ACTG2 encodes smooth muscle gamma-enteric actin (gamma-2 actin), one of the six conventional vertebrate actin isoforms and the actin of the contractile apparatus of visceral smooth muscle. Like all conventional actins it is a 376-residue ATP-binding protein of the actin fold that exists in equilibrium between a monomeric (G-actin) and a filamentous (F-actin) state: monomers add to the barbed end of a two-stranded helical polymer, hydrolyse their bound ATP shortly after incorporation, and release inorganic phosphate to leave an ADP-bound protomer, which is the state resolved in every deposited ACTG2 filament structure. Purified human ACTG2 polymerises with a critical concentration slightly higher and a polymerisation rate slightly lower than skeletal alpha-actin, is nucleated by the smooth muscle nucleator leiomodin-1, is decorated by the smooth muscle tropomyosin isoform Tpm1.4, and is propelled by smooth muscle myosin at a velocity indistinguishable from alpha-actin. In tissue, ACTG2 is the thin-filament actin of the enteric and urogenital smooth muscle contractile apparatus, where actin filaments anchored at dense bodies and dense plaques slide against smooth muscle myosin thick filaments to generate contraction. Expression is strongly restricted: abundant in the smooth muscle cells of intestinal muscularis mucosa and muscularis propria, in intestinal vascular smooth muscle, and in bladder, uterus and seminal vesicle. The protein carries the actin-family post-translational modifications, including N-terminal processing to a mature acidic N-terminus and SETD3-dependent methylation of His-74, which is resolved in its own cryo-EM structures. Heterozygous missense variants in ACTG2 are the commonest cause of visceral myopathy and of megacystis-microcolon-intestinal hypoperistalsis syndrome, acting dominant-negatively rather than by loss of function. Recurrent substitutions at a small set of conserved arginines (R40, R63, R148, R178, R211, R257) account for most cases and act by distinct biophysical routes - destabilising the monomer, blocking polymerisation, weakening intersubunit contacts within the filament, or impairing binding of actin-binding proteins - all converging on reduced visceral smooth muscle contractility.
Definition: An actin filament of the smooth muscle contractile apparatus, composed of actin and associated proteins including tropomyosin and caldesmon, anchored at dense bodies and membrane-associated dense plaques, and interdigitating with smooth muscle myosin thick filaments during contraction. Unlike the striated counterpart it lacks troponin and is not organised into sarcomeres.
Justification: GO models the smooth muscle contractile apparatus at fibre level (GO:0030485 smooth muscle contractile fiber) and at anchor level (GO:0030486 smooth muscle dense body) but has no term at filament level, while the striated equivalent exists (GO:0005865 striated muscle thin filament). The filament is exactly the level at which the smooth muscle actins ACTG2 and ACTA2, the smooth muscle tropomyosins and caldesmon act, and the only compartment terms currently available to them are either too general (GO:0015629 actin cytoskeleton) or categorically wrong, as GO:0032982 myosin filament is on the ISS row reviewed here. Placement was checked rather than guessed: GO:0036379 myofilament is not available as a parent because its definition restricts it to the smallest contractile units of a myofibril (striated muscle fiber), and GO:0005865 is itself not currently a descendant of GO:0005884 actin filament, so the natural placement is part_of GO:0030485 by analogy with GO:0030486, with the exact axiomatisation left to GO editors.
Parent term: smooth muscle contractile fiber
Supporting Evidence:
| GO Term | Evidence | Action | Reason |
|---|---|---|---|
| GO:0005576 extracellular region | IBA GO_REF:0000033 | REMOVE | Summary: Phylogenetic inference placing ACTG2 in the extracellular region, from PANTHER node PTN004322804. The whole chain behind this row was traced and none of it survives. The node has a single IBD seed. The cached PAINT table records it as RGD:621676, which resolves to rat Acta2 (P62738, Swiss-Prot, 377 aa) - the other smooth muscle actin, not ACTG2's orthologue. Three things about that seed, stated separately because the first version of this review ran them together and a cross-check against the parallel ACTA2 review caught it. First, the seed gene is also a recipient. Rat Acta2's own GO:0005576 annotation is GO:0005576 IBA GO_REF:0000033 with PANTHER:PTN004322804 and RGD:621676 - the identical row human ACTG2 carries, citing this same node and rat Acta2's own identifier. A self-referential IBA is a normal PAINT construct and is not by itself a defect, but it does mean the seed's exact-term annotation is this node's own output and cannot be counted as independent support for the term. Second, no annotation anywhere in the chain carries experimental evidence for GO:0005576 at the node. Rat Acta2's only non-IBA annotation anywhere under GO:0005576 is GO:0005604 basement membrane, IDA, from PMID:30476341 - a fetal-rat-testis retinoic-acid and mono-(2-ethylhexyl)-phthalate toxicology study in which alpha-smooth-muscle actin marks the peritubular cells bounding the seminiferous cords. The PAINT table records which gene seeded an IBD, not which of that gene's annotations justified it, so the identification of that IDA as the underlying experimental basis is an inference - a well-constrained one, since it is the only candidate, but an inference, and it is labelled as one here rather than asserted. Third, and separately, that IDA does reach human by a different route which has nothing to do with this node: human ACTA2 carries GO:0005604 basement membrane IEA from GO_REF:0000107 with UniProtKB:P62738 - Ensembl Compara, a different term, a different evidence code and a different reference from the IBA. ACTG2 carries no such row at all. The two chains are independent and are not merged here. Which node matters. Scanning all 1420 human GO:0005576 IBA annotations shows PTN004322804 reaches exactly two genes, ACTA2 and ACTG2. It is the only node in PTHR11937 that specifically covers the smooth muscle actin pair, and the only term it gives them puts them outside the cell - while GO:0005884 actin filament, which is true of both, is asserted on single-gene nodes that reach neither (see the NEW GO:0005884 row below). The reference-projection check on PMID:30476341 comes back negative and is reported because it was run: 13 annotations over 7 distinct entities, with only Acta2 receiving the localisation IDA and the other six receiving retinoic-acid-response IEP terms. So this is genuine per-protein curation at RGD, not a complex-to-subunit spread, and nothing here criticises the rat annotation. The objection is entirely to the propagation: a paralog, in a different organism and tissue, with the term broadened en route, and with a qualifier - is_active_in - that asserts ACTG2 carries out its molecular function in the extracellular region. A cytoplasmic filament protein does not. Reason: Single-seed propagation from the paralogous rat Acta2, and no non-IBA evidence for the term exists anywhere in the chain: the seed gene's own GO:0005576 row is this node's own output, and the only experimental anchor under that term anywhere is a basement-membrane immunolocalisation of the paralog in fetal rat testis - a descendant term, in another organism, in a tissue ACTG2 does not occupy. The receiving node covers only the two smooth muscle actins, so the defect is confined to and characteristic of this pair. No actin performs its function in the extracellular region, and ACTG2's only other extracellular support is bulk proteomics (rows 17-19), so nothing independent survives to keep the term. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG COMPARTMENT OR COMPLEX MISMATCH CONTEXT OR TISSUE MISMATCH GRANULARITY MISMATCH CIRCULAR PROPAGATION Sources checked: PANTHER:PTN004322804 · PANTHER ancestral node whose human descendants are exactly the two smooth muscle actins, ACTA2 and ACTG2 SOURCE BAD Not a protein, so it carries no evidence of its own; recorded because its clade composition is what makes this a smooth-muscle-actin-specific defect rather than a family-wide one. RGD:621676 · Acta2 (Rattus norvegicus, P62738, Swiss-Prot, 377 aa) SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. The IBA is GO:0005576 from this same node, citing PTN004322804 and this gene's own RGD id, so the seed gene is also a recipient and its exact-term annotation is the node's own output - which is why CIRCULAR_PROPAGATION is listed. The IDA is GO:0005604 basement membrane from PMID:30476341, three levels below the term propagated, on a paralog, in rat fetal testis; it is the only non-IBA annotation the seed holds under GO:0005576 and therefore the only candidate experimental basis for the IBD, though the PAINT table does not record which annotation was used. That IDA reaches human ACTA2 by a separate Ensembl Compara route (GO:0005604 IEA, GO_REF:0000107) which ACTG2 does not carry at all. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md reaches 2 human gene(s): ACTA2, ACTG2 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - asserted at **PTN004322804** (IBD, taxon `taxon:`, 20260528), seeds: RGD:621676 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:30476341 | 13 | 7 | GO:0005604: 1, GO:0032526: 6, GO:0080021: 6 | PMID:30476341 Humans are universally exposed to low levels of phthalate esters (phthalates), which are used to plasticize polyvinyl chloride. file:human/ACTG2/ACTG2-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm, cytoskeleton. |
