H01 — Ward Dual-Cassette Feasibility¶
Pre-registration note. The claim, assumption stack, and killshot menu below are frozen as of the first commit of this file (2026-04-24, git SHA TBD at commit time). Subsequent edits log in git with rationale in the commit message. See README.md and ../linter-design.md §6 for the convention.
Claim¶
Ward's A. awamori glucoamylase-KEX2 dual-cassette architecture, validated for single-cassette human lactoferrin (hLf) expression at >2 g/L submerged (Ward 1995, PMID 9634791), can be layered with a second expression cassette for A. flavus uricase (uaZ) in A. oryzae solid-state rice koji fermentation without silencing either heterologous protein or disrupting native metabolite production (kojic acid, ergothioneine).
This is the single gating feasibility test for the koji endgame strain thesis. If it passes, Year 2–3 koji development converges on one engineered strain delivering four molecules covering five NLRP3-pathway chokepoints. If it fails, the platform falls back to the two-strain co-ferment path (koji-endgame-strain.md §4.1) which preserves the coverage matrix at the cost of single-strain elegance.
Assumption Stack¶
-
Ward 1995 results translate from A. awamori to A. oryzae. — load-bearing. The two species share >99.5% coding-region identity for secretion-machinery genes (glucoamylase, KEX-2-family endoprotease, ER folding chaperones). A. oryzae has independently been validated as an hLf expression host at 25 mg/L in Ward 1992 (PMID 1368268) — the floor, not the ceiling — so the species-translation assumption is supported at the lower titer and extrapolated at the Ward 1995 higher titer. Evidence level: In Vitro (Ward 1992 direct; Ward 1995 by extrapolation).
-
Solid-state rice koji fermentation supports the protein-folding + secretion loads of both cassettes simultaneously. — load-bearing, no direct published precedent. Ward 1995 was submerged culture with defined iron supplementation and controlled O₂/pH/feed. Solid-state rice koji has different mass transfer, different water activity, different available iron, different O₂ gradients. No published paper demonstrates dual-heterologous-protein expression in A. oryzae solid-state koji at g/L-scale titers. Evidence level: Mechanistic Extrapolation.
-
Native A. oryzae proteases do not degrade the expressed uricase or lactoferrin at substrate-relevant rates. — load-bearing. A. oryzae secretes a suite of proteases (alkaline serine [alp], neutral metallo [npr], acid-stable aspartic) as part of its starch-degrading lifestyle. Lactoferrin is moderately protease-resistant (pepsin digestion generates active lactoferricin rather than full degradation); uricase protease susceptibility in this matrix is empirically open. Protease-knockout host strains (Δalp, Δnpr) are available as a fallback but complicate the selection / integration design. Evidence level: Mechanistic Extrapolation.
-
Iron-binding capacity of recombinant lactoferrin is preserved in solid-state rice conditions. — supporting. Published Lf work in Aspergillus is submerged culture with defined Fe³⁺ supplementation (Ward 1995; Sun 1999 PMID 10089347 confirmed native fold and iron-binding at 2.2 Å resolution). Rice grain has low free iron (~1–3 ppm bioavailable); rice bran is higher but variable. Fe supplementation of the solid-state substrate (FeCl₃ or iron citrate at 10–100 ppm) is likely the workaround if native iron is insufficient, but the assumption that solid-state-produced Lf retains apo/holo-switchable iron-binding is untested. Evidence level: In Vitro (submerged only).
-
Selection markers for the two cassettes can co-exist without metabolic competition or cassette loss during propagation. — stylistic / routine. pyrG (uracil biosynthesis), niaD (nitrate reductase), amdS (acetamidase), and ptrA (pyrithiamine resistance) are the standard food-grade-compatible marker set for A. oryzae. Pairwise combinations are routine in the filamentous-fungus literature. Cassette copy-number stability over industrial-scale propagation (~10⁹-fold expansion from master seed to production run) is a standard production-engineering question, not a scientific one. Evidence level: In Vitro (established industrial practice).
-
KEX-2 processing capacity in A. oryzae is sufficient for fusion-based cassettes at g/L-scale combined titer. — load-bearing if both cassettes use fusion architecture. If only the lactoferrin cassette is a glucoamylase-KEX2 fusion and the uricase cassette uses direct secretion with its own signal peptide, KEX-2 capacity is non-competing. If both cassettes depend on endogenous processing peptidase, KEX-2 becomes a shared resource and the dual-cassette strain could saturate it. Published A. oryzae KEX-2 capacity studies are thin. Evidence level: Mechanistic Extrapolation.
-
Fungal-style glycosylation of solid-state koji lactoferrin remains within the Ward 1995 submerged envelope. — supporting. Aspergillus hLf has simpler mannose-rich N-glycans vs. native milk hLf complex sialylated/fucosylated glycans (Almond 2012 PMID 23012214). The glycosylation difference makes recombinant Lf ~40× less immunogenic and ~200× less allergenic in BALB/c mice — potentially a feature for chronic oral dosing. Whether solid-state fermentation further shifts glycosylation (water-activity-dependent mannosyltransferase activity is documented in other Aspergillus species) is empirically open. Evidence level: Animal Model (submerged material only).
Killshot Menu¶
Ranked by score = (kill_pr × info_weight) / (cost × time_penalty) per linter-design.md §4. Sorted descending. Higher score = run first.
| # | Killshot | Cost | Weeks | kill_pr | info_weight | Failure modes | Score (rel.) |
|---|---|---|---|---|---|---|---|
| 1 | Literature/patent landscape deep-dive on A. oryzae dual-heterologous-cassette expression + solid-state koji multi-protein precedent. Scope: PubMed + Google Scholar + Espacenet patent search; adjacent precedents (Wang 2023 A. niger, Li 2024 A. oryzae multi-copy same-protein). Cheapest possible upstream move — answers whether the assumption "no published precedent" is actually correct, or whether we've missed a directly-relevant paper. If a published dual-cassette A. oryzae precedent exists, it either validates the hypothesis cheaply or reveals a failure mode we missed. | $0 | 1 | 0.3 | 0.9 (root — tests assumption 2 and partially 1) | published-literature-gap, species-gap-translation | HIGHEST — run first |
| 2 | Dual-cassette transformation + fermentation + quantified readouts. Both cassettes integrated sequentially (Cassette A lactoferrin PamyB-glucoamylase-KEX2site-hLf-TamyB with pyrG selection; Cassette B uricase PTEF1-amyB_SP-uaZ-TgpdA with niaD selection). Solid-state rice koji 48–60 h at 30°C, 35% moisture, with submerged-culture parallel control. Readouts: uricase activity (UA-disappearance assay), Lf titer (ELISA + Western), iron-binding (UV-Vis 465 nm), native metabolite panel (kojic acid HPLC, ergothioneine LC-MS), qPCR for cassette copy numbers, SDS-PAGE for truncated/unprocessed species. This is the full §3.4 protocol from koji-endgame-strain.md. | $5,000 | 12 | 0.5 | 1.0 (root — directly tests the claim) | expression/localization-mismatch, kinetics/concentration, assay-specificity | High |
| 3 | Single-cassette uricase-only control in A. oryzae RIB40 solid-state rice koji. Tests whether uricase expresses in solid-state format at all, independent of the dual-cassette question. If single-cassette uricase in solid-state koji is already <50 mg/g, the dual-cassette target is off before Ward layering even starts. This is the Year 1 starting strain per engineered-koji-protocol.md and doubles as a killshot for the endgame hypothesis. | $2,000 | 8 | 0.1 | 0.4 (leaf — tests subset of assumption 2) | expression/localization-mismatch, substrate-availability | Medium |
| 4 | Native metabolite panel pre/post engineering. Parallel WT-vs-engineered kojic-acid and ergothioneine titer comparison. If either drops >50% when the two heterologous cassettes are added, the native-metabolite-preservation component of the claim is killed even if the heterologous proteins express fine. This readout overlaps with killshot 2 but is separable — it can be run on a simpler single-cassette intermediate first. | $1,000 | 2 | 0.15 | 0.5 (midstream — tests claim's "without disrupting native metabolite production" clause) | substrate-availability, expression/localization-mismatch | Medium |
| 5 | Iron-binding functional assay on expressed lactoferrin. UV-Vis at 465 nm for apo-vs-holo ratio; optional CD spectroscopy for fold confirmation; pH-dependent iron release kinetics. Tests assumption 4 directly. If recombinant solid-state-produced Lf has lost iron-binding, the CP1b (Fenton ROS) mechanism from koji-endgame-strain.md §2.2 is compromised, though CP4 and CP6b mitophagy mechanisms may survive. Not a full kill of H01 (the cassette still expresses) but a significant narrowing of the Lf functional claim. | $500 | 2 | 0.2 | 0.3 (leaf — tests assumption 4 only) | expression/localization-mismatch, assay-specificity | Medium-low |
(Scores are qualitative for v0. Numeric scoring with decay and independence weighting is future work per linter-design.md §6.)
