Working journal for the GO annotation review. Append-only; newest sections at the bottom.
PTCH1 is a 1447-residue polytopic membrane protein with 12 transmembrane helices and two
large extracellular domains (ECD1, ECD2), plus a long cytoplasmic C-terminal tail. The
UniProt feature table lists TM1 at 100..117 and a second cluster of eleven helices from 438
onward, with the sterol-sensing domain annotated at 438..598. The cryo-EM work says the same
thing independently: PMID:29954986
It belongs to the RND-permease superfamily (patched family; UniProt SIMILARITY: Belongs to
the patched family) and is the vertebrate orthologue of Drosophila ptc. Human has a
paralogue, PTCH2 (Q9Y6C5), in the same PANTHER family; the module review
(modules/hedgehog_signaling.yaml) groups Drosophila ptc, human PTCH1 and human PTCH2 in the
receptor tier under PANTHER:PTHR46022.
Four alternative-promoter/splice isoforms are recorded (Q13635-1 "L", -2 "L'", -3 "M", -4
"S"). Nothing in the GOA set is isoform-scoped, and I found no evidence that the reviewed
functions are isoform-specific, so no isoform: fields were added.
PTCH1 is the high-affinity receptor for all three mammalian Hedgehog ligands. The original
biochemistry is unambiguous and covers the whole ligand family at once:
PMID:9811851
That single sentence is the evidential basis for three of the GOA rows (the three
GO:0097108 hedgehog family protein binding IPIs against mouse Shh/Ihh/Dhh) and for the two
GO:0005119 smoothened binding IPIs.
The structures then explain how. The dominant contact is not protein–protein at all but
lipid-in-pocket: PMID:29995851. The
second lipid does the same on the other end — PMID:31548691
Stoichiometry is two receptors per ligand, through two chemically distinct epitopes:
PMID:30139912, and at higher order PMID:31127104 This matters for the review because it
is the one defensible reading of the orthology-transferred GO:0005113 patched binding row
(see §5.1): PTCH1 genuinely engages another PTCH1.
The modern model is that PTCH1 is not a classical signal-transducing receptor but a
transporter that keeps cholesterol away from SMO, and that ligand binding switches the
transporter off. UniProt states it directly: In absence of hedgehog, acts as an inhibitor of
smoothened protein (SMO) by preventing SMO access to cholesterol, with a catalytic activity
line Reaction=cholesterol(in) = cholesterol(out).
Evidence chain:
This is exactly what modules/hedgehog_signaling.yaml encodes for the patched_receptor
annoton (MF GO:0140303 intramembrane lipid carrier activity, process
GO:0045879 negative regulation of smoothened signaling pathway, location plasma membrane),
and what the GO-CAM models record: gocams/index.tsv gives PTCH1 the same MF/BP pair with
location GO:0098804 non-motile cilium membrane in six human models
(696022cd00000812, 696022cd00000908, 696022cd00001146, 696022cd00001204, 696022cd00001246,
696022cd00001370). No conflict with the module or the family reviews was found.
One genuine tension worth recording: the 1998/2001 co-immunoprecipitation evidence that PTCH1
and SMO form a complex (PMID:9811851 above; PMID:11278759) sits awkwardly with the
catalytic/sub-stoichiometric model, in which PTCH1 acts on the sterol pool rather than on SMO
directly. I did not remove the smoothened binding rows on that basis — they are experimental
IPIs and the observation stands — but they are marked non-core, because the core molecular
function is the sterol transport, and that is also how the module and every GO-CAM model
render the PTCH1→SMO relationship (causal inhibition, not binding).
Plasma membrane and, functionally, the membrane of the non-motile primary cilium. UniProt:
Cell membrane ... Multi-pass membrane protein and Cell projection, cilium membrane, with
the note that PTCH1 localises to the primary cilium and prevents SMO from getting there. The
caveolar/raft subpool is real but came from one 2001 study:
PMID:11278759
Ligand binding triggers internalisation and degradation — UniProt PTM records ITCH
ubiquitination at Lys-1426 in the unliganded state and SMURF1/2 ubiquitination in response to
SHH, degradation is essential for PTCH1 clearance from the primary cilium and smoothened
pathway activation. Reactome R-HSA-5632677 ("PTCH is internalized") is the source of the
GO:0030666 endocytic vesicle membrane row; this is a transit compartment on the way to
degradation rather than a site of function, hence non-core.
Germline loss of function causes nevoid basal cell carcinoma (Gorlin) syndrome
PMID:8681379, with truncating mutations
predominating PMID:8981943 A nice period detail that anticipates the transport model:
PMID:8981943
UniProt also lists basal cell carcinoma (MIM:605462) and holoprosencephaly 7 (MIM:610828).
The human-genetics IMPs in GOA (neural tube patterning, limb morphogenesis, pharyngeal system
development, somite development) all trace to the Gorlin phenotype and were kept as
non-core: they are genuine consequences of losing Hedgehog repression, but they are
downstream developmental outcomes, not the protein's own activity.
