Research journal for the GO annotation review. Entries are appended; inline citations give the
PMID and the supporting text quoted from the cached publication.
smoothened was cloned as a segment-polarity gene required for cells to respond to Hedgehog,
and the sequence immediately suggested a receptor: PMID:8700230. The same
paper established the genetic requirement: PMID:8700230.
Patched, not Smo, binds the ligand; Ptc acts on Smo PMID:8898207.
UniProt topology (smo-uniprot.txt): 1036 aa, signal/precursor, seven transmembrane segments
spanning residues 259–553, and a cytoplasmic TOPO_DOM of 554–1036. That ~480-residue C-terminal
tail is the single most important structural difference from vertebrate SMO (787 aa in human, with
a much shorter tail), and it is where nearly all the fly-specific regulation happens.
Hedgehog inverts the fates of the two membrane proteins PMID:10966113. Surface
accumulation is not a side effect but the readout of activation
PMID:15616566, and the dose-response is graded
PMID:15616566.
Mass spectrometry mapped the sites: PMID:15598741, and
genetics showed the PKA/CK1 cluster architecture mirrors the one on Ci
PMID:15592457.
The mechanism of the switch is electrostatic
PMID:17960137 and
the output is dimerisation
PMID:17960137.
The three serine clusters are read in order, not in bulk
PMID:22537496.
Additional layers on the same tail:
- Gprk2 phosphorylates and also scaffolds: PMID:20844016; and it drives the opposite arm of the cycle
PMID:19850026. Gprk2 also acts indirectly through cAMP
PMID:22096079.
- Ubiquitination opposes accumulation: PMID:22253574, countered
by USP8 PMID:22253573.
- Trafficking: PMID:36083801 and, at the highest Hh levels,
PMID:36083801.
- Extracellular loops matter too: PMID:22223683.
This is the core fly-specific mechanism and the reason so many protein binding rows exist.
PMID:14597665 — binding and
function map to the same segment.
Directness was established by two-hybrid plus co-IP
PMID:14614827, with
PMID:14614827.
Independently: PMID:14636583 and
PMID:14523402.
The functional consequence is recruitment to the membrane
PMID:21844892 and, explicitly,
PMID:21844892. Concentration at the membrane is what activates Fused
PMID:21664578 and
PMID:21852395.
Curation consequence. This is why the bare GO:0005515 protein binding rows whose partner is
Costal-2 were modified to GO:0030674 protein-macromolecule adaptor activity rather than removed.
Smo is not merely in a complex with Cos2; its phosphorylated tail is the element that brings the
complex to the membrane, and deleting just that segment kills signalling. modules/hedgehog_signaling.yaml
assigns the same term to its smo_tail_recruitment annoton, so the review and the module agree.
Fused also binds Smo directly, and this is a reciprocal loop:
PMID:17182028;
PMID:17658259;
PMID:30541874; and the feedback also runs through Cos2
PMID:17671093.
PKA binds the tail as well, and Hh switches its preferred substrate:
PMID:24985345;
PMID:25289679.
PMID:18987629 and
PMID:18987629. This supports both GO:0004930 (IDA) and
GO:0007193 (IMP), and is the basis for modifying the broad InterPro GO:0004888 row.
Vertebrate SMO is activated by cholesterol/oxysterols occupying the CRD and a transmembrane channel.
In the fly the demonstrated lipid ligands are different:
UniProt's FUNCTION line carries "Activated by cholesterol in response to hedgehog (hh) morphogen"
but flags it By similarity (from human Q99835), not as fly evidence. Recorded as a
suggested_questions entry rather than asserted.
interpro/panther/PTHR11309/PTHR11309-review.yaml records assessment: UNRESOLVED for
GO:0005929 cilium at PAINT node PANTHER:PTN000885245, because the node asserts cilium for all
eumetazoan Smoothened descendants while fly Hh signalling is famously cilium-independent — yet
FB:FBgn0003444 (this gene) is one of the seeds. The family review could not inspect the fly
annotation offline and flagged it rather than pruning it.
It resolves in favour of the node. The fly GOA record carries its own experimental ciliary
annotation, GO:0005929 cilium, evidence IDA, reference PMID:24768000 (Kuzhandaivel et al.,
Cell Rep 2014), qualifier located_in. The paper is exactly on point:
UniProt independently records the location from the same study: the SUBCELLULAR LOCATION block in
smo-uniprot.txt has "Cell projection, cilium membrane {ECO:0000269|PubMed:24768000}", and TISSUE
SPECIFICITY has "Expressed in olfactory sensory neurons (at protein level)" with the same ECO.
