FERMT3

Summary: Podosomes are a supported kindlin-3 adhesion-site location.

Reason: The combined annotation joins mouse Q8K1B8, its Ensembl translation and UniProt location SL-0295. The actual mouse donor source shows Kindlin-3 at the actin-surrounding podosome ring in hematopoietic cells (PMID:16876785, Results and Figure 5F). This is a conserved location inference, not a Human Protein Atlas experiment or an assay of every human splice product.

Summary: The mouse-to-human podosome transfer is biologically supported.

Reason: The exact mouse donor has an IDA podosome annotation from PMID:16876785. Its original Results distinguish endogenous staining from the EGFP-tagged dendritic-cell localization in Figure 5F. Failure of ectopic Kindlin-3 to target fibroblast focal adhesions in the same study is cell-context dependent and does not negate the positive podosome result.

Summary: Integrin binding is a core molecular function.

Reason: The inherited integrin-binding IBD agrees with the mouse Kindlin-3 interaction with beta-integrin cytoplasmic tails in PMID:19234461 (Figure 1d,e) and the human integrin-activation mechanism. The mouse experiment uses recombinant tails and platelet lysate; it is not described as purified human full-length protein binding.

Summary: The orthology-based integrin-binding inference is supported.

Reason: Mouse Q8K1B8 has experimentally supported integrin-binding donor records, including the tail interaction in PMID:19234461, Figure 1d,e. The actual Ensembl one-to-one relation agrees with the source translation. Retain the conserved interaction without assigning a measured affinity or specificity to every human product.

Summary: Curated ortholog transfer supports integrin binding.

Reason: The literal donor is mouse Fermt3 Q8K1B8. Its positive tail-binding evidence in PMID:19234461 and the independent human functional evidence support this conserved interaction; no paralog substitution or target-specific loss is demonstrated.

Summary: The modeled extracellular FERMT3 route is real, but its source-specific protein result remains unverified.

Reason: Reactome reaction R-HSA-481007 reaches output set 481033 and the literal human Q86UX7 entity 8862985 in extracellular region. However, the exact FERMT3 result in the cited platelet-releasate study PMID:23640497 and its supplemental tables could not be inspected. Preserve that positive curated route while leaving the source-specific location unresolved. Intracellular integrin function does not exclude extracellular detection; intact P-selectin responses do not establish or negate FERMT3 release.

Summary: Kindlin-3 contributes directly to leukocyte adhesion.

Reason: The positive IBD is consistent with human leukocyte activation-dependent adhesion in PMID:19234460 and mouse neutrophil firm adhesion in PMID:19234461. The preserved upstream/selectin responses distinguish an integrin-adhesion defect from a global failure of leukocyte signaling.

Summary: The mouse ortholog supports leukocyte cell-cell adhesion.

Reason: The exact mouse donor IMP derives from PMID:19234461, whose Results and Figures 3–4 show defective firm adhesion to activated endothelium while selectin-mediated rolling remains intact. This supports the inherited human role and does not imply that kindlin-3 is an extracellular adhesion ligand.

Summary: Kindlin-3 supports integrin-dependent leukocyte adhesion.

Reason: Human affected leukocytes had normal integrin abundance but defective activation-dependent adhesion and aggregation. PMID:19234460 Figure 2B,C directly examines lymphocyte and neutrophil aggregation, while Figure 4 provides KINDLIN3 rescue and knockdown evidence for adhesion. Retain this positive leukocyte context as an output of the integrin-activation role, without extending it to all adhesion mechanisms or all splice products.

Summary: Cell-matrix adhesion is part of the kindlin-3 mechanism.

Reason: The family-level IBD agrees with the actual human patient-cell adhesion and rescue results on fibronectin/fibrinogen in PMID:19234460, Figure 4. Kindlin-3 contributes intracellular integrin activation and adhesion-site function rather than serving as the matrix substrate. The tagged rescue construct is not a natural splice-product comparison.

Summary: The InterPro inference is consistent with integrin-mediated signaling.

Reason: The target has IPR037837 PH-domain and IPR037843 kindlin-family matches, both linked to GO:0007229. This term includes ligand-triggered signaling, so inside-out activation alone would be incomplete support. Human substrate-dependent spreading in PMID:19234460 and the mouse outside-in result in PMID:18278053, Figure 4, provide the relevant additional context.

Summary: Kindlin-3 supports integrin-dependent adhesion and spreading while several upstream signaling responses remain intact.

Reason: The official term describes signaling initiated by ligand binding to integrin. PMID:19234460 establishes defective receptor activation together with adhesion-dependent spreading, substrate contacts, and their rescue; it does not identify a kinase activity of kindlin-3. Figure 1D and the signaling controls preserve P-selectin induction and several agonist responses, so this is not a global signaling defect. Keep the supported signaling/adhesion context as non-core, with integrin activation the more directly resolved process. Inside-out activation alone is not being relabeled as the entire outside-in pathway.