| GO:0015629 actin cytoskeleton | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic inference from PANTHER node PTN002631484, the deep eukaryote-level actin node. Verified independently rather than read off a sibling review: QuickGO shows this node donates GO:0015629 to eighteen human genes - ACTA1, ACTA2, ACTB, ACTBL2, ACTC1, ACTG1, ACTG2, ACTL8, ACTL9, ACTL10, ACTRT1-3 and the five POTE genes - which is the same set the merged ACTL10 review reports, confirmed here rather than inherited. The donor set is strong. Of the 25 WITH/FROM tokens, 24 are protein identifiers, all 24 resolve, and all 24 carry their own experimental-code annotation for this term or a descendant, across twelve organisms; yeast ACT1 alone carries nine IDAs. Two resolve only to unreviewed TrEMBL entries (Candida ACT1 and C. elegans act-5) - the same two the merged ACTL10 review reports for this node - and both still carry their own IDA, so the weakness is one of name provenance rather than evidence provenance. Unlike the divergent actin-like proteins that inherit this term at 33-50 per cent identity to beta-actin, ACTG2 is a conventional actin: 98.7 per cent identical to the alpha-actin of the F-actin structure used to define the protomer interface, and identical at all 38 of its interface positions. It is active in the actin cytoskeleton in the strongest available sense - its own filament has been reconstructed at 2.45 A - so is_active_in is exactly right here, where on ACTL10 it presupposes an unmeasured activity. The term is broad, but the node is deep and its human descendants span 33 to 100 per cent identity to beta-actin, so GO:0015629 is the genuine last common ancestor of the donor set and no granularity defect exists at the node. The specific structure ACTG2 occupies - the smooth muscle contractile apparatus - is captured additively on the MODIFY of row 10 and the NEW GO:0005884 row rather than by narrowing this one. Reason: Correct, and core rather than peripheral: ACTG2 is a conventional filament-forming actin whose own cryo-EM structures place it in a two-stranded actin polymer. The breadth of the term reflects the depth of the node, which is the ontology working correctly, not a curator failing to be specific. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002631484 · PANTHER eukaryote-level actin node spanning the conventional actins, the POTE actin-fusion genes and the divergent actin-like proteins SUPPORTS TRANSFER Not a protein; recorded because its clade breadth is what justifies the breadth of the donated term and rules out GRANULARITY_MISMATCH. UniProtKB:P60709 · ACTB (Homo sapiens, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx2, IDAx3, IMPx1. UniProtKB:P63261 · ACTG1 (Homo sapiens, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx2, IDAx1. UniProtKB:P68133 · ACTA1 (Homo sapiens, Swiss-Prot, 377 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx4, IDAx3, IMPx1, ISSx1. UniProtKB:P68032 · ACTC1 (Homo sapiens, Swiss-Prot, 377 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx2, IDAx2, ISSx1. UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) SUPPORTS TRANSFER Own evidence: IBAx1, IDAx1. Legitimate as one donor among 24 at a deep family node, unlike rows 11-16 where the same entry is used alone for a one-to-one ortholog transfer although chicken has its own ACTG2 (P63270). UniProtKB:Q6QAQ1 · ACTB (Sus scrofa, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx2, IEAx2, IPIx1, ISSx1. UniProtKB:Q8I4X0 · ACT1 (Plasmodium falciparum isolate 3D7, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1, IEAx1, ISSx2. MGI:MGI:87906 · Actg1 (Mus musculus, P63260, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence: IBAx2, IDAx3, IEAx1, ISOx5. The cross-reference resolves to five candidate entries; the reviewed one is used and all are recorded in RESULTS.md. MGI:MGI:87909 · Acta2 (Mus musculus, P62737, Swiss-Prot, 377 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1, IEAx2, ISOx2, ISSx1. RGD:621676 · Acta2 (Rattus norvegicus, P62738, Swiss-Prot, 377 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx2, ISOx1, ISSx1. RGD:628837 · Actb (Rattus norvegicus, P60711, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx2, IDAx2, ISOx3. RGD:1304556 · Actg1 (Rattus norvegicus, P63259, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx2, IDAx4, ISOx3. SGD:S000001855 · ACT1 (Saccharomyces cerevisiae S288c, P60010, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx9. PomBase:SPBC32H8.12c · act1 (Schizosaccharomyces pombe 972, P10989, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx4, TASx1. CGD:CAL0000191211 · ACT1 (Candida albicans SC5314, A0A1D8PFR4, TrEMBL (unreviewed), 376 aa) SUPPORTS TRANSFER Own evidence: IBAx1, IDAx1. Unreviewed entry, so its GO evidence is genuine but its protein NAME is an automatic label; counted under evidence provenance, not name provenance. FB:FBgn0011743 · Arp53D (Drosophila melanogaster, P45891, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. WB:WBGene00000064 · act-2 (Caenorhabditis elegans, P10984, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence: IBAx1, IDAx1. Resolved only by a free-text lookup on the bare identifier; WormBase gene ids are not in UniProt's xref:wormbase index, which holds transcript and protein ids. WB:WBGene00000065 · act-3 (Caenorhabditis elegans, P0DM42, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. WB:WBGene00000066 · act-4 (Caenorhabditis elegans, P10986, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. WB:WBGene00000067 · act-5 (Caenorhabditis elegans, O45815, TrEMBL (unreviewed), 375 aa) SUPPORTS TRANSFER Own evidence: IBAx1, IDAx2. Unreviewed entry, reported rather than hidden; matches the merged ACTL10 review's finding for this same node. dictyBase:DDB_G0269234 · act1 (Dictyostelium discoideum, P07830, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx3, IEAx1. dictyBase:DDB_G0275023 · act22 (Dictyostelium discoideum, Q553U6, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1, IEAx1, ISSx1. dictyBase:DDB_G0289487 · act3 (Dictyostelium discoideum, P07829, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1, IEAx1, ISSx1. dictyBase:DDB_G0289811 · act10 (Dictyostelium discoideum, Q54GX7, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx3, IEAx1. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md reaches 18 human gene(s): ACTA1, ACTA2, ACTB, ACTBL2, ACTC1, ACTG1, ACTG2, ACTL10, ACTL8, ACTL9, ACTRT1, ACTRT2, ACTRT3, POTEE, POTEF, POTEI, POTEJ, POTEKP file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md Distinct source proteins across all rows: **31** file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md ACTG2 scores **38/38** chemically compatible (38 identical, 0 conservative, 0 non-conservative, 0 gaps) and ACTA2 **38/38**. Both sit with the conventional actins file:human/ACTG2/ACTG2-uniprot.txt CC -!- SUBUNIT: Polymerization of globular actin (G-actin) leads to a |
| GO:0005200 structural constituent of cytoskeleton | IBA GO_REF:0000033 | ACCEPT | Summary: Phylogenetic inference from PANTHER node PTN000940351. This is the term the PAINT curators have been actively pruning across this family, so whether ACTG2 is inside or outside the retained set was checked against primary data rather than assumed. From the cached PAINT table, GO:0005200 is asserted once, at PTN000940351 (IBD, 10 seeds), and negated by IRD at exactly eight descendant nodes - PTN000233596, PTN000233752, PTN000233796, PTN000233887, PTN000234048, PTN001732543, PTN007551901 and PTN008986528. ACTG2 sits under none of them. QuickGO confirms the surviving reach is ten human genes: ACTA1, ACTA2, ACTC1, ACTG2, ACTL9, ACTL10, ACTR10, ACTRT1, ACTRT2 and ACTRT3. So the four conventional actins in that set are ACTA1, ACTA2, ACTC1 and ACTG2 - ACTB and ACTG1 receive the term from a different node, PTN002631586 - and ACTG2 is in the retained set. The donor set is heterogeneous by design: yeast ACT1, human ACTB, mouse and rat Actg1 and two Dictyostelium actins, but also human Arp2 and Arp3 (branched-filament nucleators) and yeast ARP1 and ARP10 (subunits of dynactin's actin-like minifilament). At that depth "structural constituent of cytoskeleton" is the genuine last common ancestor of what the donors do, so GRANULARITY_MISMATCH does not apply and no more specific term could be chosen without arbitrarily privileging one donor lineage. All ten protein donors resolve and all ten carry their own experimental-code annotation for the term. Three merged sibling reviews resolved this identical row three different ways - ACTR10 ACCEPT, ACTL10 MARK_AS_OVER_ANNOTATED, ACTRT3 MODIFY to GO:0005198 - and it is worth stating why that is not the AADACL-style inconsistency it resembles. There, three genes with identical donor sets and comparable evidence got three answers, at most one of which could be right. Here the recipients differ in precisely the property the term asserts. On the 38-position filament protomer interface, recomputed in this review's bioinformatics folder, ACTR10 scores 20 identical plus 8 conservative, ACTRT1 13 plus 8, ACTL10 3 plus 2, and ACTG2 38 plus 0. A term about contributing to the structural integrity of a cytoskeletal structure is a measured fact at 38/38 and an open question at 5/38. ACTG2 therefore accepts it on its own evidence without disturbing the divergent-gene calls. Reason: Verified to be in the retained rather than the IRD-negated set, from the PAINT table directly. The term is the true LCA of a heterogeneous donor set that mixes conventional actins with Arp2/3 and dynactin ARPs, so its breadth is correct rather than lazy, and for ACTG2 specifically it is backed by the gene's own cryo-EM filament structures and by complete conservation of the protomer interface. Core molecular function. