ARHGEF18

Existing Annotations Review

GO Term	Evidence	Action	Reason
GO:0005085 guanyl-nucleotide exchange factor activity	IBA GO_REF:0000033	ACCEPT	Summary: Phylogenetic propagation of exchange-factor activity from PANTHER node PTN002677784, seeded by the Drosophila ortholog cyst/Dp114RhoGEF, mouse Arhgef2, human AKAP13, ARHGEF28 and ARHGEF2, and the target itself. This is the core molecular function, and it is independently established by direct biochemistry on the purified human protein. Reason: Every WITH/FROM token was resolved against its own authority. FB:FBgn0032796 is Drosophila cyst/CG10188, also called Dp114RhoGEF; MGI:MGI:103264 is mouse Arhgef2; Q12802 is AKAP13/Lbc; Q8N1W1 is ARHGEF28/p190RhoGEF; Q92974 is ARHGEF2/GEF-H1. The fly gene is a genuine ortholog rather than a distant paralog, which is what makes this a legitimate deep node. Silver et al. call Cysts the Drosophila orthologue of mammalian p114RhoGEF, GEF-H1, p190RhoGEF and AKAP-13, so PTN002677784 sits ancestral to the whole Lbc clade before its vertebrate expansion. One observation is worth registering rather than repairing. Current PANTHER assigns human Q6ZSZ5 and mouse Q6P9R4 to PTHR47440:SF1, while every other seed in this list, the fly ortholog included, now sits in PTHR13944. The family label on the annotation no longer matches the family labels on its seeds, which is a consequence of PANTHER splitting the vertebrate ARHGEF18 lineage into its own family after the IBD was made, not evidence that the seeds are wrong. The propagation is in any case redundant here, because human ARHGEF18 carries an IDA for the same term and Blomquist et al. measured exchange on purified protein. The catalytic Dbl tyrosine is intact at Tyr-606, which is supporting rather than decisive; see the residue claim and the knowledge gap on substrate identity. Propagation Review Root cause: NO FAILURE CORE Sources checked: PANTHER:PTN002677784 · PTHR47440 ancestral node, Lbc/RH-RhoGEF clade SUPPORTS TRANSFER Node predates the vertebrate expansion of the Lbc clade; current PANTHER places the seeds across two families, PTHR47440 and PTHR13944. FB:FBgn0032796 · Drosophila cyst / CG10188 / Dp114RhoGEF SUPPORTS TRANSFER The single fly ortholog of ARHGEF18, GEF-H1, ARHGEF28 and AKAP13. Independently annotated GO:0005085 by IDA from PMID:36917931. UniProt entries Q9VIV0, M9PBE1 and Q8MQR4, all PANTHER PTHR13944:SF21. MGI:MGI:103264 · mouse Arhgef2 (GEF-H1) SUPPORTS TRANSFER Resolves to UniProtKB:Q60875, ARHG2_MOUSE, PANTHER PTHR13944:SF20. UniProtKB:Q12802 · AKAP13 (Lbc) SUPPORTS TRANSFER PANTHER PTHR13944:SF18. UniProtKB:Q8N1W1 · ARHGEF28 (p190RhoGEF) SUPPORTS TRANSFER PANTHER PTHR13944:SF22. UniProtKB:Q92974 · ARHGEF2 (GEF-H1) SUPPORTS TRANSFER PANTHER PTHR13944:SF20. UniProtKB:Q6ZSZ5 · human ARHGEF18, the target itself SUPPORTS TRANSFER The target's own accession, which is expected rather than circular. ARHGEF18 carries GO:0005085 by IDA from PMID:14512443, and that annotation is one of the descendant evidences the PAINT curator used to place the IBD, so the gene legitimately appears among the sources of the IBA it later receives. Its presence marks experimental grounding on the target; what the IBA then adds is that the activity is inherited from the ancestral Lbc clade rather than being lineage-specific. Supporting Evidence: PMID:11085924 p114-Rho-GEF interacted specifically with RhoA, in its nucleotide-free and guanosine 5'-[gamma-thio]triphosphate-bound states, but not with Rac1 and Cdc42, and efficiently catalysed guanine nucleotide exchange of RhoA PMID:31409654 To explore the regulation of Rho GTPases in vivo, we analyzed the Rho GTPase guanine nucleotide exchange factor (RhoGEF) Cysts, the Drosophila orthologue of mammalian p114RhoGEF, GEF-H1, p190RhoGEF, and AKAP-13 file:human/ARHGEF18/ARHGEF18-bioinformatics/RESULTS.md Bottom line: Tyr-606 is conserved in all 14 panel members including the Rac- and Cdc42-specific GEFs, so retaining it says the DH domain is of the catalytically competent type and says nothing whatever about which GTPase ARHGEF18 acts on.
