TOLL9 (Q5TWL7, AGAP006974) — Ligand and Immune-Pathway Specificity: Final Curation Report

Target gene: Anopheles gambiae TOLL9 (UniProt Q5TWL7 / AGAP006974)
Focus: function_assignment — toll9-ligand-and-immune-pathway-specificity
Source: genes/ANOGA/TOLL9/TOLL9-ai-review.yaml

Executive Judgment

Verdict: Partially supported / the strong claim is over-annotated as a direct function assignment.

The seed hypothesis proposes that A. gambiae TOLL9 (1) has pattern-recognition receptor (PRR) activity and (2) participates in innate immunity through the Toll signaling pathway. After three iterations of orthology, domain-architecture, partner-availability, and literature analysis, the defensible position is nuanced:

• What is supported (retain): TOLL9 is a canonical Toll/TLR-family transmembrane receptor. Its generic, phylogeny-transferred annotations — signaling receptor activity (GO:0038023, MF), signal transduction (GO:0007165, BP), and plasma membrane (GO:0005886, CC) — are consistent with its domain architecture (LRR ectodomain + two TM helices + cytoplasmic TIR) and with the ancestral scope of PANTHER family PTHR24365. These should be retained.

• What is NOT supported as a direct function assignment (flag as leads, do not assert): The specific claims that TOLL9 directly has PRR activity (GO:0038187) and functions as an innate-immune Toll-signaling receptor are biologically plausible but experimentally unproven in Anopheles gambiae. The only direct PRR/LPS evidence comes from the more distant lepidopteran Bombyx mori, and even there the downstream signaling was demonstrated using a chimeric receptor whose transmembrane and TIR regions are mouse TLR4 — so the native insect signaling readout was never shown. The closest genetic test, a Drosophila Toll-9 loss-of-function mutant (~36% identical to the target), shows no antibacterial defect. The Anopheles mosGILT data are correlative expression changes, not a perturbation of TOLL9 itself.

Most important caveat: A "small experimental donor set" is not automatically weak evidence — but in this case the donor evidence is both single-species and mechanistically partial (chimeric TM/TIR), and the closest ortholog contradicts a required antibacterial role. That combination is exactly what should prevent a species-level MF PRR assignment while leaving the immune hypothesis open as a curator lead.

Evidence Matrix

Key Findings (expanded)

F001 — TOLL9 carries only generic receptor annotations, transferred phylogenetically

UniProt Q5TWL7 has protein existence "Inferred from homology" (level 3) — there is no experimental characterization of this specific protein. Its domain architecture is unambiguously that of a Toll/TLR receptor: an extracellular leucine-rich-repeat (LRR) ectodomain (Pfam PF13855 LRR_8 ×2; SMART LRR_TYP ×7), two transmembrane helices (aa 29–49 and 607–631), and a cytoplasmic TIR signaling domain (aa 667–807; Pfam PF01582; PROSITE PS50104). It belongs to PANTHER PTHR24365 ("TOLL-LIKE RECEPTOR 1"), subfamily SF530. All four of its GO annotations are IBA (phylogenetically inferred via GO_Central): plasma membrane (GO:0005886), signaling receptor activity (GO:0038023), signal transduction (GO:0007165), and inflammatory response (GO:0006954). Notably absent are GO:0038187 (pattern recognition receptor activity) and any Toll-signaling or defense-response BP term. The seed hypothesis's stronger claims are therefore not currently supported by the transferred annotation set — they would be new assertions. The current leaf PTN000688561 in PTHR24365 descends from PTN002808045, which supports only generic receptor/signaling/plasma-membrane properties — consistent with the observed transfer scope.

F002 — Direct PRR/LPS evidence is Bombyx-only and mechanistically chimeric

The direct experimental basis for insect Toll9 as an LPS-recognizing PRR is PMID: 33963082: BmToll9 (Bombyx mori) recognizes LPS through two MD-2-like partners (BmMD-2A/BmMD-2B). However, the demonstration of downstream signaling used a chimeric receptor in which the BmToll9 ectodomain was fused to mouse TLR4 transmembrane and TIR domains, and the readout (LPS-induced inflammatory-factor release) occurred in murine cells. Thus the insect ectodomain's LPS/MD-2 binding is real, but native insect TM/TIR signaling and an insect functional readout were not established. Against a required immune role, the Drosophila Toll-9 complete loss-of-function mutant (PMID: 21386906) shows no defect in basal or infection-induced AMP responses, while acknowledging overexpression-induced signaling. The Anopheles mosGILT^null data (PMID: 38191283) are transcriptional co-regulation only — TOLL9 itself was never perturbed.