Sequencing logic. Run #1 first unconditionally — it's free and tests the literature-gap assumption that several other killshots presume. If #1 surfaces a published precedent (positive: validates; negative: reveals a mode we missed), reshuffle the menu. #3 and #4 can run in parallel with or before #2 as lower-risk intermediate steps — they de-risk #2's larger spend. #5 is a targeted narrowing move that matters most if #2 returns ambiguous Lf-function data.
Failure-Mode Ontology Reference¶
Tagged modes drawn from linter-design.md §5:
- published-literature-gap / training-distribution. The claim rests on a Ward 1995 precedent that's single-protein, submerged, in a different Aspergillus species. Killshot 1 probes this directly.
- species-gap translation. A. awamori → A. oryzae (assumption 1). Killshot 1 partially; killshot 2 fully.
- expression / localization mismatch. The cassettes integrate but the proteins aren't secreted, folded, or processed. Killshots 2, 3, 5.
- kinetics / concentration mismatch. The proteins express but at titers too low for the intended dose (2–3 g/day Lf at 10–15 g/day koji). Killshots 2, 3.
- assay specificity. ELISA cross-reactivity with native Aspergillus proteins; UA-disappearance interference from native urate-relevant enzymes. Killshots 2, 5.
- substrate availability / compartment mismatch. Iron for Lf folding; starch for amyB induction; N source for uricase constitutive expression. Killshots 4, 5.
Two killshots share a failure-mode vector only if the overlap is load-bearing. Killshots 2 and 3 overlap on expression/localization-mismatch but #3 is single-cassette while #2 is dual — they probe different sub-modes of the same family. Independence is imperfect but non-zero; roughly 0.4–0.6 Jaccard on failure-mode vectors.
Pre-Committed Thresholds¶
Declared before any killshot executes. These are the lines in the sand.
Alive¶
Dual-cassette A. oryzae strain in solid-state rice koji (48–60 h, 30°C, 35% moisture) produces:
- Uricase: ≥50 μmol/h/OD (equivalent to ≥100 mg/L pore-fluid / ≥10 mg/g dry koji at industrial ratios — matches ALLN-346 clinical dosing floor)
- Lactoferrin: ≥500 mg/L pore-fluid equivalent (or ≥50 mg/g dry koji — the Phase B floor per koji-endgame-strain.md §3.4)
- Native kojic acid: within 30% of WT titer (WT baseline 3–5 g/L per aspergillus-oryzae.md; floor 2.1 g/L)
- Native ergothioneine: within 30% of WT titer (WT baseline ~20 mg/g dry mycelium; floor 14 mg/g)
- Lactoferrin iron-binding: ≥40% of submerged-culture reference per UV-Vis 465 nm
Killed¶
Any one of the following:
- Either heterologous protein undetectable (uricase <10 mg/L, or Lf <100 mg/L pore-fluid equivalent) after two rounds of cassette/host optimization
- Native kojic acid or ergothioneine drops >50% from WT baseline and does not recover under iron / N supplementation
- Lactoferrin iron-binding <20% of submerged reference (indicates fold collapse, not just titer weakness)
Pending / Ambiguous¶
Intermediate outcomes that cross neither threshold:
- Uricase 10–100 mg/L, Lf 100–500 mg/L (expression works but titers are Phase C rather than Phase B)
- Native metabolite drop 30–50% (recoverable with media optimization, not yet disqualifying)
- Iron-binding 20–40% (fold is partially compromised but not lost)
Document explicitly; propose single-variable follow-up experiments per linter-design.md §7 rather than calling Alive or Killed prematurely.
Kill Switches¶
Sanity / safety stops independent of the scientific thresholds.
- Aflatoxin elevation. A. oryzae is GRAS and non-aflatoxigenic (the aflR regulator is cryptic / truncated), but engineered strains can drift. If aflatoxin (B1, B2, G1, G2) is detected above regulatory limit (20 ppb total in food matrices) at any point, halt the experiment and re-evaluate. Food-grade status is non-negotiable.
- Spore morphology drift. If sporulation fails or produces aberrant morphology (a sign of stress, marker instability, or major metabolic disruption), halt and return to single-cassette or WT baseline for diagnosis.
- Host strain drift. If qPCR indicates cassette loss >10% over three serial passages, the strain is unstable; halt and redesign integration strategy.
- Budget overrun. Hard cap at $7,500 total spend across killshots 1–5 before re-evaluating the whole program. The cheap-first sequencing is designed to surface go/no-go signal within the first $3,000.
Failure Modes Probed (Coverage Map)¶
Which failure modes does the killshot menu actually cover?
| Failure mode | Probed by | Coverage |
|---|---|---|
| published-literature-gap | K1 | Full |
| species-gap-translation | K1 (partial), K2 (full) | Full |
| expression / localization | K2, K3, K5 | Full |
| kinetics / concentration | K2, K3 | Full |
| assay specificity | K2, K5 | Partial (ELISA cross-reactivity not separately orthogonalized) |
| substrate availability | K4, K5 | Partial (iron tested via K5; N / starch not separately) |
| chokepoint collapse | — | Not applicable (H01 is an engineering-feasibility claim, not a therapeutic-mechanism claim) |
| dose-translation scaling | — | Not applicable until the therapeutic-use hypothesis H0N (future) is committed |
Coverage gaps are acceptable — H01 is a feasibility-gate hypothesis, not a mechanism claim. The therapeutic-efficacy hypothesis ("dual-cassette koji at dose X produces biomarker change Y in a gout n=1 protocol") is a separate future H0N with its own card.
