AGTPBP1 evidence notes

2026-09-27 — full annotation review

This review uses the normal source3 seed for human AGTPBP1 (Q9UPW5, HGNC:17258). The archived official HGNC subset records an approved symbol and aliases KIAA1035, Nna1 and CCP1. The coordinator independently checked current main d35dcc30b44924f79c0510b281ae824aa536848a, canonical/alias paths, task branch and separate open-PR searches before authoring. All canonical gene paths were absent on that publication base. The original seed has 26 PENDING annotations, nine references and three alternative products. No source assertion, evidence code, reference identifier, qualifier, negation flag or isoform was changed. UniProt and GOA remain byte-identical to the fetched seed.

The review and annotation-reviewer skills were applied, followed by core-function synthesis. This assigned reviewer performed the annotation review; a separate bounded consultation with annotation_aars2 independently checked the nuclear-localization sources. The genuine Falcon/fallback and initial publication-caching commands ran concurrently. Both stopped before provider/fetch dispatch because the uv build environment could not resolve PyPI dependencies (hatchling and uv-dynamic-versioning). Neither research provider returned a report. No provider-named file was created. Manual primary-source research continued, and the installed normal publication CLI was used for the finite additional reference sets. Logs: /tmp/AGTPBP1-provider.log, /tmp/AGTPBP1-initial-fetch.log, /tmp/AGTPBP1-primary-fetch.log, /tmp/AGTPBP1-donor-fetch.log.

Direct chemistry and core function

• PMID:22170066, Cytosolic carboxypeptidase 1 is involved in processing α- and β-tubulin. The local cache is abstract-only. The original full PMC article was read externally. Methods identify human CCP1-His6 expressed in Sf9; wild-type-baculovirus eluates provide preparation controls. Human HEK293T perturbations, purified porcine tubulin and polymerized HEK293T microtubules are distinct experimental components. Mass spectrometry and chelator sensitivity support terminal-glutamate processing by a metallopeptidase. The preparation is affinity-enriched, not described here as proven homogeneous. Cellular peptide-level measurements argue against bulk intracellular peptide turnover as its predominant assayed role.
• PMID:24022482, Cytosolic carboxypeptidase 5 removes α- and γ-linked glutamates from tubulin. Full PMC3798508 Methods and comparative assays were read. The title does not identify the complete experimental scope: human CCP1 and mouse CCP5 are directly compared. Synthetic glutamylated tails and tubulin distinguish terminal alpha-linked trimming by CCP1 from CCP5-mediated gamma-linked branch-point removal. The branch-point limitation is bounded to the tested tubulin substrates.
• PMID:21074048, A family of protein-deglutamylating enzymes associated with neurodegeneration. The primary PubMed abstract and DOI 10.1016/j.cell.2010.10.014 identify the Cell paper and the CCP-family chemistry. Its full body was not read here. It is a traced mouse donor source, not a substitute for the independently read human assays.
• PMID:29593216, Klf4 glutamylation is required for cell reprogramming and early embryonic development in mice. The full publisher article was read. Mouse knockout MEFs/ESCs provide most mechanistic work; human reprogramming and human CCP shRNA methods are separately described. In-vitro Klf4 deglutamylation and protein-stability experiments support non-tubulin substrate scope. The mouse embryogenesis findings do not become a new human process assertion, and the paper's discussion of possible substrate-dependent branch-point behavior is not treated as demonstrated tubulin chemistry.

The broad InterPro proteolysis row is refined to GO:0035609 C-terminal protein deglutamylation and GO:0035610 protein side chain deglutamylation. CCP1 itself performs the hydrolysis, so this is participation evidence rather than loss-of-function necessity alone. These two processes distinguish gene-encoded terminal residues from post-translational side chains. They are sibling terms under protein deglutamylation; they are not described as formal children of proteolysis. No NEW row is needed. The core keeps GO:0004181 metallocarboxypeptidase activity with the two measured processes. The current child GO:0106423 tubulin-tyrosine carboxypeptidase activity removes terminal tyrosine and must not be substituted for the CCP1 glutamate reaction.

Primary ontology checks used AmiGO GO:0004181, GO:0106423, GO:0035609, and the GO:0035608 definition and child list. The side-chain child's direct AmiGO/QuickGO/OLS requests timed out; its exact label and sibling placement were independently visible in the parent page. This access limit is not an assertion that the term is obsolete. No Q9UPW5 or MGI:2159437 entry was found in the cached GO-CAM index. The normal InterPro family CSV places human Q9UPW5 in PTHR12756:SF24, but no PAINT ancestral reconstruction was inferred from that membership file.