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN000940351 · PANTHER pan-actin node; none of its eight IRD-negated descendant nodes contains ACTG2 SUPPORTS TRANSFER Not a protein. Recorded because the composition of its donor set - conventional actins plus Arp2/Arp3 plus dynactin ARP1/ARP10 - is what makes the general term the correct LCA rather than a granularity failure. SGD:S000001855 · ACT1 (Saccharomyces cerevisiae S288c, P60010, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. SGD:S000001171 · ARP1 (Saccharomyces cerevisiae S288c, P38696, Swiss-Prot, 384 aa) - dynactin centractin SUPPORTS TRANSFER Own evidence: IDAx1. A non-conventional actin-related protein; its presence in the donor set is part of why the donated term is general rather than "actin filament". SGD:S000002513 · ARP10 (Saccharomyces cerevisiae S288c, Q04549, Swiss-Prot, 284 aa) SUPPORTS TRANSFER Own evidence: IPIx3. Actin-like protein of the dynactin pointed end; another divergent member of the donor set. UniProtKB:P60709 · ACTB (Homo sapiens, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: EXPx1, IBAx1, IDAx3, IMPx1, TASx1. UniProtKB:P61158 · ACTR3 (Homo sapiens, Swiss-Prot, 418 aa) - Arp3 SUPPORTS TRANSFER Own evidence: IDAx1 (PMID:11741539). A branched-filament nucleator, not a conventional actin. UniProtKB:P61160 · ACTR2 (Homo sapiens, Swiss-Prot, 394 aa) - Arp2 SUPPORTS TRANSFER Own evidence: IDAx1 (PMID:11741539). MGI:MGI:87906 · Actg1 (Mus musculus, P63260, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1, ISOx2. RGD:1304556 · Actg1 (Rattus norvegicus, P63259, Swiss-Prot, 375 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx2, ISOx1. dictyBase:DDB_G0269234 · act1 (Dictyostelium discoideum, P07830, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. dictyBase:DDB_G0289811 · act10 (Dictyostelium discoideum, Q54GX7, Swiss-Prot, 376 aa) SUPPORTS TRANSFER Own evidence in QuickGO for this term or a descendant: IBAx1, IDAx1. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md reaches 10 human gene(s): ACTA1, ACTA2, ACTC1, ACTG2, ACTL10, ACTL9, ACTR10, ACTRT1, ACTRT2, ACTRT3 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md ACTG2 scores **38/38** chemically compatible (38 identical, 0 conservative, 0 non-conservative, 0 gaps) and ACTA2 **38/38**. Both sit with the conventional actins file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md all 14 published columns reproduced exactly PMID:38820162 structures of WT ACTG2, R40C, and R257C filaments using helical reconstruction |
| GO:0005856 cytoskeleton | IEA GO_REF:0000044 | MARK AS OVER ANNOTATED | Summary: Automatic mapping of UniProt's own subcellular-location vocabulary term SL-0090 (Cytoskeleton) onto GO, derived from the curated SUBCELLULAR LOCATION line "Cytoplasm, cytoskeleton". The statement is correct and rests on UniProt curation rather than on a cross-species transfer, and it is resolved the same way as the merged ACTB review resolves its identical row. It is subsumed by GO:0015629 actin cytoskeleton on row 2 and by the proposed GO:0005884 actin filament, so it adds no information beyond them. Reason: Correct and curator-derived, if less specific than the terms ACTG2 already has and than those proposed here. Consistent with the merged ACTB review's handling of the same SubCell-derived row. [2026-08 SL project re-review] Uninformative parent from the UniProt subcellular-location pipeline. The sole source for this annotation is GO_REF:0000044 (UniProtKB-SubCell), and the gene already carries strictly more specific term(s) from independent evidence, namely GO:0015629 actin cytoskeleton (IBA). The term is not false; it is an under-specified location that adds no information over what is already annotated, which is the dominant failure mode of the SL pipeline (see projects/SL.md). Marked over-annotated rather than MODIFY because the more precise term is already present, so there is nothing to replace it with. Supporting Evidence: file:human/ACTG2/ACTG2-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm, cytoskeleton. |
| GO:0007010 cytoskeleton organization | IEA GO_REF:0000108 | KEEP AS NON CORE | Summary: Inter-ontology logical inference: the WITH/FROM field is the GO term GO:0005200 itself, so this row is a mechanical restatement of row 3 in the biological-process aspect rather than an independent observation. It is not wrong - a structural actin does participate in building and remodelling cytoskeletal structures, and ACTG2's polymerisation and depolymerisation kinetics have been measured directly - but it carries no evidence of its own and sits far above the process ACTG2 is actually known for. The informative process is proposed additively on the NEW GO:0006939 row rather than by editing this one, since editing a mechanically-derived row would only be undone by the next pipeline run. Reason: Correct but wholly derivative: its sole WITH/FROM entry is another GO term, so it restates the molecular function row rather than adding evidence. Kept as a true generic process annotation, marked non-core because the process ACTG2 is characterised for is visceral smooth muscle contraction. Supporting Evidence: PMID:38820162 WT ACTG2 exhibited slower polymerization compared to |
| GO:0045471 response to ethanol | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: Ensembl Compara transfer from rat Actg2 (P63269), the genuine one-to-one orthologue, so the orthology step is sound. The problem is what is being transferred. Rat Actg2's own annotation is IEP - inferred from expression pattern, its only evidence for this term - from PMID:28320086, a study of ethanol-induced liver fibrosis in rats in which the actin readout is alpha-smooth muscle actin as a marker of hepatic stellate cell activation. The reference annotates three entities in total (Tgfb1, Il6, Actg2), all IEP, all to this same term. Two independent reasons not to carry it to human ACTG2. First, a change in the abundance of a structural protein under a chemical challenge does not make that protein involved in the response; IEP records co-behaviour, not participation, and this is the classic over-annotation shape for response-to-chemical terms. Second, ACTG2 is an enteric and urogenital smooth muscle actin and is not a hepatic stellate cell protein, so the cell type in which the observation was made is not one the human gene product occupies. A separate concern is recorded in the notes and raised as a question rather than asserted here: the paper measures alpha-SMA, which is the ACTA2 gene product, and rat Acta2 received no annotation from it. Whether that is a symbol conflation at the rat end cannot be settled from the abstract, and the action taken here does not depend on it - it is a judgement about the propagation, not about the RGD curator's reading. Reason: An expression-level observation in rat hepatic stellate cells, transferred automatically to a human enteric smooth muscle actin. The orthology is right and the term is not fabricated, but neither the evidence type nor the cell type supports asserting that human ACTG2 is involved in the response to ethanol. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: SOURCE EVIDENCE WEAK CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P63269 · Actg2 (Rattus norvegicus, Swiss-Prot, 376 aa) - the true one-to-one orthologue SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IEPx1 only, from PMID:28320086, one of three entities annotated by that reference and all by IEP. ensembl:ENSRNOP00000050322 · Ensembl rat protein identifier accompanying the UniProt donor SUPPORTS SOURCE BUT NOT TARGET Identifier for the same rat gene product; carries no separate evidence. Supporting Evidence: PMID:28320086 as denoted by reducing a-smooth muscle actin (a-SMA) expression in the liver file:human/ACTG2/ACTG2-uniprot.txt CC -!- TISSUE SPECIFICITY: In the intestine, abundantly expressed in smooth |
| GO:0071354 cellular response to interleukin-6 | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: Same propagation route and same weakness as row 6, from a different paper. Rat Actg2's only evidence for this term is IEP from PMID:21294755, in which cultured rat hepatic stellate cells were treated with leptin and acetaldehyde and alpha-SMA protein was measured as an activation marker, with IL-6 neutralisation used to show the effect is IL-6 dependent. The reference-projection check gives 8 annotations over 6 distinct entities - Col1a1, Tgfb1, Il6, Timp1, Actg2 and Mapk14 - a profibrogenic marker panel measured in one experiment, with the terms spread across the whole set. Human ACTG2 is not a hepatic stellate cell protein, and an IL-6-dependent change in the level of a marker protein is not evidence that the protein participates in the cellular response to IL-6. Reason: Automatic transfer of a rat expression-pattern observation made on a marker panel in hepatic stellate cells, a cell type in which ACTG2 is not the relevant smooth muscle actin. The term describes the experiment's stimulus, not a function of the protein. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: SOURCE EVIDENCE WEAK CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P63269 · Actg2 (Rattus norvegicus, Swiss-Prot, 376 aa) SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IEPx1 only, from PMID:21294755. ensembl:ENSRNOP00000050322 · Ensembl rat protein identifier accompanying the UniProt donor SUPPORTS SOURCE BUT NOT TARGET Identifier for the same rat gene product; carries no separate evidence. Supporting Evidence: PMID:21294755 an activation marker of HSCs PMID:21294755 Rat HSCs in second or third passage were utilised. file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:21294755 | 8 | 6 | GO:0044320: 1, GO:0071354: 1, GO:1905641: 6 | |
