ENTR1 (donkey, A0A8C4MKL2) notes

Sequence and record

• Unreviewed TrEMBL entry, 356 aa, "Endosome-associated-trafficking regulator 1"; PANTHER
  PTHR31259:SF3; Pfam ENTR1 coiled-coil region (PF29363) at 178..347
  [file:EQUAS/ENTR1/ENTR1-uniprot.txt "Belongs to the ENTR1 family."].
• No literature on donkey ENTR1. All 8 GOA rows are TreeGrafter IEA (GO_REF:0000118) from
  PTN009117290.
• Pairwise comparison to human Q96C92 (435 aa): 81.41% identity over 355 paired residues; the
  donkey sequence covers only 81.61% of human
  [file:EQUAS/ENTR1/ENTR1-bioinformatics/RESULTS.md "Paired coverage is 81.61% of the human sequence and 99.72% of the donkey sequence."].
  Inspection of alignment.txt: RESULTS.md lists human residues 9-85 as unpaired;
  donkey residue 9 onward aligns from human residue ~86. The donkey sequence also begins with Gln
  ("QPPVHAAP..."), not Met, so this entry is most likely an N-terminally incomplete gene model
  (UniProt does not flag it as a fragment). Seven donkey ENTR1 entries exist in UniProt
  (351-476 aa; UniProt REST query 2026-10-14), so longer models of the same gene exist.
• Consequence: the human N-terminal region (aa 1-100) is the one shown to be necessary and
  sufficient for basal-body localization, ciliogenesis rescue and IFT88 binding
  PMID:27767179.
  Human UniProt also maps GIT1 binding to the N-terminus
  [file:EQUAS/ENTR1/ENTR1-bioinformatics/human_Q96C92-uniprot.txt "Interacts (via N-terminus) with GIT1"].
  The conserved C-terminal coiled-coil region is intact.

Human ortholog biology

• Endosomal protein (early and recycling endosomes) with roles in receptor sorting, cytokinesis
  and ciliogenesis [file:EQUAS/ENTR1/ENTR1-bioinformatics/human_Q96C92-uniprot.txt
  "Endosome-associated protein that plays a role in membrane receptor sorting, cytokinesis and ciliogenesis"].
• SNX27-retromer cargo recycling: VPS35 interactome identified SDCCAG3; depletion caused
  trafficking defects of retromer cargo PMID:25278552.
  Human GOA (QuickGO 2026-10-14) has GO:0032456 endocytic recycling (IMP, PMID:25278552) and
  GO:0030904 retromer complex only as colocalizes_with (IDA, PMID:25278552).
• Cilia: basal body localization, IFT88 interaction, PKD2 ciliary localization
  PMID:27767179.
• Cytokinesis/midbody (PMID:23108400, not cached; cited via human UniProt text).

Decisions

• Endosome localizations (early, recycling): ACCEPT, core.
• Cytokinesis, midbody, centrosome: KEEP_AS_NON_CORE.
• Cilium-related terms: KEEP_AS_NON_CORE for the gene, with the caveat that this entry lacks the
  N-terminal region that mediates these roles in human.
• NEW: GO:0032456 endocytic recycling (ISS from human IMP), as the core process.

ProtNLM2 (donkey40 snapshot)

• GO:0032456 endocytic recycling -> COR (not in donkey GOA; supported by human IMP).
• GO:0030904 retromer complex -> UNC (human evidence is colocalization/interaction, not
  subunit membership).
• Non-GO: subcellular_location "Membrane" (not scored).

2026-10-21 — Full re-review, independent of the ProtNLM2 predictions

Scope: re-examined all 8 GOA rows, the previously proposed NEW row, description and
core_functions. ProtNLM2 outputs were not opened or used.

Deep research

• just deep-research-falcon EQUAS ENTR1 --fallback perplexity-lite failed on 2026-10-21:
  falcon returned "Edison API request failed: Failed to authenticate" (403) and the
  perplexity-lite fallback returned HTTP 401 insufficient_quota. No deep-research file was
  written; the review relies on the cached literature (PMID:23108400 abstract, PMID:24980502,
  PMID:25278552 and PMID:27767179 full text) and the human UniProt record.

