Unc-93 homolog B1; ER-resident multi-pass (12-TM) membrane chaperone/trafficking factor for
nucleotide-sensing Toll-like receptors.
Falcon deep research (UNC93B1-deep-research-falcon.md) arrived after the review was written and
was compared against every decision.
- Agreement: ER-resident 12-TM MFS-like chaperone; binds TLR3/7/8/9 (and TLR5) via the
N-terminal six-helix bundle; escorts them ER -> Golgi -> endolysosomes; LoF (e.g. H412R, 3d)
traps TLRs in the ER and causes HSE susceptibility. All location, TLR-binding and transport
decisions, the KEEP_AS_NON_CORE treatment of TLR3/7/9 pathway rows, and the protein-binding
MODIFY/REMOVE decisions are consistent with the report. No annotation action changed.
- Additions (verified in primary papers): UNC93B1 also negatively regulates TLR7/TLR8 after
delivery, and gain-of-function variants cause monogenic SLE/chilblain lupus
PMID:38869500;
PMID:38780621.
Description extended with one sentence; both papers added to references.
- Additions (abstract-only, not used for annotation): UNC93B1 binds STING and delivers it to
lysosomes for degradation, attenuating cGAS-STING signalling PMID:35577759.
Single-laboratory report; raised as a suggested question rather than a NEW annotation.
- Nuance / mild tension: the report (via Song 2022, Front Immunol, PMID:35874766, not cached)
says Asn272 glycosylation of UNC93B1 is needed for MyD88 recruitment to TLR9, which qualifies the
older statement that UNC93B1 is dispensable for signal initiation PMID:18305481. Not verified
in a cached primary text; description wording left as is.
- Minor report inaccuracy: it calls the TLR3 structure "mouse" only; Ishida 2021 solved human and
mouse TLR3-UNC93B1 plus human TLR7-UNC93B1.
- Changes: description (GoF lupus sentence), 3 new references + falcon file reference, one
supported_by quote from the falcon file on a GO:0006886 row, one new suggested question.
UNC93B1 is a trafficking chaperone. Applying the CLAUDE.md participation test: the "work" of a TLR
signaling pathway (ligand binding, receptor dimerization, TIR/adaptor recruitment, kinase
activation) is performed by the TLR and downstream components, NOT by UNC93B1. UNC93B1 is
required for the pathway (knockout abolishes signalling), but this is necessity, not
participation — and the same paper that establishes necessity states UNC93B1 is dispensable for
signal initiation PMID:18305481. UNC93B1 acts upstream as the chaperone that positions the
receptor.
However, GO curators have consistently annotated both human (Q9H1C4) and mouse (Q8VCW4) UNC93B1
to GO:0034138/0034154/0034162 (TLR3/7/9 signaling pathway) with experimental (IMP) and
phylogenetic (IBA, PANTHER:PTN002928790) evidence, and the mouse rows use
acts_upstream_of_or_within. Per the "do not overrule curators from incomplete evidence" rule,
these experimental annotations are not removed. Instead they are marked KEEP_AS_NON_CORE: the core
molecular function is Toll-like receptor binding / protein carrier (trafficking) activity and the
core process is intracellular protein transport; the TLR signaling-pathway terms record the
downstream process UNC93B1 enables as an upstream chaperone, which is non-core.
Comparator: CNPY3 (Q9BT09), another TLR chaperone, carries signaling receptor binding (GO:0005102)
but is not annotated to the TLR signaling-pathway process terms in QuickGO — consistent with the
view that a chaperone's core representation is receptor binding + trafficking rather than pathway
participation. UNC93B1's pathway annotations are a curator convention grounded in strong knockout
data, retained here as non-core.
ER and ER membrane (core, IDA), Golgi membrane (transits), endosome and lysosome (relocalizes with
CpG stimulation), early phagosome / phagocytic vesicle (ortholog/subcell mapping, non-core). All
ACCEPT; consistent with the ER-to-endolysosome trafficking route.