| GO:1905641 cellular response to acetaldehyde | IEA GO_REF:0000107 | MARK AS OVER ANNOTATED | Summary: The third row of the same block, from the same paper as row 7 (PMID:21294755). This one shows the panel effect most clearly: of the six entities that reference annotates, all six carry GO:1905641 by IEP. A term that every member of a six-gene marker panel receives with identical evidence records the design of the experiment rather than a property of any one gene product, and transferring it across species by Ensembl Compara adds nothing. Reason: One rat hepatic-stellate-cell experiment gave the same IEP term to all six genes it measured; the human transfer inherits that without any ACTG2-specific support, in a cell type where ACTG2 is not expressed. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: SOURCE EVIDENCE WEAK CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P63269 · Actg2 (Rattus norvegicus, Swiss-Prot, 376 aa) SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IEPx1 only, from PMID:21294755, one of six entities receiving this same term from that reference. ensembl:ENSRNOP00000050322 · Ensembl rat protein identifier accompanying the UniProt donor SUPPORTS SOURCE BUT NOT TARGET Identifier for the same rat gene product; carries no separate evidence. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:21294755 | 8 | 6 | GO:0044320: 1, GO:0071354: 1, GO:1905641: 6 | PMID:21294755 an activation marker of HSCs |
| GO:0016887 ATP hydrolysis activity | ISS GO_REF:0000024 | ACCEPT | Summary: Curated sequence-similarity transfer from pig ACTA1 (P68137), which carries its own EXP annotation for this term from PMID:24743229, a study of Plasmodium actin structures in which the vertebrate skeletal actin serves as the characterised reference. UniProt makes the same call in its own record, giving ACTG2 EC 3.6.4.- and the full Rhea reaction with ECO:0000250 to P68137, so the GOA row and the UniProt entry agree by construction. The donor is a paralog rather than an orthologue, which on rows 11-16 is a fatal objection but is not one here, because ATP hydrolysis on polymerisation is a property of the actin fold shared by every conventional actin rather than an isoform-specific behaviour, and because the same reference independently carries IDA evidence for the activity on three Plasmodium actins. ACTG2's own data now corroborate it directly, which is worth recording even though it does not change the action: the three ACTG2 filament structures were prepared with ATP and every protomer of every one contains Mg-ADP, because hydrolysis and phosphate release happened in the tube before vitrification. That is ACTG2 hydrolysing its own bound nucleotide, and a curator reading that paper could reasonably upgrade this row from ISS. Reason: A well-founded fold-level activity, held by every conventional actin, whose donor carries real experimental evidence and whose conclusion is independently corroborated by ACTG2's own structures. Core molecular function. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:P68137 · ACTA1 (Sus scrofa, Swiss-Prot, 377 aa) - skeletal alpha-actin SUPPORTS TRANSFER Own evidence in QuickGO for this term: EXPx1 from PMID:24743229. A paralog rather than an orthologue, acceptable here because the activity is a property of the actin fold; the same reference also carries IDA for this term on three Plasmodium actins. Supporting Evidence: file:human/ACTG2/ACTG2-uniprot.txt DE EC=3.6.4.- {ECO:0000250|UniProtKB:P68137}; file:human/ACTG2/ACTG2-uniprot.txt Reaction=ATP + H2O = ADP + phosphate + H(+); Xref=Rhea:RHEA:13065, PMID:38820162 release occurred during sample preparation, before vitrification, resulting in all three structures containing Mg PMID:24743229 Structural differences explain diverse functions of Plasmodium actins. |
| GO:0032982 myosin filament | ISS GO_REF:0000024 | MODIFY | Summary: This row has the right biology attached to an impossible compartment, so it is repaired rather than removed. The donor, UniProtKB:F1P476, is an unreviewed TrEMBL chicken entry for ACTA2 - a second record for the same paralog used on rows 11-16 - carrying a single IDA for GO:0032982 from PMID:8006065, "Actin isoform compartments in chicken gizzard smooth muscle cells". The sentence behind the annotation reads that with a muscle-actin-specific antibody labelling was found generally around the myosin filaments of the contractile apparatus. Around, not within. GO:0032982 is defined as a supramolecular fiber containing myosin heavy chains in which the heavy chains are arranged into a filament; an actin cannot be located in one. What the observation describes is the actin filaments of the smooth muscle contractile apparatus interdigitating with the myosin filaments, which is GO:0030485 smooth muscle contractile fiber. Unusually for this donor, the underlying biology transfers better to ACTG2 than to ACTA2. Chicken gizzard is enteric smooth muscle, the paper states that differentiated smooth muscle contains both muscle actin isoforms, and the antibody used was pan-muscle-actin rather than alpha-specific, so gamma-enteric actin is among what was labelled. The compartment call therefore stands once the term is corrected. Reason: The term is not merely too general, it is categorically wrong: GO:0032982 is a myosin-heavy-chain polymer and cannot contain actin, and the source sentence says labelling was around the myosin filaments, not in them. GO:0030485 smooth muscle contractile fiber is what the observation supports and is the structure ACTG2 occupies in tissue. On why this row is repaired while five rows from the same chicken gene are removed: the split is not about the donor, it is about whether the observation is one that could have been made on gamma-enteric actin. Those five come from chick cardiac cushion mesenchyme using an antibody raised against alpha-smooth-muscle actin specifically, a cell type and a reagent that exclude ACTG2. This one comes from gizzard - enteric smooth muscle, where gamma-enteric actin is a major isoform - using a pan-muscle-actin antibody that does not discriminate the two. The paralogy of the donor is the same in both cases; the transferability of the observation is not. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: COMPARTMENT OR COMPLEX MISMATCH SOURCE MISCITATION Sources checked: UniProtKB:F1P476 · ACTA2 (Gallus gallus, TrEMBL (unreviewed), 377 aa) SUPPORTS SOURCE BUT NOT TARGET Unreviewed entry, and a duplicate record for the same chicken gene as the reviewed P08023 used on rows 11-16. Own evidence in QuickGO for this term: IDAx1 from PMID:8006065, whose own sentence describes labelling around rather than within the myosin filaments. The observation itself does apply to gamma-enteric actin, because the antibody was pan-muscle-actin and the tissue was gizzard. Proposed replacements: smooth muscle contractile fiber Supporting Evidence: PMID:8006065 Using an antibody specific only for muscle actin, labelling was found generally around the myosin filaments of the contractile apparatus, but was absent from the core of the dense bodies that contained beta-actin. PMID:8006065 Differentiated smooth muscle cells typically contain a mixture of muscle (alpha and gamma) and cytoplasmic (beta and gamma) actin isoforms. PMID:38820162 Actin and myosin polymers form the thin and thick filaments of muscle cells that interact and slide past one another during muscle contraction, respectively |
| GO:0005737 cytoplasm | ISS GO_REF:0000024 | KEEP AS NON CORE | Summary: First of six rows transferred by curated sequence similarity from UniProtKB:P08023, which resolves to chicken ACTA2, "Actin, aortic smooth muscle" - the other smooth muscle actin, not ACTG2's orthologue. All six of that entry's source annotations come from a single 1999 paper, PMID:10633868, on alpha-smooth-muscle actin in chick atrioventricular endocardial cushion formation. Three facts make the block a paralog transfer: chicken has its own reviewed ACTG2 (P63270, 376 aa) which was not used; the same block is already on human ACTA2, which is the correct target and holds five of the six rows from the same donor; and the cell type studied is migrating cardiac cushion mesenchyme, which is not where ACTG2 is expressed. This particular row is nevertheless kept, because the conclusion is independently true and independently sourced: UniProt's own curated SUBCELLULAR LOCATION line for ACTG2 reads "Cytoplasm, cytoskeleton". Removing a correct compartment because the inference route that happened to deliver it was flawed would be a worse outcome than keeping it and recording the flaw. It is marked non-core because "cytoplasm" says nothing about where in the cytoplasm ACTG2 acts, which is the contractile apparatus. Reason: Correct conclusion reached by a flawed route. The donor is the chicken paralog rather than the available chicken orthologue, but ACTG2 is a cytoplasmic protein on UniProt's own curation, so the term stands. Non-core because it is the least specific true statement available about this protein's location. Propagation Review Root cause: NO FAILURE NON CORE Failure modes: WRONG ORTHOLOG OR PARALOG Sources checked: UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) - chicken alpha-smooth-muscle actin SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IDAx1 from PMID:10633868. A paralog: chicken's own ACTG2 orthologue is P63270 and was not used. The transferred term is nonetheless true of ACTG2 on independent UniProt curation, which is why this row alone of the six is kept. Supporting Evidence: file:human/ACTG2/ACTG2-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm, cytoskeleton. file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - shared (**5**): GO:0010628, GO:0030027, GO:0030175, GO:0044297, GO:0090131 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md The chicken orthologue that was available and not used is **P63270** = ACTH_CHICK, gene **ACTG2**, 376 aa (Swiss-Prot). PMID:10633868 alpha-Smooth-muscle actin (SMA) is the major isoform of adult vascular tissues. |
