AKAP12

Existing Annotations Review

GO Term	Evidence	Action	Reason
GO:0005515 protein binding	IPI PMID:16642035 An anchored PKA and PDE4 complex regulates subplasmalemmal c...	MODIFY	Summary: Partner P13861 is human PRKAR2A, the RII-alpha regulatory subunit of PKA. This is AKAP12's defining activity and the assay is a direct one, so the bare term should be replaced by the specific child. Reason: P13861 resolves to reviewed canonical PRKAR2A (404 aa, Swiss-Prot), the type II-alpha regulatory subunit of protein kinase A - so GO:0034237 protein kinase A regulatory subunit binding is directly licensed by the same experiment that produced this row. IntAct records this pair under detection method "filter binding" for PMID:16642035, a direct in-vitro binding assay rather than a screen, on a partner with 231 IntAct records (moderate, not a hub), and a second, methodologically independent record ("anti tag coip", PMID:35271311 OpenCell endogenous tagging). The interaction is independently established on the human protein by PMID:9000000, which mapped high-affinity RII binding to gravin residues 1526-1780 and isolated a PKA-gravin complex from human erythroleukemia cells. Bare "protein binding" here discards the one fact that names the gene. Proposed replacements: protein kinase A regulatory subunit binding Supporting Evidence: PMID:16642035 Co-immunoprecipitation confirmed that gravin maintains a signaling complex that includes PKA and PDE4D. PMID:9000000 residues 1526-1780 of gravin bind the regulatory subunit (RII) of protein kinase A with high affinity, and residues 265-556 bind protein kinase C
GO:0005515 protein binding	IPI PMID:20029029 Regulation of epidermal growth factor receptor trafficking b...	REMOVE	Summary: An unvalidated hit from a membrane-yeast-two-hybrid screen against unliganded EGFR; the paper never discusses AKAP12, and the only other EGFR row comes from the same platform and an overlapping author group. Reason: P00533 is reviewed canonical human EGFR (1210 aa) and carries 2526 IntAct interaction records - a 20-fold hub against AKAP12's 128, so recovery in an abundance-sensitive screen is weak evidence of a specific pairing. The source is a modified membrane yeast two-hybrid (MYTH) screen that returned 87 candidate EGFR interactors; AKAP12 appears nowhere in the cached article text, and no orthogonal experiment on this pair is reported. The second EGFR row on this gene (PMID:31980649) is recorded by IntAct under the same detection method ("ub reconstruction", i.e. split-ubiquitin MYTH) and shares authors (Curak, Stagljar) with this study, so the two rows are not independent replication of the pair. AKAP12's documented receptor scaffolding is for the beta2-adrenergic receptor, a different receptor class; the one paper that examined EGF signalling on this protein concluded that the EGF receptor is probably not the direct kinase of SSeCKS, and that its mitogen-induced tyrosine phosphorylation is FAK-dependent. Removed under the generic-protein-binding policy: the row carries no functional information and no cited experiment licenses a more informative molecular function. Removal does not assert the screen hit is false - it is unreplicated, not refuted. Supporting Evidence: PMID:20029029 We used a modified membrane yeast two-hybrid system together with bioinformatics to identify 87 candidate proteins interacting with the ligand-unoccupied EGFR.
GO:0005515 protein binding	IPI PMID:20195357 A comprehensive resource of interacting protein regions for ...	REMOVE	Summary: Partner P35222 is human CTNNB1. Two orthogonal screens recover the pair, but beta-catenin is a 1138-record IntAct hub, no assay shows direct binding, and the one dedicated study of AKAP-junction interactions assigned beta-catenin to AKAP220 instead. Reason: P35222 resolves to reviewed canonical beta-catenin (781 aa, Swiss-Prot), so the identifier is sound. IntAct records the pair from two publications under two different detection methods - "display technology" (this mRNA-display resource, core data set reported at 70% verification) and "anti tag coip" (PMID:35271311, OpenCell). I first read that as enough to name the partner and proposed MODIFY to GO:0008013 beta-catenin binding; measurement withdrew it. Beta-catenin carries 1138 IntAct interaction records against AKAP12's 128, so it is a hub whose recovery in two abundance-sensitive screens is weak evidence of a specific pairing; neither method demonstrates direct binding, which is what GO:0008013 asserts; no study maps an interface or tests a consequence; and the only paper that examined AKAP binding at endothelial junctions attributed VE-cadherin and beta-catenin binding to AKAP220, not AKAP12 (PMID:25188285) - a counter-indication I had recorded and then over-ridden. Contrast PRKAR2A, whose 231 records, direct filter-binding assay and region mapping on the human protein do support the specific child. Removed under the generic-protein-binding policy rather than marked over-annotated, since the defect is absent functional information rather than a claim exceeding the evidence. The partner is real and the screens are not refuted; removal does not assert otherwise. Supporting Evidence: PMID:25188285 Furthermore, immunoprecipitation analysis demonstrated that AKAP220 interacts not only with PKA but also with VE-cadherin and ß-catenin.
GO:0005515 protein binding	IPI PMID:21900206 A directed protein interaction network for investigating int...	REMOVE	Summary: FHL1 from an automated yeast-two-hybrid interaction-mating network; replicated by two further studies but with no informative molecular-function term licensed by the evidence. Reason: Q13642 is reviewed canonical FHL1 (323 aa, Swiss-Prot), a four-and-a-half LIM domain adaptor of muscle and focal adhesions with only 97 IntAct records over 78 partners - not a hub, so its recovery is not explained by abundance. IntAct records the AKAP12-FHL1 pair from three distinct publications and two method classes - Y2H here, "2h fragment pooling" (PMID:23414517) and endogenous-tag co-IP (PMID:35271311) - so this is a replicated screen interaction rather than a singleton. No study shows which surface of FHL1 is engaged, so GO:0030274 LIM domain binding would assert more than was measured, and the pair is consistent with AKAP12's independently established focal-adhesion localisation without being a defining activity. Removed under the generic-protein-binding policy, no more informative term being licensed; removal does not assert the interaction is false.
GO:0005515 protein binding	IPI PMID:23414517 A human skeletal muscle interactome centered on proteins inv...	REMOVE	Summary: Second, independent yeast-two-hybrid recovery of FHL1, from a human skeletal-muscle library screen centred on limb-girdle muscular dystrophy proteins. Reason: Same partner as the PMID:21900206 row, from a different library and a different laboratory (Y2H with fragment pooling against a human skeletal-muscle cDNA library, 1492 binary interactions reported). Independent recovery of the same pair makes the interaction credible, but no cited experiment identifies a binding determinant or licenses a more informative molecular function. AKAP12 is not discussed individually in the paper, so the interaction remains screen-level. Removed under the generic-protein-binding policy; removal does not assert the interaction is false.
