs permeabilized).
    • WT: 65 ± 5 % surface.
    • E1659A + DMSO: 15 ± 5 % surface (impaired trafficking).
    • Rescue target: ≥ 30 % surface with compound.
  2. ER retention (co-stain BiP/GRP78) — immunofluorescence. Rescued cells show reduced BiP co-localization with stereocilin.
  3. Secretion into medium (Western blot supernatant) — confirm mature glycoform released.

Triage rank:

  • GREEN: rescue ≥ 30 % surface at 10 µM → carry to mouse OHC ex-vivo and toxicology.
  • YELLOW: rescue 15–30 % at 10 µM OR ≥ 30 % at 100 µM → medchem optimization.
  • RED: < 15 % rescue at any dose → kill compound, feed pocket info back.

Expected timing: 4 weeks cell-culture + assay dev + triage, cost ≈ $10k.

Tox pre-screen (concurrent with Gate 3b)

Per STRC h01 Fenamic Scaffold Tox Audit 2026-04-23: any parent fenamate hit must be profiled against cochlear ion-channel panel BEFORE mouse OHC work:

  • TRPM4 (patch-clamp in HEK293-TRPM4, 10 µM compound)
  • Cx50 gap junction (HEK293-Cx50, dye transfer)
  • BK channel (OHC-relevant, HEK293-BK)
  • KCNQ4 (DFNA2 gene — critical — HEK293-KCNQ4, whole-cell patch)
  • TMEM16A/ANO1 (stria vascularis)

Kill threshold: > 25 % inhibition at 10 µM on any cochlear-relevant channel.

Phase 8 output artifacts

For each candidate compound:

  • ThermoFluor ΔTm table (3 conc × 2 constructs)
  • MST Kd table (25 °C + 37 °C)
  • cryoEM map + PDB deposition (if G3a passed)
  • HEK293 rescue % surface (10/100 µM, n=6 biological replicates)
  • Ion-channel panel IC₅₀s (at least TRPM4, KCNQ4, Cx50)

Reported in proof note **STRC h01 Phase 8 Wet-Lab Triage Results** (placeholder — to be written after wet-lab).

Decision tree into Phase 9

  • 1–2 compounds full-pass G1+G2+G3 → Phase 9 mouse OHC ex-vivo + PK/PD + IND-enabling tox. Engage CRO (Charles River / Envigo).
  • 0–1 compound passes → Phase 3c v4 fragment-growing or Phase 3c v5 de novo RFdiffusion pocket design. Feed G1 SAR back to medchem.
  • All RED on G1 → pocket dead at chemistry; pivot to h01-alt (interface rescue at STRC dimer interface) or h11 (if open).

2026-04-24 refresh log

  • v3b delivered (29 YELLOW / 21 RED, 0 GREEN; ceiling Kd 4.6 µM on 3-amino-benzofuran-2-COOH cluster). Adds 15 YELLOWs to G1 candidate pool; synth Tier-2/3, CRO quote needed.
  • Phase 5e validated WT-pocket as proxy for E1659A → Gate 1 ThermoFluor on E1659A is the right validation path; mutant-only re-dock not required.
  • Phase 4h Tier-1 commercial set (tafamidis + iododiflunisal + diflunisal) elevated to “order this week, do not wait for v4”. Baseline Tm + positive-control calibration unblocks G1 the moment Sf9 protein production lands.
  • Phase 3c v4 fragment-grow launched 16:47 local 04-24 (ETA 17:57); top-N feed in within 24 h to expand G1 cohort.
  • Off-target panel queued as STRC h01 Phase 6c Off-Target Panel (cochlear ion channels TRPM4/Cx50/BK/KCNQ4/TMEM16A + COX-1/2) — runs concurrent with Gate 3b cell rescue.

Ranking delta

  • Hypothesis h01: no change; recorded for traceability.

Connections

Graph View

Backlinks