F003 — MD-2-like partners exist in Anopheles, so partner-absence does not block the mechanism

A UniProt search (taxid 7165; Pfam PF02221 ML domain / IPR003172 MD-2-related lipid-recognition) returns ≥16 A. gambiae proteins (e.g., Q7QJ15, Q7PR71, Q7QCX7, A7USS7, A7UVK2, F5HMB6, and a cluster Q7QCX4/5/7/8/9 of ~148–170 aa). These are the mosquito counterparts of Bombyx BmMD-2A/B and mammalian MD-2. The presence of this accessory class means the Bombyx-style mechanism is not excluded in Anopheles on partner-availability grounds. However, none is annotated as interacting with TOLL9/AGAP006974, and no LPS-binding assay ties any of them to the target — so this is feasibility, not evidence of function.

F004 — Anopheles TOLL9 vs Drosophila Toll-9 are ~36% identical

A Needleman–Wunsch global alignment (match +1 / mismatch −1 / gap −1) of full-length A. gambiae TOLL9 (Q5TWL7, 812 aa) against D. melanogaster Toll-9 (Q9VPH1, 900 aa) yields 35.6% identity (343 identical positions across 963 aligned columns). Both are single-copy members of the same PANTHER family/subfamily and share the LRR+TM+TIR architecture. This level of divergence is high enough that the dipteran LOF-negative is the most relevant genetic evidence available, yet imperfect enough that it cannot fully exclude a divergent function. (The Bombyx BmToll9 could not be resolved to a single reviewed UniProt accession, so an Anopheles-vs-Bombyx identity was not computed — a stated limitation.)

Mechanistic Model / Interpretation

The immediate molecular function under test is direct microbial-PAMP recognition (PRR activity) coupled to Toll-pathway signal transduction. The distinction that governs the curation call is:

 Direct gene-product activity          |  Downstream / inferred (NOT direct)
 -------------------------------------- | --------------------------------------
 - Transmembrane receptor topology  YES |  - AMP induction phenotypes (Drosophila LOF = none)
 - TIR-domain signaling capacity    ~   |  - Immune-gene co-regulation (mosGILT RNA-seq)
 - LPS/MD-2 recognition (Bombyx,        |  - "Innate immunity" as a pathway outcome
   chimeric TM/TIR)                 ~   |  - Inflammatory response (vertebrate-only, GO:0006954)
 - LPS recognition in Anopheles     NO  |

Three mechanistic separations drive the verdict:

1. Recognition of a processed host cytokine vs. microbial PAMP. Canonical Drosophila Toll binds the cleaved cytokine Spätzle, not microbes directly. The seed hypothesis correctly warns against assuming every insect Toll uses Spätzle and against equating host-cytokine recognition with GO:0038187 microbial-PAMP recognition. TOLL9's actual ligand in Anopheles is unknown.

2. Ectodomain binding vs. native signaling. In Bombyx, the ectodomain's LPS/MD-2 binding is demonstrated, but the productive signaling event used mouse TLR4 TM+TIR. This means the insect protein's own TM/TIR was not shown to transduce the signal — a critical gap when transferring an MF PRR term.

3. Correlation vs. perturbation. The Anopheles mosGILT^null induction of AGAP006974 is transcriptional co-regulation; TOLL9 was never knocked down or out. It cannot establish direct function.

Synthesizing: TOLL9 is architecturally a receptor, and the raw ingredients for an Anopheles Toll9/MD-2/LPS PRR complex exist (partner class present). But the chain of direct evidence for this species is broken at every functional link — no ligand assay, no partner interaction, no loss-of-function phenotype — and the closest genetic ortholog argues against a required antibacterial role. The proper model is a plausible-but-unproven immune receptor whose only defensible current annotations are generic.

