STRC Research

STRC Pharmacophore Model K1141 Pocket

3 min read

STRC Pharmacophore Model K1141 Pocket

A 3D pharmacophore extracted from WT-STRC Job 4 within 12 Å of the Phase-2B K1141 pocket centroid (7.7, −5.4, −41.5). The pocket offers 65 protein features the ligand must complement: 23 backbone C=O (need ligand H-bond donors), 21 backbone N–H (need ligand acceptors), 10 hydrophobic centroids, 4 aromatic ring centroids, 4 anionic (D/E) sidechains, and 3 cationic sidechains (K1141-NZ, R1169). Anchor triangle for ligand positioning: K1141-NZ, F1646 ring, G1645-O. Key design constraint: ligand needs a carboxylate/acidic anion placed 3–3.5 Å from K1141-NZ AND an aromatic centroid 5.7 ± 1.5 Å from that anion — matching the protein’s native K1141↔F1646 geometry.

Anchor triangle (protein features driving ligand orientation)

featurexyz (Å)role for ligand
K1141 NZ(4.79, −6.74, −45.37)salt-bridge anchor — lost E1659 carboxylate rescue
F1646 ring centroid(6.49, −11.02, −42.02)aromatic π-stack partner
G1645 backbone O(3.39, −6.22, −43.70)H-bond donor target

Pairwise distances:

  • K1141 NZ ↔ F1646 ring: 5.69 Å (ideal cation-π / salt-bridge-to-aromatic distance)
  • K1141 NZ ↔ G1645 O: 2.47 Å (already H-bonded in WT — ligand must NOT displace this)
  • K1141 NZ ↔ G1647 O: 8.59 Å (reaches farther subpocket)

The K1141-NZ / G1645-O H-bond (2.47 Å) is already load-bearing in WT. In E1659A, loss of the E1659 OE1/OE2 carboxylate leaves K1141 under-coordinated on one face. A pharmacochaperone that donates a carboxylate to the opposite face of K1141-NZ (3-3.5 Å, approaching from the pocket centroid direction) restores the lost contact without disturbing the G1645-O H-bond.

Ligand design rules (the target pharmacophore)

  1. Required: ionizable anion at pH 7 — carboxylate (−COO⁻), tetrazole, sulfonate, sulfonamide, acyl sulfonamide, or phosphonate.
  2. Required: anion oxygen within 3.5 Å of K1141-NZ when docked (salt bridge rescue).
  3. Preferred: aromatic ring centroid within 4.2–7.2 Å of the anion → match the K1141-NZ ↔ F1646 5.7 Å distance for simultaneous engagement.
  4. Preferred: at least one ligand H-bond donor (N–H, O–H) to reach G1645/G1647 backbone carbonyls.
  5. Size: MW 150–350 Da (pocket volume 159 ų); rule-of-3 preferred (MW <300, logP <3, HBD ≤3, HBA ≤3, rot ≤3).
  6. Polarity: TPSA 40–90 Ų (balanced for endolymph penetration, middle-ear topical delivery).
  7. Keep-out: none within 12 Å — the E1659A cavity is not a keep-out; ligand may optionally extend into it.

Feature counts (within 12 Å of pocket centroid)

feature kindcountligand counter-feature
hb_donor_target (protein C=O)23H-bond donor
hb_acceptor_target (protein N–H)21H-bond acceptor
hydrophobic_target10hydrophobic
aromatic_target4aromatic
cation_target (D/E)4donor/cation (careful — many are already D1140/D1173 self-salt-bridged)
anion_target (K/R)3ANION ← primary

Primary anchor (anion target) = K1141-NZ. Secondary anions are occupied (D1140 and D1173 form a local cluster with R1169 — ligand should not compete).

Disease-residue delta

E1659 WT side chain loses CG, CD, OE1, OE2 (4 heavy atoms, ~45 ų) in E1659A. This leaves a cavity near the loop. Phase 2B pocket centroid is 10–13 Å from E1659 Cα — the drug binds near but does NOT fill the mutation site directly. It works allosterically: stabilising the loop’s folded position against the LRR face by engaging K1141.

Replication

cd ~/STRC/models
python3 pharmacochaperone_phase3a_pharmacophore.py
# reads: job4-wildtype.cif, job3-mutant.cif
# writes: pharmacochaperone_phase3a_pharmacophore.json

Deterministic: identical input CIFs + same Python/Biopython yield identical JSON.

Files / Models

  • ~/STRC/models/pharmacochaperone_phase3a_pharmacophore.py — extractor
  • ~/STRC/models/pharmacochaperone_phase3a_pharmacophore.json — 65-feature JSON, anchor triangle, design rules
  • ~/Sites/site-strc-egor-lol/public/models/job4-wildtype.cif — input (AF3 20-job set)

Connections

Graph View

Backlinks