Protein Design

Alongside experimental structural biology, we use structure-based generative models to design de novo protein binders against membrane-protein targets (GPCRs and solute carriers), aimed at defined epitopes to serve as crystallisation/cryo-EM fiducials, purification reagents, conformational stabilisers, or starting points for functional modulators. The folds span natural antibody-like scaffolds (Ig/VHH) and genuinely novel de novo topologies with no natural structural homolog (novelty is scored per design).

Designs are generated and scored with the Boltzlab BoltzMol / BoltzProt API (boltz.bio/boltzmol-boltzprot-api).

The models below are the best-scoring candidates from recent campaigns, each shown on the full receptor/transporter: click a card to load the interactive 3D model. The designed binder is in colour; the target is in grey; for the transporters, the N-terminal bundle (light) and C-terminal bundle (dark) are shaded separately. These are predicted designs (no experimental validation yet); reported affinities are computational proxies.

Designed Ig-fold binder (magenta) docked on the full GPR75 receptor (grey)
GPR75 (orphan class-A GPCR): Ig-fold binder tOZRH at the intracellular (G-protein-coupling) face, on the full receptor (cofold). Design ipTM 0.92.
Designed helical mini-binder (blue) at the intracellular face of the full viral GPCR US28 (grey)
US28 (viral HCMV GPCR): mini-binder Yq1Cxu with a novel de novo fold (~64% novel) at the intracellular Nb7 (G-protein-coupling) face, on the full receptor. Co-fold ipTM 0.96, ~1100 Ų buried, ΔG −9.9 kcal/mol (~57 nM). An intracellular conformational stabilizer, not an extracellular ligand.
Designed beta-sheet binder (amber) on the full SPNS2 transporter, with N-terminal (light) and C-terminal (dark) bundles
SPNS2 (S1P transporter): β-sheet binder at the nbD12 epitope on the intracellular (cytoplasmic) face, modeled onto the cryo-EM structure (8QV6, inward-open). ipSAE 0.73, ΔSASA ~1290 Ų.
Designed helical bridging bundle (green) clamping the full SLC15A4 transporter, with N-terminal (light) and C-terminal (dark) bundles
SLC15A4 (lysosomal histidine/peptide transporter): novel de novo helical bundle (~75% novel) designed to cross both N- and C-terminal bundles and lock the transporter in its outward-open conformation, deliberately trading away interface score for that constraint (ipTM 0.78), on the luminal (lysosomal) side; modeled onto 8QSK.
Designed Ig-fold binder (violet) on the full SLC15A4 transporter, with N-terminal (light) and C-terminal (dark) bundles
SLC15A4 (lysosomal histidine/peptide transporter): Ig-fold binder (~71% novel sequence) optimized for maximal predicted binding at the Fab107 luminal (lysosomal) epitope, modeled onto 8QSK (outward-open). ipTM 0.88, ΔSASA ~1710 Ų, ΔG −12.9 kcal/mol (~0.3 nM).
SLC15A4 dimer (two grey protomers) each carrying a violet Ig-fold binder on its distal face, with the dimer interface unobstructed
SLC15A4 dimer (two protomers, grey; 8QSK, outward-open): the Ig-fold binder modeled on both protomers (violet) engages the distal faces, leaving the dimer interface unobstructed, so binding is predicted not to perturb dimerisation.