GO:0005085 guanyl-nucleotide exchange factor activity	IDA PMID:14512443 G Protein betagamma subunits stimulate p114RhoGEF, a guanine...	ACCEPT	Summary: Direct assay of exchange-factor activity. Niu et al. used pull-downs and dominant-negative GTPases in cells and found activation of RhoA and Rac1 but not Cdc42; Blomquist et al. had already shown exchange on purified RhoA. This is the gene's core molecular function and its only curated molecular-function evidence. Reason: The term is correct and is the most specific one GO offers. It is worth being explicit that this is not a granularity choice. GO:0005089 Rho guanyl-nucleotide exchange factor activity was merged into GO:0005085 (obsolete, term replaced by GO:0005085, obsolescence reason IAO:0000227), as were the Rac, Ras, ARF, Rab and Ran equivalents, and GO:0005085 now has no is_a children. Checked in QuickGO, OLS4 and the GO API, because QuickGO silently resolves merges and returns the replacement record as non-obsolete. Substrate identity is therefore carried in core_functions.substrates and recorded as an ontology knowledge gap rather than proposed as a new term. Two caveats on the evidence itself, both left in place rather than acted on. First, this paper's Rac1 result is contradicted by the purified-protein experiment in PMID:11085924 and by the in-cell depletion in PMID:21258369; UniProt carries both readings and so does this review. Second, Reactome places ARHGEF18 in a generic GEFs catalyst set for a p75NTR reaction whose stated outputs are RAC1 and Cdc42, which is a set-level assignment rather than a measurement on this protein. Supporting Evidence: PMID:14512443 we have determined that p114RhoGEF activated RhoA and Rac1 but not Cdc42 proteins PMID:11085924 p114-Rho-GEF interacted specifically with RhoA, in its nucleotide-free and guanosine 5'-[gamma-thio]triphosphate-bound states, but not with Rac1 and Cdc42, and efficiently catalysed guanine nucleotide exchange of RhoA PMID:21258369 p114RhoGEF is a widely expressed gene and, in vitro, functions as a specific activator of RhoA17-19. Accordingly, depletion of p114RhoGEF resulted in reduced levels of active RhoA in HCE cells without significantly affecting Rac and Cdc42 file:human/ARHGEF18/ARHGEF18-deep-research-affinage.md Through its DH/PH catalytic module it activates RhoA (and Rac1, but not Cdc42) Knowledge gap: Whether ARHGEF18 is a RhoA-only exchange factor or acts on Rac1 as well is unresolved, and the answer is not expressible in GO either way. OPEN BIOLOGYONTOLOGY MF_DARK Resolve: Biologically, real-time exchange kinetics on the human DH-PH module against a RhoA, Rac1, Cdc42, RhoB and RhoC panel would settle it. The one systematic Dbl-family panel of this kind does not include p114RhoGEF. Ontologically nothing can be proposed, since GO:0005089, GO:0030676, GO:0005088, GO:0005086, GO:0017112 and GO:0005087 were all merged into GO:0005085, which now has zero is_a children in both OLS4 and QuickGO, exactly as the RhoGAP terms were merged into GO:0005096.
GO:0005085 guanyl-nucleotide exchange factor activity	IEA GO_REF:0000120	ACCEPT	Summary: Electronic assignment from the DH domain signature (IPR000219) and the ARHGEF18-specific PH signature (IPR037744). Correct and fully redundant with the IDA and IBA rows for the same term. Reason: The domain assignment it rests on is real and matches UniProt's own feature table, which places the DH domain at 447-644 and the PH domain at 684-786. The IPR037744 signature is ARHGEF18-specific rather than generic, so this is not a family-wide over-call. It adds nothing the IDA does not already carry.
GO:0005515 protein binding	IPI PMID:25753039 ZO-1 controls endothelial adherens junctions, cell-cell tens...	KEEP AS NON CORE	Summary: Co-immunoprecipitation with CGNL1 (paracingulin, called JACOP in the paper) from human dermal microvascular endothelial cell extracts. The interaction is real and mechanistically important, but protein binding as a term says nothing about what it does. Reason: Q0VF96 resolves to CGNL1, cingulin-like protein 1. The finding this row stands for is that paracingulin, downstream of ZO-1, is what puts ARHGEF18 at endothelial junctions, which is the recruitment arm of the protein's core function rather than an accessory contact. Kept rather than removed because the interaction is directly observed; non-core because the informative content is a localisation dependency and GO:0005515 cannot express it. The same paper's two IMP rows do carry that content. A more informative molecular-function term is not available, since GO:0017048 Rho GTPase binding, the nearest analogue for the partner class, was itself merged into GO:0031267 small GTPase binding, and that would in any case be the wrong partner here. Supporting Evidence: PMID:25753039 p114RhoGEF and JACOP coimmunoprecipitated from endothelial cell extracts, and both ZO-1 and JACOP were required for junctional recruitment of p114RhoGEF
GO:0005515 protein binding	IPI PMID:35271311 OpenCell: Endogenous tagging for the cartography of human ce...	KEEP AS NON CORE	Summary: Interaction with AKAP13 detected by endogenous tagging in the OpenCell high-throughput interactome, curated by IntAct. Q12802 is AKAP13/Lbc, a paralogous RhoGEF of the same Lbc clade. Reason: A single high-throughput pull-down between two members of the same RhoGEF clade, with no follow-up anywhere in the literature and no functional consequence reported. Kept because the dataset is a curated interactome rather than a prediction and the accession resolves correctly; non-core because nothing is known about what the contact does, and because paralogs sharing a compartment are a well-known source of proximity-tagging signal. Worth noting for the future that AKAP13 is also one of the PAINT seeds for this gene's exchange-factor IBD, so the two proteins are linked both phylogenetically and physically without either link being explained.