GO:0005515 protein binding	IPI PMID:31980649 Extensive rewiring of the EGFR network in colorectal cancer ...	REMOVE	Summary: The second EGFR row, from the same split-ubiquitin membrane-Y2H platform and an overlapping author group as PMID:20029029; not an independent replication. Reason: IntAct records this pair under detection method "ub reconstruction" (split-ubiquitin MYTH), the same platform used for the PMID:20029029 EGFR row, and the two author lists overlap (Curak J, Stagljar I). Counting these as two independent supports for an AKAP12-EGFR interaction repeats the NbExp error - one assay technology logged twice. AKAP12 is not named anywhere in the cached article text, which is a large systems-level survey of EGFR network rewiring in KRAS(G13D) colorectal cells. Removed under the generic-protein-binding policy rather than marked over-annotated, since the problem is that the row carries no functional information rather than that it overstates the evidence. The bait is a real human protein and the hit is not contradicted, only unconfirmed.
GO:0005515 protein binding	IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce...	MODIFY	Summary: OpenCell endogenous-tag IP-MS recovery of PRKAR2A - the same partner and the same conclusion as the PMID:16642035 row, by a completely different method. Reason: OpenCell tags proteins at their endogenous loci and reads interactions by IP-MS, so this record avoids the over-expression artefact that weakens most large interactome rows; the reference nonetheless carries 2872 GO:0005515 annotations, so any single edge is a high-throughput call. What makes this one usable is convergence: the same partner (reviewed canonical PRKAR2A, P13861) was measured by direct filter binding in PMID:16642035 and mapped to a defined region of the human protein in PMID:9000000. Given an identified PKA regulatory subunit, GO:0034237 is the term the evidence supports. Proposed replacements: protein kinase A regulatory subunit binding Supporting Evidence: PMID:9000000 a PKA-gravin complex can be isolated from human erythroleukemia cells
GO:0005515 protein binding	IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce...	REMOVE	Summary: OpenCell endogenous-tag IP-MS recovery of beta-catenin. Resolved the same way as the PMID:20195357 row, because the two rows describe one partner and must not receive opposite verdicts. Reason: Endogenous tagging is a genuine methodological improvement over the mRNA-display row - no over-expressed constructs - but it does not address what actually weakens this pair: beta-catenin's 1138 IntAct records make it a hub that co-purifies readily, IP-MS cannot distinguish direct binding from shared complex membership, and this single reference supplies 2872 GO:0005515 annotations so no individual edge is validated. Deliberately given the same action as the other CTNNB1 row: splitting a partner across two verdicts would be incoherent, and the AADACL2/3/4 lesson is that one row resolved several ways is a defect even when each call is individually arguable.
GO:0005515 protein binding	IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce...	REMOVE	Summary: Third independent recovery of FHL1, here by endogenous-tag co-immunoprecipitation. Reason: The only non-Y2H support for the AKAP12-FHL1 pair, and what makes the interaction credible. Still screen-level: OpenCell contributes 2872 GO:0005515 annotations from one publication, so the edge is not individually validated, and no study addresses what the association does. Removed under the generic-protein-binding policy: the evidence identifies the partner but not the binding determinant, so no more informative term is licensed. Removal does not assert the interaction is false.
GO:0005516 calmodulin binding	IEA GO_REF:0000002	ACCEPT	Summary: A motif-level InterPro signature whose defining feature is the calmodulin-binding element itself, present in three copies in the human protein and mapped biochemically in the rodent orthologue. Reason: IPR001573 is "A kinase-anchoring protein AKAP5 and AKAP12, calmodulin (CaM)-binding motif" - the WSK/WXSXK motif - i.e. the signature is the binding element, not a bare fold, which is the case in which a signature-derived molecular function is best founded. UniProt annotates three instances in human AKAP12 (residues 607-627, 756-776 and 801-821; PROSITE-ProRule PRU01241), and PMID:11820772 mapped four 1-5-10-motif CaM-binding sites biochemically on the rodent orthologue and showed that PKC prephosphorylation antagonises CaM binding - which is the phosphorylation gate that makes this a real regulatory interaction rather than a sequence coincidence. UniProt also carries the Calmodulin-binding keyword. Supporting Evidence: PMID:11820772 Four CaM binding sites were identified by binding assays that conform to the so-called 1-5-10 motif. PMID:11820772 Notably, CaM binding was antagonized by prephosphorylation of SSeCKS by PKC.
GO:0005737 cytoplasm	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: A phylogenetic inference whose non-self donor carries its own experimental localisation to the same term; sound, but too general to be one of this gene's informative locations. Reason: Resolving the WITH/FROM set: PANTHER:PTN002755149 is a tree node, not a protein; UniProtKB:Q02952 is AKAP12 itself, so that token is self-referential and records a PAINT curator judging the term applicable rather than a circular inference; and ZFIN:ZDB-GENE-030131-9753 is zebrafish akap12b, a teleost AKAP12 orthologue whose only UniProt entries (F2Z4T4, A0A0R4IC31) are unreviewed TrEMBL. Querying that donor's own evidence shows it holds GO:0005737 by two independent IDA annotations (PMID:17575056, PMID:18725198), so this is not a family-level guess propagating without measurement. The term stays general because that is where the donors agree, and it is kept as non-core because the human protein's informative compartments - plasma membrane, cell cortex, cytosol - are annotated separately and are where its anchoring function is exercised. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: PANTHER:PTN002755149 · PANTHER tree node (AKAP12 orthogroup) SUPPORTS TRANSFER Internal PAINT node, not a protein; PTHR23209 contains only AKAP12 orthologues (1185 proteins, a single subfamily). UniProtKB:Q02952 · AKAP12 (human) - the annotation target itself SUPPORTS TRANSFER Self-referential WITH/FROM token, which records a PAINT curator's judgement that the term applies at this node rather than a circular inference. ZFIN:ZDB-GENE-030131-9753 · akap12b (Danio rerio); UniProt entries F2Z4T4 / A0A0R4IC31 are both unreviewed (TrEMBL) SUPPORTS TRANSFER QuickGO shows this donor carries GO:0005737 by its own IDA twice (PMID:17575056, PMID:18725198), so the propagated term rests on measurement.
GO:0005737 cytoplasm	TAS PMID:10858453 Gravin-mediated formation of signaling complexes in beta 2-a...	KEEP AS NON CORE	Summary: Author-statement localisation from the beta2AR desensitisation paper; correct but strictly less informative than the cytosol IDA and the cell-cortex and plasma-membrane rows on the same gene. Reason: TAS on a primary research article is a weak code, and "cytoplasm" adds nothing that the GO:0005829 cytosol IDA (Human Protein Atlas, a main location call) and the GO:0005938 cell cortex row do not already state more precisely. It is not wrong - gravin was originally described as a cytoplasmic autoantigen (PMID:1522245) - so it is kept rather than removed, but it is not part of the protein's informative localisation.
GO:0005829 cytosol	IDA GO_REF:0000052	ACCEPT	Summary: Immunofluorescence-based curation; the Human Protein Atlas lists Cytosol as one of only two main subcellular locations for AKAP12. Reason: GO_REF:0000052 is GO annotation from curated immunofluorescence data. The Human Protein Atlas API returns exactly two subcellular main locations for AKAP12, "Plasma membrane" and "Cytosol", with no additional locations - so both GO_REF:0000052 rows on this gene are main calls rather than weak secondary ones, and they agree with the classical description of gravin as a 250 kDa cytoplasmic antigen giving a trabecular staining network (PMID:1522245). A large soluble pool is expected for a myristoylated peripheral scaffold that cycles on and off the membrane.