| GO:0010628 positive regulation of gene expression | ISS GO_REF:0000024 | REMOVE | Summary: Second of the six chicken-ACTA2 ISS rows, and the one whose term is hardest to reconcile with its source. PMID:10633868 uses antisense oligodeoxynucleotides to knock down alpha-smooth-muscle actin and reports that this reduced SMA expression and mesenchymal formation; SMA in that paper is the target of TGF-beta-dependent induction, not an agent that raises the expression of anything else. Nothing in the abstract supports smooth-muscle actin positively regulating gene expression. Only the abstract of that 1999 paper was available, so the AgBase IDA on chicken ACTA2 is not declared wrong here; it is raised as a question for AgBase in suggested_questions. The removal of the human row does not rest on it. It rests on the same grounds as rows 13-16: the donor is the chicken paralog while chicken's own ACTG2 orthologue exists and was not used, the observation is in chick cardiac cushion mesenchyme rather than visceral smooth muscle, and human ACTA2 - the correct ISS target - already carries the identical row from the identical donor. Reason: Paralog transfer of a process term that the source paper does not support even for the paralog, into a gene whose expression is restricted to tissues the paper did not study. The correct human recipient, ACTA2, already holds the same annotation from the same donor. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG CONTEXT OR TISSUE MISMATCH ROLE CONFLATION Sources checked: UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) - chicken alpha-smooth-muscle actin SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IDAx1 from PMID:10633868. Chicken's own ACTG2 orthologue P63270 exists and was not used, and human ACTA2 already carries this same ISS row from this same donor. Supporting Evidence: PMID:10633868 Antisense oligodeoxynucleotide (ODNs) specific for SMA reduced both SMA expression and mesenchymal formation in AV endothelial cells cultured with myocardium on a collagen gel lattice. PMID:10633868 We examined its role in the formation of chicken atrioventricular (AV) endocardial cushion tissue. file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:10633868 | 8 | 2 | file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - shared (**5**): GO:0010628, GO:0030027, GO:0030175, GO:0044297, GO:0090131 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md The chicken orthologue that was available and not used is **P63270** = ACTH_CHICK, gene **ACTG2**, 376 aa (Swiss-Prot). |
| GO:0030027 lamellipodium | ISS GO_REF:0000024 | REMOVE | Summary: Third of the six chicken-ACTA2 ISS rows. The observation is that alpha-smooth-muscle actin is distributed in a punctate manner in the lamellipodia and filopodia of invading mesenchymal cells during chick atrioventricular cushion formation. That is a real localisation of the chicken paralog in a migrating embryonic mesenchymal cell. It is not a localisation of human enteric gamma-actin, which is expressed in the differentiated smooth muscle of intestinal muscularis mucosa and muscularis propria. Chicken's own ACTG2 orthologue (P63270) was available and was not used, and human ACTA2 already carries this identical row from this identical donor. Reason: Paralog transfer of a cell-type-specific localisation from migrating chick cardiac cushion mesenchyme to a differentiated visceral smooth muscle actin. The correct human orthologue of the donor, ACTA2, already holds it. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) - chicken alpha-smooth-muscle actin SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IDAx1 from PMID:10633868. Chicken ACTG2 (P63270) is the unused orthologue; human ACTA2 already carries the same ISS row. Supporting Evidence: PMID:10633868 distributed in a punctate manner in the lamellipodia/filopodia of invading mesenchymal cells file:human/ACTG2/ACTG2-uniprot.txt CC muscle cells of muscularis mucosa and muscularis propria. Also detected file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - shared (**5**): GO:0010628, GO:0030027, GO:0030175, GO:0044297, GO:0090131 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md The chicken orthologue that was available and not used is **P63270** = ACTH_CHICK, gene **ACTG2**, 376 aa (Swiss-Prot). |
| GO:0030175 filopodium | ISS GO_REF:0000024 | REMOVE | Summary: Fourth of the six chicken-ACTA2 ISS rows, from the same sentence as row 13 and removed for the same reasons: paralog donor while the chicken orthologue exists, embryonic cardiac-cushion mesenchyme rather than visceral smooth muscle, and the annotation already correctly placed on human ACTA2. Filopodia are protrusive structures of migrating cells built largely by cytoplasmic actins; there is no report of enteric gamma-actin in one. Reason: Paralog transfer of a migrating-cell localisation into a differentiated smooth muscle actin, duplicating an annotation that already sits correctly on ACTA2. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) - chicken alpha-smooth-muscle actin SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IDAx1 from PMID:10633868. Supporting Evidence: PMID:10633868 distributed in a punctate manner in the lamellipodia/filopodia of invading mesenchymal cells file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - shared (**5**): GO:0010628, GO:0030027, GO:0030175, GO:0044297, GO:0090131 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md The chicken orthologue that was available and not used is **P63270** = ACTH_CHICK, gene **ACTG2**, 376 aa (Swiss-Prot). |
| GO:0044297 cell body | ISS GO_REF:0000024 | REMOVE | Summary: Fifth of the six chicken-ACTA2 ISS rows, again from PMID:10633868 and again in transforming chick endothelial and mesenchymal cells - the term distinguishes the cell body from the migratory appendages described in rows 13 and 14. Removed on the same grounds. It is also the least informative of the six: for a protein that is cytoplasmic throughout the cell, "cell body" adds nothing that row 11 does not already say, and it carries the misleading implication of a neuronal or otherwise polarised cell. Reason: Paralog transfer from an embryonic chick cell type, adding no information beyond the cytoplasm row and implying a cellular polarity ACTG2 does not have. Already held by ACTA2, the donor's true orthologue. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) - chicken alpha-smooth-muscle actin SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IDAx1 from PMID:10633868. Supporting Evidence: PMID:10633868 We examined its role in the formation of chicken atrioventricular (AV) endocardial cushion tissue. file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - shared (**5**): GO:0010628, GO:0030027, GO:0030175, GO:0044297, GO:0090131 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md The chicken orthologue that was available and not used is **P63270** = ACTH_CHICK, gene **ACTG2**, 376 aa (Swiss-Prot). |
| GO:0090131 mesenchyme migration | ISS GO_REF:0000024 | REMOVE | Summary: Last of the six chicken-ACTA2 ISS rows and the only one whose donor evidence is IMP rather than IDA: antisense knockdown of alpha-smooth-muscle actin in chick atrioventricular endothelial cells reduced mesenchymal formation and blocked TGF-beta-inducible migratory appendage formation. That is a genuine developmental role for chicken alpha-SMA, and human ACTA2 carries the annotation from the same donor. It is not a role of human ACTG2. Enteric gamma-actin is expressed in differentiated visceral smooth muscle; the process here is endothelial-to-mesenchymal transformation in the embryonic heart, in a species where the true ACTG2 orthologue (P63270) exists and was not used. This is the clearest single case of the block being aimed at the wrong smooth muscle actin. Reason: A developmental process demonstrated for chicken alpha-smooth-muscle actin and transferred to the wrong human paralog, when the chicken ACTG2 orthologue was available and when human ACTA2 already holds the annotation. No evidence links ACTG2 to mesenchyme migration in any species. Propagation Review Root cause: PROPAGATION BAD Failure modes: WRONG ORTHOLOG OR PARALOG CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:P08023 · ACTA2 (Gallus gallus, Swiss-Prot, 377 aa) - chicken alpha-smooth-muscle actin SUPPORTS SOURCE BUT NOT TARGET Own evidence in QuickGO for this term: IMPx1 from PMID:10633868, from antisense knockdown of alpha-SMA. Real evidence about the paralog; nothing about ACTG2. Supporting Evidence: PMID:10633868 Antisense oligodeoxynucleotide (ODNs) specific for SMA reduced both SMA expression and mesenchymal formation in AV endothelial cells cultured with myocardium on a collagen gel lattice. file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md - shared (**5**): GO:0010628, GO:0030027, GO:0030175, GO:0044297, GO:0090131 file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md The chicken orthologue that was available and not used is **P63270** = ACTH_CHICK, gene **ACTG2**, 376 aa (Swiss-Prot). PMID:10633868 alpha-Smooth-muscle actin (SMA) is the major isoform of adult vascular tissues. |