GO:0005737 cytoplasm	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: Mapped from UniProt's Cytoplasm subcellular-location statement, which is attributed to PMID:15558029, where endogenous SA-RhoGEF colocalises with Sept9b along actin stress fibres in REF52 fibroblasts. Reason: True but uninformative. A cytoplasmic pool is expected for a protein that cycles onto junctions, and Terry et al. describe exactly that, an antibody staining both the junctional complex and a cytoplasmic pool. The functionally meaningful locations are the apical plasma membrane and the junctional complex.
GO:0005829 cytosol	IDA GO_REF:0000052	KEEP AS NON CORE	Summary: Human Protein Atlas immunofluorescence. Consistent with the diffuse cytoplasmic pool seen by antibody staining in epithelial cells. Reason: An antibody-based localisation in cultured cells that are not necessarily forming mature junctions will report the soluble pool, which is where the protein waits rather than where it acts. Kept because it is a direct observation; non-core for the same reason the cytoplasm row is non-core.
GO:0005829 cytosol	TAS Reactome:R-HSA-205039	KEEP AS NON CORE	Summary: One of two Reactome cytosol rows. This one comes from the reaction p75NTR indirectly activates RAC and Cdc42 via a guanyl-nucleotide exchange factor, in which ARHGEF18 is a member of a generic GEFs catalyst set rather than a named participant. Reason: Querying the Reactome content service confirms both the participation and its nature. ARHGEF18 is returned among the reaction's reference entities, and the catalyst is guanyl-nucleotide exchange factor activity of GEFs, a set. The reaction's declared input and output are RAC1 bound to GDP and RAC1 bound to GTP, and its display name additionally invokes Cdc42, a GTPase that both primary papers agree ARHGEF18 does not act on. The compartment assignment is therefore inherited from set membership, not measured on this protein. Kept rather than removed because a cytosolic pool is independently real, and because the Rac1 half of the reaction is not contradicted by PMID:14512443. Marked non-core, and the Cdc42 over-reach is recorded here rather than acted on, since the annotation under review is a localisation. Propagation Review Root cause: SOURCE WEAK OR INFERRED Failure modes: ROLE CONFLATION Sources checked: Reactome:R-HSA-205039 · p75NTR indirectly activates RAC and Cdc42 via a guanyl-nucleotide exchange factor SOURCE WEAK OR INFERRED ARHGEF18 enters this reaction as an unnamed member of a generic GEFs catalyst set; the compartment is a participant-compartment assignment, and the reaction name invokes Cdc42, which ARHGEF18 does not act on.
GO:0005829 cytosol	TAS Reactome:R-HSA-419166	KEEP AS NON CORE	Summary: The second Reactome cytosol row, from the reaction GEFs activate RhoA,B,C. ARHGEF18 is again a member of a generic GEFs catalyst set, but here the reaction substrate matches the protein's established activity. Reason: Same mechanism as the other Reactome cytosol row, and the same verdict, but this reaction is the biologically apt one. Its input and output are RhoA, RhoB and RhoC bound to GDP and to GTP, which is what this protein is documented to do. The compartment is still a set-level participant assignment rather than an observation, so it stays non-core.
GO:0005856 cytoskeleton	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: Mapped from UniProt's Cytoplasm, cytoskeleton statement, attributed to PMID:15558029, where SA-RhoGEF and Sept9b colocalise along actin stress fibres in REF52 cells and SA-RhoGEF fragments alter endogenous septin filament structures. Reason: Supported by the cited experiment, but the observation is in fibroblasts, not in the junction-forming epithelia where this protein's characterised function lies, and stress-fibre association there is the phenotype the protein suppresses once junctions form. Kept because the underlying immunofluorescence is real; non-core because cytoskeleton is a very general compartment and the association is indirect, mediated by septins. Supporting Evidence: PMID:15558029 We have identified a Rho-guanine nucleotide exchange factor (GEF) as a binding partner for a mammalian septin Sept9b using yeast two-hybrid screening
GO:0005886 plasma membrane	EXP PMID:29601110 Expression of novel "LOCGEF" isoforms of ARHGEF18 in eosinop...	KEEP AS NON CORE	Summary: Turton et al. show that the LOCGEF isoforms of ARHGEF18 expressed in human eosinophils are distributed perimembranously in resting cells and relocate to the tip of the nucleopod on activation. Reason: A genuine experimental localisation in a primary human cell type, and the only one outside epithelium and endothelium. Non-core because plasma membrane is the general parent of the functionally informative location, which is the apical plasma membrane and the junctional complex; and because this experiment concerns isoforms 1 to 3, which carry alternative N-termini, while the p114 isoform is not detected in eosinophils, so it is not obvious that the junctional mechanism applies there. Supporting Evidence: PMID:29601110 Immunofluorescence microscopy revealed striking relocalization of LOCGEFs from the plasma membrane to the two poles upon polarization of eosinophils by IL5, IL33 and CCL11 PMID:29601110 No p114 transcript or protein band was detected in eosinophils.
