Ice nucleation without a membrane anchor
Two candidates, both soluble. Lead A (DBM-INP1) is a C3-symmetric bacterial homotrimer presenting three 120° ice-binding faces to solution. Lead B (DBM-INP2) (from Fusarium avenaceum) is a fungal β-hairpin sandwich whose 6.5 Å cleft is modeled to template 2D bilayer hexagonal ice. Both structures are predicted models assessed in explicit-solvent MD; no wet-lab assay data is reported here.
Development stage gates
Standardized discovery progression gates across the platform.
- CRO Synthesis Queued
- Lead A, the bacterial trimer (DBM-INP1). Expression cassettes designed, codon-optimized, and queued for commercial CRO gene synthesis and expression (U.S. provisional application 64/165,496).
- USPTO Patent Filed
- Lead B, the Fusarium ice press (DBM-INP2). Priority patent application filed with the USPTO (application 64/162,890) covering novel ice-templating architecture.
Lead A: soluble C3-symmetric homotrimer
Bacterial INPs such as Pseudomonas syringae InaZ rely on an insoluble outer-membrane anchor (MNLDKALVLR...), which drives inclusion-body aggregation in recombinant hosts. Candidate DBM-INP1 carries no such anchor:
1. Three radial ice-binding faces at 120°
Three identical 514-AA protomers assemble along a central C3 axis. The hydrophobic backbones interlock in the core, leaving three solvent-exposed TxT and SxL ice-nucleating planes at 120° offsets.
2. Self-capping stalk (residues 466–514)
Three 49-AA C-terminal α-helices twist into a central 3-helix coiled-coil pin. In the model this shields the hydrophobic core from proteolysis and should limit rung fraying at snow-gun pressures (300–800 psi).
- Top apex (1–25 AA ×3) Three-fold crown
- Central core (26–465 AA ×3) 3× 120° radial ice faces
- Base pin (466–514 AA ×3) 3-helix coiled-coil stalk
- Predicted Tonset -1.7°C to -2.5°C (Class A)
Lead A biophysics: DBM-INP1 C3 homotrimer
| Biophysical parameter | Predicted specification | Conventional InaZ (Snomax) | Engineering rationale |
|---|---|---|---|
| Quaternary architecture | Soluble C3 homotrimer (3× 514 AA = 1,542 AA) | Membrane-anchored monomer (~150 kDa) | Soluble assembly; no membrane fraction required. |
| Radial ice-binding faces | 3× 120° radial TxT / SxL planes | Single planar face (requires membrane lipid pool) | Three ice-binding faces per macromolecule instead of one. |
| Terminal capping stalk | 49-AA 3-helix coiled-coil pin (466–514) | Unprotected β-solenoid ends | Locks the C-terminus; expected to reduce rung fraying in high-shear nozzles. |
| Membrane anchor dependency | No transmembrane segment | Obligate hydrophobic N-anchor (MNLDKALVLR) | Allows secretory expression in K. phaffii or cytosolic E. coli BL21(DE3). |
| Nucleation onset temperature | Tonset = -1.7°C to -2.5°C | -3.0°C to -5.0°C (killed cells) | Class A onset range without membrane scaffolding. |
>DBM-INP1_Mature_Monomer [U.S. Provisional Application 64/165,496, 514 AA]
[514 AA COMPOSITION-OF-MATTER & C3 HOMOTRIMER ATOMIC COORDINATES PROTECTED UNDER BILATERAL DELAWARE CDA // ACCESS VIA FORMAL EVALUATION OPTION]
Genomic synteny and 30-kb island (U.S. provisional application 64/165,496)
Candidate DBM-INP1 comes from an environmental metagenome-assembled genome, on a contiguous >1.4 Mb scaffold with no truncation. The gene sits in a 30-kilobase ice nucleation island:
porV, sprF, gldA-H) for native extracellular export.Lead B biophysics: Fusarium bilayer ice press (DBM-INP2)
What the ordered peptide is meant to settle. Ice nucleation in Fusarium is a documented phenomenon — Pouleur et al. (1992) found it across 11 species — and the published measurements already describe a small nucleator: a ~5,300 Da MALDI-TOF peak, SEC fractions at 12 kDa and above 660 kDa, cell-free activity in filtered broth down to 1.41 ng/mL, and nucleation between −2 °C and −5 °C. What this platform adds is an unannotated 32–40 residue sORF whose 3,961 Da peptide plus eight mannoses comes to 5,258 Da, within 0.8% of that measured mass, and a model of how its 6.5 Å cleft would template bilayer hexagonal ice (Koga, Zeng & Tanaka, Nature 388, 1997). The size is not the anomaly; it is the match.
Three questions stay open, and the synthesis is ordered to put them to a bench. Whether this sORF is in fact the molecule behind the measured activity — the identification is computational, however well the mass agrees. Whether a synthetic peptide reproduces the behavior of a form that is natively secreted and glycosylated, which solid-phase synthesis does not provide. And whether the peptide self-assembles into the modeled bilayer under assay conditions: the MD here models the assembled state rather than testing whether assembly happens. Those are the experiment, not caveats attached to a result.
