EHD3 curation notes

Deep research status

• just deep-research-falcon human EHD3 was run (unavailable perplexity-lite fallback, then --timeout 2400). See the end of this file for the outcome. The review was written from cached primary literature.

Key findings (with provenance)

• Resides in recycling tubules/vesicles; interacts with EHD1 PMID:12121420
• Endosome-to-Golgi transport and Golgi morphology PMID:19139087; Golgi effects are indirect PMID:19139087
• Recycling PMID:17233914; stabilizes tubular recycling endosomes PMID:27189942
• Cardiac ankyrin-B-associated membrane protein targeting PMID:20489164
• Ciliogenesis, partly redundant with EHD1 PMID:25686250; PMID:25686250; PMID:25686250
• The 2000 cloning paper (PMID:10673336) only predicts a nucleotide-binding site and an NLS; nucleic acid binding and nucleus TAS rows were REMOVEd.

Curation decisions (summary)

• GTP binding IEA: MODIFY -> ATP binding (family evidence PMID:15710626). Ciliary membrane IEA: MODIFY -> ciliary pocket membrane.
• Protein binding x16, nucleic acid binding TAS, nucleus TAS: REMOVE.
• Recycling, endosome-to-Golgi, cilium assembly, ciliary pocket, HPA rows, recycling endosome membrane: ACCEPT.
• Cardiac physiology terms, Golgi-to-lysosome transport, regulation of Golgi organization: KEEP_AS_NON_CORE.
• NEW ciliary vesicle assembly (GO:1905556, IMP, PMID:25686250), noting redundancy with EHD1.

HPA cilium atlas vs module role

• Module: stage 2_ciliary_vesicle, "ciliary vesicle formation", membrane-shaping ATPase with EHD1.
• HPA v25: Primary cilium (S), Primary cilium transition zone (S); main location plasma membrane. GOA HPA rows (cilium, ciliary transition zone, plasma membrane) accepted.
• Assessment: consistent with the module, with a caveat: EHD3's contribution is partly redundant with EHD1 and cell-type dependent (dispensable in RPE1, where EHD1 is >5-fold more abundant). core_functions keep ciliary vesicle assembly but list it alongside the non-ciliary recycling/endosome-to-Golgi and cardiac targeting roles. Unlike EHD1, I did not assert an ATP hydrolysis MF for EHD3 because direct EHD3 ATPase data were not available in the cache.

Deep research outcome

• Falcon succeeded on retry (--timeout 2400): EHD3-deep-research-falcon.md. Consistent with the review: EHD3 favors tubulation/stabilization (EHD1 and EHD4 favor vesiculation), early endosome -> ERC and endosome-to-Golgi transport, preciliary membrane/ciliary pocket role with EHD1, cardiac NCX1/CaV1.2/CaV3 targeting via ankyrin-B, glomerular endothelial VEGFR2 trafficking (with EHD4 compensation). It flags that no EHD3-specific ATP-hydrolysis data were retrieved, supporting the decision not to assert an ATPase MF for EHD3 in core_functions.