GO:0005886 plasma membrane	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Phylogenetic propagation of plasma-membrane localisation from PANTHER node PTN002677825, restricted to taxon 117571. Both seeds are human ARHGEF18. Reason: Both seeds are the gene being annotated. UniProtKB:Q6ZSZ5 is the reviewed human ARHGEF18 entry and UniProtKB:A0A590UK10 is a 1089-residue unreviewed entry for the same human gene, assigned to the same PANTHER subfamily PTHR47440:SF1. That is expected rather than circular, because ARHGEF18 carries this compartment experimentally, by EXP from PMID:29601110 and by IDA from the Human Protein Atlas, and such descendant evidence is exactly what a PAINT curator uses to place an ancestral call. What follows is only that the row adds no evidence the target did not already have. Kept rather than removed because the conclusion is correct and is_active_in is arguably the better qualifier for a GEF that must be membrane proximal to load its substrate; non-core because the term is the general parent of apical plasma membrane. Worth recording for the family rather than for this row: the node is restricted to taxon 117571 and no non-human member has yet contributed to it, so the inheritance claim it encodes is currently untested outside human. Propagation Review Root cause: EVIDENCE CIRCULAR OR REDUNDANT Sources checked: PANTHER:PTN002677825 · PTHR47440 node restricted to taxon 117571 (Euteleostomi) SUPPORTS TRANSFER Node placement rests on the target's own experimental localisation; no non-human member of the clade has contributed evidence to it yet. UniProtKB:Q6ZSZ5 · human ARHGEF18, reviewed entry, the target itself SUPPORTS TRANSFER Carries GO:0005886 by EXP from PMID:29601110 and by IDA from GO_REF:0000052. Self-appearance marks experimental grounding on the target. UniProtKB:A0A590UK10 · human ARHGEF18, unreviewed 1089-residue entry for the same gene SUPPORTS TRANSFER Resolved against UniProt. Gene name ARHGEF18, organism Homo sapiens, PANTHER PTHR47440:SF1. The same gene as the target under a second accession, so it is not an independent ortholog.
GO:0005886 plasma membrane	IDA GO_REF:0000052	KEEP AS NON CORE	Summary: Human Protein Atlas immunofluorescence placing the protein at the plasma membrane, consistent with the EXP row and with junctional staining in epithelia. Reason: Direct and consistent with three independent lines of evidence, but at the general parent of the informative term. The functionally meaningful statement is the apical plasma membrane and the apical junction complex, both of which are covered separately.
GO:0005886 plasma membrane	IEA GO_REF:0000044	KEEP AS NON CORE	Summary: Mapped from UniProt's Cell membrane statement, attributed to PMID:29601110. The fourth row for this term. Reason: Correct and redundant with the EXP row derived from the same publication. Non-core for the same reason as the other plasma-membrane rows.
GO:0070160 tight junction	IDA PMID:21258369 Spatially restricted activation of RhoA signalling at epithe...	NEW	Summary: Proposed new annotation. Terry et al. image endogenous p114RhoGEF against tight-junction and adherens-junction markers in Caco-2 and human corneal epithelial cells and find it overlapping occludin but not E-cadherin. Reason: The most specific compartment this gene has direct evidence for, and it is absent from GOA entirely. The marker comparison is the informative part, because overlapping occludin and not E-cadherin distinguishes the tight junction from the adherens junction rather than placing the protein at junctions in general. It is corroborated in a second tissue by Tornavaca et al., where ZO-1 and paracingulin are both required to recruit the protein to endothelial junctions. GOA currently records only GO:0030054 cell junction, six times over, from Reactome complex membership. The parent term is deliberate. GO:0005923 bicellular tight junction is the legacy term most existing annotations sit on, but it asserts a subtype this experiment cannot resolve, since confocal colocalisation along a junctional belt does not separate bicellular from tricellular contacts. Occludin is not a marker that excludes tricellular contacts either, though the point is weaker than it first looks and is stated here at the strength the source supports. PMID:40878853 finds occludin in close proximity to tricellulin at cell-cell junctions and shows that occludin knockout perturbs tricellulin localisation at tricellular tight junctions, while also reporting that occludin itself does not accumulate at them as prominently as tricellulin does. So occludin is predominantly bicellular but is present and functionally required at tricellular contacts, which makes it unable to license the subtype on its own. The resolution argument is the load-bearing one and stands without this. Choosing the parent also makes the companion GO:0043296 annotation non-redundant, because GO:0005923 has GO:0043296 among its parents in both OLS4 and QuickGO while GO:0070160 does not, so located_in would have propagated one from the other. Supporting Evidence: PMID:21258369 Junctional p114RhoGEF overlapped with occludin, a tight junction protein, but not E-cadherin, an adherens junction component, indicating that p114RhoGEF associates with the apical junctional complex and tight junctions PMID:40878853 BioID proteomics identified several proximity partners of tricellulin, and knockout studies on angulin-1/LSR, occludin and afadin provided evidence that these proteins control tricellulin accumulation to tTJs to different extents and mechanisms. PMID:40878853 tricellulin is in close proximity to occludin and afadin at cell-cell junctions, but neither one of these other proteins displays similar prominent accumulation to tTJs as tricellulin
GO:0007264 small GTPase-mediated signal transduction	IBA GO_REF:0000033	KEEP AS NON CORE	Summary: Phylogenetic propagation from PTN002677784 with the target itself as the only seed. Reason: The sole gene seed is UniProtKB:Q6ZSZ5, the gene being annotated, and that is expected rather than circular. ARHGEF18 carries GO:0007264 by IDA from PMID:14512443, and a PAINT curator placing an ancestral call uses exactly such descendant evidence, so the target legitimately appears among the sources of the IBA it receives. A short donor list is likewise not weak support. What follows is only that the row adds no evidence the target did not already have, which is a reason to treat it as non-core rather than as a defect. Non-core also because the term is one of the broadest in the process ontology, and because the informative content, that the small GTPase is RhoA, is what GO cannot express. Propagation Review Root cause: EVIDENCE CIRCULAR OR REDUNDANT Sources checked: PANTHER:PTN002677784 · PTHR47440 ancestral node, Lbc/RH-RhoGEF clade SUPPORTS TRANSFER The same node carries a well-seeded GO:0005085 IBD across six accessions, so the node placement itself is multi-species; for this term the curator cited the target's own experimental annotation. UniProtKB:Q6ZSZ5 · human ARHGEF18, the target itself and the sole gene seed SUPPORTS TRANSFER Carries GO:0007264 by IDA from PMID:14512443. Self-appearance in the WITH/FROM marks experimental grounding on the target and is not circularity.