GO:0005856 cytoskeleton	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: Derived from a UniProt subcellular-location line that UniProt itself hedges; the protein associates with a cortical cytoskeletal matrix rather than being a cytoskeletal component. Reason: The SL-0090 mapping traces to UniProt's "Cytoplasm, cytoskeleton" line, which carries ECO:0000305 (curator inference) and is annotated with the explicit caveat that the protein may be part of the cortical cytoskeleton. The underlying observations are rodent and describe association rather than membership - SSeCKS induction elaborates a detergent-resistant cortical cytoskeletal matrix (PMID:9744295), and in mesangial cells SSeCKS decorates a network that the authors state is distinct from F-actin (PMID:9885289). Kept because the association is real and reproduced in several systems, but not core: GO:0005938 cell cortex, also annotated, is the better-supported description, and no measurement places AKAP12 within a filament system. Supporting Evidence: PMID:9744295 The induction of SSeCKS (removal of tet) caused significant cell flattening and the elaboration of an SSeCKS-associated cortical cytoskeletal matrix resistant to Triton X-100 extraction. PMID:9885289 enrichment of SSeCKS in mesangial cell podosomes and along a cytoskeletal network distinct from F-actin
GO:0005886 plasma membrane	IDA GO_REF:0000052	ACCEPT	Summary: A Human Protein Atlas main location that agrees with the experimentally established N-myristoylation and with the subplasmalemmal signalling module the protein organises. Reason: This is one of the two HPA main location calls for AKAP12. It is corroborated by independent lines: UniProt records N-myristoyl glycine at position 2 with ECO:0000269 evidence from three proteomic studies, so the membrane anchor is experimentally demonstrated on the human protein rather than predicted; PMID:16642035 identifies gravin as the organiser of a subplasmalemmal PDE4 module among membrane-associated AKAPs; and PMID:9000000 found human gravin concentrated at the cell periphery. This is the compartment in which the gene's core anchoring function is exercised. Supporting Evidence: PMID:9000000 gravin is concentrated at the cell periphery and is enriched in filopodia
GO:0005925 focal adhesion	HDA PMID:21423176 Analysis of the myosin-II-responsive focal adhesion proteome...	KEEP AS NON CORE	Summary: A mass-spectrometry adhesion proteome that assigned focal adhesion to 285 proteins, including cytoplasmic ribosomal subunits; plausible for AKAP12 on independent grounds but not individually validated. Reason: Querying QuickGO by reference shows PMID:21423176 supplies GO:0005925 to 285 annotations, all HDA and all assigned by UniProt. The set includes RPLP0 (P05388), RPLP1 (P05386) and RPLP2 (P05387) - three large-ribosomal-subunit P proteins - which is the signature of abundant co-purifying material in an isolated-adhesion preparation rather than per-protein validation. The call is nonetheless retained because AKAP12 has independent adhesion-related cell biology: its induction reorganises vinculin-associated adhesion plaques and increases integrin-independent FAK tyrosine phosphorylation (PMID:9744295), and ablating it repositions vinculin-associated focal complexes (PMID:9885289). Non-core because the localisation is dynamic and secondary to the cortical and plasma-membrane pool.
GO:0005938 cell cortex	IEA GO_REF:0000044	ACCEPT	Summary: The best-supported of the three UniProt subcellular-location transfers; the cortical pool is where the protein's anchoring function is exercised. Reason: SL-0138 maps UniProt's "Cytoplasm, cell cortex" line. Although that line carries ECO:0000305, the underlying evidence is broad and includes the human protein: gravin is concentrated at the cell periphery and enriched in filopodia (PMID:9000000, human erythroleukemia cells), SSeCKS stains the cell edge and a cortical cytoskeletal matrix (PMID:12083796), and the induced cortical matrix is detergent-resistant (PMID:9744295). GO:0005938 is defined as the region just beneath the plasma membrane, which is exactly where a myristoylated AKAP anchoring a subplasmalemmal cAMP module acts. Supporting Evidence: PMID:12083796 in FAK-/- cells, SSeCKS costains along F-actin stress fibers, in contrast to FAK+/+ cells, where most SSeCKS stains at the cell edge and along a cortical cytoskeletal matrix
GO:0007165 signal transduction	IBA GO_REF:0000033	ACCEPT	Summary: A very general term, but the granularity check comes back negative - the donors genuinely disagree on which signalling pathway, so the parent is the correct least-common-ancestor call. Reason: The reflex on a root-adjacent BP term is GRANULARITY_MISMATCH, and that requires the donors to agree. They do not. The zebrafish orthologue under this node carries its own IMP annotations for cytoskeletal and Rho-axis processes (GO:0035024 negative regulation of Rho protein signal transduction, GO:0032060 bleb assembly, GO:0060028 convergent extension involved in axis elongation), whereas the human protein's measured signalling is the cAMP/PKA and PKC axis already captured by GO:0141156 and GO:0090036. Pushing the node term down to either branch would arbitrarily pick one donor over the other, so the general parent is the honest LCA. Recorded as a negative result of a check that was run, not skipped. The reciprocal half is a coverage gap and is raised in suggested_questions rather than acted on here. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: PANTHER:PTN002755149 · PANTHER tree node (AKAP12 orthogroup) SUPPORTS TRANSFER Tree node rather than a protein. PTHR23209 is a single-gene orthogroup integrated as IPR028540, so every member is an AKAP12 orthologue. UniProtKB:Q02952 · AKAP12 (human) - the annotation target itself SUPPORTS TRANSFER Self-referential token recording a PAINT curator judging the term applicable at the node, not a circular inference.
GO:0007186 G protein-coupled receptor signaling pathway	TAS PMID:10858453 Gravin-mediated formation of signaling complexes in beta 2-a...	KEEP AS NON CORE	Summary: Correct - gravin scaffolds the beta2-adrenergic receptor complex - but recorded as an author statement at a granularity well above what the paper actually measured. Reason: The cited work shows gravin association with the beta2-adrenergic receptor increasing on agonist stimulation, and antisense suppression of gravin disrupting receptor recruitment of GRK2, beta-arrestin and clathrin and blocking recovery from desensitisation. That supports participation in GPCR signalling, so the term is retained. It is non-core because the specific, measurable contribution - restoring an internalised receptor to the surface - is better captured by the GO:0001921 row proposed below, and because TAS understates evidence that the same paper reports experimentally. Supporting Evidence: PMID:10858453 Suppression of gravin expression via antisense oligodeoxynucleotides disrupted agonist-induced association of the receptor with G-protein-linked receptor kinase-2, beta-arrestin, and clathrin as well as receptor recovery from desensitization.