Summary: The isolated kindlin-3 F1 loop binds acidic phospholipid vesicles.

Reason: PMID:27101375 Figure 3 and its Results show NMR signal attenuation with POPS and POPG vesicles, with substantially weaker changes for POPC. GO:0005543 phospholipid binding captures the actual substrate class more precisely than lipid binding while covering both tested acidic phospholipids. The assay used the recombinant Leu142–Gln224 loop and a separate synthetic peptide, not a comparison of intact natural splice products. The proposed full-length membrane orientation in Figure 6 remains a hypothesis.

Proposed replacements: phospholipid binding

Summary: Human kindlin-3 is recruited to the plasma membrane during adhesion; refine the broad membrane annotation using independent cellular evidence.

Reason: The original NK-cell HDA record is retained unchanged. Its final published supplement remains inaccessible; the related author thesis has positive ENSP00000279227/Kindlin-3 membrane-fraction detections but is not an established identical dataset. Independent human evidence supports the more precise gene-level plasma-membrane location: PMID:19234460 Figure 4e shows KINDLIN3-EGFP at adherent B-cell contacts, and PMID:42574321 Figure 2B–F directly follows full-length human short and IPRR-containing long proteins at the plasma-membrane contact zone in rescued neutrophil-like HL-60 cells. Both forms recruit during arrest, with greater recruitment of the long form in that setting. PMID:27101375 supplies complementary isolated F1-loop lipid-binding evidence, not an independent full-length localization measurement. GO:0005886 includes membrane-associated proteins and does not require an integral topology. This refinement uses matching independent location evidence; it does not retrospectively establish the final NK-cell target join or its detected natural splice product.

Proposed replacements: plasma membrane

Summary: Kindlin-3 acts at cell-substrate adhesion contacts.

Reason: The cell-substrate-junction IBD agrees with tagged Kindlin-3 at adherent human patient-cell contact sites in PMID:19234460, Figure 4e. Mouse podosome-ring localization in PMID:16876785 supplies a separate adhesion-site observation. The 2026-10-26 QuickGO is_a/part_of query places neither podosome GO:0002102 nor cell-substrate junction GO:0030055 in the other term's ancestor set. Both locations remain represented; these observations do not define every contact as a classical fibroblast focal adhesion.

Summary: Kindlin-3 restores adhesion-dependent migration of deficient human B cells.

Reason: PMID:19234463 Figure 4B,C measures restored movement on ICAM-1 after kindlin-3 complementation of human LAD-III B cells. Methods explicitly identifies the rescue construct as mouse EGFP-KINDLIN3; the human genetic lesions and recipient-cell phenotype remain distinct from that cross-species complementation. Retain migration as a positive cellular consequence of the integrin-adhesion function, without inferring a separate motility motor or a human splice-product-specific activity.

Summary: Positive regulation of migration is a supported cellular consequence of the adhesion mechanism.

Reason: ARBA00028907 explicitly emits GO:0030335; its current exact target-matching condition branch was not independently reconstructed. PMID:19234463 shows migration of human patient B cells on ICAM-1 restored by mouse EGFP-KINDLIN3. Retain this bounded adhesion-dependent migration role as non-core, without treating the rescue as purified human-protein evidence or claiming migration in all cell types.

Summary: The modeled granule-lumen pool is unverified, and its membrane-routing mechanism is unresolved.

Reason: Reactome R-HSA-481007 places human FERMT3 Q86UX7, entity 8862993, in the alpha-granule-lumen input set 481003. This lumen is separated from the cytosol by a membrane. The UniProt record has no SIGNAL or TRANSMEM feature, consistent with the established cytoplasmic integrin-tail adaptor role and providing no classical secretory-routing explanation. Those feature absences do not establish that FERMT3 cannot occur in a minor luminal pool; no alternative entry route has been demonstrated in the evidence inspected here. The accessible first page of PMID:23640497 describes depletion-based platelet-release profiling, which does not by itself resolve membrane sidedness. Its exact FERMT3 result, complete Methods and supporting compartment evidence remain unread. Retain source-specific uncertainty: neither a luminal location nor its exclusion is established.

Summary: Integrin activation is a central kindlin-3 role.

Reason: The IBD agrees with human deficiency, activation-specific ligand/antibody responses and rescue evidence in PMID:19234460, and with the donor mouse evidence in PMID:19234461.

Summary: Kindlin-3 is required for productive activation of leukocyte integrin adhesion.