GO:0007264 small GTPase-mediated signal transduction	IDA PMID:14512443 G Protein betagamma subunits stimulate p114RhoGEF, a guanine...	KEEP AS NON CORE	Summary: Direct evidence that expressing p114RhoGEF changes the GTP-loading state of Rho-family GTPases in cells, from the same experiments that support the exchange-factor IDA. Reason: Correct but at the top of the process branch. The specific content of the experiment is RhoA and Rac1 loading, and the informative process term for a GEF is the positive-regulation one, GO:0035025, which this review proposes on the IBA row for GO:0035023. The acts_upstream_of_or_within qualifier is appropriate for a readout measured on the downstream GTPase rather than on the GEF. Supporting Evidence: PMID:14512443 we have determined that p114RhoGEF activated RhoA and Rac1 but not Cdc42 proteins
GO:0007264 small GTPase-mediated signal transduction	IEA GO_REF:0000002	KEEP AS NON CORE	Summary: InterPro2GO mapping from IPR037744, the ARHGEF18-specific PH-domain signature. Reason: The signature is specific to this protein family rather than a generic PH match, so the mapping is sound. Redundant with the IDA and IBA rows for the same term, and equally general.
GO:0008360 regulation of cell shape	IDA PMID:14512443 G Protein betagamma subunits stimulate p114RhoGEF, a guanine...	KEEP AS NON CORE	Summary: Expressing p114RhoGEF induces cell rounding and stress-fibre formation, which Niu et al. used as the functional readout of RhoA activation. Reason: A real gain-of-function phenotype, reproduced independently in J82 and HEK-293 cells by Blomquist et al. Non-core because it is a downstream consequence of RhoA activation in a sparse overexpression setting rather than a physiological role, and because the same protein produces the opposite cortical phenotype in a junction-forming monolayer. Cell shape is also where the disease and tissue biology converges without being the molecular function, since the apical constriction, actomyosin belt and lumen phenotypes are all shape changes downstream of the same exchange activity. Supporting Evidence: PMID:11085924 the overexpression of p114-Rho-GEF in J82 and HEK-293 cells induced the formation of actin stress fibres and stimulated serum-response-factor-mediated gene transcription in a Rho-dependent manner
GO:0016324 apical plasma membrane	EXP PMID:22006950 Lulu2 regulates the circumferential actomyosin tensile syste...	ACCEPT	Summary: Nakajima and Tanoue localise p114RhoGEF to apical cell-cell boundaries in epithelial cells, where Patj recruits it through a PDZ-mediated interaction and Lulu2 activates it. Reason: This is the functionally informative location and the one the mechanism requires. It is corroborated by three independent groups. Terry et al. report that junctional p114RhoGEF overlaps with occludin but not E-cadherin, placing it at the apical junctional complex rather than the adherens junction; Tornavaca et al. show that ZO-1 and paracingulin are both required for junctional recruitment in endothelium; and the medaka mutant loses apicobasal polarity of the retinal neuroepithelium. Kept at core level because localisation here is not incidental, it is the substrate of the protein's defining property, which is spatially restricted RhoA activation. Supporting Evidence: PMID:22006950 We further found that Patj, an apical cell polarity regulator, recruits p114RhoGEF to apical cell-cell boundaries via PDZ (PSD-95/Dlg/ZO-1) domain-mediated interaction PMID:21258369 Junctional p114RhoGEF overlapped with occludin, a tight junction protein, but not E-cadherin, an adherens junction component, indicating that p114RhoGEF associates with the apical junctional complex and tight junctions
GO:0016324 apical plasma membrane	IEA GO_REF:0000044	ACCEPT	Summary: Mapped from UniProt's Apical cell membrane statement, which is attributed to the same publication as the EXP row. Reason: Correct, and redundant with the EXP row it derives from. Accepted rather than demoted because the term itself is the functionally informative one.