GO:0007193 adenylate cyclase-inhibiting G protein-coupled receptor signaling pathway	IMP PMID:17081159 Role of protein kinase C-dependent A-kinase anchoring protei...	MODIFY	Summary: The siRNA experiment is real and human, but it shows gravin coupling adenylate cyclase to inhibitory signals in young cells and to stimulatory signals in senescent cells; the annotation keeps only the inhibitory half. Reason: Gravin knockdown in human diploid fibroblasts greatly reduced basal cAMP and reversed the cAMP response to lysophosphatidic acid, so involvement in an adenylate-cyclase pathway is well supported and the IMP code is species-correct. What the paper does not support is the directional restriction: its own conclusion is explicitly bidirectional and age-dependent, coupling the AC system to inhibitory Gi-alpha signals in young cells and to stimulatory PKC signals in senescent cells, and the knockdown itself lowered basal cAMP, which is the behaviour of a positive rather than an inhibitory component. GO:0007188 adenylate cyclase-modulating G protein-coupled receptor signaling pathway is the parent that covers activation or inhibition and is the term the data actually license. This is a specificity correction, not a challenge to the curator's reading of the experiment. Proposed replacements: adenylate cyclase-modulating G protein-coupled receptor signaling pathway Supporting Evidence: PMID:17081159 When the expression of Gravin and AKAP79 was blocked by small interference RNA transfection, the basal level of cAMP was greatly reduced and the cAMP status after LPA treatment was also reversed. PMID:17081159 may play important roles in maintaining the basal AC activity and in coupling the AC systems to inhibitory signals such as Gialpha in young cells, and to stimulatory signals such as PKCs in senescent cells
GO:0008179 adenylate cyclase binding	IPI PMID:17081159 Role of protein kinase C-dependent A-kinase anchoring protei...	ACCEPT	Summary: Co-immunoprecipitation of human gravin with adenylate cyclase 6 in human diploid fibroblasts, with AC2 as an internal negative control. Reason: O43306 is reviewed canonical human ADCY6 (1168 aa). The experiment is in human cells, so IPI is the species-correct code, and it is discriminating rather than a blanket pull-down: gravin co-immunoprecipitated AC4 and AC6 but explicitly not AC2, which is an internal specificity control of the kind most interactome rows lack. Anchoring the cyclase alongside PKA and PDE4D is what makes AKAP12 a cAMP microdomain organiser rather than merely a PKA tether. Supporting Evidence: PMID:17081159 Co-immunoprecipitation experiments revealed that Gravin and AKAP79 do not associate with adenylyl cyclase type 2 (AC2) but bind to AC4/6
GO:0008179 adenylate cyclase binding	IPI PMID:17081159 Role of protein kinase C-dependent A-kinase anchoring protei...	ACCEPT	Summary: The AC4 half of the same human co-immunoprecipitation experiment. Reason: Q8NFM4 is reviewed canonical human ADCY4 (1077 aa). Same assay, same human cells and the same AC2 negative control as the ADCY6 row; GOA correctly split the two partners into separate rows. Isoform-level cyclase selectivity is a substantive fact about how AKAP12 builds its microdomain, and it is worth keeping both accessions rather than collapsing them.
GO:0016020 membrane	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: The lipid-anchor claim behind this mapping is experimentally established, but the term is the uninformative parent of the plasma membrane IDA already on this gene. Reason: SL-0162 maps UniProt's "Membrane; Lipid-anchor" line. The anchor itself is not merely inferred: UniProt records N-myristoyl glycine at residue 2 with ECO:0000269 evidence from three independent experimental studies, so the protein really is lipid-anchored. The term is nonetheless the general parent of GO:0005886 plasma membrane, which this gene already carries by IDA from immunofluorescence curation and which is the compartment every functional study places it in. Kept as correct-but-redundant rather than removed.
GO:0032496 response to lipopolysaccharide	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Orthology transfer from rat Akap12, whose expression is LPS-responsive in two independent systems; a canonical use of IEP for a "response to" term, but rodent and peripheral to the gene's core function. Reason: Q5QD51 is reviewed canonical rat Akap12 (1687 aa) and a true 1:1 orthologue - it sits in the same single-gene PANTHER orthogroup PTHR23209:SF4 as human AKAP12, so this is an ortholog transfer, not a paralog transfer. Querying the donor's own evidence, the rat row is IEP from PMID:23912647 (LPS-induced acute lung injury), and the same direction is seen independently in rat astrocytes where LPS induces SSeCKS translocation and siRNA against it blunts downstream signalling (PMID:17873284). "Response to" terms are exactly what IEP is for, so the annotation is properly formed. Non-core: this is an inflammatory-context expression response in rodent tissue, not an activity of the human protein, and no human experiment tests it. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER True 1:1 orthologue (PTHR23209:SF4, same subfamily as human AKAP12). QuickGO shows the donor holds GO:0032496 by IEP from PMID:23912647. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:17873284 By using siRNA knockdown of SSeCKS expression, LPS-induced signaling events were partly inhibited
GO:0032760 positive regulation of tumor necrosis factor production	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: The best-supported of the rodent inflammatory transfers - two independent rat siRNA experiments, in the same cell type, both reducing TNF-alpha output. Reason: The rat donor holds this term by IMP from two separate papers (PMID:17873284, PMID:19937403), both knocking SSeCKS down in rat astrocytes and both reducing cytokine production; a third study reproduces it in the EAE model (PMID:20155814). That is a genuinely replicated perturbation result rather than an expression correlation, which is why this row is kept while GO:0035733 and GO:0061870 from the same transfer batch are not. Non-core rather than core: the effect is astrocyte-specific rodent immunobiology with no human counterpart, and it is downstream of the PKC signalling that GO:0090036 already captures. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER QuickGO shows the donor carries two independent IMP annotations to GO:0032760 (PMID:17873284, PMID:19937403). ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:19937403 SSeCKS gene silencing by siRNA inhibited the up-regulation of p-ERK and the pro-inflammation cytokines
GO:0035733 hepatic stellate cell activation	IEA GO_REF:0000107	MARK AS OVER ANNOTATED	Summary: The rat source measured only that Akap12 expression rises while stellate cells activate; nothing was perturbed, and the authors' own conclusion is explicitly speculative. Reason: The donor annotation is IEP from PMID:23925424, whose entire content is expression measurement - SSeCKS mRNA in healthy versus CCl4-injured rat liver and protein in hepatic stellate cells activated on plastic. "Involved in hepatic stellate cell activation" asserts participation in the process; an expression rise during the process is correlation. The paper's own closing sentence hedges the functional claim twice, as a "possibility" reached "perhaps". The discriminator is internal to the same reference: the curators also drew GO:0061870 from it as IMP, so they read one row as measured and one as correlative - yet no perturbation is described in the abstract for either. Marked over-annotated rather than removed because the human row is an electronic transfer whose donor is a real experimental annotation I could not read in full; a curator with the full text should confirm whether any loss-of-function experiment exists. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: SOURCE EVIDENCE WEAK CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SOURCE WEAK OR INFERRED A true orthologue, but its GO:0035733 annotation is IEP from PMID:23925424, an expression-only study in rat liver with no loss-of-function experiment reported in the abstract. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SOURCE WEAK OR INFERRED Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:23925424 The upregulation of SSeCKS protein expression in rat HSCs during activation in vitro and in vivo suggested the possibility of SSeCKS, an important part of function of the activated HSCs, perhaps through modulation of liver regeneration or formation of liver fibrosis after various injuries.