Compartments and mitochondrial process

• PMID:17244818, A novel subfamily of mouse cytosolic carboxypeptidases. The immutable cache contains the primary abstract. It explicitly reports cytosolic CCP1 in a mouse-family study. Full construct details were not recovered. The curated human EXP row is not rejected from the title; its broad cytoplasmic assertion is consistent with independent human enzymology. The early detyrosination hypothesis does not replace later direct deglutamylation evidence.
• PMID:23085998, Functional segregation and emerging role of cilia-related cytosolic carboxypeptidases (CCPs). The local cache and accessible publisher route are abstract-only. The family-level cilia/microscopy summary does not resolve the CCP1-specific nuclear panels. The immutable human UniProt record explicitly attributes nucleus and cytoplasm to this paper, including interphase/dividing-cell distribution. Curator deference is appropriate for these supported compartments; no other paralog's ciliary location is silently transferred.
• The primary HPA subcellular page, read 2026-09-27, rates nucleoplasm as supported. Antibody HPA071094 is listed for SH-SY5Y, SK-MEL-30 and U2OS nuclear staining. Additional plasma-membrane/ciliary locations have uncertain reliability. Nuclear/nucleoplasmic rows are retained non-core without manufacturing a new nuclear mechanism.
• PMID:23242554, Identification of CRM1-dependent Nuclear Export Cargos Using Quantitative Mass Spectrometry. The indexed primary PMC3591659 and author PDF resolve tagged CCP1 redistribution after leptomycin B, NES/truncation experiments and N-terminal CRM1 binding. The inspected Methods identify a gifted CCP1 plasmid but do not establish its species. Thus these results corroborate trafficking without being certified as direct human-protein assays. CCP1 is the export cargo, not the transport executor; no nuclear-export process is added.
• PMID:20620870, Mitochondrial dysfunction in NnaD mutant flies and Purkinje cell degeneration mice reveals a role for Nna proteins in neuronal bioenergetics. Indexed full PMC3101252 Figure 5A studies rat brain/spinal-cord fractionation; Figure 5B studies mouse liver subfractions with compartment markers. This is positive rodent localization evidence. It supports the mouse donor and a bounded whole-organelle human similarity transfer; neither human submitochondrial topology nor contamination is inferred.
• PMID:30337352, CCP1 promotes mitochondrial fusion and motility to prevent Purkinje cell neuron loss in pcd mice. The full publisher PDF was read. Figures 2–4 directly analyze human RPE1 CCP1-null cells and mitochondrial morphology/fusion. Figure 5 compares enzymatically active and inactive CCP1 under stress. Mouse neurons provide motility measurements. Reduced Drp1 dosage rescues fly phenotypes but the tested partial mouse reduction does not rescue pcd degeneration. The work supports a non-core organization role; its proposed tubulin/motility-to-fusion mechanism is not proof of an intramitochondrial catalytic substrate or a CCP1 fusion-GTPase activity.

The cytoplasmic/cytosolic/microtubule assertions remain core at their source resolution. Mitochondrial location and organization remain contextual. A protein may modify cytoskeletal substrates and also possess a mitochondrial-associated pool; those facts are not mutually exclusive.

Mouse donor provenance and developmental scope

The MGI Agtpbp1 GO graph was inspected as the historical 2023-03-10 snapshot it identifies, not presented as a current annotation release. Q641K1 is mouse Agtpbp1; the MGI sequence page and primary records resolve its source chains.

PMID:11884758, primary PubMed record, identifies Nna1 mutations in mouse pcd alleles and adult-onset neuronal degeneration. The full Science paper was not recovered. PubMed links an erratum in Science 2002;297:337, but its content was not recovered; no retraction, identifier remapping or biological correction is inferred. PMID:16465590, primary PubMed record, identifies the mouse pcd5J insertion and protein destabilization. Publisher/author routes did not yield the full paper. Both references are identity-verified but leave the developmental evidence unresolved.

PMID:30420557, Loss of tubulin deglutamylase CCP1 causes infantile-onset neurodegeneration. The full author manuscript supplies human patient fibroblast/muscle data and central/peripheral disease findings. Figure 3E/F and Methods use mouse CCP1-YFP variants in human HEK293 cells. The authors explicitly lacked direct human Purkinje histology and discussed a retinal phenotype difference. These boundaries matter for differentiation transfers. For the much broader GO:0050905 neuromuscular process, conserved enzymology and the human/mouse nerve/muscle evidence justify non-core retention; the term is not restricted to neuromuscular-junction formation.