Status¶
Survived Killshot #1 (Literature/patent landscape — academic, 2026-05-05) and Killshot #1.5 (Patent landscape deep-dive — industrial IP, 2026-05-05). Both produced partial-validates / partial-confirms-novelty outcomes converging on the same conclusion: dual-cassette heterologous protein co-expression in A. oryzae under submerged conditions is well-precedented in both academic literature and industrial IP, but no patent or paper in any of the major databases (Google Patents, USPTO, EPO/Espacenet, JPO, CNIPA, Lens.org) discloses two heterologous proteins co-expressed in A. oryzae solid-state rice koji at therapeutic-grade titers. The §1.9 wet-lab feasibility test retains its novelty premise on the solid-state-format axis, with the architectural and species-translation premises now well-supported by both academic and patent precedent. Assumption 1 (species translation A. awamori → A. oryzae) and Assumption 6 (KEX-2 capacity for fusion cassettes) remain upgraded to In Vitro (multiple precedents). Assumption 2 (solid-state dual-protein) remains Mechanistic Extrapolation with adjacent flanking precedents (Wakai 2019 — three-cassette submerged in A. oryzae; Shinkawa 2020 — single-cassette heterologous SSF in A. oryzae; Senoo/Tezuka 2024 — three-to-four-enzyme self-cloning SSF in A. oryzae) bracketing the gap from both directions. Killshot #2 (the wet-lab dual-cassette experiment) remains the next gating move.
Survival count: 2.
Survival score: 0.45 (two killshots survived; both partial-validates outcomes scored at the lower end of their respective kill_pr ranges to reflect that the residual solid-state-specific question is unanswered, plus a confidence boost from Killshot #1.5 narrowing the residual industrial-IP risk from ~30% to <10%).
Log¶
| Date | Killshot | Outcome | Notes |
|---|---|---|---|
| 2026-05-05 | #1 — Lit/patent landscape (academic) | VALIDATES (architecture) + CONFIRMS NOVELTY (solid-state format) | See Killshot #1 Findings below. Sub-thresholds: assumption 1 partially validated (Ward-style dual-fusion architecture demonstrated for adalimumab heavy+light chain at niaD/sC loci in A. oryzae, Huynh 2020 PMC7257131); assumption 6 partially validated (KEX-2-like KRGGG cleavage of two simultaneous AmyB-fusion cassettes works in A. oryzae); assumption 2 (solid-state dual-protein) un-falsified but unprobed by literature. No published failure mode that would kill H01 was surfaced. |
| 2026-05-05 | #1.5 — Patent landscape deep-dive (industrial IP) | VALIDATES (architecture) + CONFIRMS NOVELTY (solid-state format) | See Killshot #1.5 Findings below. Direct industrial-IP precedent for A. oryzae dual/multi-heterologous co-expression exists (WO2017211803 Novozymes; CN104004760B East China Univ of Sci & Tech; pTYGS academic system) but all are submerged-only or not titer-quantified for the dual case. No patent in any database (Google Patents, USPTO, EPO, JPO, CNIPA, Lens.org) discloses solid-state rice koji × dual heterologous protein × therapeutic-grade titer. The residual "~30% probability of unpublished industrial IP" risk from Killshot #1 is reduced to <10% — the remaining tail is unrecoverable trade secrets at Novonesis / DSM-Firmenich / Genencor. |
Killshot #1 Findings (2026-05-05)¶
Headline¶
Outcome category: VALIDATES (architecture-level) + CONFIRMS NOVELTY (solid-state-format-level).
The assumption "no published precedent for A. oryzae dual-heterologous-cassette expression" is partially false. Multiple peer-reviewed papers demonstrate two-or-more heterologous protein cassettes co-expressed in A. oryzae under submerged conditions, including a direct architectural precedent of the Ward 1995 glucoamylase-KEX2-fusion design applied to two simultaneous heterologous payloads (Huynh 2020 [1]). The species-translation assumption (A. awamori → A. oryzae) is independently validated by full-length antibody expression in both species (Ward 2004 A. awamori [2]; Huynh 2020 A. oryzae [1]). However, no paper reports two heterologous proteins simultaneously expressed in A. oryzae solid-state rice koji at therapeutic-grade titers, so the §1.9 experiment retains its novelty premise on the format axis.
Top precedents (ranked by direct relevance to H01)¶
| # | Citation | Year | PMID/PMCID | Study type | Key finding (1 sentence) | Relevance to H01 |
|---|---|---|---|---|---|---|
| 1 | Huynh et al. — Functional production of human antibody by the filamentous fungus Aspergillus oryzae [1] | 2020 | PMC7257131 (PMID 32514366) | In Vitro (submerged) | Adalimumab heavy chain and light chain co-expressed as two separate AmyB-KRGGG-fusion cassettes integrated at niaD and sC loci in A. oryzae RIB40-derived host (NSlD-ΔP10 ten-protease deletion strain), yielding 39.7 mg/L functional full-length IgG with antigen-binding equivalent to commercial Humira®. | Direct architectural precedent. Same cassette design (P_amyB::amyB-KRGGG-payload::T_amyB), same host, same KEX-2-like cleavage strategy that H01 hypothesizes for layering uricase onto Ward Lf cassette. Validates assumptions 1, 3 (with caveat of needing protease-deletion host), and 6. Format = submerged DPY medium; not solid-state. |
| 2 | Ward et al. — Characterization of humanized antibodies secreted by Aspergillus niger var. awamori [2] | 2004 | PMID 15128505 | In Vitro (submerged, 14 L bioreactor) | Trastuzumab IgG produced at 900 mg/L in A. awamori using glucoamylase fusion architecture — the original Ward 1995 design extended to dual-chain antibody. | Validates assumption 1 at high-titer scale. Demonstrates the Ward fusion architecture scales to two simultaneously-expressed heterologous payloads. Same caveat (submerged, A. awamori not A. oryzae — but Huynh 2020 closes the species gap). |
| 3 | Wakai et al. — Modified expression of multi-cellulases in a filamentous fungus Aspergillus oryzae [3] | 2019 | PMID 30623869 | In Vitro (submerged) | Three distinct heterologous cellulase genes (cellobiohydrolase, endoglucanase, β-glucosidase) co-expressed in A. oryzae with multi-copy integration (5–16 copies/gene); P-sodM/T-glaB promoter/terminator set was strongest per copy; optimized strain achieved ~40-fold cellulolytic activity over single-integration baseline. | Three-payload precedent. If the Ward-style architecture works for two heterologous proteins (Huynh 2020), Wakai 2019 demonstrates A. oryzae tolerates three. Validates assumption 6 (no saturation of secretion machinery at 2 cassettes). Format = submerged. |
| 4 | Oikawa — Heterologous production of fungal natural products: Reconstitution of biosynthetic gene clusters in model host Aspergillus oryzae [4] | 2020 | PMC7725655 (PMID 33177296) | Review of In Vitro work | Quadruple-auxotrophic A. oryzae host NSAR1 (niaD⁻, sC⁻, ΔargB, adeA⁻) plus ptrA pyrithiamine-resistance marker provides 5 simultaneous integration slots (vectors pTAex3, pUNA, pUSA, pAdeA, pPTRI17). Used to reconstitute fungal biosynthetic gene clusters of up to 17 genes (penitrem, indole diterpenes); produces aphidicolin/solanapyrone at ~100 mg/kg in solid medium. | Validates assumption 5 (multiple selection markers co-existing without conflict — established industrial practice). The 5-marker NSAR1 platform is directly usable for the H01 design. Aphidicolin/solanapyrone "in solid medium" is the closest published touch-point on solid-substrate heterologous production, though those are small-molecule metabolites not secreted proteins. |