| GO:0070062 extracellular exosome | HDA PMID:23533145 In-depth proteomic analyses of exosomes isolated from expres... | MARK AS OVER ANNOTATED | Summary: Bulk proteomics of exosomes from expressed prostatic secretions in urine. The reference-projection check gives 1046 annotations over 1046 distinct gene products, all to this single term, with the full result enumerated across all six pages rather than read off page one. A term assigned once to every protein detected in a preparation records the preparation, not a targeted localisation of any one protein, and a highly abundant cytoplasmic filament protein is exactly what carries over into such preparations. ACTG2 has no signal peptide and no reported secretion mechanism. Resolved the same way as the merged ACTR1A and ACTR1B reviews resolve their identical GO:0070062 HDA rows. Reason: One term applied to all 1046 proteins detected in a single proteomic preparation. Not false as a detection record, but it does not support a localisation claim for a cytoplasmic actin, and it is not evidence of any function. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:23533145 | 1046 | 1046 | GO:0070062: 1046 | file:human/ACTG2/ACTG2-uniprot.txt CC -!- SUBCELLULAR LOCATION: Cytoplasm, cytoskeleton. |
| GO:0072562 blood microparticle | HDA PMID:22516433 Proteomic analysis of microvesicles from plasma of healthy d... | MARK AS OVER ANNOTATED | Summary: Same shape as row 17, from plasma microvesicle proteomics: 141 annotations over 141 distinct gene products, one term each, enumeration complete. The paper's own title reports high individual variability in the protein content of these preparations. Actin is among the most abundant proteins in any cell lysate and its appearance in a circulating vesicle preparation is uninformative about ACTG2 in particular. Reason: A single term given to every protein detected in one plasma microvesicle preparation. Records detection, not localisation of a cytoplasmic smooth muscle actin. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:22516433 | 141 | 141 | GO:0072562: 141 | |
| GO:0005576 extracellular region | HDA PMID:23580065 Shotgun proteomics reveals specific modulated protein patter... | MARK AS OVER ANNOTATED | Summary: Shotgun proteomics of tears from glaucoma patients: 95 annotations over 95 distinct gene products, one term each, enumeration complete. Together with rows 17 and 18 and the IBA on row 1, this is the whole of ACTG2's extracellular record - three bulk proteomic preparations and one paralogous rat immunostain. None of it identifies a mechanism by which a cytoplasmic actin would reach the extracellular space. Reason: One term applied to every protein detected in a tear proteome. Detection in a body fluid by shotgun proteomics does not localise a cytoplasmic actin to the extracellular region. Supporting Evidence: file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md | PMID:23580065 | 95 | 95 | GO:0005576: 95 | |
| GO:0005829 cytosol | TAS Reactome:R-HSA-445699 | KEEP AS NON CORE | Summary: First of eight identical GO:0005829 rows exported one per Reactome reaction (R-HSA-445699, -445700, -445704, -445705, -9914537, -9934294, -9934410, -9934486, spanning Smooth Muscle Contraction, Regulation of CDH1 Function and the dystrophin-glycoprotein complex). They are one curatorial claim - that ACTG2 participates in these reactions as a cytosolic entity - rendered eight times by the export, not eight independent findings, and the reasoning is given once here and cross-referenced on rows 21 to 27. The claim is true and, for this gene, more than usually meaningful. Actin exists in a monomer-polymer equilibrium, the monomeric G-actin pool is cytosolic, and that pool is the disease-relevant variable here: every characterised ACTG2 variant is assayed as a shift in the G-actin/F-actin ratio or in the critical concentration. It is nevertheless not the core location. ACTG2's function is discharged in the filament, in the contractile apparatus of visceral smooth muscle; cytosol is where the unpolymerised reserve sits. Reason: True, and more meaningful for this gene than for most, since the cytosolic G-actin pool is what disease variants perturb. Non-core because the protein does its job in the filament, and because the eight identical rows are one Reactome export rather than eight findings. Supporting Evidence: PMID:38820162 WT ACTG2 exhibited slower polymerization compared to PMID:36264152 The Actg2R257C heterozygous variant impairs SMCs contraction by interfering with actin polymerization, leading to GI motility disorders. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-445700 | KEEP AS NON CORE | Summary: Second of the eight per-reaction Reactome cytosol rows; see row 20 for the reasoning, which is identical and is not restated per reaction. Reason: True but generic, and one of eight renderings of a single Reactome claim. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-445704 | KEEP AS NON CORE | Summary: Third of the eight per-reaction Reactome cytosol rows; see row 20. Reason: True but generic, and one of eight renderings of a single Reactome claim. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-445705 | KEEP AS NON CORE | Summary: Fourth of the eight per-reaction Reactome cytosol rows; see row 20. Reason: True but generic, and one of eight renderings of a single Reactome claim. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-9914537 | KEEP AS NON CORE | Summary: Fifth of the eight per-reaction Reactome cytosol rows; see row 20. This one belongs to the dystrophin-glycoprotein complex pathway rather than to smooth muscle contraction, but the compartment claim it exports is the same. Reason: True but generic, and one of eight renderings of a single Reactome claim. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-9934294 | KEEP AS NON CORE | Summary: Sixth of the eight per-reaction Reactome cytosol rows; see row 20. From the Regulation of CDH1 Function pathway. Reason: True but generic, and one of eight renderings of a single Reactome claim. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-9934410 | KEEP AS NON CORE | Summary: Seventh of the eight per-reaction Reactome cytosol rows; see row 20. Reason: True but generic, and one of eight renderings of a single Reactome claim. |
| GO:0005829 cytosol | TAS Reactome:R-HSA-9934486 | KEEP AS NON CORE | Summary: Eighth and last of the per-reaction Reactome cytosol rows; see row 20. This reaction is "CDH1-associated CTNNA1 binds F-actin", i.e. the one place in the export where ACTG2 is being used as filamentous rather than monomeric actin, which makes cytosol the least apt of the eight. Reason: True but generic, and one of eight renderings of a single Reactome claim; in this reaction ACTG2 stands in for F-actin, so the filament terms proposed below fit the underlying claim better than cytosol does. |
| GO:0005884 actin filament | IDA PMID:38820162 Molecular mechanisms linking missense ACTG2 mutations to vis... | NEW | Summary: Proposed new annotation, and the central gap in ACTG2's record. ACTG2 has no annotation to GO:0005884 or to any descendant of it, while every other conventional human actin has one. That is not a judgement anyone made about ACTG2; it is an artefact of node placement. In PANTHER family PTHR11937, GO:0005884 is asserted at exactly three nodes, all on 2019-03-01: PTN000233075, whose sole IBD seed is ACTA1 and which reaches ACTA1 alone; PTN000748220, whose sole seed is ACTC1 and which reaches ACTC1 alone; and PTN002631586, seeded from ACTG1 and mouse Actg1, which reaches ACTB, ACTBL2, ACTG1, ACTL8 and the five POTE genes. Two singleton nodes plus one cytoplasmic-clade node, and no node covering the smooth muscle actins - so ACTA2 and ACTG2 both fall through. The only node in the family that does specifically cover the smooth muscle pair is PTN004322804, and what it gives them is extracellular region (row 1). For ACTG2 the gap can be closed with the gene's own data rather than with another inference. Ceron et al. expressed human ACTG2 in human cells, purified it, polymerised it, and solved three cryo-EM helical reconstructions of ACTG2 filaments at 2.45 to 2.72 A (PDB 8V2O wild type, 8V2Z R257C, 8V30 R40C), reporting that the structures are very similar to that of alpha-actin with no change in helical twist or rise. Independently, the 38 protomer-interface positions computed from an F-actin structure are 38/38 identical in ACTG2, tying ACTA1 and ACTC1 at the top of a 16-protein panel whose bottom is ACTL10 at 5/38; that panel is only reported because it first reproduces all 14 columns published in the merged ACTL8 review exactly. is_active_in rather than located_in, because ACTG2's molecular function - contributing structural integrity to a cytoskeletal structure - is discharged in the filament, and because that is the qualifier the conventional actins already carrying this term use. Reason: A concrete, structurally demonstrated annotation gap affecting exactly the two smooth muscle actins, traceable to which PANTHER node the term was attached to rather than to any assessment of the genes. Supported here by ACTG2's own cryo-EM filament structures, not by family inference, so it can be entered as IDA. Supporting Evidence: PMID:38820162 structures of WT ACTG2, R40C, and R257C filaments using helical reconstruction PMID:38820162 The three structures are very similar overall and also similar to that of PMID:38820162 In the WT structure, residue R257 forms an important intrasubunit salt bridge with E196, located on a loop that contacts a protomer from the opposite long-pitch helix file:human/ACTG2/ACTG2-uniprot.txt DR PDB; 8V2O; EM; 2.45 A; A/B/C/D/E=1-376. file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md ACTG2 scores **38/38** chemically compatible (38 identical, 0 conservative, 0 non-conservative, 0 gaps) and ACTA2 **38/38**. Both sit with the conventional actins file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md all 14 published columns reproduced exactly |