GO:0030036 actin cytoskeleton organization	IDA PMID:14512443 G Protein betagamma subunits stimulate p114RhoGEF, a guanine...	KEEP AS NON CORE	Summary: Stress-fibre induction on p114RhoGEF expression, the actin arm of the same gain-of-function experiment behind the cell-shape row. Reason: True, and true for essentially every RhoA GEF, which is what makes it uninformative at this level. The specific and reproducible statement about this protein is that it builds the junctional actomyosin belt and, in doing so, suppresses basal stress fibres; that content sits in the GO:0051497 and GO:0150105 rows and in core_functions. Kept because the experiment is direct. Supporting Evidence: PMID:11085924 the overexpression of p114-Rho-GEF in J82 and HEK-293 cells induced the formation of actin stress fibres and stimulated serum-response-factor-mediated gene transcription in a Rho-dependent manner
GO:0030054 cell junction	TAS Reactome:R-HSA-2160931	KEEP AS NON CORE	Summary: One of six identical Reactome cell-junction rows. ARHGEF18 is a component of Reactome's Tight Junction Complex entity, and every reaction that consumes or produces that complex assigns the cell-junction compartment to all of its participants. This reaction is Disassembly of tight junctions. Reason: The underlying biology is well supported by direct experiment, so the row is not wrong, since ARHGEF18 really is a junctional protein. But the row is a participant-compartment assignment attached to a TGF-beta, PARD6A and SMURF1 reaction in which ARHGEF18 plays no described part, confirmed by querying the Reactome content service for the reaction's participants. Non-core also because GO:0030054 is the general parent of the informative term. Direct evidence supports the tight junction and apical junction complex specifically, and that is what the two proposed new compartment annotations record. Supporting Evidence: PMID:21258369 Junctional p114RhoGEF overlapped with occludin, a tight junction protein, but not E-cadherin, an adherens junction component, indicating that p114RhoGEF associates with the apical junctional complex and tight junctions
GO:0030054 cell junction	TAS Reactome:R-HSA-2160932	KEEP AS NON CORE	Summary: One of six identical Reactome cell-junction rows, from the reaction SMURF1 binds phosphorylated PARD6A, whose inputs and outputs contain the Tight Junction Complex entity of which ARHGEF18 is a member. Reason: Same participant-compartment mechanism and the same verdict as the other five Reactome cell-junction rows. Kept because junctional localisation is independently established by experiment; non-core because the term is general and the row carries no information specific to this gene.
GO:0030054 cell junction	TAS Reactome:R-HSA-2160935	KEEP AS NON CORE	Summary: One of six identical Reactome cell-junction rows, from the reaction SMURF1 ubiquitinates RHOA within the tight-junction-associated complex. Reason: Same participant-compartment mechanism and the same verdict. Worth one remark because RHOA appears here as the SMURF1 substrate rather than as the GEF's substrate, so the co-occurrence of ARHGEF18 and RHOA in this reaction is not evidence about exchange activity.
GO:0030054 cell junction	TAS Reactome:R-NUL-2161147	KEEP AS NON CORE	Summary: One of six identical Reactome cell-junction rows. The R-NUL prefix marks an inferred, non-human-specific event, here TGFBR1 is recruited to tight junction by binding Pard6a. Reason: Same participant-compartment mechanism as the R-HSA rows, one step further removed because the event itself is an inference rather than a human-curated reaction. Kept for the same reason and at the same level.
GO:0030054 cell junction	TAS Reactome:R-NUL-2161160	KEEP AS NON CORE	Summary: One of six identical Reactome cell-junction rows, from the inferred event TGFBR2 is recruited to tight junctions-associated, Pard6a-bound, TGFBR1 after TGF-beta stimulation. Reason: Same participant-compartment mechanism and the same verdict as the other five.
GO:0030054 cell junction	TAS Reactome:R-NUL-2161165	KEEP AS NON CORE	Summary: One of six identical Reactome cell-junction rows, from the inferred event TGFBR2 phosphorylates Pard6a. Reason: Same participant-compartment mechanism and the same verdict as the other five. Six rows asserting one fact, none of which is an observation about ARHGEF18.
GO:0031267 small GTPase binding	ISS PMID:29876405 Crystal structures of the PH domains from Lbc family of RhoG...	NEW	Summary: Proposed new annotation, and the only molecular function of this protein that has been solved structurally. The PH domain of the mouse ortholog was crystallised at 1.4 angstrom bound to RhoA loaded with a GTP analogue, PDB 6BCB, and the same interface is a general property of the Lbc-family RhoGEFs, where it acts as a positive-feedback mechanism rather than as the catalytic site. Reason: This is a distinct activity from the exchange reaction, on a different domain and in the opposite nucleotide state, and GO records none of it for this gene in any species. The transfer is by sequence similarity rather than IPI because the solved construct is the mouse ortholog UniProtKB:Q6P9R4; the ten RhoA-contacting residues computed from 6BCB at 4 angstrom map to human Ile-740, Arg-742, Glu-743, Val-744, Ala-745, Asn-746, Phe-752, Ile-754, Pro-761 and Met-763, all identical, so the interface is conserved without exception. GO:0031267 is the most specific term available, because GO:0017048 Rho GTPase binding was merged into it and it now has zero is_a children in both OLS4 and QuickGO, so neither the partner identity nor the GTP-loaded state can be expressed in the term itself. The reference field is deliberately the paper rather than GO_REF:0000024. That GO_REF covers manual transfer of experimentally-verified annotation data to orthologs, which presupposes an annotation on the donor, and the mouse record carries no GO:0031267 to transfer. Mouse Arhgef18 has 23 GO annotations and not one of them is experimental, ten being ISO projections from human. Citing PMID:29876405 directly keeps the row consistent with that fact, since the paper is where the similarity-bearing evidence actually lives. The cleaner two-step route, if MGI will take it, is to enter GO:0031267 on the mouse record first by IPI from this paper with RHOA in the WITH/FROM field, and let the human row follow as an ISS behind it. Supporting Evidence: PMID:29876405 Here we present two crystal structures at resolutions of 1.4 Å and 2.0 Å of RhoA complexed with the PH domain from p114RhoGEF (PDB access code 6BCB) and AKAP-LbcRhoGEF (PDB access code 6BCA), respectively PMID:23493395 This demonstrates feasibility of the hypothesis that binding of activated RhoA to the PH domains acts as a positive feedback mechanism file:human/ARHGEF18/ARHGEF18-bioinformatics/RESULTS.md Bottom line: all ten RhoA-contacting residues of the mouse PH domain are identical in human ARHGEF18 and all ten lie inside the human PH domain (684-786), so the RhoA-binding interface is conserved without exception across the orthologs.