Four human rows come by Ensembl-Compara transfer from mouse Ptch1 (Q61115) annotations whose
source is PMID:10500113, "Sonic hedgehog protein signals not as a hydrolytic enzyme but as an
apparent ligand for patched" (checked on PubMed). In mouse those rows are
GO:0005113 patched binding (IDA), GO:0008201 heparin binding (IDA),
GO:0001841 neural tube formation (IDA) and GO:0007165 signal transduction (IDA).
GO:0042802 identical protein binding ratherGO:0001843 neural tube closure IMP in mouse). Kept, non-core.GO:0045879 is directlyTwelve human IEA rows carry rat Ptch1 (Q6UY90 / ENSRNOP00000026287) in WITH/FROM. Tracing
them through the GO API, the rat originals split cleanly:
| rat term | evidence | rat source |
|---|---|---|
| spermatid development, prostate gland development, liver regeneration, response to xenobiotic stimulus / mechanical stimulus / estradiol / retinoic acid / alkaloid | IEP | 8 separate expression studies |
| dendritic growth cone, axonal growth cone, postsynaptic membrane | IDA | PMID:21618238 (subcellular localisation of Patched and Smoothened in hippocampal neurons) |
| commissural neuron axon guidance | IMP | PMID:19946319 |
The IEP block is the weakest material in the whole GOA set: PTCH1 is the canonical
transcriptional readout of Hedgehog pathway activity, so its mRNA moves in essentially any
tissue where the pathway is engaged. "Responds to X" derived from that is a statement about
pathway activity, not about PTCH1's function, and it was transferred to human with no human
evidence at all. All eight marked over-annotated. The three IDA localisations and the axon
guidance IMP rest on real perturbation/imaging experiments and were kept as non-core.
Two human IMPs come from this paper. The abstract carries the transcription claim
PMID:16229683 and the full text carries the differentiation assay — figure legend
PMID:16229683 with the
text stating that PTCH1 significantly inhibited GLI1-induced AP staining in C3H10T1/2 cells.
Both rows are supported, but both are transfection/overexpression readouts of a non-canonical
arm, and PTCH1 is not itself a transcriptional regulator, so both are non-core.
Source of one IDA (GO:0045177 apical part of cell) and two IEPs (GO:0048745 smooth muscle
tissue development, GO:0072205 metanephric collecting duct development). The IDA is direct
and quotable PMID:17850284. The two IEPs are expression correlations in fixed human tissue,
and the authors say so themselves PMID:17850284 — hypothesis-generating, hence over-annotated.
GO:0005515 protein binding (7 rows)Per CLAUDE.md, only MODIFY / REMOVE / UNDECIDED are available. Partner identities were taken
from the UniProt INTERACTION block and each accession checked against UniProt:
GO:0030332 cyclin binding, whichGO:0019901 protein kinase binding; GRK2 is a kinase and the interaction is the paper'sGO:0097108 hedgehog family protein binding.GO:0140597 protein carrier activity in GO-CAM 696022cd00000908.Removal here never means the interaction is false — only that the bare term records nothing
about what PTCH1 does.
GO:0016485 protein processing (ISS from mouse)The mouse source is an IMP in PMID:16061793 ("Cilia and Hedgehog responsiveness in the
mouse"), qualified acts_upstream_of_or_within — i.e. GLI processing changes when Ptch1 is
perturbed. The human row asserts involved_in. PTCH1 has no proteolytic activity and cleaves
nothing; the GLI3 processing effect is a downstream consequence of pathway repression that
GO:0045879 already captures. Removed on the participation test.
GO:0010875 positive regulation of cholesterol efflux (IDA, PMID:21931618)Kept in substance but retermed. The paper's own data have Patched performing the movement
(purified protein binds cholesterol; heterologous expression in yeast boosts efflux), not
regulating someone else's transporter, and the direction/destination ("out of the cell") is
precisely the part that later work reframed as leaflet-level accessibility. Modified to
GO:0034204 lipid translocation, which is what the structures and simulations support and
which GOA already carries here as an inter-ontology IEA from GO:0140303.
GO:0072659 protein localization to plasma membrane (IDA, PMID:11278759)The supporting observation is PMID:11278759 — note the authors' own "seems to be". The direction of PTCH1's effect on SMO
localisation is now understood the other way round (PTCH1 restricts SMO accumulation at the
ciliary membrane; UniProt: prevents cilium localization of SMO). Marked over-annotated
rather than removed, since PTCH1 does control SMO's membrane distribution — just not by
delivering it there. The companion general term GO:0032880 regulation of protein
localization was modified to GO:1903565 negative regulation of protein localization to
cilium, which states the established direction.
Checked against the participation and comparator tests. The module's PTCH1 annoton
(GO:0140303 / GO:0045879 / plasma membrane) and all six GO-CAM activities in
gocams/index.tsv use terms that are already present in the GOA set, including
GO:0098804 non-motile cilium membrane. The candidate I did consider,
GO:1903565 negative regulation of protein localization to cilium, is introduced as a
MODIFY replacement for an existing row rather than as a new assertion, so nothing is added
that a curator has not already been in a position to add.