So the apparent contradiction dissolves once the cell type is specified. Drosophila is not an
unciliated animal — it has ciliated sensory neurons and sperm; what is true is that the
developmental epithelia (imaginal discs, embryonic ectoderm) where the classical pathway was worked
out are unciliated, and there the Cos2/Fu complex substitutes for the ciliary compartment. In the
ciliated OSNs, Smo goes to the cilium and signals there. The eumetazoan node placement is therefore
grounded in protostome evidence as well as vertebrate evidence, and is not a vertebrate assertion
leaking into flies.
Two knock-on points:
- The same paper grounds the GO:0030425 dendrite (IDA) row, which is the fly seed for the
other UNRESOLVED row at PTN000885245. That assessment resolves the same way. In an insect
OSN the sensory cilium is elaborated from the distal tip of the dendrite, so ciliary and
dendritic Smo are the same observation at two resolutions.
- By contrast, GO:0071679 commissural neuron axon guidance at the same node has no fly seed —
the WITH/FROM is PANTHER:PTN000885245|RGD:3726|ZFIN:ZDB-GENE-980526-89, i.e. rat and zebrafish
only. Verified: RGD:3726 is rat Smo ("smoothened, frizzled class receptor"),
ZFIN:ZDB-GENE-980526-89 is zebrafish smo. The process presupposes a CNS midline that axons
cross. That row was actioned REMOVE with a propagation_review, consistent with the family
review's TOO_DEEP assessment and its observation that a vertebrate-scoped sibling node
(PTN002601731) already exists.
This is a useful contrast to keep straight: three rows at one node, two of which are sound because
the fly is genuinely a seed, and one of which over-reaches because it is not.
No contradictions found. The family review's PTHR11309:SF35 entry states that Smoothened
"does not bind Wnt and is not a Wnt receptor", that PAINT records losses of Wnt receptor activity,
Wnt-protein binding, and canonical/non-canonical Wnt signalling at PTN000885245, and that the node
asserts patched binding, smoothened signalling pathway, cilium and pattern specification. The fly
GOA record contains no Wnt terms at all, which is exactly what the recorded losses predict. The
GO:0007389 pattern specification process row was kept as non-core, matching the family review's
SOUND assessment and the human SMO review.
genes/human/SMO/SMO-ai-review.yaml takes the same actions on the shared rows: MODIFY
GO:0004888→GO:0004930, MODIFY GO:0007166→GO:0007224, MODIFY GO:0016020→plasma
membrane/ciliary membrane, ACCEPT plasma membrane / cilium / GO:0004930 / GO:0007193 /
GO:0007224, KEEP_AS_NON_CORE GO:0005113 patched binding, GO:0007389, GO:0030425 dendrite.
This review follows all of those.
Genuine fly/vertebrate differences recorded in the core functions:
1. The long C-tail adaptor function (GO:0030674) has no counterpart in the human core functions;
the human protein instead uses ciliary compartmentalisation and a PKI pseudosubstrate motif that
directly inhibits PKA-C.
2. Human core functions include GO:0015485 cholesterol binding and
GO:0004862 cAMP-dependent protein kinase inhibitor activity. Neither is demonstrated for the
fly protein in this annotation set — fly Smo instead binds PKAc as a substrate-enzyme pair
(PMID:24985345), which is close to the opposite relationship, and its demonstrated lipid ligands
are PI(4)P and PA.
3. Both share Gi coupling and cAMP lowering.
GO:0042981 regulation of apoptotic process (IGI, hid) → MARK_AS_OVER_ANNOTATED. The citedGO:0005515 with Sex-lethal (PMID:17284519) → REMOVE. The evidence is co-IP/co-fractionationGO:0005515 with Usp8 (PMID:22253573, PMID:22253574) → REMOVE. Here Smo is the substrateNEW-inflation the project rules warnGO:0005515 with cos via PMID:15691767 → MODIFY (not UNDECIDED). The cached record isreference_review rather than used as groundsGO:0042803 (IDA, PMID:21852395). The cached full text of that paper documents Smo-Cos2 FRETGO:0042802 identical protein binding on both. ACCEPTed; the curator readGO:0007224 that never name Smo in the abstract (PMID:20435030,NEW terms proposed. The obvious candidate would have been something for the ciliaryGO:0005929 and GO:0060170 already cover the compartment and GO:0007224 covers thesuggested_questionsGO:0071679 at PTN000885245 be moved to the vertebrate-scoped sibling node PTN002601731?