GO:0035023 regulation of Rho protein signal transduction	IBA GO_REF:0000033	MODIFY	Summary: Phylogenetic propagation from PTN002677784, seeded by mouse Arhgef2 and human AKAP13. The conclusion is right but the term is sign-agnostic, and everything known about this protein says the sign is positive. Reason: Every primary experiment on ARHGEF18 shows activation of Rho signalling, never inhibition. Purified protein catalyses exchange on RhoA; expressing it raises active RhoA and phosphorylated myosin light chain; depleting it lowers junctional active RhoA; and a patient missense allele in the DH domain reduces RHOA activation. UniProt already phrases the variant effect as decreased function in positive regulation of Rho protein signal transduction. GO:0035025 positive regulation of Rho protein signal transduction is current and non-obsolete in QuickGO, OLS4 and the GO API, and it is an is_a child of the term annotated here, so this is a straightforward refinement rather than a change of meaning. The refinement is warranted at clade level and not only from target data, which matters because an IBA encodes a judgment about the clade rather than about this gene. Both seeds are themselves exchange factors, mouse Arhgef2 being GEF-H1 and Q12802 being AKAP13/Lbc, and an exchange factor is by definition a positive regulator of its GTPase when active, so the ancestral call the node encodes already carries the positive sign; the target-specific results agree with it rather than substituting for it. Two further observations do not change the action. The seeds are paralogs rather than the target or its fly ortholog; and Terry et al. show that ARHGEF18 and GEF-H1 are functionally opposed at junctions, GEF-H1 being recruited and switched off where ARHGEF18 is recruited and switched on, which is a reason to be careful about transferring regulatory detail between them even though both positively regulate Rho when active. Propagation Review Root cause: TERM SCOPING PROBLEM Failure modes: GRANULARITY MISMATCH Sources checked: PANTHER:PTN002677784 · PTHR47440 ancestral node, Lbc/RH-RhoGEF clade SUPPORTS TRANSFER The clade-level inference is sound; only the sign-agnostic term is at issue. MGI:MGI:103264 · mouse Arhgef2 (GEF-H1) SUPPORTS TRANSFER A paralog, not the ortholog. Terry et al. show the two proteins are oppositely regulated at epithelial junctions. UniProtKB:Q12802 · AKAP13 (Lbc) SUPPORTS TRANSFER A paralog. Its PH domain binds activated RhoA in the same way this protein's does, in two structures from one study, PDB 6BCA and 6BCB. Proposed replacements: positive regulation of Rho protein signal transduction Supporting Evidence: PMID:28132693 the p.Thr270Ala missense variant affects a highly conserved residue in the DBL homology domain, which is required for the interaction and activation of RHOA PMID:23698346 We show that ArhGEF18-mediated activation of the small GTPase RhoA is required to maintain apicobasal polarity at the onset of retinal differentiation and to control the ratio of neurogenic to proliferative cell divisions
GO:0043296 apical junction complex	IDA PMID:21258369 Spatially restricted activation of RhoA signalling at epithe...	NEW	Summary: Proposed new annotation, from the same imaging that supports the tight junction row. Terry et al. state the conclusion in these terms, that p114RhoGEF associates with the apical junctional complex. Reason: This is the compartment the mechanism is defined in. The protein is not simply at junctions, it is at the apical junctional complex, where it generates the spatially restricted pool of active RhoA that drives junctional myosin activation. GOA has nothing between GO:0016324 apical plasma membrane and GO:0030054 cell junction, and this term is exactly the intersection the evidence supports. Proposed alongside GO:0070160 rather than instead of it, and the two are genuinely complementary rather than one implying the other. That depended on the term choice and was checked, not assumed. GO:0005923 bicellular tight junction has GO:0043296 among its parents in both OLS4 and QuickGO, so had the tight-junction row been entered at that level this one would have been inferred by propagation up part_of; GO:0070160 tight junction has no such edge, its only hierarchical parent being GO:0005911 cell-cell junction. Supporting Evidence: PMID:21258369 Junctional p114RhoGEF overlapped with occludin, a tight junction protein, but not E-cadherin, an adherens junction component, indicating that p114RhoGEF associates with the apical junctional complex and tight junctions
GO:0045177 apical part of cell	IDA PMID:22006950 Lulu2 regulates the circumferential actomyosin tensile syste...	KEEP AS NON CORE	Summary: The same Nakajima and Tanoue immunofluorescence that supports the apical plasma membrane row, curated by MGI at a more general term. Reason: Correct but strictly less informative than GO:0016324 from the identical experiment, which is annotated separately as EXP. Kept because it is a direct observation and GOA legitimately carries both granularities; non-core because the specific term is the one that matters.
