Atlastin-1 couples GTP hydrolysis, self-association and conformational change to ER membrane tethering and fusion. The review contains 70 original assertions: 34 ACCEPT, 18 KEEP_AS_NON_CORE and 18 UNDECIDED. Both deposited ATL1 products are retained. No NEW annotation, extra broad process or separate generic-binding/GTPase core is proposed. The single core uses GO:0140523, GO:0098826 and the existing fusion process GO:0016320.
The 18 unresolved PMID:32814053 screen pairs have positive curator-recorded interaction counts. Their exact original pair tables, constructs and validations were not accessible. This is a source-access boundary, not a finding that the interactions are false or nonspecific. The supported spastin, reticulon, REEP and other contextual associations are retained as non-core. No more specific molecular activity is invented from coimmunoprecipitation alone. The user action definitions govern these decisions; lack of functional information alone is not treated as biological contradiction.
PMID:38509071 separates human ATL1 cytosolic 1–447 single-molecule/GTPase measurements, human full-length 1–558 yeast complementation and Drosophila 1–541 lipid mixing. S398Y/N440T are disease variants; H443P is engineered. Its Figure 3 caption says Drosophila K418P while the Results body says H418P. This source discrepancy is retained. No full-length human lipid-mixing or content-mixing experiment is attributed to that paper. PMID:34817557 provides complementary selected original evidence of canonical human ATL1-1 lipid mixing, and its recovered XML record has now been read as documented below. ATL2 splice-form results are not assigned to ATL1 product 2.
PMID:19665976 distinguishes human/recombinant ATL1 perturbation and association experiments, native rat tissue co-IP, HeLa ATL2/3 depletion and Xenopus network assays. PMID:27619977 distinguishes tagged ATL1 in U2OS cells from Xenopus extracts and ATL2/3 junctional enrichment. K80A and its dimerization-defective controls support the network mechanism, with expression-dependent limitations. Earlier Golgi/cis-Golgi and oligomer-size observations are retained in context; they are not converted into a constant physiological tetramer or a primary Golgi-fission activity.
IBA rows are retained in agreement with direct human evidence. The PAINT node/MSA placements were not replayed. Target self-inclusion is legitimate descendant grounding, and donor-list length is not an evidence score. The source objects, partner isoforms, qualifiers and evidence codes are unchanged. The local GO-CAM index previously showed no ATL1/Q8WXF7 entry; no broader absence is inferred.
The normal primary seed and nine auxiliary PMID imports are complete. All 14 original PMID records now exist canonically. The ordinary provider attempt failed and produced no report; manual primary-source work is documented here without a fabricated provider output. The ordinary non-force publication attempt produced no new records before the reviewed recovery import. Earlier source records remain byte-identical.
The additional PMID:34817557 source has now been read. Its full-text availability, title and literal supporting passage are resolved without changing annotation actions or source objects. The separately validated source import precedes canonical application.
The authoring preparation independently reread all 14 canonical abstracts. Detailed external and cached Results/Methods below are attributed to the prior primary-source consultation, not claimed as newly reread in their entirety. Figure pixels, full supplements and raw interaction matrices remain uninspected.
The preserved core-source and post-seed reading records provide additional assay detail and the explicit Drosophila residue-notation correction. They remain part of the preparation inputs.
The exact recovered normal cache was read from publications/PMID_34817557.md inside tmp/source64-verified-artifact.zip from the validated acquisition archive. Its SHA-256 is c091ef752d8ed4fd8d1b64eac2dcb24a2aa002aeb28040653de58a0d4126bf2c. Source acquisition and import are recorded separately from the scientific reading.
Source: Crosby et al., Reconstitution of human atlastin fusion activity reveals autoinhibition by the C terminus, PMID:34817557(https://pubmed.ncbi.nlm.nih.gov/34817557/), doi:10.1083/jcb.202107070. The cache identifies PMC8624677 and full-text XML extraction.
Actual reading comprised the complete cached abstract/frontmatter, the introductory canonical-ATL1 paragraph, ATL1 fusion Results and Figure 1 caption, and the construct, purification, GTPase and lipid-mixing Methods (cache lines 84, 103, 106, 111, 237, 243, 249, 255). Human canonical ATL1-1, purified from HEK293-derived Expi293 cells with N-terminal tags, was reconstituted into donor/acceptor liposomes. GTP-dependent lipid dequenching supports direct ATL1 membrane fusion. R77E and amphipathic-tail alterations impaired this readout. The Methods distinguish lipid mixing from inner-leaflet assays. I did not inspect figure pixels, all supplements, or the whole paper, and do not assign ATL2-specific inner-leaflet/splice-form conclusions to ATL1.
This resolves the supporting-reference boundary for the existing fusogenic core and row 51 without changing any of the 70 annotation actions, their original evidence/source objects, or the two deposited products. No new annotation or new isoform-specific claim is proposed.
ATL1 is an ER membrane fusogen whose cytoplasmic GTPase and three-helix bundle couple nucleotide state to homotypic association. The human soluble-domain studies support nucleotide-dependent dimerization and tethering: "homotypic interactions are timed by GTP binding and
hydrolysis." PMID:23334294, Structural basis for conformational switching and GTP loading of the large G protein atlastin. These fragments establish conformational steps; they do not by themselves constitute a membrane-fusion assay.
The 2024 single-molecule study refines the ordering: "ATLcyto can form a loose
crossover dimer upon GTP binding, which is tightened by GTP hydrolysis" PMID:38509071, Dissecting the mechanism of atlastin-mediated homotypic membrane fusion at the single-molecule level. Thus the force-generating transition should not be reduced to a model in which crossover begins only after hydrolysis. The study links phosphate release to dimer disassembly and GDP dissociation to resetting the monomer. Its human cytosolic measurements and Drosophila full-length lipid-mixing experiments remain separate, as detailed above.
Direct human membrane-fusion support comes from the complementary proteoliposome experiment: "ATL1 catalyzed lipid mixing in a GTP-dependent manner" PMID:34817557, Reconstitution of human atlastin fusion activity reveals autoinhibition by the C terminus. The endpoint is lipid dequenching. This establishes the human canonical ATL1 fusion contribution without assigning ATL2-specific splice-form findings or an uninspected content-mixing assay to ATL1.
GO:1990809 ER tubular network organization and GO:0007029 ER organization remain accepted consequences of ATL1 performing ER fusion. They are represented by the more specific fusion process in the single core-function synthesis, rather than repeated as separate core activities. No annotation action or source assertion changed in this follow-up.
The CPT1C paper (PMID:25751282) supplies a relevant ATL1 association but principally studies CPT1C; its reference relevance is now MEDIUM. The broad interactome paper (PMID:32814053) has LOW relevance to the present ATL1 functional synthesis because its exact ATL1 pair evidence remains inaccessible. Its identifier was confirmed against PubMed, while its pair-specific correctness remains UNVERIFIED and the 18 annotation decisions remain UNDECIDED. These are distinct judgments about identity, relevance and evidential support.
PR #3545 requests REMOVE for ten supported generic protein-binding records to satisfy the repository policy. The user's action definitions reserve removal for incorrect annotations, so genericity alone does not justify that change here. The ten records remain KEEP_AS_NON_CORE, the ten validation warnings remain disclosed, and the review remains DRAFT pending resolution of that policy difference. This follow-up does not claim that the requested change has been accepted or the PR approved.