SNX-BAR dimers directly bind to the CI-MPR tail to promote retrograde sorting to the TGN. (A) The GFP-tagged CI-MPR tail and GFP only were lentivirally expressed in SILAC-labeled HEK293 cells and precipitated with GFP-trap beads, and then interacting proteins were identified by LC-MS/MS. The panel shows a STRING network analysis of CI-MPR tail–interacting proteins. The numbers beneath each protein indicate the fold enrichment of the protein when compared with the GFP-only control. (B) Western blot–based verification of the interacting proteins. (C) GFP-trap IPs of the indicated GFP-tagged SNX-BAR proteins precipitating the endogenous CI-MPR from HEK293 cells. (D) GFP-trap IPs of the GFP-tagged WT CI-MPR tail and of the indicated mutant CI-MPR tail constructs were probed for the presence of the indicated endogenous SNXs and endogenous VPS35 by Western blotting. (E) Direct recombinant interaction assay with GST-tagged CI-MPR tail produced in bacteria and the indicated individual His-tagged SNX-BAR proteins produced as secreted proteins in HEK293 cells. (F) Direct recombinant interaction assay with GST-tagged CI-MPR tail produced in bacteria, and the indicated His-tagged SNX-BAR proteins coexpressed as secreted proteins in HEK293 cells. (G) Three CRISPR constructs per indicated gene targeting distinct genomic regions of the respective gene were pooled and cotransfected with a puromycin resistance–encoding plasmid. After puromycin selection and recovery for 3 d, cells were seeded onto coverslips, and CI-MPR localization to the TGN was analyzed by immunofluorescent staining of endogenous CI-MPR (red) and endogenous TGN46 (green). The screen was performed twice, and the colocalization (coloc) between CI-MPR and TGN46 was quantified over 20 images per condition acquired in the two experiments. The efficiency of the respective CRISPR-Cas9 treatment was verified by Western blotting against selected targets. Bars, 10 µm. Error bars indicate SD. *, P < 0.05 compared with control conditions in a t test.