# EoE Lead Binder Spec Sheets

## CCL26 neutralizing binder — spec sheet

**Design ID:** ccl26_design_39_T0.3  ·  **RFdiffusion→SolubleMPNN→Boltz-2 pipeline**

| Property | Value |
|---|---|
| Binder length | 67 aa |
| Interface ipTM | **0.942** |
| Complex pLDDT | 0.779 |
| pTM | 0.956 |
| Boltz confidence | 0.811 |
| Buried surface area | 1220.2 Ų |
| Target interface residues | 25 |
| Binder interface residues | 27 |
| Epitope coverage | 4/5 hotspots (16, 54, 55, 56) |
| On-target score | 0.848 |
| SolubleMPNN score | 1.528 |

**Binder sequence:**
```
SEQLKEEEEKKKLEEERREEEEEAREEEASRKAAAQRAAVDPEGANKELEELERARKEAEELYKKAR
```

**Mechanism:** engages CCL26's C-terminal α-helix basic cluster (R54/K55/K56) plus the N-loop (H16) — the surface that mediates both CCR3 receptor engagement and glycosaminoglycan binding (PMID 35742962). Occluding this face is designed to block eosinophil recruitment at the chemokine-presentation step. The binder is acidic/helical, which is electrostatically complementary to CCL26's basic GAG-binding surface — mechanistically consistent, but see caveats.

---

## POSTN neutralizing binder — spec sheet

**Design ID:** postn_design_39_T0.1  ·  **RFdiffusion→SolubleMPNN→Boltz-2 pipeline**

| Property | Value |
|---|---|
| Binder length | 81 aa |
| Interface ipTM | **0.901** |
| Complex pLDDT | 0.911 |
| pTM | 0.941 |
| Boltz confidence | 0.909 |
| Buried surface area | 842.2 Ų |
| Target interface residues | 22 |
| Binder interface residues | 20 |
| Epitope coverage | 3/4 hotspots (83, 85, 108) |
| On-target score | 0.788 |
| SolubleMPNN score | 0.943 |

**Binder sequence:**
```
MTKEQNEAEAEAMKLMLEARRLAAEGAAGPGVEELTKKALEKLKEAIEKAKKTGKERREKLIKQAKDLEEDAKARAAAKLA
```

**Mechanism:** engages the exposed integrin-interaction patch on POSTN's FAS1-IV domain (residues 83/85/108). Blocking this surface is designed to interrupt periostin-integrin signaling that drives esophageal remodeling and the DSG1-loss→POSTN feed-forward loop — an anti-remodeling mechanism addressing disease that persists in histologic remission (PMID 39343172).

---

### Shared caveats
- **In silico only.** ipTM/pLDDT are model confidences, not measured affinities. These designs require
  wet-lab validation: expression, SEC/SPR binding assays, and functional (chemotaxis / remodeling) readouts.
- **Sequence composition.** Both leads are helical and charged-rich (a SolubleMPNN bias on these RFdiffusion
  backbones). This favors solubility but warrants a specificity check — confirm they do not bind off-target
  acidic/basic surfaces — and possibly a second design round with tighter compositional constraints.
- **Target-form note.** CCL26 target = mature NMR chemokine (1G2S). POSTN target = isolated FAS1-IV domain
  (5WT7); full-length periostin has multiple FAS1 domains, so a binder to one domain is a starting point,
  not a validated whole-protein neutralizer.