| GO:0043531 ADP binding | IDA PMID:38820162 Molecular mechanisms linking missense ACTG2 mutations to vis... | NEW | Summary: Proposed new annotation. ACTG2 currently carries GO:0016887 ATP hydrolysis activity but no nucleotide-binding term at all - GOA holds neither GO:0005524 nor GO:0043531 for it. The gap is not specific to ACTG2: after the retirement of UniProt-keyword-derived annotations, ACTA2, ACTB and ACTG1 also carry no nucleotide-binding term, and among the conventional actins only ACTA1 (TAS) and ACTC1 (IDA) do. The term chosen is the ligand actually observed, following the principle the merged ACTR1A, ACTR1B and ACTR5 reviews used. Reading the coordinates of all three ACTG2 structures gives ADP and Mg in every protomer of every one, and the authors state why: hydrolysis and phosphate release happened during sample preparation, before vitrification. No structure resolves ATP-bound ACTG2. ATP binding is equally true of the protein - UniProt gives ACTG2 EC 3.6.4.- and the full ATP + H2O reaction, and hydrolysis presupposes binding - but it rests on inference from the catalytic-activity statement rather than on a resolved observation. It is therefore proposed as a separate companion row below at the weaker evidence code it earns, ISS, rather than folded into this one, so that a curator can see which of the two nucleotide terms is observed and which is inferred. ACTA1 likewise carries both. Reason: A real gap: an ATP-hydrolysing actin with no nucleotide-binding annotation. Grounded in the ligand resolved in ACTG2's own structures rather than in the actin fold, which makes ADP binding the directly-evidenced call and ATP binding the inferred companion. Supporting Evidence: PMID:38820162 release occurred during sample preparation, before vitrification, resulting in all three structures containing Mg file:human/ACTG2/ACTG2-bioinformatics/RESULTS.md Every ACTG2 filament structure carries **ADP**, not ATP, in every protomer. file:human/ACTG2/ACTG2-uniprot.txt Reaction=ATP + H2O = ADP + phosphate + H(+); Xref=Rhea:RHEA:13065, |
| GO:0005524 ATP binding | ISS GO_REF:0000024 | NEW | Summary: Proposed new annotation, the companion to the ADP binding row above, and the term that makes ACTG2's nucleotide record internally consistent. GOA currently asserts that ACTG2 hydrolyses ATP (GO:0016887, row 9) while giving it no ATP-binding annotation at all, which cannot be right: a protein that hydrolyses a nucleotide binds it. The gap is family-wide rather than a judgement about ACTG2. Of the six conventional human actins, only ACTA1 (GO:0005524 and GO:0043531, both TAS) and ACTC1 (GO:0005524 IDA) carry any term under GO:0000166 nucleotide binding; ACTB, ACTG1, ACTA2 and ACTG2 carry none. That is the expected consequence of the retirement of UniProt-keyword-derived annotations, which is where ACTG2's own ATP-binding row used to come from - UniProt's DR block still lists GO:0005524 IEA:UniProtKB-KW for this entry, but GOA no longer exports it. ISS rather than IDA, because unlike ADP binding no ACTG2 structure resolves bound ATP. The donor is human ACTC1 (P68032), which holds GO:0005524 by IDA. It is deliberately NOT the pig ACTA1 (P68137) used for the GO:0016887 hydrolysis row: P68137 carries no GO:0005524, and in fact no annotation anywhere under GO:0000166 nucleotide binding, so an ISS from it would transfer a term the donor does not hold. The merged ACTA2 review adjudicates the identical question and reaches the same substitution. Caveat carried from that review: ACTC1 is a paralogue rather than an ortholog of ACTG2, and GO_REF:0000024 is worded for orthologs; the transfer rests on the two proteins being 377/376-aa conventional actins identical across the nucleotide pocket. The claim is binding only; no additional ATP-dependent activity is asserted beyond the GO:0016887 already accepted on row 9. Reason: Restores internal consistency to a record that asserts ATP hydrolysis with no ATP binding, a gap created by the retirement of keyword-derived annotations rather than by any assessment of the protein. Entered as ISS, one code weaker than the ADP binding row, because ATP-bound ACTG2 has not been resolved - the distinction between the observed and the inferred ligand is kept visible rather than collapsed. Propagation Review Root cause: NO FAILURE CORE Sources checked: UniProtKB:P68032 · ACTC1 (Homo sapiens, Swiss-Prot, 377 aa) - cardiac alpha-actin SUPPORTS TRANSFER Holds GO:0005524 by IDA, so the donated term is one the donor actually carries. Chosen over the GO:0016887 donor P68137 (pig ACTA1), which holds no GO:0005524 and nothing under GO:0000166; the merged ACTA2 review makes the same substitution for the same reason. Paralogue rather than ortholog, so the GO_REF:0000024 wording is a stretch. Supporting Evidence: file:human/ACTG2/ACTG2-uniprot.txt Reaction=ATP + H2O = ADP + phosphate + H(+); Xref=Rhea:RHEA:13065, file:human/ACTG2/ACTG2-uniprot.txt DR GO; GO:0005524; F:ATP binding; IEA:UniProtKB-KW. |
| GO:0017022 myosin binding | IDA PMID:38820162 Molecular mechanisms linking missense ACTG2 mutations to vis... | NEW | Summary: Proposed new annotation, and the second term stranded by the same node-placement problem as GO:0005884. GO:0017022 is asserted in PTHR11937 at PTN000748220 only, seeded from ACTC1, and reaches ACTC1 alone; the same node also carries GO:0033275 actin-myosin filament sliding and GO:0007015 actin filament organization, each likewise reaching only ACTC1. Nothing about myosin engagement is specific to cardiac actin. ACTG2's own evidence is direct. Ceron et al. ran in-vitro motility assays in which phalloidin-stabilised ACTG2 filaments were propelled across coverslips coated with a smooth muscle myosin construct comprising the motor domain and essential light chain MYL6, and found the gliding velocity indistinguishable from that of alpha-actin, with and without the smooth muscle tropomyosin Tpm1.4. Productive myosin-driven translocation of a filament requires the myosin motor to bind it. The same paper independently places one of the disease-mutated residues, R148, on the myosin-binding interface. Deliberately claimed at binding level only. The gliding assay uses a myosin fragment on a surface, so it demonstrates the interaction and its motor output but does not license a tissue-level term on its own; smooth muscle contraction is proposed separately, on genetic evidence. On the evidence code: IPI would be the convention if the interacting partner were identified, but the paper does not identify it. SMM-S1 is described only as a "smooth muscle myosin construct" from a baculovirus provided by a collaborator, coexpressed in Sf9 cells with the essential light chain MYL6; no heavy-chain gene or species is given, and MYH11 appears in the paper only in the introduction, as a visceral-myopathy disease gene, not as the identity of this construct. MYL6 is named but is a light chain, not the actin-binding moiety, so citing it as the interactor would misdescribe what was assayed. IDA is therefore used for a direct in vitro assay whose partner is characterised functionally but not by accession; a curator who can establish the construct's heavy chain from the cited methods reference should convert this to IPI. Reason: Direct in-vitro demonstration that smooth muscle myosin engages and translocates ACTG2 filaments at wild-type velocity, filling a term that PAINT currently attaches only to ACTC1's own node despite it being a general property of muscle actins. Supporting Evidence: PMID:38820162 The filament gliding velocity was similar for PMID:38820162 we used smooth muscle myosin construct SMM-S1, comprising the motor domain and essential light chain (MYL6) PMID:38820162 R148 also forms part of the myosin-binding interface |