| Biophysical parameter | Predicted specification | Conventional open-surface INPs | Physical mechanism |
|---|---|---|---|
| Monomer size | 39 amino acids (3,961 Da) | 1,200 – 1,500 AA (120–160 kDa) | Compact β-hairpin sandwich with few exposed cleavage sites. |
| Catalytic cavity | 6.5 ± 0.3 Å internal cleft | Open 2D planar sheet | Confinement width matches 2D bilayer hexagonal ice (6.44 Å). |
| Antiparallel Thr registry | i + j = 31 (Thr5–Thr26, Thr7–Thr24, Thr9–Thr22) | Parallel TxT ladders | Opposing threonine hydroxyls orient water dipoles into a hexagonal ice seed. |
| Ice lattice alignment | 4.79 Å lateral • 6.89 Å longitudinal pitch | 4.52 Å × 7.36 Å (ice Ih prism face) | Opposed biaxial strain (+6.0% / -6.3%), modeled to accelerate 2D freezing. |
| Nucleation onset temperature | Tonset = -2.0°C to -4.0°C | -5.0°C to -8.0°C (fragmented) | Nanoconfinement is modeled to pre-pay part of the entropic cost of freezing. |
>DBM-INP2_Mature_39AA_Ice_Press [U.S. Provisional Application 64/162,890]
[39 AA COMPOSITION-OF-MATTER & 6.5 Å BILAYER CLEFT COORDINATES PROTECTED UNDER BILATERAL DELAWARE CDA // ACCESS VIA FORMAL EVALUATION OPTION]
12× O-mannosylation brush and tetrahedral order (qtet)
Filamentous fungi protect short β-hairpin scaffolds from amyloid aggregation by O-mannosylation. All-atom GLYCAM-06 simulation of DBM-INP2 shows two effects:
Explicit-solvent molecular dynamics (OpenMM 8.0 / AMBER14 / TIP4P/Ice)
In explicit solvent the cleft hydrates and expands on its own, and the scaffold keeps its registry on a hexagonal ice slab. The behavior is therefore not an artifact of the in vacuo model. All values below are simulation output, not assay data.
| Metric | Simulation conditions | MD observation | Interpretation |
|---|---|---|---|
| Basal epitaxial H-bonds | 979-water ice Ih basal slab at 265.15 K | 182.8 persistent basal H-bonds | Continuous coordination pins liquid water to the ice lattice. |
| Spontaneous cleft expansion | Explicit water solvation at 268.15 K | 2.84 Å → 3.77 Å (+33%) | Reverses the in vacuo collapse and opens a sub-nanometer slit pore. |
| Internal cavity hydration | 50 ns solvated trajectory | 237.5 waters (peak 249) | The internal water population orders into 2D bilayer hexagonal ice. |
| Multimer stability | Hexameric / tetrameric assembly | Backbone RMSD: 1.31 Å | The β-scaffold holds its fold over 50 ns at sub-zero temperature. |
| O-mannosylation brush | 12× O-glycosylation | >71.55 Å projection into bulk solvent | The brush is expected to block cross-β aggregation; the cleft stays 100% accessible in the simulation. |
Full 50 ns explicit-solvent trajectory checkpoints (.dcd / .xtc), water dipole orientation maps, and free-energy profiles (ΔGnuc) are available to prospective licensing partners.
Food labeling and regulatory position
| Compliance dimension | Bacterial Snomax (P. syringae) | DarkBiome Fusarium candidate | Commercial implication |
|---|---|---|---|
| FDA regulatory precedent | Class 1 plant pathogen; not approved for food | FDA GRAS Notice GRN 000091 | The genus has food precedent (Quorn mycoprotein); this protein would still need its own review. |
| Endotoxin (LPS) content | High Gram-negative lipopolysaccharide load | 0% LPS (eukaryotic host) | Avoids the depyrogenation filtration train ($250k+). |
| Downstream processing cost | High (cell lysis, membrane fraction isolation) | 60%–75% lower DSP CAPEX (projected) | Secretion allows direct microfiltration and tangential-flow concentration. |
| Fermentation | Low-titer batch bacterial cultures | >5 g/L target in Aspergillus / Pichia | Airlift fermentation yields mycoprotein biomass plus secreted INP in the supernatant. |
Applications and licensing framework
| Sector | Market size (TAM) | Problem | Candidate fit |
|---|---|---|---|
| Lyophilization excipients | $2.5B | Stochastic ice nucleation gives vial-to-vial variability when freeze-drying mRNA lipid nanoparticles and monoclonal antibodies. | Lead B (DBM-INP2): controlled nucleation between -2.0°C and -4.0°C. |
| Ski-resort snowmaking | $1.2B – $2.5B | Snowmaking fails at marginal temperatures (-2°C to 0°C), and Snomax is restricted in some sensitive watersheds. | Lead A (DBM-INP1): predicted onset -1.7°C, capped stalk, non-pathogenic source organism. |
| Food cold chain | $1.8B | Large ice crystals rupture cell membranes in frozen beef, fish, and berries, causing drip loss and texture damage. | Lead B (DBM-INP2): nucleating warmer should give smaller, more uniform crystals. |
| Cloud seeding | $600M | Silver iodide is ineffective in warm clouds (T > -6°C) and accumulates in reservoirs. | Leads A and B: protein nucleators with predicted onset between -1.7°C and -4.0°C. |
Request a 90-day INP evaluation option
The option suits food-freezing, pharmaceutical formulation and snowmaking groups that want to measure freezing onset on their own droplet-freezing arrays before taking a license.
- Lead A (DBM-INP1 trimer) or Lead B (DBM-INP2 ice press) purified protein aliquots (50 mg to 250 mg), synthesized to order.
- Droplet-freezing assay protocol for measuring Tonset.
- 100% of the evaluation package fee credited toward the upfront license fee.