GO Curation Implications

Lead recommendation (requires curator verification):

The evidence does not justify promoting TOLL9 from a generic signaling receptor to a specific microbial-PAMP pattern-recognition receptor at the species level. "Protein binding" is explicitly avoided as a recommendation; the informative and defensible MF term remains signaling receptor activity (GO:0038023).

Evidence Base

• PMID: 33963082 — Toll9 from Bombyx mori functions as a pattern recognition receptor sharing features with TLR4. Establishes the LPS/MD-2-like PRR property for insect Toll9 and situates Toll9 with the mammalian TLR4 group. Verified snippet: "Toll9 members differ from other insect Tolls by clustering with the mammalian TLR4 group, which recognizes lipopolysaccharide (LPS) through interaction with myeloid differentiation-2 (MD-2)-like proteins." Crucially, the functional signaling assay used a chimeric receptor with mouse TLR4 TM+TIR, so it demonstrates ectodomain ligand recognition but not native insect signal transduction.

• PMID: 21386906 — Lack of an antibacterial response defect in Drosophila Toll-9 mutant. Loss-of-function genetics in the closest dipteran ortholog directly weakens a required antibacterial immune role. Verified snippet: "Our results suggest that Toll-9 is neither required to maintain a basal anti-microbial response nor to mount an efficient immune response to bacterial infection."

• PMID: 35238667 — Review-level statement confirming the Toll9/MD-2/LPS PRR property is a Bombyx-specific experimental finding: "the Toll9/MD-2-like protein complex in the silkworm, Bombyx mori, acts as an innate pattern-recognition receptor that recognizes lipopolysaccharide (LPS)." Reinforces that this is a single donor species.

• PMID: 38191283 — Reports AGAP006974 induction after mosGILT disruption in A. gambiae (RNA-seq). This is expression correlation, not a TOLL9 perturbation, and cannot establish direct function.

• PMID: 25573502 — Crayfish PcToll (clusters with Drosophila/Anopheles Toll9 and human Tolls) is bacteria-inducible; RNAi and S2 overexpression implicate it in AMP regulation. A competing/context line of evidence from a distant species that shows the Toll9 clade can be immune-associated, but does not establish direct PRR activity for the mosquito ortholog.

Conflicts and Alternatives

• Paralog/clade confusion: Cross-species Toll9-clade immune associations (crayfish PcToll, PMID: 25573502) are suggestive but come from distant taxa and overexpression/RNAi, not the mosquito ortholog. They should not be transferred as direct evidence.
• Organism-specific divergence: The strongest genetic test (Drosophila LOF) is negative for antibacterial defense, and the strongest biochemical test (Bombyx) is chimeric and single-species. These pull in opposite directions and together argue for a species-level "unresolved/lead" status rather than assertion.
• Correlation ≠ causation: mosGILT^null induction of AGAP006974 is a co-regulation signal, easily produced by generic stress/immune transcriptional programs without TOLL9 having direct PRR activity.
• Database carry-over risk: Adding GO:0038187 by phylogeny would propagate a Bombyx-specific, chimeric-signaling finding across the family — precisely the over-annotation the review should avoid.
• Ligand-class ambiguity: Whether TOLL9's ligand is a microbial PAMP (→ GO:0038187) or a processed host cytokine (Spätzle-like) is unknown; conflating the two would mis-annotate the MF.

Limitations and Knowledge Gaps

1. No Anopheles-specific functional data. There is no LPS/PAMP binding assay, no ligand identification, no interaction assay with any MD-2-like partner, and no loss-of-function phenotype for AGAP006974. Everything about TOLL9's ligand and immune role in Anopheles is inferred. What would resolve it: direct LPS/microbial-binding assay and RNAi/CRISPR knockdown with infection challenge.

2. Chimeric-signaling caveat unresolved for any insect. Even the strongest donor (Bombyx) did not show native insect TM/TIR signaling. What would resolve it: reconstitute full-length insect Toll9 signaling in an insect cell system with an insect readout (e.g., Drosomycin-luciferase in S2 cells).

3. Bombyx–Anopheles distance not quantified. BmToll9 could not be mapped to a single reviewed accession, so the identity to the actual donor of the experimental property was not computed. What would resolve it: obtain the exact BmToll9 sequence from PMID33963082 and align.

