GPR50 (Q13585, MTR1L_HUMAN) — curation notes

What the gene is

617-aa class A (rhodopsin-like) 7TM receptor, cloned from human pituitary as "H9"
and named for its ~45% identity to the MT1/MT2 melatonin receptors
PMID:8647286.
Unusual features: no N-linked glycosylation sites and an exceptionally long
(>300 aa) cytoplasmic C-terminal tail. Expression is concentrated in hypothalamus
and pituitary (UniProt TISSUE SPECIFICITY), notably in tanycytes and the
dorsomedial hypothalamus.

The melatonin question — settled, and settled against melatonin

This is the one contested-ligand question for GPR50 that CAN be resolved.

The founding paper reports the negative result directly:
PMID:8647286
Clement et al. 2017 dissected why: swapping the second extracellular loop (E2)
residues between MT1 and GPR50 produced reciprocal loss- and gain-of-melatonin
responses, i.e. GPR50's E2 loop is the reason it is melatonin-blind
PMID:28898928
The 2026 cryo-EM paper restates it and provides the structural comparison
PMID:41666959
The Jockers lab states it flatly in the 2025 KO paper
PMID:40091563

Curation consequence. GOA carries GO:0008502 melatonin receptor activity
(IEA, GO_REF:0000002) with WITH/FROM = InterPro:IPR000025. IPR000025 is the
melatonin receptor family signature — GPR50 matches it by descent, not by
pharmacology. This is a textbook family-signature over-propagation and the
primary literature, including the paper that first expressed the protein,
contradicts it. Action: REMOVE.

Is GPR50 a G-protein-coupled receptor at all? Yes — constitutively.

Cryo-EM of the ligand-free receptor at 3.4 A plus TRUPATH BRET2 basal-activity
profiling: PMID:41666959
and PMID:41666959
Independently, GPR50-KO mice show a ligand-independent G12/13-RhoA output:
PMID:40091563

So the receptor activity and the G12/13-RhoA coupling are supported by two
laboratories using orthogonal methods (structure/BRET vs. KO mice + RhoA assays),
and both describe activity in the absence of an added agonist.

The contested deorphanization: L-LEN

PMID:41495223
L-LEN is a 10-aa peptide from the ProSAAS (PCSK1N) precursor, captured from mouse
hypothalamic extract by a genetically-encoded photo-cross-linker sited in the GPR50
binding interface
PMID:41495223
with Gi-type coupling
PMID:41495223

Why this is not yet settled.
1. The cryo-EM paper, published five weeks later, still describes the receptor as
having no characterized agonist:
PMID:41666959
It does not cite or engage the L-LEN result. Two 2026 papers asserting opposite
states of knowledge, neither addressing the other, is not consensus.
2. The pathway assignments disagree. L-LEN => Gai/cAMP inhibition; cryo-EM and the
KO mice => Ga12/13-RhoA (with Gai/o only "moderate"). A single receptor can be
promiscuous, but nobody has reconciled the two readouts in the same system.
3. L-LEN is a single-laboratory result. No independent replication yet.

Honest action: UNDECIDED is not applicable here because there is no GOA row for
an L-LEN pairing — nothing to adjudicate. The point is recorded in
suggested_questions and in the top-level description as a stated dispute.

Non-receptor / receptor-adjacent biology that IS well supported

Negative regulation of MT1 by heterodimerization. GPR50 heterodimerizes
constitutively with MT1 and MT2; in the MT1 heterodimer it abolishes
high-affinity agonist binding and G-protein coupling
PMID:16778767.
Note the irony: GPR50's best-documented connection to melatonin biology is as an
inhibitor of a melatonin receptor, which is the opposite of "melatonin receptor
activity".
Calpain cleavage and nuclear translocation of the C-terminal domain. CAPN1
cleaves between Phe-408 and Ser-409; the released CTD enters the nucleus and
binds GTF2I/TFII-I to regulate FOS transcription
PMID:31900622.
This is what the GO:0005634 nucleus EXP annotation records — it is the cleaved
fragment, not the intact 7TM receptor, that is nuclear.
NOGO-A (RTN4) interaction and synaptic enrichment
PMID:19699797
— the source of GO:0014069 postsynaptic density. Note the 2009 paper reported
GPR50 increasing neurite length on overexpression, whereas the 2025 KO paper
reports GPR50 restraining neurite outgrowth; the KO/loss-of-function direction
is the more reliable one.
KAT5/TIP60 and glucocorticoid receptor signalling PMID:21858214.
ADAM17 interaction / Notch in HCC PMID:32405532.

Actions taken

Term	Evidence	Action
GO:0008502 melatonin receptor activity	IEA (InterPro:IPR000025)	REMOVE
GO:0004930 GPCR activity	IBA, IEA, TAS	ACCEPT (constitutive)
GO:0007186 GPCR signaling pathway	IBA, IEA, TAS	ACCEPT
GO:0005886 plasma membrane	IBA, IEA, IDA, EXP, TAS	ACCEPT
GO:0005634 nucleus	IEA, EXP	ACCEPT (cleaved CTD)
GO:0014069 postsynaptic density	IEA, EXP	ACCEPT
GO:0016020 membrane	IEA	ACCEPT (redundant with plasma membrane)
GO:0005515 protein binding	IPI x3	MARK_AS_OVER_ANNOTATED
GO:0007267 cell-cell signaling	TAS (PMID:8647286)	MARK_AS_OVER_ANNOTATED

No NOT/negated qualifiers are present in the GOA file.