| 5 | Li, Lu, Zhang et al. — CRISPR/Cas9-Mediated Multiplexed Genome Editing in Aspergillus oryzae [5] | 2023 | PMC9864741 | In Vitro (submerged) | CRISPR/Cas9 multiplexed editing in A. oryzae RIB40 with up to four simultaneous targets (yA + three amylase copies); single-step integration of Thermomyces lanuginosus lipase (TLL) at amyB and amyC loci with 9.9% knock-in efficiency for two simultaneous integrations; double-copy strains showed elevated TLL activity. | Tooling precedent. Demonstrates that the molecular biology of inserting two cassettes at two loci in one transformation in A. oryzae is now routine; CRISPR/Cas9 with the yA morphological marker reaches 100% editing efficiency under colour selection. De-risks the engineering side of §1.9. |
| 6 | Li, Zhang, Li et al. — Characterization of Aspergillus oryzae mutant and its application in heterologous lipase expression [6] | 2024 | PMC11742560 (PMID 39830075; named in H01 assumption stack) | In Vitro (submerged) | Three-locus integration of TLL lipase at amyA, amyB, and amyC in A. oryzae C19 mutant chassis cell yielded 3.3-fold expression increase over single-copy (113.6 U/L shake flask, 125.6 U/L 5-L bioreactor); α-amylase locus identified as efficient heterologous-protein integration site. | Validates the multi-copy / multi-locus integration strategy for the same protein. With Wakai 2019 (different proteins) and Huynh 2020 (different proteins), brackets the H01 dual-cassette claim from both directions. |
| 7 | Karaman et al. — Large-Scale Production of Anti-RNase A VHH Expressed in pyrG Auxotrophic Aspergillus oryzae [7] | 2023 | PMC10297652 | In Vitro (submerged, 6 L fermenter) | Single-cassette VHH antibody fragment under glucoamylase promoter in pyrG⁻ A. oryzae yielded 44 mg/L (shake flask) and 1.4 g/L (fermenter) — confirming that the glucoamylase-fusion A. oryzae expression platform reaches gram-per-liter titers at scale. | Titer-scaling reference. Establishes the upper end of single-cassette glucoamylase-fusion A. oryzae productivity is ~1 g/L in standard fermenter conditions, supporting the >2 g/L Ward 1995 claim is not anomalous. Submerged. |
| 8 | Shinkawa & Mitsuzawa — Feasibility study of on-site solid-state enzyme production by Aspergillus oryzae [8] | 2020 | PMC7045521 | In Vitro (solid-state, ammonia-treated rice straw) | Constructed a pyrG⁻ ligD⁻ A. oryzae HO2 platform strain; transformed with three separate single-cassette constructs (endoxylanase from Talaromyces aurantiacus, β-glucosidase and cellobiohydrolase from T. cellulolyticus); each strain confirmed secretion of its respective heterologous enzyme in solid-state biomass culture. | Closest solid-state heterologous precedent. Confirms that A. oryzae can secrete heterologous enzymes in a solid-state format with biomass substrate. Limitation: each enzyme was expressed in a separate strain, not co-expressed in one strain; titers not g/L-quantified; substrate is rice straw not polished koji rice. The dual-cassette-on-rice question is one inferential step away. |
| 9 | Sun et al. — Aspergillus oryzae as a Cell Factory (review) [9] | 2024 | PMC11051239 | Review | Comprehensive review of A. oryzae as industrial chassis. Notes critical caveat: "there are certain proteins that are not secreted in solid-state culture, unlike submerged culture, such as the glucoamylase-encoding gene glaB" — solid-state and submerged secretion patterns differ in protein-specific ways. | Material caveat. Some proteins secrete in submerged but not solid-state. This is the primary open risk for H01: even if Lf and uricase express well in submerged dual-cassette format (per Huynh 2020 + Wakai 2019), solid-state-specific silencing or secretion failure for one or both is empirically untested. |
| 10 | Rendsvig, Workman, Hoof — Bidirectional histone-gene promoters in Aspergillus: characterization and application for multi-gene expression [10] | 2019 | PMC6900853 | In Vitro (multiple Aspergillus species) | Characterized the conserved bidirectional H4.1/H3 histone intergenic region as a strong constitutive bidirectional promoter (P_h4h3) in five Aspergillus species; expressed mRFP1 + mCitrine simultaneously from a single locus. | Architectural alternative to the Ward two-locus design. If two-locus integration proves problematic in solid-state, a single-locus bidirectional cassette is a fallback. Not yet tested with secreted proteins at industrial titers. |
What this means for assumption stack¶
Synthesis of literature against H01's assumption stack:
| Assumption | Pre-K1 evidence level | Post-K1 evidence level | Change rationale |
|---|---|---|---|
| 1. Ward 1995 A. awamori → A. oryzae translation | In Vitro (Ward 1992 single-cassette only) | In Vitro (multiple precedents) | Huynh 2020 [1] + Karaman 2023 [7] + Wakai 2019 [3] + Li 2024 [6] all demonstrate Ward-style glucoamylase/amylase-promoter heterologous expression in A. oryzae. The species translation is no longer an extrapolation. |
| 2. Solid-state dual-protein supports both folding/secretion loads | Mechanistic Extrapolation | Mechanistic Extrapolation (with adjacent precedents) | Bracketed by Wakai 2019 (three cassettes submerged) and Shinkawa 2020 (single cassette solid-state). The dual-cassette × solid-state intersection is the un-tested cell. Sun 2024 [9] caveat re: glaB silencing in solid-state is a flagged risk. |
| 3. Native proteases don't degrade Lf/uricase at lethal rates | Mechanistic Extrapolation | In Vitro (with caveat) | Huynh 2020 [1] shows endogenous proteases are the dominant loss mechanism for heterologous antibody in A. oryzae — a 10-protease deletion strain (NSlD-ΔP10) was required for 39.7 mg/L titer. Implication for H01: a protease-deletion host (or inducer-controlled timing) is likely needed to hit the 500 mg/L Lf threshold. The pepE/dppIV/dppV/alpA/pepA/AopepAa/AopepAd/cpI deletion set is a documented starting point. |
| 4. Solid-state Lf retains iron-binding | In Vitro (submerged only) | In Vitro (submerged only) — no change | No solid-state recombinant Lf paper found. Killshot #5 (iron-binding assay) remains the way to test this. |
| 5. Multiple selection markers co-exist without conflict | In Vitro (industrial practice) | In Vitro (multiple precedents) | Oikawa 2020 [4] documents the NSAR1 quadruple-auxotrophic platform plus ptrA = 5 simultaneous markers in A. oryzae (niaD, sC, argB, adeA, ptrA), used routinely for ≥17-gene cluster reconstitutions. The 2-marker H01 design is well within precedent. |
| 6. KEX-2 capacity sufficient for two fusion cassettes | Mechanistic Extrapolation | In Vitro (direct precedent) | Huynh 2020 [1] demonstrates simultaneous KEX-2 cleavage of two distinct AmyB-KRGGG-payload fusions (heavy chain + light chain) producing both correctly-cleaved proteins. KEX-2 capacity is not the bottleneck at the 39.7 mg/L scale; whether it saturates at the 500 mg/L Lf + 100 mg/L uricase H01 targets is empirically open but no published evidence suggests it would. |
| 7. Solid-state Lf glycosylation within submerged envelope | Animal Model (submerged only) | Animal Model (submerged only) — no change | No published comparative glycan analysis of A. oryzae solid-state vs. submerged secreted protein. Future Killshot #6 (glycan profiling) would test this. |
What this means for §1.9 wet-lab framing¶
§1.9 (Ward 1995 dual-cassette feasibility test) is not eliminated by literature but its framing changes:
- The dual-cassette engineering is no longer the load-bearing risk — it's textbook in A. oryzae under submerged conditions. The §1.9 experimental design should explicitly include a submerged-culture parallel control (already in killshot #2 spec) so that solid-state-specific failure modes can be cleanly attributed to the format rather than the architecture.
- Protease-deletion host is now elevated from "fallback" to "default." Huynh 2020 [1] shows ten-protease-deletion (NSlD-ΔP10) was necessary to detect functional antibody at all; the wild-type RIB40 background may be a non-starter for the Lf side of the dual cassette. The §1.9 design should consider starting from NSlD-ΔP10 or an equivalent industrial protease-knockout chassis.