The three UNDECIDED developmental rows do not assert that the curator was wrong, that human assays are required for every ISS transfer, or that degeneration excludes a developmental role. They identify the exact source-access and process-scope questions that remain. PAINT source blocks name only PTN ancestral nodes, preserve positive human self-evidence, and distinguish unverified node reconstruction from independently supported target function.

Provisional outcome and cache gates

All 26 source assertions are adjudicated: 14 ACCEPT, eight KEEP_AS_NON_CORE, three UNDECIDED and one MODIFY. There are no NEW annotations. One integrated core describes the deglutamylase chemistry. Three alternative products are retained without unsupported isoform-specific assignments.

The three original PMID caches are immutable and abstract-only. Nine additional cited PMIDs require normal cache recovery: 11884758, 16465590, 20620870, 21074048, 23242554, 24022482, 29593216, 30337352 and 30420557. Both normal CLI attempts terminated with exit code 1: zero of four primary requests and zero of five donor/nuclear requests were recovered, all due to DNS. No publication bytes were created or edited. External full-source access is documented above and does not imply a recovered local cache. No Reactome records or provider artifacts are cited. The YAML remains DRAFT while source-cache warnings persist.

Independent review and integrity checkpoint

The coordinator independently read all 26 rationales, 20 reference assessments, the core and questions, and rechecked the human enzyme/chelator/tubulin experiments, CCP1-versus-CCP5 linkage specificity and human RPE1 experiments. The biological review passed with no requested changes. The separate nuclear-source consultation agreed that the family-focused abstract does not refute curated human localization. The preservation check confirms all 26 source objects, all nine original reference ID/title pairs, all three alternative products and both machine files. All 30 cached quote occurrences match their sources; no uncached quote is used. The recursive authored/provider census contains 12 PMIDs, nine missing caches, no Reactome citations and no generated provider artifact.

Final validation

The full targeted validator terminated successfully (exit code 0). Its sole grouped warning lists the nine missing PMID caches above; all schema, ontology and best-practice checks passed. History validation and HTML rendering passed. The final manifest contains the four authored files, two exact source3 machine records and three exact source3 publication records, all absent from the verified publication base. The review remains DRAFT until the nine cache gates close.

2026-09-27 — source11 recovery and PR #3292 evidence follow-up

This entry supersedes the earlier cache-access statements without rewriting their history. The coordinator verified all five canonical gene files against published head 3b5465177521704acbf0395eb6b8fda9c04d27bd. Formal review 5329768612 and linked comment 5854670491 were read in full. All 26 seeded objects, 26 decisions, three alternative products and the integrated core are preserved. No NEW annotation or altered machine qualifier is introduced.

Source11 supplied nine exact normal-fetched records, independently verified before canonical import (receipt tmp/source11-canonical-import-receipt.json, SHA256 ab4bf2ada68208f29d9f0f618d80b33ee55c8956df0597002f8cec95d1b4f7a4). The new local full records are PMID:20620870, PMID:29593216 and PMID:30337352. Their extracted Results and Methods were reread; this does not claim that every supplementary image is present. PMID:11884758, PMID:16465590, PMID:21074048, PMID:23242554, PMID:24022482 and PMID:30420557 remain abstract-only records. The corresponding full_text_unavailable fields describe the repository cache, while reference notes separately identify externally read full originals. All cached titles match the fetched metadata exactly, including Greek letters. Citation identity remains VERIFIED where primary PubMed and source text were actually checked; unread full text is documented rather than conflated with an unverified identifier.

Evidence attached to the claim it supports

The repeated purified-tubulin sentence is retained for enzyme/substrate assertions, but replaced for compartment and physiological claims. Cytoplasm and cytosol rows now cite the actual cytosolic-distribution statement from PMID:17244818 and the curated UniProt subcellular statement. The human source record is not represented as a second independent experiment. The zinc row now cites the full UniProt catalytic zinc-binding feature at residue 920; residues 923 and 1017 have the same PROSITE-ProRule evidence. These sequence-based assignments and chelator-sensitive catalysis support the retained binding assertion, but do not constitute direct measurements of human metal occupancy or stoichiometry.

The HPA nucleoplasm row no longer cites its own UniProt GO cross-reference. Primary-source receipt, reread 2026-09-27: HPA AGTPBP1 subcellular page, “Localized to the Nucleoplasm (supported)”. The page lists antibody HPA071094 and nucleoplasmic staining in SH-SY5Y, SK-MEL-30 and U2OS; additional compartments are rated uncertain. The short quote is preserved here solely as a receipt of that primary page, not as independent evidence generated by these notes. Microscopy pixels were not newly reanalyzed.