4. Partner interaction unproven. ≥16 ML-domain proteins exist in Anopheles, but which (if any) partners TOLL9 is unknown. What would resolve it: co-IP / pull-down or AlphaFold-Multimer screening of the TOLL9 ectodomain against the ML-domain set.

5. Ligand identity. Microbial PAMP (GO:0038187) vs. processed host cytokine (Spätzle-like) is undetermined — these map to different GO terms. What would resolve it: ligand pull-down and Spätzle-processing assays.

Discriminating Tests

1. RNAi / CRISPR knockdown of AGAP006974 in A. gambiae followed by bacterial (Gram-negative/LPS) and fungal challenge, measuring AMP induction and survival — directly tests the immune/Toll-pathway claim in the target species.
2. Direct LPS/PAMP binding assay with recombinant TOLL9 ectodomain ± Anopheles ML-domain partners — tests GO:0038187 PRR activity directly.
3. Native insect signaling reconstitution (full-length TOLL9, not a mouse-TLR4 chimera) in S2 cells with a Drosomycin/AMP reporter — resolves the chimeric-TM/TIR caveat.
4. TOLL9–ML-domain interaction screen (co-IP or AlphaFold-Multimer across the ≥16 candidates) — tests the accessory-partner requirement.
5. Ligand-class discrimination: microbial PAMP vs. processed Spätzle-like cytokine — determines the correct GO term (GO:0038187 vs. cytokine-receptor terms).

Proposed Follow-up Actions / Curation Leads (require curator verification)

Action changes:
- Retain generic IBA annotations: GO:0038023 (signaling receptor activity, MF), GO:0007165 (signal transduction, BP), GO:0005886 (plasma membrane, CC).
- Do NOT assert GO:0038187 (PRR activity) or Toll-signaling/defense BP terms as species-level annotations. Record them as hypothesis leads with the Bombyx chimeric caveat and the Drosophila LOF-negative explicitly noted.
- GO:0006954 (inflammatory response) is out of scope here (already resolved by the vertebrate-only constraint).

Candidate references with snippets to verify:
- PMID: 33963082: "Toll9 members differ from other insect Tolls by clustering with the mammalian TLR4 group, which recognizes lipopolysaccharide (LPS) through interaction with myeloid differentiation-2 (MD-2)-like proteins" — supports PRR/LPS property, but note chimeric TM/TIR.
- PMID: 21386906: "Toll-9 is neither required to maintain a basal anti-microbial response nor to mount an efficient immune response to bacterial infection" — refutes required antibacterial role in the closest ortholog.
- PMID: 35238667: confirms the PRR property is Bombyx-specific (single donor).

Suggested curator questions:
- Is a chimeric-receptor demonstration (mouse TLR4 TM+TIR) sufficient to transfer an MF PRR term to a distant ortholog? (Recommended answer: no.)
- Should GO:0038187 be added only after an Anopheles-specific binding or LOF assay exists? (Recommended: yes.)

Suggested experiments: as listed under Discriminating Tests.

Summary

A. gambiae TOLL9 is a bona fide Toll/TLR-family transmembrane receptor whose generic annotations — signaling receptor activity (GO:0038023), signal transduction (GO:0007165), and plasma membrane (GO:0005886) — are well justified by domain architecture and the ancestral PTHR24365 transfer scope and should be retained. The seed hypothesis's stronger claims — that TOLL9 directly has pattern-recognition receptor activity (GO:0038187) and functions as an innate-immune Toll-signaling receptor — are biologically plausible but experimentally unproven in this species. The direct PRR/LPS evidence is confined to the more distant Bombyx mori and was demonstrated through a receptor chimera bearing mouse TLR4 transmembrane and TIR domains; the closest genetic test (Drosophila Toll-9 LOF, ~36% identical) shows no antibacterial defect; and the Anopheles mosGILT signal is merely correlative. The MD-2-like accessory partner class does exist in Anopheles (≥16 ML-domain proteins), so the mechanism is not excluded — but no TOLL9 ligand, interaction, or loss-of-function phenotype has been demonstrated. The correct curation posture is therefore to retain the generic receptor terms and treat PRR/immune terms as leads, not supported assignments.

Artifacts

• OpenScientist artifact evidence matrix
• OpenScientist artifact go decision table
• OpenScientist final report
• OpenScientist final report