STRC Research

Polymer-Linked Ligand Dimers Spanning Allosteric Binding Sites — Kramer & Karpen 1998

3 min read

Polymer-Linked Ligand Dimers Spanning Allosteric Binding Sites — Kramer & Karpen 1998

Citation: Kramer, R.H. & Karpen, J.W. (1998). Spanning binding sites on allosteric proteins with polymer-linked ligand dimers. Nature, 395, 710-713. DOI: 10.1038/28021

Authors: University of Miami / University of Colorado School of Medicine.

The Strategy

Polymer-linked ligand dimers (PLDs): two identical ligands joined by a PEG chain of variable length. The idea: if a protein has two binding sites for the same ligand, a PLD where the linker length matches the site spacing will bind with dramatically higher affinity than either the monomer or a too-short/too-long PLD.

The System Studied

cGMP-gated (CNG) channels and cGMP-dependent protein kinase (PKG) — each have 4 cGMP binding sites. PLDs used: two cGMP moieties connected via 8-thio-cGMP to bifunctional PEG vinylsulphone. PEG MW ranged 282–20,000 Da, rms lengths 15–123 Å.

Key Results

Olfactory CNG channel (OLF):

  • cGMP K₁/₂ = 3.1 µM
  • 3400PEG-(cGMP)₂: K₁/₂ = 12 nM — 260-fold improvement
  • 20000PEG-(cGMP)₂: K₁/₂ = 250 nM (linker too long, lowers C_eff)
  • Optimal rms length: ~39 Å (2000PEG)
  • Hill coefficient drops from ~1.7 (cGMP) to ~1.1 (optimal PLD) — consistent with spanning two sites

Rod photoreceptor CNG channel (RET):

  • Optimal PLD: 1200PEG-(cGMP)₂, rms ~30 Å — shorter than OLF channel, indicating closer site spacing
  • Partial agonist behavior — polymer chain slightly interferes with gating

PKG:

  • Optimal: 282PEG-(cGMP)₂ (shortest tested, rms 15 Å)
  • K₁/₂ = 4 nM vs 120 nM for cGMP — 30-fold improvement
  • Indicates homologous PKG sites across subunits are <20 Å apart (not the 26 Å heterologous intrasubunit sites from crystal structure)

Overall: up to 1000× potency increase vs cGMP

Mechanism Confirmed

Dissociation kinetics: 3400PEG-(cGMP)₂ off-rate is 3400-fold slower than cGMP from OLF channel. When first cGMP dissociates, second has C_eff ~13.4 mM at 39 Å rms length — vastly higher than Kd (4 µM). Rebinding is near-certain before partner can also detach.

Competition experiment: adding free cGMP accelerates PLD dissociation (must displace the tethered ligand) — direct proof of bivalent simultaneous site occupancy.

The C_eff Formula

C_eff = 1000 / (N_A × 2/3 × π × r³)

At r = 39 Å: C_eff = 13.4 mM. Compare to Kd ~4 µM → C_eff/Kd ≈ 3400 → essentially guarantees rebinding = 3400× slower off rate = 3400× affinity enhancement.

Design Takeaway

You do not need to know the crystal structure of the target protein to find optimal PLD length — you scan empirically across a PEG length series. The potency peak tells you the approximate site-to-site distance. Selectivity between related proteins (OLF vs RET vs PKG) comes naturally because each has a different optimal PLD length.

Relevance to STRC

STRC protein has known interaction domains with tectorin and collagen. If any of these domains present two binding sites within 15-120 Å of each other, a PLD strategy could yield high-affinity stabilizers with 100–1000× potency improvement over monovalent peptide fragments. No crystal structure required — just a PEG length scan plus functional assay.

Connections

Graph View

Backlinks