Fig 2: The COR1C CC limits bud actin to facilitate ER contact, endosome fission, and CI-M6PR sorting. (A) Representative images of LE buds stable for duration of acquisition in conditions that did not rescued fission rate (C). COS-7 cells were cotransfected with COR1A/1B/1C siRNAs to deplete all type I coronins and with GFP-Rab7 (LE, magenta), mCh-FAM21 (WASH complex, green), and with either Halo E-vec, siRES COR1C ΔCC-Halo, or siRES COR1C ACT– ΔCC-Halo. Arrows indicate bud of interest. (B) Representative images of LE fission events in conditions that rescued fission rate (C). COS-7 cells were cotransfected with COR1A/1B/1C siRNAs to deplete all type I coronins and with GFP-Rab7 (LE, magenta), mCh-FAM21 (WASH complex, green), and with either siRES COR1C-Halo, siRES COR1C ACT–-Halo, or siRES COR1C CC-Halo. Arrows indicate bud of interest. (C) Quantification of data in A and B. Graph shows percentage of FAM21-labeled LE buds that underwent fission per cell during a 2-min time lapse. Note that only constructs containing the CC were able to restore fission. Data for graph from Halo E-vec: 139 endosomes in n = 15 cells; siRES COR1C: 122 endosomes in n = 15 cells; siRES COR1C ACT–: 213 endosomes in n = 17 cells; siRES COR1C ΔCC: 281 endosomes in n = 18 cells; siRES COR1C ACT– ΔCC: 296 endosomes n = 17 cells; and siRES COR1C CC: 331 endosomes in n = 22 cells, performed in triplicate. (D) Representative images of the M6PR trafficking assay. The relative fluorescence intensity of internalized anti-CI-MPR antibody immunostaining reveals trafficking of internalized anti-CI-MPR to TGN in Cos7 cells cotransfected with control siRNA, COR1A/1B/1C siRNAs to deplete all type I coronins, or FAM21 siRNA and with either GFP E-vec, siRES COR1C-GFP, or siRES COR1C ΔCC-GFP (not depicted). Cells were stained to mark CI-M6PR (green) and Giantin (Golgi, magenta). Dispersed vesicular CI-M6PR signal is indicative of failure to recycle, whereas concentrated CI-M6PR signal at the Golgi indicates normal retrograde sorting. (E) Quantification of data in D. Graph shows the background-corrected ratio of CI-M6PR signal localized at the Golgi relative to the vesicular signal in the cytoplasm such that larger values indicate less efficient retrograde recycling. Data for graph from control siRNAs: n = 24 cells; FAM21 siRNAs: n = 23 cells; COR1A/1B/1C siRNAs + E-vec: n = 25; COR1A/1B/1C siRNAs + siRES COR1C: n = 24 cells; COR1A/1B/1C siRNAs + siRES COR1C ΔCC: n = 24 cells; COR1A/1B/1C siRNAs + siRES COR1C CC: n = 26 cells, performed in triplicate. (F) Representative images of COS-7 cells cotransfected with COR1A/1B/1C siRNAs to deplete all type I coronins and with GFP-Rab7 (LE, gray), mCh-FAM21 (WASH complex, green), BFP-Sec61β (ER, magenta), and either Halo E-vec, siRES COR1C-Halo, siRES COR1C ACT–-Halo, siRES COR1C ΔCC-Halo, siRES COR1C ACT– ΔCC-Halo, or siRES COR1C CC-Halo (gray, left panel). Magnified inset (5 × 5 μm) on right show representative examples of ER contact with endosomes. Note: Vacuolar contact with ER is always preserved, whereas ER contact with the FAM21 labeled bud is not. (G) Line scan analysis of dashed lines shown in D are positioned from the rear vacuolar contact across the length of the FAM21-labeled buds. Double ER peaks (shaded purple) are observed when bud contact is rescued. The first peak is always present and corresponds to the vacuolar contact. The second ER peak which aligns with the FAM21 indicates proper ER recruitment to the bud. Note that ER contact with bud is dependent on the presence of the CC domain. (H) Quantification of data in D. All FAM21-positive LE buds in areas with resolvable ER were tracked, and ER contact was scored as the percentage of time during a 2-min video that contact is maintained. Data for graph from Halo E-vec: 109 endosomes in n = 14 cells; COR1C: 134 endosomes in n = 17 cells; COR1C ACT–: 135 endosomes in n = 17 cells; COR1C ΔCC: 123 endosomes n = 17 cells; COR1C ACT– ΔCC: 105 endosomes in n = 15 cells; or COR1C CC: 137 endosomes in n = 19 cells, performed in triplicate. X indicates mean, and line indicates median. Statistical analyses were performed with one-way ANOVA, P value from Tukey’s test: *, P < 0.05; ***, P < 0.001. Scale bars for whole cell = 5 μm; insets = 1 μm.

Fig 3: Aflibercept is largely co-localized with CI-M6PR. iBREC exposed to aflibercept were immunostained to detect aflibercept (red) and either the marker of multivesicular endosomes CI-M6PR (a, green) or Rab7 (b, green), specific for late endosomes and lysosomes. a The prominent CI-M6PR-specific staining close to the nucleus broadly overlapped with that of aflibercept (yellow arrows). b In contrast, aflibercept-specific signals and those specific for Rab7 gave rise to a mosaic pattern with few overlaps (yellow arrowheads). Scale bar, 10 μm

Fig 4: Bevacizumab is somewhat co-localized with CI-M6PR. After treatment of iBREC with the humanized IgG bevacizumab for 4 h, cells were immunostained to visualize the IgG (red) or CI-M6PR (green). The prominent CI-M6PR-specific signals close to the nucleus only to a small extent overlapped with those of bevacizumab (yellow arrowheads). Scale bar, 10 μm

Fig 5: Immunoblot of FAM21 depletion and type I coronin depletion and rescue in COS-7 cells for CI-M6PR sorting. Immunoblots probed with antibodies indicated show efficient depletion of COR1A/1B/1C or FAM21 for experiments performed in Fig. 5, D and E. Blots were also probed for GFP to show exogenous rescue expression and GAPDH as loading control. Source data are available for this figure: SourceData FS5.