2026-09-28. Human ARL13B, UniProt Q3SXY8, HGNC:25419. All 176 original annotations and three alternative products are preserved. The review distinguishes GTP hydrolysis, guanine-nucleotide exchange on ARL3, and tubulin binding. It proposes the latter two molecular functions; it adds no biological-process annotation.
PMID:26551564(https://pubmed.ncbi.nlm.nih.gov/26551564/) establishes ARL13B as an ARL3 exchange factor. The initial yeast two-hybrid constructs are murine, and much of the structure and interface work uses Chlamydomonas proteins. The decisive human biochemical experiment instead uses human ARL13B residues 18–278 purified from insect cells and murine ARL3, with human ARL2 as a negative comparator. Human full-length ARL13B expression also increases active ARL3 in HEK293 and mouse IMCD3 cells. The paper states, “the human Arl13B also exhibits strong GEF activity for Arl3 (Figure 6A).” This supports GO:0005085 directly, without relabelling algal structures as human experiments. The proposed ciliary ARL3-GTP gradient is a mechanistic model, not a directly measured endogenous gradient in every tissue.
PMID:33438581(https://pubmed.ncbi.nlm.nih.gov/33438581/) adds a distinct reaction context: human ARL13B and ARL2BP/BART cooperate in ARL3 nucleotide exchange, and BART also stabilizes active ARL3 until displacement by other effectors. The GDP-to-GDP control distinguishes exchange stimulation from merely binding the GTP-loaded product. The source says, “BART was able to accelerate the nucleotide exchange approximately 70-fold”. This does not establish an autonomous BART GEF, a stable BART–ARL13B complex, or uniform exchange efficiency under all nucleotide conditions. Those distinctions also prevent treating the 2015 and 2021 results as contradictory binary classifications of GEF activity.
PMID:28487361(https://pubmed.ncbi.nlm.nih.gov/28487361/) examines purified full-length mouse and human ARL13B from mammalian expression cells. The abstract foregrounds murine protein; that does not mean every experiment is murine. CK2 phosphorylates the murine preparation but not the human protein, and CK2 can use GTP as a phosphate donor. The actual nucleotide-binding and GTP-hydrolysis controls therefore matter. Solution-based binding avoids confusing phosphate transfer with reversible nucleotide binding. The weak intrinsic hydrolysis signal may include contaminating nucleotidase activity, as the authors explicitly acknowledge. Bovine-brain extract stimulates ARL13B-dependent turnover with no-protein and GST controls, excess cold nucleotide, and reduced activity of K34A; the stimulating factor was not isolated. Retain GTPase activity with those limits rather than claiming a well-established high autonomous turnover rate.
The same 2017 study finds that T35N retains nucleotide binding and intrinsic/stimulated hydrolysis but loses ARL3 exchange-factor activity. G75Q is also not established as a constitutively active GTPase mutant. Thus the older labels “GDP-locked” or “GTPase-inactive” cannot be used as premises when interpreting later cellular rescues. The physiological hydrolysis-stimulating factor and the coupling of nucleotide state to individual effector functions remain unresolved.
PMID:29592971(https://pubmed.ncbi.nlm.nih.gov/29592971/) supplies direct evidence for GO:0015631 tubulin binding. The original Results distinguish recombinant human GST-ARL13B binding to purified bovine tubulin from the initial mouse TAP bait in human RPE1 cells. GST-only controls and domain mapping support a G-domain interaction. The source states, “ARL13B can directly interact with tubulin dimers.” Human fragments were also tested with Chlamydomonas axonemes, while cellular rescue used mouse fibroblasts. This supports binding and a role in membrane–axoneme organization; it does not establish a motor or polymerase activity. Retained tubulin binding of T35N alone cannot prove independence from GTP hydrolysis, because the earlier 2017 biochemical measurements show preserved hydrolysis. No conflicting experimental observation is discarded to preserve the mutant label.
The tubulin study separately measures ciliary length and the distributions of SMO and SSTR3 in Arl13b-null mouse cells. Wild-type human or mouse ARL13B restores ciliary length. G-domain deletion disrupts membrane-protein distributions, and T35N separates some distribution phenotypes from complete ciliary-length rescue. These findings corroborate existing ciliary assembly and receptor-localization annotations, with mouse experimental provenance explicit. The physical mechanism of membrane–axoneme anchoring remains a supported model rather than a solved molecular bridge.
