FGD4 encodes frabin, a DH-domain exchange factor with an N-terminal actin-associated region, two PH domains and a FYVE zinc finger. Its conserved molecular role is regulation of CDC42 signaling. Cytoskeletal protrusions and cell-shape effects are downstream consequences of this activity; they do not establish autonomous filament-polymerizing activity.
The actual PANTHER version-19 tree places human Q96M96 at PTN002489344 beneath the two positive IBD nodes used by the current annotations. The ungapped target sequence exactly matches all 766 canonical residues. Human FGD1 is a different leaf and subfamily. Consequently, the ISS donors P98174 and P52734 must be described as FGD1 paralogs, even though the shared catalytic architecture and target evidence support several transfers. The evidence sidecar records the genuine acquisition URLs, hashes, node path and literal IBD joins.
The mechanistic donor evidence is strongest for nucleotide exchange. PMID:8969170, original Methods and exchange Results with Figures 1–3, tests an engineered human FGD1 DH–PH fragment, residues 330–710. GDP release/GTP loading differs from GTP hydrolysis, but the current ontology explicitly includes the exchange-defined descendant GO:0090630 activation of GTPase activity beneath GO:0043087. The target Reactome CDC42 exchange event and actual InterPro DH–PH matches support that refinement without a GAP or hydrolysis claim. The core is narrowed to GO:0032489 regulation of Cdc42 protein signal transduction, with the existing broader GO:0051056 rows preserved. Ontology definitions and typed relations document the current terms without inferring historical merge events.
Localization requires a different comparison. PMID:12913069, localization Results accompanying Figure 4, places tagged mouse Fgd1 in cytoplasmic/Golgi pools and recruited ruffles or lamellipodia in COS-7 cells. Its N-terminal cortactin/mAbp1-targeting mechanism is not interchangeable with FGD4’s N-terminal actin-associated region. Positive target cytosol and actin-rich protrusion evidence does not settle Golgi, ruffle or lamellipodial residence. The latter three remain unresolved without asserting their biological absence.
PMID:17564959, Methods “Expression constructs” and Results with Figures 5–7, explicitly uses rat full-length Frabin-GFP and engineered 1–297/1–249 truncations. Full-length protein colocalizes with F-actin in microspikes and promotes their formation; truncations retain colocalization but induce fewer protrusions. The human patient fibroblast comparison found no significant microfilament/microtubule/centrosome differences under its tested conditions. This is a context-specific negative, not a universal loss of cytoskeletal function. PMID:17564972, functional Results and Figure 3, reports rodent nerve/rat Schwann-cell expression and frabin-induced filopodia and lamellipodia; construct species is not explicit in that passage. Formation of a lamellipodium is not the same claim as frabin localization within it. Neither paper independently measures purified exchange specificity or direct actin binding.
The four Reactome sources have complete target routes. FGD4 entity 3000215 represents Q96M96 residues 1–766 in cytosol. It is an established member of CDC42-GEF set 9013227, reached through catalyst 9013228 from event 9013159 and the CDC42 subpath of 9012999. The older 205039 and 419166 events instead use generic GEF set 194849. Their imported cytosol claims can be retained without assigning FGD4 all RAC or RhoA/B/C activities in those generic sets.
The electronic mechanisms are explicit. ARBA00027280 branch 4 matches the actual target DH FunFam and Primates lineage. InterPro records DH 206–393, FGD4 PH 423–516 and FYVE 551–620, with their literal GO mappings. Reviewed UniProt gives its own slightly different PH/FYVE boundaries, which are retained rather than normalized. FYVE metal binding is a supporting structural inference; PH-family membership alone is not a measured phosphoinositide preference.
The three natural splice objects remain exactly as acquired. Q96M96-2 reconstructs to 280 residues and Q96M96-3 to 188 residues; neither retains the complete canonical catalytic architecture. No activity, stability, tissue abundance or equivalence to the engineered rat truncations is asserted. The donor papers’ fragments are not renamed as human FGD4 isoforms.
The genuine provider invocation did not produce a research report. No synthetic provider report or new publication quotation was created. Complete manual reference scopes distinguish original-text reads from abstract-only normal caches. Fifteen accepted normal files remain byte-exact; eventual publication must honor the FGD1 shared-source dependency and a fresh ownership gate.