Read on 2026-09-20/21. This is a reviewer record, not a provider-generated report or a replacement publication cache.
The main article is subscription content and PMC13434118 is embargoed until 2027-08-01. The PubMed abstract, public main-figure images, supplementary PDF and supplementary legend DOCX were accessible. Figure interpretations below use those actual sources rather than an inferred full-text claim.
The public main Figure 8A shows ARHGAP36-EGFP with F-actin at pseudopodia and tail regions; 8B shows the spatial actin/pMLC2/ARHGAP36 context during cell-in-cell formation. Supplementary Figure S6A locates ARHGAP36 at E-cadherin-positive cell junctions. S6B-C reports catenins, myosins and tropomyosin among co-purifying proteins; co-purification is not proof of direct binary binding. S7C compares full-length protein and truncation/motif constructs at E-cadherin/F-actin-positive junctions. S8A compares control and ARHGAP36-knockdown MDA-MB-436-10 cells, with altered polarized F-actin and pMLC2 distribution. The supplementary legend explicitly identifies these polarized distributions. Main Figure 7 shows forced nuclear relocalization and reduced cell-in-cell formation. These observations support a contextual actin-associated pool and spatial organization, independently of catalytic RhoGAP activity. They do not demonstrate purified actin binding or a GTP-hydrolysis reaction.
Re-executed the existing supplementary-workbook parser for PMID:32203420. ARHGAP36 was negative for RhoA, Rac1 and Cdc42, with ARHGAP35 and ARHGAP1 positive controls. The library identifies human isoform 2. The 15/65 all-negative fraction is not an estimated false-negative rate: the true activities of all rows are not known. ARHGAP17 provides a concrete example of a characterized GAP missed by this screen. The negative screen therefore corroborates, rather than independently proves, loss of conventional GAP catalysis.
Current UniProt sequence positions were rechecked: Q6ZRI8 sequence version 1 T258; structural anchor Q07960 version 1 R282; PAINT-source O54834 version 3 R435; mouse ortholog B1AUC7 version 1 T246. The existing domain alignments support their correspondence. A shared human/mouse threonine supports divergence in that ortholog branch, not an assertion that every family member lost every associated process. ND root annotations record missing specific experimental annotation, not a negative assay or contradiction of IBA.
The actual PTHR12635 tree places Q6ZRI8 leaf PTN002489201 below positive IBD PTN000973894 for GAP activity, actin organization and actin cytoskeleton. Node placement was inspected directly, not inferred from a subfamily label or source count. GO:0007015 includes spatial arrangement of actin structures; GO:0015629 includes associated proteins and complexes. Neither claim requires retention of the donor's catalytic GAP mechanism. Exact lineage, GO definitions, screen results, sequence versions and source hashes are retained in the project JSON artifact.