PMID:26582389(https://pubmed.ncbi.nlm.nih.gov/26582389/) reports nucleotide-sensitive exocyst association in mouse IMCD3/NIH3T3 cells and nonciliated human HeLa cells. It says, “Sec8 and Sec5 are also enriched in the precipitates in a GTPγS-sensitive manner”. Human RPE1 and canine MDCK imaging are separate experiments. Purified murine GST-ARL13B associates with translated Sec8 and Sec5; Exo70 does not show the same direct binding in that assay. The sequence species of the translated partners was not resolved in the inspected Methods. This supports the existing GTP-dependent interaction function without asserting that ARL13B is an exocyst subunit. Conditional mouse renal phenotypes remain tissue context.
PMID:20643351(https://pubmed.ncbi.nlm.nih.gov/20643351/) uses mouse and human tagged PIFO to recover endogenous human ARL13B from HEK293T cells. Both original PIFO interaction assertions are supported as coassociation; a purified binary interface was not shown in the inspected experiment. They remain non-core rather than being rejected merely because the paper foregrounds PIFO. The explicit user ActionEnum defines REMOVE by likely biological error; genericity alone does not establish such error. This precedence decision is recorded here rather than placed into every biological reason.
PMID:24421332(https://pubmed.ncbi.nlm.nih.gov/24421332/) explicitly uses ARL13B among four ciliary markers in serum-starved human RPE1 cells. It supports the location assertion even though Talpid3 is the main experimental target. The Girdin study PMID:27623382(https://pubmed.ncbi.nlm.nih.gov/27623382/) was read only at abstract level; its positive cilium annotation is retained with independent corroboration, without claiming its microscopy was re-inspected. The preserved UniProt record supplies direct, checkable ciliary-membrane and plasma-membrane statements. Its note excluding the transition zone in particular localization studies prompts examination of the original HPA image, rather than disproving that separate experiment without seeing it. The exact HPA transition-zone and intercellular-bridge images remain unread.
PMID:23223633(https://pubmed.ncbi.nlm.nih.gov/23223633/) provides a nonciliary trafficking context in human HeLa cells: silencing/rescue, CD1a recycling, tagged-protein plasma-membrane/tubular localization and actin association. FLAG-versus-EGFP localization controls were inspected in the original indexed Results. Complete construct-species Methods and supplementary material were not available in that read. The study's negative tubulin co-immunoprecipitation in its conditions does not refute the later purified-tubulin experiment. Direct actin binding is not asserted. The normal cache is abstract-only; the detailed external source reading does not change that flag.
PMID:23150559(https://pubmed.ncbi.nlm.nih.gov/23150559/) examines INPP5E targeting, ARL13B interaction, and disease variants in human RPE1/HEK293T and mouse IMCD3 contexts. Targeted original Results include co-immunoprecipitation and purified-GST/translated-partner assays; complete construct Methods and supplements remain unread. Its negative ARL2/3-depletion result is conditional on that experiment and does not overturn the later reconstituted ARL3 activation mechanism. Its use of the T35N GDP-locked label has the same biochemical caveat. The normal cache is abstract-only.
Two Reactome entries place ARL13B in an INPP5E ciliary-targeting model. Cilium localization is independently supported even though each modeled release step has not been audited. Seven other entries concern cytosolic handling of misfolded ciliary proteins. The followup graph inspection below resolves their exact ARL13B membership. These are retained as non-core locations in a modeled disposal context. Being a substrate in disposal does not make ARL13B a ubiquitination enzyme, motor or autophagy effector.
The 128 remaining interaction-screen rows retain their exact original partner identifiers, isoforms and sources. PMID:32296183(https://pubmed.ncbi.nlm.nih.gov/32296183/), PMID:32541000(https://pubmed.ncbi.nlm.nih.gov/32541000/) and PMID:40355756(https://pubmed.ncbi.nlm.nih.gov/40355756/) establish article-level context, but the exact pair experiments and controls have not been adjudicated. A missing target-name hit in extracted article prose is not evidence that the pair was never tested. No guessed supplementary-table location or wrong-gene accusation is made.
Neural patterning, migration, radial-glial and Smoothened-signaling assertions remain non-core developmental contexts. The review preserves curator transfers and does not infer the PAINT ancestral placement from donor counts or from the target occurring in its own evidence list. The motile-cilium ancestral activity-site assertion and the separate electronic mouse-derived motile-cilium location remain unresolved for different provenance reasons. No matching ARL13B/Q3SXY8 entry was found in the GO-CAM index.