- The solid-state question is the real load-bearing test. Sun 2024 [9] explicitly notes glaB-type proteins fail to secrete in solid-state despite secreting in submerged. Whether AmyB-fused Lf and AmyB-fused uricase fall on the "secretes in both" or "secretes in submerged only" side of this divide is the actual experimental novelty.
- The 5-marker NSAR1 platform (Oikawa 2020 [4]) is the suggested host platform — it's already industrial-practice and accommodates the 2-cassette H01 design with three slots to spare for downstream additions (e.g., kojic-acid-pathway enhancement, ergothioneine pathway maintenance).
- No urgency change for §1.9 priority — it remains the #1 gate. The literature deep-dive narrowed the question but didn't answer it.
comp-010 Follow-Up: Cassette Compatibility (2026-05-05)¶
comp-010 is in silico design support — not a killshot and not a survived killshot. It addresses sequence-level cassette-design questions that Killshot #1 (literature audit) could not: KEX2 internal-site geometry, codon-optimization burden, disulfide load, and secretion-targeting signals. Key findings for the H01 assumption stack:
- Assumption 6 (KEX-2 capacity): No high-risk KEX2 internal sites in lactoferrin (2 sites: 1 abolished P1'=D, 1 moderate P1'=K). Uricase has 1 high-risk site (pos 128, P1'=N) but only load-bearing if uricase is in a fusion architecture — the §1.9 proposed design uses direct secretion for uricase, making this moot. Assumption 6 remains In Vitro (direct precedent) per Killshot #1; comp-010 adds no new evidence level.
- Assumption 3 (protease load, H01 only): Disulfide load analysis confirms dual-cassette ER burden = 1.06× Huynh 2020 baseline (17 disulfides, all on Lf). Within demonstrated capacity. No change to assumption-3 evidence level.
- Uricase C-terminal SKL: A potential PTS1 peroxisomal signal in the uricase sequence. Verify secretion in §1.9. If misrouted, append 3×Ala to mask. Pre-experiment design note, not a H01 assumption change.
Full analysis: wiki/cassette-compatibility-computational.md | wiki/etc/experiments/comp-010-cassette-compatibility/
Most material caveat / what couldn't be verified¶
- Patent landscape was not directly searchable in Killshot #1. Google Patents returns JavaScript-only pages that WebFetch cannot parse, and Espacenet rejected the unauthenticated request (HTTP 403). Three industrial actors (Novozymes, DSM, Genencor/Danisco) have ≥30 years of Aspergillus multi-protein engineering history that is largely unpublished except as patent claims. Original estimated ~30% probability that one of them has filed (or holds expired/abandoned) IP on a directly-relevant A. oryzae dual-cassette koji process. This caveat was directly addressed by Killshot #1.5 (Patent landscape deep-dive, 2026-05-05), which used Google Patents indirection via WebSearch to surface 13 directly-relevant patents and patent-adjacent disclosures across US, EP, WO, JP, and CN jurisdictions. Residual industrial-IP risk reduced to <10% (unrecoverable trade-secret tail). See Killshot #1.5 Findings below.
- No published A. oryzae dual-heterologous-protein × solid-state × therapeutic-titer paper exists. This is the actual gap H01 is testing. The Wakai 2019 + Shinkawa 2020 bracket suggests the gap is closeable but not closed.
- Glycoprotein-specific solid-state secretion patterns are under-characterized. The Sun 2024 [9] glaB observation is one data point; whether it generalizes to AmyB-fusion architectures or only to native glaB is unclear.
- The Wang 2023 A. niger multi-locus paper named in the H01 sources frontmatter (PMID 37807677) was not in the Paperclip corpus and could not be verified directly. Its inclusion in the assumption stack as an adjacent precedent should be re-checked when accessible.
- Ward 1995 (PMID 9634791) itself was not in the Paperclip corpus — relied on the Ward 2004 follow-up (PMID 15128505) and downstream citation chain to characterize the architecture.
References (Paperclip citation block)¶
[1] Huynh HH, Morita N, Sakamoto T, Katayama T, Miyakawa T, Tanokura M, Fushinobu S, Maruyama J. "Functional production of human antibody by the filamentous fungus Aspergillus oryzae." Fungal Biology and Biotechnology 7:7 (2020). PMC7257131. doi:10.1186/s40694-020-00098-w https://citations.gxl.ai/papers/PMC7257131#L25
[2] Ward M, Lin C, Victoria DC, Fox BP, Fox JA, Wong DL, Meerman HJ, Pucci JP, Fong RB, Heng MH, Tsurushita N, Gieswein C, Park M, Wang H. "Characterization of humanized antibodies secreted by Aspergillus niger." Applied and Environmental Microbiology 70:2567-2576 (2004). PMID 15128505. (Cited in [1] as ref 13; not in Paperclip corpus directly)
[3] Wakai S, Nakashima N, Ogino C, Tsutsumi H, Hata Y, Kondo A. "Modified expression of multi-cellulases in a filamentous fungus Aspergillus oryzae." Bioresource Technology 276:146-153 (2019). PMID 30623869. doi:10.1016/j.biortech.2018.12.117 (Verified via PubMed abstract; not in Paperclip corpus directly)
[4] Oikawa H. "Heterologous production of fungal natural products: Reconstitution of biosynthetic gene clusters in model host Aspergillus oryzae." Proceedings of the Japan Academy. Series B, Physical and Biological Sciences 96:420-430 (2020). PMC7725655. doi:10.2183/pjab.96.031 https://citations.gxl.ai/papers/PMC7725655#L18
[5] Li Q, Lu J, Zhang G, Zhou J, Li J, Du G, Chen J. "CRISPR/Cas9-Mediated Multiplexed Genome Editing in Aspergillus oryzae." Journal of Fungi 9(1):109 (2023). PMC9864741. doi:10.3390/jof9010109 https://citations.gxl.ai/papers/PMC9864741#L42
[6] Li Q, Zhang C, Li J, Du G, Li Z, Zhou J, Zhang G. "Characterization of Aspergillus oryzae mutant and its application in heterologous lipase expression." Synthetic and Systems Biotechnology 10(2):365-372 (2024). PMC11742560. doi:10.1016/j.synbio.2024.12.005 https://citations.gxl.ai/papers/PMC11742560#L52
[7] Karaman E, Eyüpoğlu AE, Mahmoudi Azar L, Uysal S. "Large-Scale Production of Anti-RNase A VHH Expressed in pyrG Auxotrophic Aspergillus oryzae." Current Issues in Molecular Biology 45(6):4778-4791 (2023). PMC10297652. doi:10.3390/cimb45060304 https://citations.gxl.ai/papers/PMC10297652#L54
[8] Shinkawa S, Mitsuzawa S. "Feasibility study of on-site solid-state enzyme production by Aspergillus oryzae." Biotechnology for Biofuels 13:36 (2020). PMC7045521. doi:10.1186/s13068-020-1669-3 https://citations.gxl.ai/papers/PMC7045521#L40
[9] Sun Z, Wu Y, Long S, Feng S, Jia X, Hu Y, Ma M, Liu J, Zeng B. "Aspergillus oryzae as a Cell Factory: Research and Applications in Industrial Production." Journal of Fungi 10(4):248 (2024). PMC11051239. doi:10.3390/jof10040248 https://citations.gxl.ai/papers/PMC11051239#L28
[10] Rendsvig JKH, Workman CT, Hoof JB. "Bidirectional histone-gene promoters in Aspergillus: characterization and application for multi-gene expression." Fungal Biology and Biotechnology 6:24 (2019). PMC6900853. doi:10.1186/s40694-019-0088-3 https://citations.gxl.ai/papers/PMC6900853#L7
Killshot #1.5 Findings (2026-05-05) — Patent Landscape¶
Headline¶
Outcome category: VALIDATES (architecture-level) + CONFIRMS NOVELTY (solid-state-format-level).