GO:0043025 neuronal cell body	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: Backed by a real rat immunofluorescence IDA in which labelled sensory-neuron cell bodies were counted; a restricted, tissue-specific location rather than a general one. Reason: The rat donor holds GO:0043025 by IDA from PMID:11814414, an immunolabelling survey of rat brain, spinal cord and dorsal root ganglia in which SSeCKS-labelled cell bodies were scored against substance P, CGRP and fluoride-resistant acid phosphatase. That is a direct localisation measurement, so the transfer is sound. Non-core: the paper's own conclusion is that expression is restricted to a distinct nociceptive sensory-neuron population, so this is a cell-type-specific rodent location, not a general property of the human protein. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER QuickGO shows the donor carries GO:0043025 by IDA from PMID:11814414. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:11814414 A small portion of the SSeCKS-labeled cell bodies appeared to represent a subpopulation of substance P (4.8%) and CGRP (4.7%) containing neurons
GO:0043116 negative regulation of vascular permeability	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: The term is right for the human protein, but it is right for reasons the transfer never used - the rat donor annotation is sign-inverted relative to its own cited paper, which reports SSeCKS promoting, not restraining, endothelial permeability. Reason: Two separate findings have to be kept apart here. First, the propagation is defective: the rat donor holds GO:0043116 by IMP from PMID:20454828, and that paper reports that depleting SSeCKS attenuated the cytokine-induced fall in transendothelial electrical resistance and rise in dextran flux - i.e. removing the protein protected the barrier - and concludes in terms that SSeCKS is involved in endothelial hyperpermeability. On its own evidence that source annotation should be GO:0043117 positive regulation of vascular permeability. Second, the term nevertheless holds for AKAP12, on evidence the transfer did not use: mouse-specific siRNA against Akap12 in MyEnd endothelial cells impairs barrier function and abolishes cAMP-mediated barrier stabilisation (PMID:25188285). The row is therefore kept, with the provenance defect recorded rather than silently inherited. The two results are also reconcilable - AKAP12 appears to be barrier-protective basally and permeability-promoting under inflammatory cytokine stimulation - which makes GO:0043117 a plausible additional annotation rather than a replacement. Propagation Review Root cause: SOURCE BAD Failure modes: REGULATORY SIGN INVERSION Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SOURCE BAD A true orthologue, but its GO:0043116 IMP from PMID:20454828 has the opposite sign to that paper's stated result and conclusion; GO:0043117 is what the experiment supports. Worth reporting to RGD. Verified against the abstract, not the full text. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SOURCE BAD Pipeline identifier accompanying the UniProt accession; it inherits the donor's sign error rather than introducing one. Supporting Evidence: PMID:20454828 Depletion of endogenous SSeCKS in RPMVEC significantly attenuated cytokine-induced decrease in TER and increase in P (d), but not to the basal levels. PMID:20454828 SSeCKS is involved in the endothelial hyperpermeability induced by IL-1β and TNF-α in inflammatory process. PMID:25188285 siRNA-mediated AKAP12 and AKAP220 knockdown significantly impaired endothelial barrier function.
GO:0051018 protein kinase A binding	IEA GO_REF:0000002	ACCEPT	Summary: An orthogroup-specific InterPro family signature, not a bare fold; the mapped activity is the one directly measured on the human protein. Reason: IPR028540 is integrated with PANTHER PTHR23209, which is a single-gene orthogroup - 1185 proteins, one subfamily, and the only three reviewed members are human, rat and mouse AKAP12. A GO mapping from a signature of that specificity is about as well-founded as an IEA gets, and the term it supplies is exactly what direct experiment shows: human gravin residues 1526-1780 bind PKA RII with high affinity and a PKA-gravin complex is isolable from human cells (PMID:9000000), and UniProt annotates a corresponding RII-binding region. Accepted as core. Supporting Evidence: PMID:9000000 We have identified and characterized another mammalian scaffold protein which coordinates the location of protein kinase A and protein kinase C
GO:0051018 protein kinase A binding	TAS PMID:10858453 Gravin-mediated formation of signaling complexes in beta 2-a...	ACCEPT	Summary: Correct, and at the right granularity for this particular reference, which co-purified whole PKA holoenzyme with the receptor-gravin complex rather than resolving a subunit. Reason: The cited experiment shows agonist-induced association of the receptor-gravin complex with protein kinases A and C. Because it detects the kinase in a complex rather than identifying which subunit is contacted, GO:0051018 - whose GO comment notes it spans binding to either the catalytic or the regulatory subunit - is the right term for this row, and it should not be pushed to GO:0034237. The more specific child is instead introduced on the two PRKAR2A rows, where a named regulatory subunit was actually measured. Kept additive rather than corrective. Supporting Evidence: PMID:10858453 Agonist stimulation also induced a robust association of the receptor-gravin complex with protein kinases A and C, G-protein-linked receptor kinase-2, beta-arrestin, and clathrin.
GO:0051602 response to electrical stimulus	IEA GO_REF:0000107	MARK AS OVER ANNOTATED	Summary: The electrical stimulus was applied to a rat's vagus nerve; Akap12 was measured in lung tissue several causal steps downstream. This is an anti-inflammatory reflex readout, not a response of the protein's cell to an electrical stimulus. Reason: The donor annotation is IEP from PMID:23912647, in which rats received lipopolysaccharide to induce acute lung injury and vagus nerve stimulation as treatment; stimulation reduced lung oedema, neutrophil infiltration and cytokine levels and, among other things, lowered SSeCKS protein in lung tissue. The causal chain from electrode to annotation runs nerve stimulation to cholinergic anti-inflammatory reflex to reduced pulmonary inflammation to reduced Akap12 - the pulmonary microvascular endothelial cell never experienced an electrical stimulus. The same datum is already represented, more defensibly, by the GO:0032496 response to lipopolysaccharide row from the identical paper. Over-annotated rather than removed because GO:0051602's definition does admit organism-level responses, so the row is technically satisfiable even though it conveys nothing about AKAP12. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: SOURCE EVIDENCE WEAK ROLE CONFLATION Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SOURCE WEAK OR INFERRED True orthologue, but the GO:0051602 annotation is IEP from a whole-animal vagus-nerve-stimulation experiment in which the stimulus was not delivered to the cells in which Akap12 was measured. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SOURCE WEAK OR INFERRED Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:23912647 vagus nerve stimulation not only decreased the expressions of Src-suppressed C kinase substrate and E-selectin proteins in lung tissue
