STRC Research

STRC Pharmacochaperone K1141 Fragment Pocket

4 min read

STRC Pharmacochaperone K1141 Fragment Pocket

A druggable allosteric subpocket sits 3.2 Å from K1141 NZ on WT STRC (AF3 Job 4), 9 Å from the mutation-responsive loop 1642-1651. Combined volume 159 ų (two adjacent subpockets, 99 + 60 ų), druggability 0.86. A ligand docked here directly replaces the lost E1659 carboxylate contribution by donating a H-bond acceptor to K1141 NZ — textbook VX-809 mode.

Scan method

Stand-alone LIGSITE-style grid cavity detection (no external tools) on WT Job 4 CIF, 18 Å sphere around the centroid of loop 1639-1651. Parameters: 0.8 Å grid, 14-ray enclosure, 9 of 14 rays must hit protein within 9 Å (moderate buriedness — the loop-LRR interface is a cleft, not a buried cavity; strict 11/14 thresholds missed it). Primary clusters split into subpockets via depth-maximum seed expansion. Scoring: volume (peak 250 ų), hydrophobic fraction (0.4-0.7 optimum), lining-residue count (10-22 optimum), H-bond donors + acceptors, mean burial depth.

Result: 5 primary clusters → 19 subpockets → 9 loop-facing → top druggability 0.86.

The two top subpockets merge into one fragment-sized site

metricsubpocket #1subpocket #2
volume99 ų60 ų
druggability0.860.80
hydrophobic frac0.440.57
depth max4.7 Å5.4 Å
d → loop centroid9.3 Å8.9 Å
d → E1659 CA10.8 Å13.0 Å
d → K1141 NZ7.3 Å3.2 Å

Centroids are 5.1 Å apart → the two subpockets are adjacent sub-domains of one binding site, combined V = 159 ų (fragment-merge sized). Subpocket #2 wraps the K1141 side-chain tip.

Docking box

  • Center (WT Job 4 frame): (7.7, −5.4, −41.5) Å
  • Box size: 18 × 18 × 18 Å (covers both subpockets + 5 Å tolerance)
  • Core anchor residues (in both subpockets): D1140, K1141, R1169, D1173, G1645, F1646, G1647
  • Full lining (23 residues union): 1140, 1141, 1142, 1143, 1144, 1168, 1169, 1170, 1172, 1173, 1612, 1642-1648, 1651-1656

The site spans three structural regions: LRR concave face (1140-1173), linker (1612), and mutation-responsive loop (1642-1656). Binding here caps the loop against its LRR-face anchor — stabilises the WT-like conformation in the MUT background.

Pharmacophore hypothesis

  • 1 H-bond acceptor / carboxylate → K1141 NZ (core salt-bridge rescue)
  • 1 H-bond donor or acceptor → D1140 / D1173 (supporting contacts)
  • 1 aromatic / hydrophobic cluster → F1646 (directional π-stacking)
  • Flexible linker spanning 6-10 Å between pharmacophores
  • MW 250-350, logP 2-4 (fragment-merge starting point)

Why this is fragment-mergeable, not HTS-ready

V = 159 ų is on the small end of lead-drug territory (leads typically 200-500 ų). The site is a cleft open to solvent (not deeply buried — max depth 4.7-5.4 Å), so an initial fragment hit at V ≈ 100 ų can be elaborated by growing into the adjacent subpocket across 5 Å. Fragment library of choice: DSi-Poised Library (2 k compounds, Diamond Light Source) — designed for growable scaffolds — or Enamine Fragment Collection (40 k).

Next steps

  • Phase 3 docking: AutoDock Vina + GNINA CNN rescore against box (7.7, −5.4, −41.5) Å, 18³ ų. Filter: K1141-NZ H-bond geometry (≤3.5 Å, acceptor angle 100-140°) + F1646 contact. Top 50.
  • Phase 4 MD validation: 50 ns × 3 GROMACS ff19SB+GAFF2+TIP3P. Metrics: ligand RMSD <3 Å vs docked pose, K1141 contact persistence >60% frames, loop (1642-1651) Cα-RMSD back <2 Å in MUT background.
  • Phase 5 FEP+: target ΔΔG_rescue ≤ -4 kcal/mol (restores ≥50% of 8.4 kcal/mol lost binding energy).
  • Selectivity panel: hERG, CYP3A4, HSA.

Files / Models

  • ~/STRC/models/pharmacochaperone_phase2_pocket_scan.py — initial broad scan (8/14 rays, superseded by Phase 2B)
  • ~/STRC/models/pharmacochaperone_phase2b_subpockets.py — the LIGSITE-style scan that produced the numbers here
  • ~/STRC/models/pharmacochaperone_phase2b_results.json — full cluster + subpocket table
  • ~/STRC/models/pharmacochaperone_phase2b_top_subpocket.pdb — top subpocket as dummy atoms (PyMOL load)
  • ~/STRC/models/pharmacochaperone_phase2b.png — figure

Connections

Graph View

Backlinks