The Killshot #1 caveat — "~30% probability of unpublished industrial IP from Novonesis / DSM-Firmenich / Genencor that an authenticated patent search could surface" — has been directly probed across Google Patents, USPTO, EPO/Espacenet (via WebSearch indirection — direct fetch returned HTTP 403), J-PlatPat (via Google Patents JP corpus), CNIPA (via Google Patents CN corpus), and Lens.org. Patent precedent for A. oryzae dual-heterologous-protein expression does exist under submerged conditions (Novozymes WO2017211803, East China Univ of Sci & Tech CN104004760B, plus the academic pTYGS plasmid system), validating the architectural premise of H01. No patent in any of the databases checked discloses two heterologous proteins co-expressed in A. oryzae solid-state rice koji at therapeutic-grade titers in any language. The residual industrial-IP risk is now estimated at <10% — the remaining tail is unrecoverable trade secrets behind corporate firewalls, which the patent literature by definition cannot illuminate. The §1.9 wet-lab feasibility test retains its novelty premise on the format axis without modification.
Top patents and patent-adjacent disclosures (ranked by direct relevance to H01)¶
| # | Patent / Disclosure | Year | Assignee | Status | Key claim or finding (1 sentence) | Relevance to H01 |
|---|---|---|---|---|---|---|
| 1 | WO2017211803A1 — Co-expression of heterologous polypeptides to increase yield | 2017 | Novozymes A/S | Ceased (2018-12-07) | Method of producing a poorly expressed first heterologous polypeptide in a host filamentous fungus by co-expressing it with a second heterologous polypeptide that is well-expressed; integrates two heterologous polynucleotides at different genomic loci via FLP-mediated recombination into a pyrG-minus strain. Examples: arabinofuranosidase + Trametes amyloglucosidase in A. niger shake flask, MU-1 medium, 30°C × 7 days. | Most directly-relevant industrial-IP precedent. Validates the dual-cassette architecture in filamentous fungi as a Novozymes-claimed method (now lapsed). Examples are A. niger not A. oryzae, and submerged shake flask not solid-state — but the broader claim language encompasses A. oryzae. The patent has lapsed (ceased 2018-12-07), so the architecture is in the public domain. Validates assumption 1 and assumption 6 from the industrial-IP side. |
| 2 | CN104004760B — Aspergillus oryzae expression cassette and its engineering bacterium for secreting foreign protein | 2014 | East China University of Science and Technology | Active until 2034-06-12 | Single-cassette A. oryzae expression vector for foreign-protein secretion. Specification text states "this expression cassette can be combined with other expression cassettes to simultaneously express two foreign proteins" but only GFP is experimentally demonstrated; no quantitative titer reported. | Theoretical dual-cassette claim, no experimental validation. Active patent in China — would be a freedom-to-operate (FTO) consideration for any Open Enzyme wet-lab work that ends up using this exact cassette architecture in mainland China. The vague dual-cassette mention is unlikely to support enforcement claims given the absence of dual-cassette examples. |
| 3 | pTYGS plasmid system (Bristol/Cox lab, academic) | 2011–2020 | University of Bristol (academic, unpatented) | N/A | Series of expression vectors for A. oryzae NSAR1 with four pre-assembled promoters (PgpdA, Padh, Peno, PamyB) on a single plasmid, allowing simultaneous transfer and expression of four heterologous genes. Used to reconstitute fungal biosynthetic gene clusters of up to 17 genes. | Closest published 4-gene single-plasmid architecture. Open-access academic tool, no IP encumbrance. Demonstrates that the quadruple-cassette extension of H01 is also tractable, should the dual cassette validate. All published applications are submerged. |
| 4 | EP0238023A2 / EP0238023B2 — Process for the production of protein products in Aspergillus oryzae and a promoter for use in Aspergillus | 1987 (filed); 2002 (granted B2) | Novo Nordisk / Novozymes A/S | Expired 2007-03-16 | Foundational A. oryzae heterologous expression patent: TAKA-amylase promoter, glucoamylase promoter, argB / amdS / pyr4 / DHFR markers. Single-cassette only; cotransformation with separate marker vector mentioned but not multi-payload co-expression. | Foundational and expired. All claims now public domain. The TAKA-amylase promoter (P-amyB) used in the H01 design originates here. |
| 5 | US5571697A — Expression of processed recombinant lactoferrin and lactoferrin polypeptide fragments from a fusion product in Aspergillus | 1994 (filed); 1996 (granted) | Baylor College of Medicine / Agennix | Expired 2013-11-05 | The Ward 1995 cassette as patented: glucoamylase or α-amylase 5' fragment + KEX-2 cleavage linker + lactoferrin coding sequence; expressed in A. awamori, A. niger, A. oryzae; submerged 80-L stirred-tank with Rushton impellers. Single-cassette only (lactoferrin alone). | The Ward-1995 patent itself. Expired and in public domain. Confirms that the architecture H01 ladders on top of has no remaining IP claims. |
| 6 | US5364770A — Heterologous polypeptides expressed in Aspergillus | 1989 (filed); 1994 (granted) | Genencor International / Danisco US | Expired 2011-11-15 | Foundational filamentous-fungus heterologous expression patent (signal sequence, promoter, terminator, selection marker). Single-cassette; experimental work in A. nidulans with chymosin, glucoamylase, proteases. | Foundational, expired, public domain. No dual-cassette claim; primarily A. nidulans-focused. |
| 7 | US5536661 / US5695985 — Process for production of protein products in Aspergillus / Thermophilic fungal expression system | 1990s | Novozymes / Novo Nordisk | Expired | Variants of the foundational TAKA-amylase / glucoamylase expression system; all single-cassette. | Expired. No new structural information beyond EP0238023. |
| 8 | JP5140832B2 — Aspergillus oryzae mutant strains for high production of heterologous proteins | ~2010 | Japanese assignee (industrial Japanese) | Granted JP | Double protease-disruption + UV mutagenesis of A. oryzae (NS-tApE → AUT-1 to AUT-7). Single-protein expression only (human lysozyme, calf chymosin); submerged 5× DPY medium; titers 26–97 mg/L. | Single-cassette only. The protease-deletion strategy is now standard practice and well-precedented by Yoon et al. ten-protease-deletion strain (referenced in Killshot #1 Huynh 2020). |
| 9 | JP2001046078A — High expression system for proteins | 2000 | Gekkeikan Sake Co Ltd | JP filed | melO promoter for A. oryzae glucoamylase (glaB) — single-cassette, submerged. Titers 756–1200 U/mL glucoamylase activity. Notes: melO promoter gives 100× higher activity in liquid vs. wild-type, and 2× higher than solid-state koji fermentation. | Single-cassette. Notable ancillary point: the patent's preference for liquid culture over solid-state for the glaB system aligns with the Sun 2024 [9] caveat from Killshot #1 — glaB-class proteins underperform in solid-state. Doesn't affect AmyB-driven cassettes (different promoter / different secretion route) but reinforces that solid-state secretion is protein-specific. |
| 10 | JP2018191551A — Multi-step multi-mutation production by genome editing in filamentous fungi | 2017 | University of Tokyo | JP filed | CRISPR/Cas9 platform vector for sequential multi-mutation in A. oryzae; demonstrated on melB/melO tyrosinase genes. Genome-editing tool, not heterologous expression. | Tooling precedent only. Confirms that multiplexed integration in A. oryzae is a U Tokyo-led academic standard; no IP barrier to multi-locus integration in §1.9 design. |