GO:0061870 positive regulation of hepatic stellate cell migration	IEA GO_REF:0000107	MARK AS OVER ANNOTATED	Summary: Transferred from a rat IMP drawn from a paper that reports no migration assay and no perturbation of Akap12 - only expression measurements in injured liver and in activating stellate cells. Reason: The donor annotation is IMP from PMID:23925424. Everything the abstract describes is measurement of SSeCKS mRNA and protein - in healthy versus CCl4-injured rat liver, and in primary hepatic stellate cells activating on plastic. No knockdown, no over-expression and no migration readout is reported, yet an IMP asserts that a perturbation of this gene changed hepatic stellate cell migration. The stated conclusion is speculative in its own words. Because the human row is an electronic orthology transfer, arguing against it does not overrule an experimental call on the human protein; and because I could only read the source abstract, this is marked over-annotated rather than removed, with the specific question - does the full text contain a migration assay? - passed to a curator. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: SOURCE EVIDENCE WEAK CONTEXT OR TISSUE MISMATCH Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SOURCE WEAK OR INFERRED True orthologue, but the GO:0061870 IMP rests on PMID:23925424, whose abstract reports only expression measurement - no perturbation and no migration assay. Assessed from the abstract; full text not accessible. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SOURCE WEAK OR INFERRED Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:23925424 The upregulation of SSeCKS protein expression in rat HSCs during activation in vitro and in vivo suggested the possibility of SSeCKS, an important part of function of the activated HSCs, perhaps through modulation of liver regeneration or formation of liver fibrosis after various injuries.
GO:0070374 positive regulation of ERK1 and ERK2 cascade	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: A real rat siRNA result, but the sign of AKAP12's effect on ERK is genuinely context-dependent and runs the other way in fibroblasts. Reason: The rat donor holds this term by IMP from PMID:19937403, where silencing SSeCKS in spinal astrocytes blocked the rise in phospho-ERK after stimulation by damaged sensory neurons - a genuine loss-of-function result, so the row is retained. It should not be read as a general property: in rat MAT-LyLu prostate carcinoma cells SSeCKS suppresses serum-induced Raf/MEK/ERK activation, and its suppression of phorbol-ester-induced ERK1/2 activity requires the PKC-binding domain (PMID:20018890). AKAP12 modulates the cascade in a cell-type- and stimulus-dependent direction rather than positively regulating it as a rule, so this is recorded as non-core with the tension stated rather than silently resolved in one direction. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER QuickGO shows the donor carries GO:0070374 by IMP from PMID:19937403 (rat astrocyte siRNA). ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:19937403 SSeCKS gene silencing by siRNA inhibited the up-regulation of p-ERK and the pro-inflammation cytokines PMID:20018890 SSeCKS suppressed serum-induced activation of the Raf/MEK/ERK pathway, resulting in down-regulation of matrix metalloproteinase-2 expression.
GO:0071347 cellular response to interleukin-1	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: IL-1beta applied directly to rat pulmonary microvascular endothelial cells raised Akap12 message and protein - the textbook-correct use of IEP for a cellular-response term. Reason: The donor annotation is IEP from PMID:20454828, in which the cytokine was applied to the same cells in which the expression change was measured. That is the situation GO:0071347 is written for, and it contrasts instructively with the GO:0051602 row from a different paper, where the stimulus reached the annotated cell only through a whole-animal reflex arc. Kept as a non-core rodent expression response: it says the gene is cytokine-inducible in endothelium, not what the human protein does. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER QuickGO shows the donor carries GO:0071347 by IEP from PMID:20454828, where the cytokine was applied directly to the cells assayed. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:20454828 IL-1β and TNF-α activated the PKC signaling pathway in RPMVEC, and up-regulated the gene and protein expression of SSeCKS.
GO:0071356 cellular response to tumor necrosis factor	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: The TNF-alpha half of the same directly-applied-cytokine experiment as the IL-1 row. Reason: Same donor evidence and same reasoning as GO:0071347: TNF-alpha was applied to rat pulmonary microvascular endothelial cells and Akap12 message and protein rose, which is a proper IEP basis for a cellular-response term. It also sits coherently alongside GO:0032760 - AKAP12 is both induced by TNF-alpha and required for its production in rodent inflammatory settings, which is a positive-feedback arrangement rather than two unrelated annotations. Non-core for the same species and context reasons. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER QuickGO shows the donor carries GO:0071356 by IEP from PMID:20454828. ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:20454828 IL-1β and TNF-α activated the PKC signaling pathway in RPMVEC, and up-regulated the gene and protein expression of SSeCKS.
GO:0090036 regulation of protein kinase C signaling	IEA GO_REF:0000002	ACCEPT	Summary: From the same orthogroup-specific signature as the PKA binding row, and independently demonstrated - AKAP12 binding to PKC-alpha lowers its kinase activity. Reason: IPR028540 covers only AKAP12 orthologues, so this is a family-specific rather than a fold-level mapping. The process is directly supported: SSeCKS/gravin/AKAP12 binds PKC-alpha through two defined motifs and that binding decreases kinase activity, knockout fibroblasts show elevated basal and phorbol-ester-induced PKC activity, and the defect is rescued by full-length but not PKC-binding-domain-deleted protein (PMID:21903576). Regulation by sequestration is the mechanism, which is why the regulation term rather than an activity term is the right shape. Core. Supporting Evidence: PMID:21903576 SSeCKS binding to PKCα decreased kinase activity and was dependent on the two PKC-binding motifs.
GO:0141156 cAMP/PKA signal transduction	IEA GO_REF:0000002	ACCEPT	Summary: The orthogroup signature supplies this electronically, but it is one of the few terms on this gene with direct human loss-of-function support. Reason: An IEA that happens to be the gene's best-evidenced process. RNA interference against membrane-associated AKAPs in human HEK293 cells identified gravin as the organiser of the PDE4 module that terminates subplasmalemmal cAMP transients, and co-immunoprecipitation confirmed a gravin complex containing PKA and PDE4D (PMID:16642035); in human diploid fibroblasts gravin knockdown greatly reduced basal cAMP (PMID:17081159). The electronic code understates the evidence - the term itself is correct and core. Reported here rather than proposed as a NEW row because the term is already present in GOA. Supporting Evidence: PMID:16642035 Knockdown of specific membrane-associated AKAPs using RNAi identified gravin (AKAP250) as the central organizer of the PDE4 complex.