For mitochondrial location, recovered PMID:20620870 Results explicitly distinguish rat brain/spinal-cord fractions, and Methods identify mouse liver submitochondrial fractions; the Figure 5B result reports outer-membrane enrichment. Both the method and result are now attached, keeping the human transfer at whole-organelle resolution. For organization, PMID:30337352 Figure 4 directly scores mito-mEOS.2-labeled fission/fusion events in human RPE1 CCP1-null cells. The lower measured fusion rate is attached to the non-core process row, with the existing microtubule-to-fusion mechanistic limitation retained. For the broad neuromuscular process, the new cached PMID:30420557 abstract supplies human cerebellar, spinal-motor-neuron and peripheral-nerve disease together with corresponding mouse pathology. This replaces an isolated in-vitro tubulin quote; it does not manufacture neurotransmission or developmental specificity.

Original 2002 main paper and remaining developmental limits

The original PMID:11884758 main article was newly recovered through Richard Sidman's author-uploaded copy, Science 295:1904–1906. The inspected original article begins with its title after the preceding article's end and ends before ResearchGate's related-publication list; the related articles are not treated as this paper's Results. It establishes the mouse pcd allele connection and postweaning degeneration. Figure 5 is in-situ expression in one-month-old mouse brain, retina and testis, including olfactory mitral cells. The main article separates its degenerative model from mutants with developmental neuronal loss. These positive results do not by themselves establish the formation-to-maturity progression defined by GO:0021772.

The linked supplementary material was not recovered. PubMed and the original PDF cover identify a correction in Science 2002;297:337 and the exact legacy publisher notice URL. Direct publisher routes did not expose the notice body, and no separate correction PMID or DOI has been established. This is an explicit source-access limit, not a retraction, remapping, or inferred change to a particular experiment. The olfactory-development row remains UNDECIDED with this narrower source-specific rationale. The photoreceptor and Purkinje differentiation rows also remain UNDECIDED because the actual developmental evidence in the full PMID:16465590 and PMID:21074048 studies is unresolved. Positive degeneration and maintenance data are retained, not treated as proof that developmental participation is impossible.

Optional suggestions and completeness

The description now mentions reported MYLK1 substrate scope, supported by the recovered PMID:21074048 abstract, alongside KLF4. The KLF4 source's mouse experiments, human reprogramming component and in-vitro deglutamylation remain distinct. The core MF stays GO:0004181 with explicit terminal-glutamate chemistry and the two precise process terms. A further ontology search did not establish a validated EC 3.4.17.24-specific MF child; GO:0106423 remains the wrong tyrosine-removal reaction. No ontology-wide absence claim or guessed replacement is made, and a new term is not proposed merely to fill the optional slot. Broad accepted compartments and integrated tubulin/zinc binding do not require duplicate cores.

The recursive census covers current YAML, appended notes and every gene-directory artifact. No provider report was generated or edited. All 12 explicitly used PMIDs now have exact caches; all citation DOIs resolve to these same records. There are no Reactome citations. The unverified separate 2002 correction remains a bounded primary-notice access issue, not a fabricated PMID/cache. The review is eligible to leave draft only if targeted validation has no remaining warnings; substantive uncertainty in a decided UNDECIDED row is retained regardless of cache availability.

Follow-up validation and preservation

The final full targeted validator passed with exit code 0 and no curation warnings. The review status is now COMPLETE; the three source-specific UNDECIDED decisions remain. History validation and HTML rendering passed. Integrity checks preserve all 26 source objects and actions, all 20 reference identifier/title pairs, three alternative products, the integrated core, both machine files and the complete prior notes prefix. All 37 evidence quotations match the cited local text, including the explicitly attributed HPA receipt. The final recursive census contains 12 cached PMIDs, five normalized DOI strings mapping to those same cached records, no Reactome IDs and no provider/hypothesis artifact. The manifest includes four authored files and nine byte-identical source11 publication dependencies, with exact published-head base blobs. The inaccessible correction notice remains documented separately from cache completeness.

The coordinator independently read the complete changed annotation/reference/description/notes content and accepted its biological scope. Final wording distinguishes the unread full bodies of PMID:21074048 and PMID:16465590 from externally recovered papers. The description uses the protein name myosin light-chain kinase; MYLK1 is retained above only as the historical source terminology, not an assertion of a newly verified human construct. No action or core change resulted.