| GO:0006939 smooth muscle contraction | IMP PMID:31769566 Recurrent arginine substitutions in the ACTG2 gene are the p... | NEW | Summary: Proposed new annotation. ACTG2 has no biological-process annotation describing what it is for, and what it is for is the contraction of visceral smooth muscle. The care needed here is to distinguish the protein's function from the consequences of mutating it, since almost the entire ACTG2 literature is disease genetics: the claim made is that ACTG2 participates in smooth muscle contraction, not that it participates in visceral myopathy. The evidence is a convergent human genetic series. Heterozygous missense variants at a small set of conserved arginines account for about half of a 53-family visceral myopathy cohort and are invariably the alleles behind the worst outcomes; the phenotype is loss of contractile function in bowel, bladder and uterus. Independently, two CRISPR knock-in mouse lines carrying human variants (R257C, D245G) reproduce it - reduced smooth muscle cell contraction in collagen gel assays, prolonged gastrointestinal transit, impaired voluntary urination - and in both the proximal defect is impaired actin polymerisation. The mechanism is dominant-negative rather than loss of function, which is why the annotation is stated at process-participation level: heterozygous truncating and frameshift ACTG2 alleles occur in healthy people, so a null allele is tolerated and the variants act by poisoning the filament. That does not weaken the inference that ACTG2 is the thin-filament actin of visceral smooth muscle - it is why poisoning it is so damaging - but it is the reason no haploinsufficiency-based claim is made. Reason: The gene's defining biological process, absent from GOA. Supported by a large human genotype-phenotype series and by two independent knock-in mouse models, with the biophysical route - impaired polymerisation, reduced contractility - measured in both. Stated as participation in contraction rather than as a disease term. Supporting Evidence: PMID:31769566 We ascertained 53 families, with visceral myopathy based on megacystis, functional bladder/gastrointestinal obstruction, or microcolon. PMID:31769566 poor outcomes (characterized by total parenteral nutrition dependence, death, or transplantation) were invariably due to one of the arginine missense alleles PMID:36264152 The Actg2R257C heterozygous variant impairs SMCs contraction by interfering with actin polymerization, leading to GI motility disorders. PMID:40617346 The Actg2D245G mutation affects intestinal motility by impairing actin polymerization and reducing cell contraction. PMID:26647307 the identified variants not only impair ACTG2 polymerization, but also contribute to reduced cell contractility PMID:38820162 These mutations are thought to cause disease through dominant-negative rather than loss-of-function mechanisms since heterozygous ACTG2 truncations and frameshifts have been documented in healthy individuals |
Loading supporting content…
Download this section (compressed HTML)Q: For PAINT, stated once with all affected genes named. Inside PTHR11937, GO:0005884 actin filament is asserted at three nodes - PTN000233075 (seed ACTA1, reaches ACTA1 only), PTN000748220 (seed ACTC1, reaches ACTC1 only) and PTN002631586 (seeds ACTG1 and mouse Actg1, reaches ACTB, ACTBL2, ACTG1, ACTL8 and the five POTE genes) - so the two smooth muscle actins ACTA2 and ACTG2 receive it from nowhere and hold zero annotations to it or any descendant. The same node PTN000748220 likewise confines GO:0017022 myosin binding, GO:0033275 actin-myosin filament sliding and GO:0007015 actin filament organization to ACTC1 alone. Would PAINT consider moving these four terms to a node that covers the muscle actins, or asserting them at a smooth-muscle node? One edit would fix two genes for four terms.
Q: For PAINT, the reciprocal half of the same finding. PTN004322804 is the only node in PTHR11937 whose human reach is exactly the smooth muscle actin pair, and the only term it carries is GO:0005576 extracellular region, from a single IBD seed, RGD:621676 = rat Acta2. No non-IBA evidence for that exact term exists anywhere in the chain: the seed gene is also a recipient of the resulting IBA - its GO:0005576 row cites this node and its own identifier - and its only non-IBA annotation anywhere under GO:0005576 is a basement-membrane IDA in fetal rat testis, three levels below the term asserted. Should that assertion be withdrawn? It is the only node-level statement PAINT makes specifically about the smooth muscle actins, and it places them outside the cell.
Q: For UniProt, on the AgBase-derived ISS block. Six ACTG2 rows are transferred from UniProtKB:P08023 (chicken ACTA2) although chicken has its own reviewed ACTG2 orthologue, P63270, and although human ACTA2 already carries five of the same rows from the same donor. Should the ACTG2 copies be withdrawn, and should GO_REF:0000024 transfers be constrained to the closest available orthologue when one exists in the source organism?
Q: For AgBase, a question rather than an assertion because only the abstract of the 1999 source was available. PMID:10633868 knocks down chick alpha-smooth-muscle actin with antisense oligonucleotides and reports reduced SMA expression and reduced mesenchymal formation; SMA is the target of TGF-beta-dependent induction throughout. On what basis was GO:0010628 positive regulation of gene expression assigned to chicken ACTA2 by IDA from that paper?
Q: For RGD. Rat Actg2 (P63269) carries GO:0045471, GO:0071354 and GO:1905641 by IEP from PMID:28320086 and PMID:21294755. Both papers measure alpha-smooth-muscle actin as a hepatic stellate cell activation marker - the ACTA2 gene product - and rat Acta2 (P62738) received no annotation from either. Were these intended for Acta2?
Q: For GOA and UniProt. ACTG2 carried GO:0016887 ATP hydrolysis activity but no nucleotide-binding term of any kind, and after the retirement of keyword-derived annotations the same is true of ACTA2, ACTB and ACTG1; only ACTA1 and ACTC1 among the conventional actins carry one, although UniProt's own DR block still lists GO:0005524 IEA:UniProtKB-KW for ACTG2. Both are proposed here - GO:0043531 as IDA on the ligand resolved in ACTG2's own structures and GO:0005524 as ISS on the catalytic-activity statement. Is the same gap worth closing across ACTA2, ACTB and ACTG1, which have identical evidence and identical silence?
Q: For GO ontology development. Is there appetite for a smooth muscle thin filament term? The smooth muscle branch is modelled at fibre and dense-body level but not at filament level, while the striated branch has GO:0005865, and the gap forces smooth muscle actins, tropomyosins and caldesmon onto terms that are either generic or wrong.
Q: For the ACTG2 field. Every characterised variant is assayed as a shift in polymerisation kinetics, yet ACTG2 filaments are propelled by smooth muscle myosin at wild-type velocity even when they carry R40C or R257C. Which downstream property of the mutant filament - fragility under load, tropomyosin occupancy, or nucleator engagement - dominates in intact muscle?
Experiment: Reconstitute the visceral smooth muscle thin filament from purified human components - ACTG2, Tpm1.4, caldesmon and calponin - and measure force and duty ratio against smooth muscle myosin in an optical-trap assay, for wild type and for the four common arginine variants at the 50/50 wild-type/mutant ratio that heterozygotes actually have. The gliding assays reported so far measure unloaded velocity, which is precisely the parameter the mutants do not change.
Hypothesis: The ACTG2 disease variants impair contraction under load rather than unloaded sliding, which is why gliding velocity is normal while patients lose peristalsis.
Type: single-molecule mechanics on a reconstituted thin filament
Experiment: Determine whether ACTG2 and ACTA2 co-polymerise into common filaments in visceral smooth muscle or segregate into separate filament populations, using isoform-specific antibodies with expansion microscopy or immuno-electron tomography on human intestinal muscularis. Much of the annotation record treats the two genes as interchangeable, and the answer determines whether a dominant-negative ACTG2 allele can also poison ACTA2 filaments.
Hypothesis: ACTG2 and ACTA2 co-polymerise into common thin filaments in visceral smooth muscle, so a dominant-negative ACTG2 allele can poison filaments containing either isoform.
Type: isoform-resolved super-resolution imaging of human tissue
Experiment: Test whether the SETD3-dependent His-74 methylation resolved in the ACTG2 cryo-EM structures is required for visceral smooth muscle function, by comparing gastrointestinal transit and bladder emptying in Setd3-null mice with the phenotype of the Actg2 knock-in lines, and by measuring polymerisation kinetics of unmethylated recombinant ACTG2 expressed in SETD3-deficient cells.
Hypothesis: SETD3-dependent His-74 methylation of ACTG2 is required for normal visceral smooth muscle contraction, not only for uterine contraction at parturition.
Type: mouse phenotyping plus in vitro polymerisation kinetics
Experiment: Establish which cells in human intestine express ACTG2 versus ACTA2 at protein level, by single-cell proteomics or isoform-resolved imaging across muscularis mucosa, muscularis propria and intestinal vasculature. Several of the annotations reviewed here rest on treating the two as one protein, and UniProt's tissue-specificity statement rests on a single study.
Hypothesis: ACTG2 and ACTA2 occupy distinct cell populations within the intestinal wall, so annotations transferred between them are transferred between different cell types.
Type: single-cell proteomics and isoform-resolved imaging
Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)Loading supporting content…
Download this section (compressed HTML)