| 11 | WO2007039990A1 — Process for producing liquid koji by using Aspergillus oryzae | 2006 | Asahi Breweries Ltd | Ceased | Liquid koji production with high native α-amylase / glucoamylase activity. Mentions heterologous proteins as ancillary benefit; no heterologous expression demonstrated. | Negative result for the H01 search. Even the Asahi (Asahi Breweries / Asahi Kasei lineage) industrial-koji patent does not claim heterologous protein co-expression. |
| 12 | Senoo et al. 2024 (PMID 38242757) — Construction of self-cloning Aspergillus oryzae strains with high production of multiple biomass-degrading enzymes on solid-state culture | 2024 | Tohoku University + Fujiwara Techno-Art (academic + industrial collaboration; not patented to public knowledge) | Published, not patented | Industrial A. oryzae AOK11 strain engineered with combinations of xylanase (xynG1), phytase (phyA), pectin lyase (pelA), polygalacturonase (pgaB) under amyB promoter; solid-state culture; transcript levels >100× parent strain. Genes are self-cloning (native or near-native), not heterologous. | Closest published multi-cassette × solid-state precedent — but self-cloning, not heterologous. Brackets the H01 question from the multi-cassette × solid-state side but does not close the heterologous-protein-specific gap. The Wakai 2019 (multi-heterologous × submerged) + Senoo 2024 (multi-self-cloning × solid-state) + Shinkawa 2020 (single-heterologous × solid-state) trio jointly bracket the H01 cell from three directions without occupying it. |
| 13 | Nielsen, Meyer, Hansen, Arnau 2025 — Enhanced Production of Bovine β-Lactoglobulin in an Industrial Aspergillus oryzae Host | 2025 | Novonesis (Novozymes lineage) — published, not yet known to be patented | Published 2025 | First disclosure of an industrial Novonesis A. oryzae lineage producing single heterologous protein (bovine β-lactoglobulin) at industrial scale. Strain derived from BECh1 lineage (large chromosomal deletion removing aflatoxin + cyclopiazonic acid clusters). Submerged. | Industrial-strain precedent for single-heterologous expression at scale. Confirms the Novonesis A. oryzae engineering platform is publicly disclosed at single-cassette level. No dual-cassette disclosure in this paper. |
Patent search coverage¶
| Database | Method | Coverage achieved | Gap |
|---|---|---|---|
| Google Patents | WebSearch (Google index) + WebFetch on individual patent pages | Full text accessible for all patents fetched (EP, US, WO, JP, CN); deep-dive on 7 patents | None material — Google Patents indexes ~120 jurisdictions including JP/CN |
| USPTO PatFT/AppFT | WebSearch + Google Patents indirection | Covered — US patents accessible via Google Patents | None material |
| EPO / Espacenet | WebSearch + Google Patents indirection (direct WebFetch returned HTTP 403) | EP patents accessible via Google Patents JS-rendered pages | Direct Espacenet API blocked — but EP patent corpus fully indexed in Google Patents |
| JPO / J-PlatPat | Google Patents JP corpus + Japanese-language WebSearch (per CLAUDE.md global-multilingual default) | Full-text Japanese patents accessible via Google Patents JP/EN; multiple JP patents reviewed (JP5140832B2, JP2001046078A, JP2018191551A, WO2007039990A1) | J-PlatPat direct UI requires JS — but Google Patents indexes JP corpus including the major industrial Japanese assignees (Asahi, Gekkeikan, U Tokyo, Riken) |
| CNIPA | Google Patents CN corpus + Chinese-language WebSearch | CN104004760B + Jiangnan University academic corpus accessible | None material — Google Patents CN coverage is comprehensive for granted patents post-2000 |
| Lens.org | WebFetch on REST search URL | Failed — Lens.org returns empty shell to non-JS-capable scrapers | Workable workaround: the same patent records surfaced via Google Patents indirection. Lens.org adds value for citation graphs and assignee analytics, neither of which was load-bearing here. |
| DPMA (Germany) | WebSearch | Spot-checked; no German-only Aspergillus oryzae dual-cassette patent surfaced | Low priority — the major fungal-engineering IP is held by Danish (Novozymes), Dutch (DSM), American (Genencor/Danisco/IFF), Japanese, and Chinese assignees, not German |
What this means for the assumption stack¶
| Assumption | Evidence level pre-K1.5 | Evidence level post-K1.5 | Change rationale |
|---|---|---|---|
| 1. Ward 1995 A. awamori → A. oryzae translation | In Vitro (multiple academic precedents) | In Vitro (multiple academic + industrial-IP precedents) | Industrial-IP confirmation: WO2017211803 (Novozymes 2017) claims co-expression in filamentous fungi including A. oryzae; CN104004760B (2014) is A. oryzae-specific (single demonstration). |
| 2. Solid-state dual-protein supports both folding/secretion loads | Mechanistic Extrapolation (with adjacent academic precedents) | Mechanistic Extrapolation (with adjacent academic + industrial precedents) | Senoo 2024 (multi-self-cloning × solid-state in industrial A. oryzae AOK11) added as third bracketing precedent. The dual-heterologous × solid-state cell remains experimentally unoccupied. |
| 3. Native proteases don't degrade Lf/uricase at lethal rates | In Vitro (with caveat — protease-deletion host needed) | In Vitro (with caveat — protease-deletion host needed) | No change. JP5140832B2 confirms the protease-double-disruption strategy is industry-standard; the ten-protease-deletion (Huynh 2020 NSlD-ΔP10) is the strong-form fallback. |
| 4. Solid-state Lf retains iron-binding | In Vitro (submerged only) | In Vitro (submerged only) — no change | No solid-state recombinant Lf patent or paper found. Killshot #5 (iron-binding assay) remains the way to test this. |
| 5. Multiple selection markers co-exist without conflict | In Vitro (multiple academic precedents) | In Vitro (multiple academic + industrial precedents) | EP0238023 (now expired) established argB/amdS/pyr4/DHFR marker set as foundational industrial practice. CN104004760B + JP5140832B2 add CN/JP industrial confirmation. |
| 6. KEX-2 capacity sufficient for two fusion cassettes | In Vitro (direct academic precedent — Huynh 2020) | In Vitro (direct academic precedent + industrial-IP framework) | WO2017211803's general two-heterologous-polypeptide framework supports the KEX-2 / signal-peptide processing assumption from the industrial side, though its examples don't use KEX-2 specifically. |
| 7. Solid-state Lf glycosylation within submerged envelope | Animal Model (submerged only) | Animal Model (submerged only) — no change | Patent literature does not address comparative glycan analysis. |
What this means for §1.9 wet-lab framing¶
§1.9 (Ward 1995 dual-cassette feasibility test) is not eliminated by patent literature — and the framing implications converge with Killshot #1's:
- The §1.9 design has a clean freedom-to-operate (FTO) profile. All foundational patents (Ward 1995 = US5571697; Novozymes EP0238023 / US5536661 / US5695985; Genencor US5364770) are expired. WO2017211803 (Novozymes 2017 dual-cassette method) is ceased (lapsed 2018-12-07). The only active patent surfaced is CN104004760B (East China Univ of Sci & Tech, expires 2034) which is China-specific and demonstrates only single-protein GFP — unlikely to constitute an enforcement risk for any non-China wet-lab work, and its dual-cassette claim is unsupported by examples.
