The decision table preserves each imported source object and its zero-based annotation index. The two cell-adhesion records are refined to cell-matrix adhesion; focal adhesion and positive regulation of integrin activation are retained as central functions. Cytoplasmic and tissue-junction locations, proliferation, migration, polarity and cultured wound closure remain positive contextual annotations.
Lai-Cheong et al., 2008, PMID:18528435 was read in the complete cached main text. The Results sections on confocal microscopy and Figures 3a/4 describe endogenous kindlin-1 at vinculin-positive focal adhesions in normal human keratinocytes and tagged corroboration in HaCaT cells. Figure 3a also records a diffuse cytoplasmic pool. The skin-immunofluorescence Results and Figure 7a show basal keratinocyte peripheral enrichment in normal human skin. Co-immunoprecipitation in Figure 2 establishes cellular association with kindlin-2 and migfilin, not purified binary binding. The unknown approximately 65-kDa band and residual mutant-cell staining do not identify a natural splice product. Relevant antibody, co-immunoprecipitation and microscopy Methods and the Discussion were read; figure pixels were not independently inspected.
Herz et al., 2006, PMID:17012746 is abstract-only in the immutable cache. Operative original text was recovered from the author-uploaded article and indexed publisher PDF text; direct publisher downloads returned 403, so no local PDF or figure-pixel read is claimed.
Experimental Procedures cover human primary-cell adhesion and migration assays, immunostaining and immunoelectron microscopy. Original Results and Figure 2C–D locate kindlin-1 intracellularly near the basal membrane between hemidesmosomes. This supports the curated cytosolic and cell–matrix-junction interpretation; it does not place the protein inside hemidesmosomes, prove a freely soluble fraction, or establish cell–cell-junction identity. The official cytosol definition includes particulate protein complexes, and cell junction includes cell–matrix connections.
Figures 4C–D and 6C concern proliferation; Figure 5 concerns polarized collagen XVII/integrin distributions; Figure 6D measures attachment to fibronectin and laminin 332. Figures 7–8 and their Results distinguish slower, less directional movement from altered membrane protrusions. These observations support the existing contextual processes without assigning cell-cycle chemistry or motor activity. COS-7 recombinant expression was used for antibody validation. The 74/78-kDa forms in Figure 3 are phosphorylation states, not evidence that the four annotated splice products have distinct functions.
He et al., 2011, PMID:21356350 has a misleadingly complete cache flag: the cached body contains abstract and Discussion, omitting Methods and Results. The original PMC Methods, Results and Figure 1/2 captions were recovered and read separately. The human normal and Kindler-syndrome-derived immortalized cell lines with kindlin-2 depletion preserve the intended genetic-interaction context.
Figure 2A measures active β1 integrin normalized to total surface receptor. Figure 2B tests matrix attachment: single-deficient effects are small and not significant, whereas combined deficiency is substantial. Figures 2D–F and migration Methods measure cultured gap closure and directionality. Kindlin-1 and kindlin-2 partially compensate; kindlin-2 has stronger cell–cell-junction effects in this system. These data support an adhesion-associated contribution and positive regulation of activation, without proving purified integrin binding, universal integrin activation, complete paralog equivalence or all phases of in vivo wound repair. No figure-pixel or supplement read is claimed.
GO:0007160 specifically covers cell attachment to matrix via adhesion molecules and is a strict descendant of the imported generic cell-adhesion term. Both measured attachment assays justify this replacement. Broader positive location records retain useful tissue or non-adhesion-pool context; a more specific focal-adhesion annotation does not make those observations false. This guide adds no quotation or new annotation and assigns no activity to an untested splice product.