Reason: PMID:19234463 identifies human FERMT3 lesions with loss of kindlin-3 protein and shows recovery of LFA-1-dependent adhesive contacts and migration after expression of mouse EGFP-KINDLIN3 in human patient B cells. This supports the gene-level integrin-activation role, reinforced by the independent human activation-specific antibody and ligand-binding measurements in PMID:19234460. It does not establish that kindlin-3 alone activates purified integrin or define a unique human splice product.

Summary: The ortholog-based integrin-activation annotation is supported.

Reason: The mouse donor carries matching IMP evidence, including PMID:19234461. The actual source distinguishes defective inside-out activation from preserved integrin abundance and bypassable ligand binding. Human FERMT3 deficiency independently supports conservation of this role.

Summary: Human kindlin-3 supports agonist-dependent activation of several integrin subfamilies.

Reason: PMID:19234460 measures loss of activation-dependent ligand or antibody binding despite preserved integrin abundance in patient platelets and leukocytes. Bypass activation remains possible, and KINDLIN3 expression restores adhesion/spreading while knockdown reproduces the cellular defect. These experiments support kindlin-3 participation in the machinery that enables integrin activation; they do not assign a catalytic activity or imply that all upstream agonist signaling is absent.

Summary: Kindlin-3 regulates integrin-mediated cell-cell adhesion.

Reason: The IBD is supported by human leukocyte activation-dependent adhesion in PMID:19234460 and the mouse endothelial-adhesion mechanism in PMID:19234461. This is an intracellular regulatory contribution, not a claim that kindlin-3 directly binds the opposing cell or is itself an integrin.

Summary: Kindlin-3 regulates integrin-mediated adhesive contacts in leukocytes.

Reason: Human LAD-III cells in PMID:19234463 lose kindlin-3 and show poor LFA-1-mediated attachment to ICAM-1; mouse EGFP-KINDLIN3 restores contact area and migration in the human recipient cells. The assay uses immobilized ICAM-1 rather than a second living cell. Independent human leukocyte aggregation in PMID:19234460 Figure 2B,C supports the gene-level cell-cell adhesion context. Retain this regulation as non-core downstream coverage of integrin activation, with both assay boundaries explicit.

Summary: The mouse ortholog supports regulation of integrin-mediated cell-cell adhesion.

Reason: The exact mouse donor has the matching IMP from PMID:19234461. Its neutrophil ligand-binding and endothelial arrest results identify an integrin-activation step. Human patient-cell evidence is concordant, with the mouse construct used for complementation kept distinct from human endogenous observations.

Summary: Kindlin-3 supports spreading after adhesion to integrin substrates.

Reason: PMID:19234460 directly reports impaired spreading of human platelets and patient-derived cells, and rescue of adherent contacts, ruffling and spreading after KINDLIN3 expression. Figures 1B and 4C–F support this morphological process. Retain it as a positive cellular output of the integrin-associated machinery, without treating deficient spreading alone as evidence for a new enzymatic or cytoskeletal motor activity.

Summary: FERMT3 detection in the reported B-cell exosome preparation remains unverified.

Reason: The actual author-posted PMID:20458337 Methods and Results describe human RN B-cell exosome purification and identify 539 proteins in Supplementary Table 1. The exact FERMT3/Q86UX7 entry could not be recovered because the published workbook remained inaccessible. Broad membrane association does not independently establish membership in these exosomes. Preserve uncertainty and defer to the original curator rather than infer a false positive. Total-exosome detection would also not by itself establish membership in the smaller MHC-II-associated subset or a role in exosome biogenesis.

Summary: Kindlin-3 contributes to the integrin-dependent platelet aggregation context.

Reason: The original PMID:23382103 Table I identifies URP2_HUMAN/Kindlin-3 (spot 6) among several proteins reduced in MDS platelets; Figure 8 measures aggregation defects. The study does not selectively perturb FERMT3, and its FRET assay measures talin-1 rather than kindlin-3. Keep the positive platelet role as non-core contextual evidence, corroborated by the human FERMT3-deficiency and platelet activation/aggregation measurements in PMID:19234460 Figure 1. This does not assign the MDS phenotype uniquely to kindlin-3 or prove a natural isoform effect.

Summary: Kindlin-3 has a conserved platelet-aggregation role.

Reason: The positive IBD agrees with the mouse aggregation defect in PMID:18278053 and independent human platelet dysfunction in PMID:19234460. The role follows integrin activation, not a general loss of platelet secretion.

Summary: The orthology-based platelet-aggregation annotation is supported.

Reason: The exact mouse donor IMP from PMID:18278053 supports aggregation through integrin activation; human platelet evidence in PMID:19234460 is concordant. Retain the function without assigning it experimentally to both natural human products.

Summary: Curated ortholog transfer supports platelet aggregation.

Reason: Mouse Q8K1B8 has the matching experimental annotation from PMID:18278053. Its activation-dependent aggregation result and the independent human phenotype support conservation of this platelet role.