GO:1900143 positive regulation of oligodendrocyte apoptotic process	IEA GO_REF:0000107	KEEP AS NON CORE	Summary: A real rat knockdown result, but a non-cell-autonomous one - Akap12 acts in astrocytes and the apoptosis occurs in a different cell type, via secreted mediators, in a disease model. Reason: The donor holds this term by IMP from PMID:20155814: siRNA against SSeCKS in rat astrocytes reduced TNF-alpha and nitric oxide output, and conditioned medium from those astrocytes no longer killed oligodendrocytes. The perturbation is genuine, which is why the row is kept. What makes it firmly non-core is that it is three steps removed from anything AKAP12 does molecularly - astrocyte scaffold to cytokine and nitric oxide release to death of a neighbouring cell type - and it was observed only in experimental autoimmune encephalomyelitis. It is a downstream consequence of the GO:0032760 cytokine phenotype rather than an independent function. Propagation Review Root cause: NO FAILURE NON CORE Sources checked: UniProtKB:Q5QD51 · Akap12 (Rattus norvegicus), reviewed (Swiss-Prot), 1687 aa SUPPORTS TRANSFER QuickGO shows the donor carries GO:1900143 by IMP from PMID:20155814 (astrocyte siRNA, conditioned-medium readout on oligodendrocytes). ensembl:ENSRNOP00000070062 · Ensembl Compara rat protein identifier for the same Akap12 gene SUPPORTS TRANSFER Pipeline identifier accompanying the UniProt accession, not an independent source. Supporting Evidence: PMID:20155814 siRNA targeted to SSeCKS in astrocytes inhibited oligodendrocyte cell death
GO:0005080 protein kinase C binding	IDA PMID:9000000 Gravin, an autoantigen recognized by serum from myasthenia g...	NEW	Summary: Direct in-vitro binding of protein kinase C to a defined region of the human protein, demonstrated in the paper that named gravin a kinase scaffold; GOA carries no PKC binding term at all. Reason: Half of what UniProt states this protein does - compartmentation of protein kinase A and protein kinase C - has no molecular function term in GOA. PKA binding is represented twice; PKC binding is absent. The evidence is on the human protein and direct: residues 265-556 of gravin bind protein kinase C in vitro (PMID:9000000), which is the observation UniProt encodes as its PKC-binding region 266-557. It is independently supported by a separate determinant mapped in the rodent orthologue, where two homologous motifs mediate phosphatidylserine-independent PKC binding (PMID:21903576) - and those published motifs match the human sequence exactly, at residues 605-633 and 754-782 (see the notes file), so the second determinant is present in human as well. IDA rather than IPI because the assay identifies PKC as the ligand without naming a specific isoform accession, and asserting one would claim an interaction that was not measured. Supporting Evidence: PMID:9000000 residues 1526-1780 of gravin bind the regulatory subunit (RII) of protein kinase A with high affinity, and residues 265-556 bind protein kinase C PMID:21903576 phosphatidylserine-independent binding of PKC by SSeCKS is facilitated by two homologous SSeCKS motifs file:genes/human/AKAP12/AKAP12-bioinformatics/RESULTS.md So the human protein has three WSK motifs, of which two also match the PKC-binding consensus, and the third (801-821) does not.
GO:0035591 signaling adaptor activity	IMP PMID:16642035 An anchored PKA and PDE4 complex regulates subplasmalemmal c...	NEW	Summary: The bridging activity itself, measured in human cells: RNAi against gravin disassembles a subplasmalemmal module whose components it co-precipitates. Reason: GOA records the individual partners AKAP12 binds but never the activity of bringing them together, which is the whole point of a scaffold and the only molecular function that is not simply an inventory of its clients. GO:0035591 is defined as bringing together two or more molecules in a signalling pathway so that they act in a coordinated way, with the explicit note that scaffold proteins tether components and localise them to specific areas of the cell - which is what was demonstrated here in human HEK293 cells by RNA interference against membrane-associated AKAPs, with co-immunoprecipitation confirming a gravin complex containing PKA and PDE4D. Choosing the adaptor term rather than importing the clients' activities is deliberate: PKA's kinase activity and PDE4D's hydrolase activity are not AKAP12's, and this term states what AKAP12 contributes without borrowing them. IMP because the demonstration that the module requires gravin is a knockdown. Supporting Evidence: PMID:16642035 Knockdown of specific membrane-associated AKAPs using RNAi identified gravin (AKAP250) as the central organizer of the PDE4 complex. PMID:16642035 Co-immunoprecipitation confirmed that gravin maintains a signaling complex that includes PKA and PDE4D.
GO:0001921 positive regulation of receptor recycling	IMP PMID:19055733 G-protein-coupled receptor-associated A-kinase anchoring pro...	NEW	Summary: AKAP12 knockdown in two human cell lines abolishes beta2-adrenergic receptor resensitisation, and the same study shows the effect is specific to AKAP12 rather than shared with AKAP5. Reason: GOA represents this gene's receptor biology only as GO:0007186 by TAS, which is far above what has been measured. The experiment is clean and human: suppressing AKAP12 in either A431 or HEK293 cells abolishes the ability to resensitise the beta2-adrenergic receptor, while suppressing AKAP5 in the same cells does not - and the reciprocal readout, ERK1/2 activation, depends on AKAP5 and not AKAP12. A double dissociation between two AKAPs in the same cells is much stronger than a single knockdown, because it excludes a generic anchoring artefact. The earlier antisense work reaches the same conclusion (PMID:10858453). IMP with human cell lines, so the experimental code is species-correct. Supporting Evidence: PMID:19055733 Suppression of the expression of AKAP12 in either cell line leads to loss of the ability of these cells to resensitize the beta2-adrenergic receptor. PMID:19055733 A431 cells were found to highly express AKAP12, but little of AKAP5.
GO:0061028 establishment of endothelial barrier	ISS PMID:25188285 PKA compartmentalization via AKAP220 and AKAP12 contributes ...	NEW	Summary: Barrier formation is one of AKAP12's best-replicated biological roles, but the knockdown in the cited study used mouse-specific siRNA in mouse endothelial cells, so the code is ISS and not IMP. Reason: GOA's only barrier-related row is GO:0043116, whose rodent provenance is sign-inverted (see that row). This term states the positive claim directly and rests on evidence that the Compara transfer did not use. The species check matters: although the study analysed human, mouse and rat material, the methods state that down-regulation used mouse-specific siRNA and the knockdown barrier measurements were made in MyEnd mouse endothelial cells, so IMP would assert a human experiment that was not performed. The supporting entity is mouse Akap12 (Q9WTQ5, reviewed, 1684 aa), a 1:1 orthologue in the same PTHR23209:SF4 subfamily. Independent corroboration comes from a separate laboratory and a separate vascular bed: siRNA against AKAP12 in cultured cerebral endothelial cells raises permeability and dysregulates ZO-1 and claudin-5, and Akap12 knockout mice sustain more blood-brain barrier damage after stroke (PMID:33260683) - also a mouse system, so it corroborates the ISS rather than converting it to an IMP. Supporting Evidence: PMID:25188285 siRNA-mediated AKAP12 and AKAP220 knockdown significantly impaired endothelial barrier function. PMID:25188285 Down-regulation of mouse specific AKAP12- and AKAP220 mRNA was obtained by using ON-Target SMARTpool siRNA PMID:33260683 When AKAP12 expression was suppressed by an siRNA approach in cultured endothelial cells, endothelial permeability was increased along with the dysregulation of ZO-1/Claudin 5 expression.