Identity of the benchmark accession

• A0A8C4MKL2 (356 aa, starts "QPPVHAAP", no Met) vs canonical human Q96C92 (435 aa): lacks the
  region aligned to human residues 9-85, i.e. most of the N-terminal 1-100 segment that is
  necessary and sufficient for basal-body localization and IFT88 binding
  PMID:27767179.
  Coiled-coil Pfam PF29363 region intact. Conclusion unchanged: an N-terminally incomplete
  Ensembl gene model (ENSEASP00005027020.2) of the single donkey ENTR1 gene; six other donkey
  ENTR1 models exist (351-476 aa, e.g. A0A9L0K3Z4 476 aa, A0A8C4MT65 433 aa), all in SF3. No
  reviewed donkey ENTR1 entry exists, so there is no separate canonical donkey product to review.

Evolutionary placement (new analysis)

• ENTR1-bioinformatics/panther_subfamily_rereview.py -> RESULTS-rereview.md
  (UniProt REST, 2026-10-21). PTHR31259 has 4 subfamilies: SF3 (1520 entries, ~900 organisms,
  vertebrates to Dictyostelium; includes donkey, horse, mouse A2AIW0, rat, bovine) and three
  great-ape-only splits: SF4 (4), SF5 (3), SF6 (19; Homo, Pan, Gorilla, Pongo).
• Human canonical ENTR1 Q96C92 is in SF6, but a human ENTR1 fragment (A0A0A0MRP9, 223 aa) scores
  into SF3 [file:EQUAS/ENTR1/ENTR1-bioinformatics/RESULTS-rereview.md "| A0A0A0MRP9 | Homo sapiens (Human) | 223 | unreviewed | PTHR31259;PTHR31259:SF3; |"].
  ENTR1 is single-copy in mammals, so SF6 vs SF3 is an HMM partition inside one orthologous
  lineage, not a paralog split. Transfer of human experimental functions is acceptable at the
  ortholog level; reported as a subfamily-placement discrepancy.
• Human IBAs (QuickGO 2026-10-21): early endosome, recycling endosome, midbody, centrosome,
  regulation of cytokinesis from PTN002039065 (2019-04-10); ciliary basal body, positive
  regulation of cilium assembly, positive regulation of protein localization to cilium from
  PTN004793274 (2024-02-29). The single donor is human Q96C92. The 8 donkey TreeGrafter rows
  (PTN009117290) are exactly this set.
• The human IMP to GO:0032456 endocytic recycling (PMID:25278552) dates from 2014-11-25, so it
  was available at both PAINT passes and was not propagated; nor was colocalizes_with retromer
  complex. Mouse carries GO:0032456 only via Ensembl Compara (GO_REF:0000107) and ISO.

Changes made

• WITHDREW the previously proposed NEW GO:0032456 endocytic recycling (ISS) and removed it from
  core_functions.directly_involved_in. Reasons: (1) PAINT curators twice declined to propagate
  this human term while propagating all others, which the campaign treats as deliberate; (2)
  the human evidence is RNAi necessity data and cargo-selective
  PMID:25278552
  PMID:25278552;
  (3) ENTR1 associates with retromer indirectly, via WASH
  PMID:24980502.
  The recycling role is kept in prose (description, core function text) and raised as a
  suggested question.
• Added PMID:23108400 (now cached, abstract-only) as direct support for endosome, midbody and
  cytokinesis rows, and PMID:24980502 for the WASH-dependent retromer association.
• Actions on the 8 GOA rows unchanged (2 ACCEPT: early and recycling endosome; 6
  KEEP_AS_NON_CORE: centrosome, midbody, regulation of cytokinesis, ciliary basal body, two
  cilium regulation terms); reasons now cite the PAINT nodes and the SF3/SF6 placement.
• Modules: no module in modules/ includes ENTR1 (grep for ENTR1/SDCCAG3/Q96C92); the only
  retromer module is the Drosophila dmel_retromer.yaml, which does not include ENTR1. ENTR1 is
  not in gocams/index.tsv.
• Relative for next review: human ENTR1 Q96C92 (ortholog; sole experimental donor for the PAINT
  nodes behind every donkey row). No genes/human/ENTR1 review exists. Mouse Entr1 A2AIW0 shares
  SF3 with donkey but has only one experimental row (cytoplasm, EXP, PMID:16332174).