- No industrial-IP failure-mode disclosure was surfaced. The hypothetical "Novonesis or DSM-Firmenich filed a patent disclosing systemic silencing of co-expressed cassettes in A. oryzae" scenario did not materialize. If such a failure mode is known industrial trade-secret knowledge, it has not been disclosed in any published patent, which weakly suggests either (a) it doesn't exist, or (b) it exists but the assignees prefer trade-secret protection over patent-claim disclosure (the standard biotech-process-IP preference). Either way, the §1.9 design is not deflected by this search.
- No industrial-IP solid-state-format precedent was surfaced. Despite searching across Japanese (Asahi Breweries, Gekkeikan, U Tokyo), Chinese (East China Univ of Sci & Tech, Jiangnan), and Western (Novozymes, DSM, Genencor) assignees, no patent was found that demonstrates two heterologous proteins simultaneously expressed in A. oryzae solid-state at therapeutic-grade titers. The Senoo 2024 publication is the closest disclosure (multi-enzyme × solid-state) but is self-cloning, not heterologous, and academic rather than IP-protected. The H01 §1.9 architecture is genuinely first-in-class on the heterologous × solid-state × dual-cassette × therapeutic-titer axis.
- Industrial Novonesis lineage is the de facto strain platform. The Nielsen 2025 paper publicly discloses that the Novonesis A. oryzae BECh1 lineage (with large chromosomal deletion removing aflatoxin + cyclopiazonic acid clusters) is the industry-standard heterologous-protein chassis. This is now the natural reference strain for §1.9 — though the published academic NSAR1 / NSlD-ΔP10 strains remain the practical starting point for an academic collaborator (Maruyama or a Role 2 / Pharma Translation collaborator) without Novonesis access.
- No urgency change for §1.9 priority. The patent landscape narrowed the residual industrial-IP risk from ~30% to <10% but did not surface either a validating direct-precedent patent (which would have de-risked §1.9) or a killing failure-mode patent (which would have forced a fallback). §1.9 remains the #1 gate.
Most material caveat / what couldn't be verified¶
- Trade-secret industrial process knowledge is by definition outside the patent literature. The remaining ~10% residual risk is unrecoverable without a Novonesis / DSM-Firmenich / Genencor / Ajinomoto insider source. This is a hard limit of any patent-database search and not a deficiency of this particular execution.
- Espacenet direct API access was blocked (HTTP 403 unauthenticated). Workaround via Google Patents indirection captures the EP patent corpus, so this is not a substantive coverage gap, but for any future patent-FTO opinion (e.g., before commercial commitments) an authenticated Espacenet / Lens.org / Derwent World Patents Index pass via Brian's institutional access (Emory library or external IP counsel) would tighten confidence.
- CNIPA full-text Chinese-language patents may be incompletely indexed in Google Patents — the major active assignees (Jiangnan University, Tianjin Institute of Industrial Biotechnology, Wang Shihua at Fujian Agriculture and Forestry Univ) have prolific publication records that may include unindexed patents. A native Chinese-speaker pass via the CNIPA UI would be the gold standard for this segment.
- The Nielsen 2025 Novonesis β-lactoglobulin paper hints at undisclosed industrial dual-cassette work — Novonesis publishing single-heterologous data in 2025 is consistent with the company having dual-cassette proof-of-concept internally but choosing not to disclose. This is the most plausible scenario for the residual <10% trade-secret risk.
References (Patent and patent-adjacent disclosure list)¶
[P1] WO2017211803A1 — Co-expression of heterologous polypeptides to increase yield (Novozymes A/S, filed 2017-06-06, ceased 2018-12-07). https://patents.google.com/patent/WO2017211803A1/en
[P2] CN104004760B — Aspergillus oryzae expression cassette and engineering bacterium for secreting foreign protein (East China University of Science and Technology, filed 2014-06-12, active until 2034). https://patents.google.com/patent/CN104004760B/en
[P3] EP0238023A2 / EP0238023B2 — Process for the production of protein products in Aspergillus oryzae and a promoter for use in Aspergillus (Novo Nordisk / Novozymes A/S, filed 1987-03-16, expired 2007-03-16). https://patents.google.com/patent/EP0238023A2/en
[P4] US5571697A — Expression of processed recombinant lactoferrin and lactoferrin polypeptide fragments from a fusion product in Aspergillus (Baylor College of Medicine / Agennix, filed 1994-11-02, expired 2013-11-05). https://patents.google.com/patent/US5571697A/en
[P5] US5364770A — Heterologous polypeptides expressed in Aspergillus (Genencor International / Danisco US, filed 1989-09-25, expired 2011-11-15). https://patents.google.com/patent/US5364770A/en
[P6] JP5140832B2 — Aspergillus oryzae mutant strains for high production of heterologous proteins (Japanese assignee, granted JP). https://patents.google.com/patent/JP5140832B2/en
[P7] JP2001046078A — High expression system for proteins (Gekkeikan Sake Co Ltd, filed 2000-02-15). https://patents.google.com/patent/JP2001046078A/en
[P8] JP2018191551A — Multi-step multi-mutation production by genome editing in filamentous fungi (University of Tokyo, filed 2017-05-16). https://patents.google.com/patent/JP2018191551A/en
[P9] WO2007039990A1 — Process for producing liquid koji by using Aspergillus oryzae (Asahi Breweries Ltd, filed 2006-08-17, ceased). https://patents.google.com/patent/WO2007039990A1/en
[P10] Senoo et al. Construction of self-cloning Aspergillus oryzae strains with high production of multiple biomass-degrading enzymes on solid-state culture. J Biosci Bioeng 137(3) (2024). PMID 38242757. (Academic publication — closest multi-cassette × solid-state precedent; self-cloning, not heterologous.)
[P11] Nielsen MB, Meyer AS, Hansen K, Arnau J. Enhanced Production of Bovine β-Lactoglobulin in an Industrial Aspergillus oryzae Host: A Step Forward in Alternative Protein Biomanufacturing. (2025, Novonesis lineage paper.) https://journals.sagepub.com/doi/10.1177/15509087251398318
[P12] Cox lab pTYGS plasmid system (academic, unpatented): four-promoter (PgpdA, Padh, Peno, PamyB) single-plasmid heterologous expression in A. oryzae NSAR1. Documented in Lazarus et al. 2020 (PMID 35524043).
Retraction History¶
None.
Cross-References¶
- koji-endgame-strain.md — the platform-thesis page this hypothesis gates. §3 (Ward 1995 architecture layering), §3.4 (protocol sketch), §4 (fallback paths if H01 is killed).
- engineered-koji-protocol.md §16 — the single-cassette lactoferrin co-expression module this hypothesis ladders on top of. Also §02–14 (Year 1 uricase-only starting strain, which is killshot 3 reframed as a standalone deliverable).
- lactoferrin.md — full dossier for the Lf side of the dual cassette; §7 discusses the Open Enzyme feasibility bet.
- uricase-variant-selection.md — source-gene analysis (A. flavus uaZ primary, C. utilis secondary).
- validation-experiments.md §1.9 — the experiment-queue entry for this feasibility test; this hypothesis is the formalization of that entry.
- aspergillus-oryzae.md — host biology, native kojic acid + ergothioneine baselines.
- linter-design.md — schema and rationale for the Falsification Card format used here.
This is the seed hypothesis for the wiki/hypotheses/ directory. The Falsification Card format will be refined as subsequent hypotheses are committed; deviations from this template should be justified in the new hypothesis's commit message and, if structural, propagated back to linter-design.md §4.