GPR75 review notes

Gene: human GPR75 (UniProt O95800), class A rhodopsin-like GPCR, 540 aa, chr 2p16.
Reviewed as part of a contested-function batch (see also GPR158, GPR25).

The contested annotation

GOA carries GO:0016493 C-C chemokine receptor activity three times on human GPR75,
with IBA, IEA (Ensembl ortholog transfer) and ISS — and no experimental
evidence code anywhere. The paired process term GO:0070098 chemokine-mediated
signaling pathway is carried on the same footing (IEA + ISS).

Where the IBA actually comes from (checked, not assumed)

The brief's premise was that the IBA has no experimental donor. That is not quite
right, and it matters. The WITH/FROM field of the human IBA is
MGI:MGI:2441843|PANTHER:PTN002796002. MGI:2441843 resolves to mouse Gpr75
(confirmed at informatics.jax.org), and QuickGO shows that mouse Gpr75 carries
GO:0016493 with IDA from PMID:17001303. So the IBD behind this IBA is
experimentally grounded — in exactly one descendant, the mouse orthologue.

Two consequences:

1. Per CLAUDE.md, donor count is not a proxy for evidential strength, and a node
   seeded by one well-characterised MOD gene can be sound. So the propagation is not
   mechanically defective, and REMOVE on "no experimental donor" grounds would have
   been wrong.
2. But because the only experimental descendant cited is the mouse orthologue, the IBA
   carries no information beyond the mouse-to-human transfer already recorded twice
   as ISS (UniProtKB:Q6X632) and as the Ensembl Compara IEA (also Q6X632). All four
   rows reduce to a single 2006 heterologous-expression study.

That study:
PMID:17001303 and
PMID:17001303

It was later supported in human neuroblastoma cells lacking the canonical CCL5 receptors:
PMID:29772059 and
PMID:29772059, concluding
PMID:29772059

So the claim is not fabricated. It is, however, a proposal that the field has declined
to adopt.

Why it is nevertheless an over-annotation

1. IUPHAR still calls GPR75 an orphan.
PMID:40362321 and
PMID:40362321

2. Both candidate agonists are explicitly contested in 2025-2026 reviews.
PMID:40757922
PMID:40362321

3. The 2026 cryo-EM structures argue against any orthosteric ligand, chemokine
included.
PMID:41545757 revealing
PMID:41545757, and the receptor is constitutively active without a ligand:
PMID:41545757

A collapsed extracellular domain is a harder problem for an 8 kDa chemokine than for a
small molecule: chemokine receptors engage their ligands over a large N-terminal /
extracellular-loop surface (CRS1) before the ligand N-terminus inserts into the pocket
(CRS2). Neither site appears to exist here.

4. GPR75 is not a chemokine receptor by descent. Its closest relatives by
transmembrane sequence are peptide receptors of other families, not CCRs:
PMID:40362321

What GPR75 is solidly

A plasma-membrane, Gq-coupled class A GPCR:
PMID:40362321, with a Gq-bound cryo-EM structure
(PMID:41545757), and the only thing GOA should confidently assert about its molecular
function is GO:0004930.

Its genetics are strong and independent of the ligand question: human loss-of-function
variants protect against obesity, and rodent knockouts are lean and insulin-sensitive.
Those are organismal phenotypes and do not license a molecular-function claim.

Curation position taken

• GO:0016493 C-C chemokine receptor activity (IBA, IEA, ISS — all three) →
  MARK_AS_OVER_ANNOTATED, with a propagation_review on the IBA recording that the
  IBD's only experimental descendant is mouse Gpr75 itself, so the phylogenetic
  annotation adds nothing over the ISS/Ensembl transfers it duplicates.
  Not REMOVE: the mouse IDA is a real experimental annotation whose full text was not
  read here, and three groups have reported CCL5-dependent, GPR75-dependent responses.
• GO:0070098 chemokine-mediated signaling pathway (IEA, ISS) →
  MARK_AS_OVER_ANNOTATED, same footing, same reasoning.
• GO:0004930 G protein-coupled receptor activity (IEA, ISS) → ACCEPT, core.
  Deliberately generic: this is the correct level of claim for a receptor IUPHAR still
  lists as orphan, and the Gq-coupled structure confirms it.
• GO:0007186 GPCR signaling pathway (IBA, IEA, ISS) → ACCEPT, core.
• GO:0005886 plasma membrane (IBA, IEA, IDA/HPA, IC, TAS) → ACCEPT.
• GO:0016020 membrane (IEA) → ACCEPT (generic but correct).

Unresolved

• Does CCL5 act on GPR75 directly, or through a co-receptor / transactivation route that
  would explain GPR75-dependence without GPR75 being the chemokine's receptor?
• If the orthosteric pocket is collapsed and the receptor is constitutively active, is
  GPR75 regulated by an allosteric ligand, by a protein partner, or by expression level
  alone?
• The 20-HETE claim is not annotated in GOA for human GPR75 and is not adjudicated here,
  but it is under the same cloud as the CCL5 claim.