PMID:18674751(https://pubmed.ncbi.nlm.nih.gov/18674751/) links ARL13B variants to two human Joubert-syndrome families and reports human wild-type rescue in zebrafish, alongside murine neuronal localization. The complete abstract and targeted figure captions were read; full clinical Methods were not. PMID:18554500(https://pubmed.ncbi.nlm.nih.gov/18554500/) supports nucleotide binding and ciliary localization at abstract level. Its proposed self-association mechanism is not elevated into a new homodimerization function without inspecting the original assays and the UniProt caution.
The original seven cached PMID abstracts and nine complete Reactome event summaries were read, with targeted primary Results/Methods for PIFO, Talpid3 and exocyst as specified above. The official eLife 2015 PDF supplied main Results, Discussion, Methods and captions before normal source recovery; image pixels and raw structural data were not inspected. For the 2017 and 2018 papers, the targeted biochemical, binding and rescue Results/Methods and relevant Discussion were inspected, including the normal recovered text. This is not a claim that every supplement or every experiment in all cited papers was reviewed.
Four cited records remain abstract-only (18554500, 18674751, 23150559, 23223633); 26551564 has XML-derived full text and 28487361/29592971 have HTML-derived text. The existing record for 33438581 is preserved. A full-text availability flag records the cache, not how much a reviewer read. Source-retrieval provenance is recorded in the curation history.
The final decisions comprise 25 ACCEPT, 19 KEEP_AS_NON_CORE and 132 UNDECIDED original rows, plus two NEW molecular functions. Three core functions distinguish GTPase activity, ARL3 exchange and tubulin binding. Specific activity locations are not separately asserted in the cores where protein localization alone does not establish the site of that particular reaction. Original qualifiers, alternative products, source files and publication bytes remain unchanged. Independent consultation and validation are recorded in the accompanying curation history.
2026-09-28. The seven event pages and the nested entity/set/complex paths were inspected. Misfolded ARL13B is explicitly human UniProt Q3SXY8, chain 1–428, cytosol. It belongs to misfolded ciliary proteins, within set 9641091 and complex 9641099. This verifies the target in PRKN-complex recruitment and ubiquitin transfer.
Polyubiquitinated ARL13B, also Q3SXY8 in cytosol, belongs to ciliary subset 9641108 and polyubiquitinated-protein set 9641120. The inspected paths place it in polymerization, release from PRKN and dynein recruitment. For vimentin association, the containing complexes are at the microtubule organizing center but their nested HDAC6/cargo components retain cytosol labels. Aggresome disassembly releases a cytosolic vimentin/HDAC6/cargo complex. These explicit modeled labels justify retaining the location assertions with their disposal-state limitation; they are not seven new target-specific localization experiments. No process annotation is proposed from cargo membership.
The GTPase core now cites the stimulated-turnover experiment alone. Nucleotide-sensitive exocyst association remains supported in its existing annotation and discussion, without being presented as a GTP-hydrolysis assay. The distinct ARL3 exchange and tubulin-binding cores are preserved.
The review of PR #3374 at 93cbb42a9 correctly identifies the repository default: an uninformative generic binding annotation can be excluded even when its reported interaction is real. That informational-exclusion criterion is distinct from concluding that the experimental evidence is false. The earlier notes should not be read as saying that the default policy requires biological falsity or that these touched annotations qualify for a legacy exception.
This task explicitly directs that generic binding not be removed solely for informativeness, while preserving uncertainty when the relevant experiment cannot be adjudicated. The retained decisions are a scoped application of those task instructions. They are not a claim of compliance with the default generic-binding policy, a global policy revision, or a reason to silence its advisory warnings.
The two retained PIFO rows preserve distinct human and mouse partner identifiers and the common human-cell pull-down context. These are co-complex associations, not purified binary binding or a distinct ARL13B catalytic mechanism. The 128 unadjudicated interaction-screen rows remain UNDECIDED; they are not promoted to retention because the PIFO experiment was accessible. The seven graph-grounded cytosol annotations retain their disposal-state limitations. No GTPase, ARL3-exchange or tubulin-binding core claim changes.
This addendum changes only the explanation of the decision. It adds no primary-source reading, assay verification, quotation, annotation or core function. All current annotation decisions, products, evidence limitations and source objects remain unchanged.