GO:0051497 negative regulation of stress fiber assembly	IMP PMID:25753039 ZO-1 controls endothelial adherens junctions, cell-cell tens...	KEEP AS NON CORE	Summary: Depleting p114RhoGEF in human dermal microvascular endothelial cells induces stress fibres and vinculin-positive focal adhesions, so the protein is required to keep them down. The sign looks inverted for a RhoA activator, and the resolution is that the two observations are made in different systems. Reason: The annotation is supported and reproducible. The same loss-of-function phenotype appears independently in epithelium, where Terry et al. report reduced perijunctional f-actin and increased formation of stress fibres. It is not in conflict with the gain-of-function result that overexpressed p114RhoGEF makes stress fibres in sparse J82 and HEK-293 cells, because confluence is the variable. Terry et al. state the mechanism directly, that RhoA activity is downregulated on confluence and interfering with junction formation stimulates RhoA signalling elsewhere in the cell. So this is a real but indirect effect, the cortical consequence of failing to build a junction rather than a molecular action on stress fibres, and involved_in slightly overstates the directness. Kept as non-core rather than modified, because no GO term expresses the causal shape, which is redistribution of a GTPase activity pool from the basal cortex to the junction. Supporting Evidence: PMID:25753039 p114RhoGEF depletion led to a loss of junctional vinculin and induction of vinculin-stained focal adhesions and stress fibers comparable to depletions of ZO-1 and JACOP PMID:21258369 reduced perijunctional f-actin and increased formation of stress fibres PMID:21258369 RhoA activity is downregulated in response to cell confluence; hence, interfering with junction formation stimulates RhoA signalling in the rest of the cell
GO:0070062 extracellular exosome	HDA PMID:18570454 Proteomic analysis of exosomes from human neural stem cells ...	MARK AS OVER ANNOTATED	Summary: A single identification in a mass-spectrometry survey of exosomes from human neural stem cells, one of 66 proteins annotated from that paper. Reason: ARHGEF18 is a 1361-residue cytoplasmic protein with no signal peptide and no transmembrane segment, whose characterised function requires it to be on the cytoplasmic face of a junction. A single bulk-proteomics identification in one cell type is the classic signature of co-purification with vesicle preparations rather than evidence of a secreted pool, and nothing in the twenty-odd primary papers on this gene reports an extracellular form. Marked as over-annotated rather than removed because the measurement itself was made and HDA is honest about what it is. The term is not contradicted; it is unsupported as a statement about where this protein does anything.
GO:0120192 tight junction assembly	IMP PMID:21258369 Spatially restricted activation of RhoA signalling at epithe...	NEW	Summary: Proposed new annotation. Depleting p114RhoGEF leaves spreading and initial contact formation intact but blocks junctional maturation, so ZO-1 staining stays discontinuous, and transepithelial resistance reaches only 40 per cent of control while 4 and 70 kilodalton dextran continues to leak. Reason: This is the process the gene is best characterised for and GOA does not record it at all. The experiment is a clean loss-of-function with a quantitative functional readout, and it is stage-resolved, since initial adhesion is unaffected and only maturation fails, which is what makes the assembly term rather than a maintenance term correct. It is corroborated in three further systems, bronchial epithelium through LKB1, endothelium through ZO-1 and paracingulin, and MDCK tubulogenesis where lumen consolidation specifically fails. The same paper supplies the two proposed compartment annotations, and it has produced no GO annotation in any species, which is the single largest curation gap on this gene. Supporting Evidence: PMID:21258369 Depletion of p114RhoGEF strongly attenuated barrier formation and only reached 40% of the values of control cultures PMID:21258369 Thus, p114RhoGEF is not required for spreading and initiation of adhesion, but for junctional maturation. PMID:23648482 Together, LKB1 and p114RhoGEF control RhoA activity in these cells to promote apical junction assembly
GO:0150105 protein localization to cell-cell junction	IMP PMID:25753039 ZO-1 controls endothelial adherens junctions, cell-cell tens...	ACCEPT	Summary: Depleting p114RhoGEF in endothelial cells causes loss of junctional vinculin, and p114RhoGEF immunoprecipitates contain vinculin while ZO-1 immunoprecipitates do not. P18206 resolves to human vinculin. Reason: The best-supported process annotation this gene has. It is specific, it names the cargo through the WITH/FROM field, and it is the process form of what the protein actually does, which is to organise the junctional cortex so that mechanotransducers are recruited. It generalises beyond vinculin and beyond endothelium, since ARHGEF18 depletion also disrupts junctional recruitment of myosin IIA in epithelium and of ZO-1 and other tight-junction markers, and since EPB41L5 and the Crumbs complex recruit it in podocytes and in Drosophila respectively. Accepted at core level. Supporting Evidence: PMID:25753039 p114RhoGEF depletion led to a loss of junctional vinculin and induction of vinculin-stained focal adhesions and stress fibers comparable to depletions of ZO-1 and JACOP PMID:25753039 p114RhoGEF, and not ZO-1, immunoprecipitates contained vinculin

Gene Description

Existing Annotations Review

Core Functions

References

Suggested Questions for Experts

Suggested Experiments

Deep Research

Affinage

📚 Additional Documentation

Notes

Bioinformatics Results

📄 View Raw YAML