NTR: ER membrane protein retrieval receptor activity

[ValWood]

Please provide as much information as you can:

• Suggested term label:

ER membrane protein retrieval receptor activi

• Definition (free text)
  Combining with a retrieval signal (such as a C-terminal KKXX/KXKX-type dilysine motif, or an exposed transmembrane domain determinant) present on integral membrane proteins that have escaped from the endoplasmic reticulum (ER) to the Golgi, and mediating their return to the ER.

• Reference, in format PMID:#######

PMID: 12972550
Sato K, Sato M, Nakano A. "Rer1p, a retrieval receptor for ER membrane proteins, recognizes transmembrane domains in multiple modes." Mol Biol Cell. 2003;14(9):3605–16.

• Gene product name and ID to be annotated to this term

https://www.pombase.org/gene/SPAC22E12.05c

• Parent term(s)

GO:0005048
signal sequence receptor activity
sibling
GO:0046923 ER lumen protein retrieval receptor activity

• Children terms (if applicable) Should any existing terms that should be moved underneath this new proposed term?

• Synonyms (please specify, EXACT, BROAD, NARROW or RELATED)

• Cross-references

• For enzymes, please provide RHEA and/or EC numbers.

• Can also provide MetaCyc, KEGG, Wikipedia, and other links.

• Any other information

[pgaudet]

See also #31939

[ValWood]

Note that the correct BP for this MF is
GO:0006890 retrograde vesicle-mediated transport, Golgi to endoplasmic reticulum

[ValWood]

Claude:

Cargo receptor is the better fit, though "retrieval receptor" is more precise still — and the "signal sequence receptor" label is actively misleading here.

The trouble starts with the grouping itself. Signal sequence receptor is a term already spoken for: SSRα/β, now usually called the TRAP complex, sitting alongside Sec61 in the ER membrane. Together with SRP and the SRP receptor, those proteins recognize N-terminal signal peptides on nascent chains and commit them to translocation. That's a targeting-and-insertion logic — transient recognition of a feature on an unfolded polypeptide, with no vesicle involved.

KDELR does something categorically different. It reads a C-terminal tetrapeptide on a fully folded, soluble lumenal protein, and its job is to package that protein into a transport vesicle. Mechanistically it's the same device as ERGIC-53/LMAN1, the p24 family, or Vps10 and the mannose-6-phosphate receptors: a transmembrane protein with a lumenal binding site for cargo and a cytosolic determinant that engages coat machinery — in this case ARFGAP1 and COPI — plus a cycle of loading and unloading driven by the pH gradient (binds in the mildly acidic Golgi, releases in the more neutral ER) and a retrieval signal on its own tail so the receptor itself gets recycled. Bind cargo, recruit coat, ride the vesicle, release, return. That's the cargo receptor pattern exactly.

Two caveats worth carrying:

The direction and the cargo are unusual. KDELR handles retrograde flux, and its ligands aren't cargo in the secretory sense — they're escaped residents being repatriated. Calling it a retrieval or escapee receptor keeps that visible where flat "cargo receptor" doesn't.

It's also a signal transducer. KDELR is seven-pass, and ligand binding triggers Src-family kinase and PKA signaling that feeds back onto Golgi trafficking and secretory capacity. That moonlighting role is genuinely GPCR-like and sits outside the cargo receptor frame entirely.

If you want an umbrella that covers KDELR alongside LMAN1 and the MPRs without importing the translocation machinery, sorting receptor is the cleanest.

[ValWood]

So the existing placement seems to be incorrect for these.

They should be moved under GO:0038024 (https://www.ebi.ac.uk/QuickGO/services/ontology/go/terms/GO:0038024/complete)
cargo receptor activity

and probably add "sorting receptor" as synonym (or primary name).

The definition of
signal sequence receptor activity
Binding to a signal sequence, a short stretch of amino acids found in a protein that acts as a signal for its proper localization in the cell.

might be too loose because it doesn't exclude all cargo receptors.

[ValWood]

Also,

should KDEL receptors be included?

Strictly speaking, no. KDEL is a retrieval signal, not a signal sequence, and the distinction is worth preserving:

[ValWood]

Finally:
Cargo receptors. Unambiguously, and the reason is short.

The ligand doesn't determine the placement — the activity does. KDEL is a signal sequence in GO's broad sense, so "binds a signal sequence" is a true statement about KDELR. But it's true of every cargo receptor with a peptide ligand, and it describes what the protein recognizes rather than what it does. If ligand identity drove MF classification, the whole cargo receptor branch would collapse back into binding.

What KDELR does is vectorial: binds ligand in the Golgi, recruits COPI through its cytosolic tail, releases on the pH shift, cycles back. Recognition is one step in a transport cycle, not the function itself. That's GO:0038024's definition with no strain.

A genuine signal sequence receptor — SSR/TRAP, SRP — stops at recognition and hands off to a translocon. No vesicle, no cargo delivery, no cycle. Different activity, correctly a different branch.

[raymond91125]

@ai4c-agent please propose a new term for ER membrane protein retrieval receptor activity.

[ai4c-agent]

🤖 Working on it...

Follow along: View workflow run

— @ai4c-agent

[ai4c-agent]

Proposed new term in PR #32392.

GO:7770094 — ER membrane protein retrieval receptor activity (molecular_function)

Definition: "Binding to a retrieval signal, such as a C-terminal KKXX/KXKX-type dilysine motif or an exposed transmembrane domain determinant, present on integral membrane proteins that have escaped from the endoplasmic reticulum (ER) to the Golgi, and mediating their return to the ER." [PMID:12972550]

Parent: is_a GO:0005048 signal sequence receptor activity — placed as a sibling of GO:0046923 ER lumen protein retrieval receptor activity, mirroring your original request.

Synonyms: dilysine motif binding (NARROW), KKXX motif binding (NARROW), KKXX signal receptor activity (NARROW), endoplasmic reticulum membrane protein retrieval receptor activity (EXACT).

This is the minimally scoped change you asked for. On the broader question raised further up this issue (and in #31939) — whether both retrieval-receptor terms should be re-parented under GO:0038024 cargo receptor activity — I've deliberately left that alone: keeping the two ER retrieval-receptor terms as siblings under the same parent means whatever you decide there can be applied to both at once, in a separate ticket. If you'd rather I re-parent this one now under cargo receptor activity (and follow up with GO:0046923 too), say the word and I'll open a second PR.

Ontology validates cleanly (make travis_build).

---

🤖 Generated by @ai4c-agent

• Model: claude-opus-4-7
• Agent harness: claude-code
• Triggered by: @raymond91125
• Run: View workflow run