Gene Description

Proposed New Ontology Terms

cyclic-nucleotide phosphodiesterase binding

Existing Annotations Review

Core Functions

Anchors the type II regulatory subunit dimer of protein kinase A through a C-terminal anchoring region, holding the kinase in a defined subplasmalemmal location rather than letting it act throughout the cytoplasm.

Binds protein kinase C through determinants in its central region and thereby damps the kinase's activity, coupling the scaffold to PKC-driven cytoskeletal and proliferative signalling.

Assembles adenylate cyclase, protein kinase A and phosphodiesterase 4D into a single membrane-proximal module, so that cAMP is produced, read and destroyed within one restricted microdomain; this bridging is the activity, not the catalysis of the components it holds.

Binds adenylate cyclase isoforms 4 and 6, but not isoform 2, recruiting the source of cAMP into the same scaffold as its effector and its terminator.

Binds calmodulin through three copies of the AKAP WSK motif, in a manner abolished by prior PKC phosphorylation, which makes the scaffold's calcium-sensing arm switchable by the kinase it also anchors.

Required, non-redundantly with AKAP5, for the internalised beta2-adrenergic receptor to be resensitised and returned to the cell surface.

References

Suggested Questions for Experts

Suggested Experiments

Deep Research

Affinage

📚 Additional Documentation

Notes

Bioinformatics Results

📄 View Raw YAML