STRC Research

STRC K1141 Fragment Construct Strategy

3 min read

STRC K1141 Fragment Construct Strategy

Distilled from DR3 — CRO Wet-Lab Vendor Menu §I.2. Independent of vendor; constrains every downstream biophysics + structural in-silico job on H01.

Why a fragment, not full-length

Full-length STRC = 1775 aa, GPI-anchored, heavily glycosylated. Three intrinsic problems for any direct binding measurement:

  1. GPI signal sequence drives micelle/inclusion-body aggregation on cellular extraction.
  2. MW asymmetry — small ligand (~350 Da) on a 200 kDa target collapses SPR/BLI ΔRI signal-to-noise.
  3. Disordered flanking regions confound thermal-shift and crystallographic packing.

Truncation to the K1141 pocket domain (~150 aa) removes all three. Folds as a discrete soluble globular unit; preserves the K1141/E1659 spatial unit; stays inside HEK293/Sf9 secretory throughput.

Implication for STRC computational pipeline

Same fragment is the right MD/AF3 unit. Phase 5d full-length AF3 mutant MD ran 651k atoms and capped at 2 ns wall-feasible (14 ns/day OpenMM Metal). A K1141-pocket ~150-aa fragment at the same solvation density is ~4× smaller (~150-180k atoms) → expect ~50-60 ns/day on the same Mac silicon → 100 ns trajectories become local-overnight, not GPU-rental territory.

Construct boundaries are the open question. The pocket residues (K1141 + ring residues in STRC Pharmacochaperone K1141 Fragment Pocket) need to span enough secondary-structure context that the fold is stable. Phase 5c established K1141 Cα RMSF 0.62 Å in full-length context; a fragment that drops below ~120 aa risks losing that rigidity. The right ~150-aa boundary is a Phase 5l candidate: AF3 a 120/150/180/220-aa fragment ladder, score by pLDDT in the pocket region + 50-ns MD K1141 RMSF vs full-length reference (target ≤ 0.8 Å).

Construct engineering specifics for downstream biophysics

If/when biophysics happens, the fragment must carry:

  • N-term mammalian secretion signal (IgK or tPA leader) → secretory-pathway folding.
  • C-term cleavable His6 + AviTag (GLNDIFEAQKIEWHE) → BirA-catalyzed single-point biotinylation → uniform streptavidin-capture orientation on SPR sensor → K1141 pocket guaranteed solvent-exposed.

Without AviTag, random amine-coupling presents the K1141 face down on the chip in ~50% of immobilized molecules — the resulting heterogeneity inflates apparent K_d and corrupts k_off fit.

Expression-system selection drives glycan environment

  • HEK293 transient (HD) — human-like complex glycans; ~4-6 wk to mg-scale. Default if K1141 pocket is sterically bordered by an N-linked glycan.
  • Baculovirus/Sf9 — paucimannose; cleaner pocket but non-human glycan periphery. Fallback if HEK293 pool yields are <1 mg/L.
  • E. coli — disqualified. No PTMs, no disulfide isomerase, fragment will refold incorrectly even with Origami strain.

The choice matters computationally too: if HEK293 is the eventual expression route, then any future glycan-aware MD on the fragment should add the human-pattern N-glycan tree at the relevant Asn rather than naked surface.

Connections

Graph View

Backlinks