Fig 2: Insulin localization is disrupted in Eipr1 KO cells. (A) Representative images of 832/13 (WT), Eipr1 KO 832/13 (Eipr1 KO), and Eipr1 KO 832/13 cells expressing WT Eipr1 (EIPR1[+]) costained for endogenous insulin and TGN38. In WT and EIPR1(+) cells, insulin is spread throughout the cytoplasm, but in Eipr1 KO cells insulin accumulates in a perinuclear region that partially overlaps with TGN38. The experiment was repeated three times, and the experimenter was blinded to the genotypes of the stained cells. Scale bars = 5 µm. (B) Pearson’s correlation coefficient was measured to quantify the localization between insulin and the TGN marker TGN38. n = 11 for WT, n = 13 for Eipr1 KO, and n = 15 for EIPR1(+); ***, p < 0.001; *, p < 0.05; ns, p > 0.05; error bars = SEM. The experiment was repeated three times. (C) Eipr1 KO cells have an increased Golgi/cytoplasmic ratio of insulin relative to WT cells. Fluorescence of a region of interest that includes the TGN divided by the fluorescence of a region of the same size in the cytoplasm, in WT and Eipr1 KO. (n = 25 for WT and Eipr1 KO; error bars = SEM; ***, p < 0.001). The data shown for the WT are the same as shown in Figure 2C of Cattin-Ortolá, Topalidou, et al. (2019) because these experiments were run in parallel with the same WT control. (D) Eipr1 KO cells have a slightly increased amount of insulin localized at or near the TGN. Fluorescence of a region of interest that includes the TGN divided by the fluorescence of a region of the same size in the background, in WT and Eipr1 KO. (n = 20 for WT and n = 30 for Eipr1 KO; error bars = SEM; *, p < 0.05). (E) Eipr1 KO cells have a decreased amount of insulin localized to the cytoplasm. Fluorescence of a region of interest in the cytoplasm divided by the fluorescence of a region of the same size in the background (n = 28 for WT and Eipr1 KO; error bars = SEM;***, p < 0.001). (F) Representative images of 832/13 (WT) and Eipr1 KO 832/13 (Eipr1 KO) cells costained for endogenous proinsulin and TGN38. In both WT and Eipr1 KO cells, proinsulin is localized in a perinuclear region that partially overlaps with TGN38. Scale bars = 5 µm. (G) Pearson’s correlation coefficient was measured to quantify the localization between proinsulin and the TGN marker TGN38. n = 18 for WT and n = 17 for Eipr1 KO; *, p < 0.05; error bars = SEM. The experiment was repeated three times.

Fig 3: DCV cargoes exit the TGN in Eipr1 KO cells. (A) Representative images of WT and Eipr1 KO 832/13 cells transfected with ANF::GFP and costained with anti-GFP and anti-TGN38 antibodies. In WT cells, ANF::GFP is distributed in cytoplasmic puncta, but in Eipr1 KO cells, ANF::GFP is restricted to perinuclear puncta. Not all cells in this field of view are transfected with ANF::GFP. Scale bars = 10 µm. (B) Schematic of the pulse-chase experiment. Cells transiently transfected with ANF::GFP were incubated at 20°C for 2 h to cause the accumulation of DCV cargoes at the TGN (pulse). Before the end of the temperature block (30 min), cyclohexamide was added to block protein translation. At the end of the temperature block, cells were returned to 37°C and incubated for various times (chase) before fixation and immunostaining. (C) Representative images of the cell categories used for qualitative assessment of TGN exit: 1) ANF::GFP concentrated at the TGN region (Golgi-like), 2) ANF::GFP distributed both at the TGN and at the cell periphery (Intermediate), or 3) ANF::GFP excluded from the TGN (Periphery). Scale bar = 5 µm. (D) Percentage of WT and Eipr1 KO cells with the indicated ANF::GFP distribution at the indicated time points. For each data point and genotype, 50–100 cells were counted blindly. The experiment was repeated three times with similar results. The data shown for the WT are the same as shown in Supplemental Figure S5D of Cattin-Ortolá, Topalidou, et al. (2019) because these experiments were run in parallel with the same WT control.

Fig 4: The density of DCVs is not affected in Eipr1 KO cells. (A) Postnuclear supernatants from WT and Eipr1 KO cells were separated by equilibrium sedimentation through 0.6–1.9 M sucrose. Fractions were blotted with an antibody against PC1/3 (mature DCV marker). The data shown are from one representative experiment of three independent experiments with similar results. (B) Band intensity for each fraction was quantified using Fiji, presented as a percentage of total band intensity, and plotted against the sucrose concentration of the fraction.

Fig 5: Membrane association and stability of EARP is disrupted in Eipr1 KO insulin-secreting cells. (A) VPS50 associates with membranes in an EIPR1-dependent manner. In WT cell fractions, endogenous VPS50 was found primarily in the postnuclear P100 membrane fraction. In Eipr1 KO 832/13 cells, VPS50 was equally distributed between the P100 membrane fraction and the S100 cytosolic fraction. ß-Tubulin served as a control soluble protein. S100, P100: supernatant and pellet fractions obtained by a 100,000 × g spin of the cell lysate, containing cytosolic and membrane-associated proteins, respectively. (B) VPS51 associates with membranes in an EIPR1-dependent manner. In WT cell fractions, VPS51::Myc was roughly equally distributed between the postnuclear P100 membrane fraction and the S100 cytosolic fraction, and endogenous VPS51 was found primarily in the P100 membrane fraction. In Eipr1 KO 832/13 cells (Eipr1 KO), VPS51::Myc was found mostly in the soluble fraction and endogenous VPS51 was equally distributed between the membrane fraction and the cytosolic fraction. (C) VPS54 associates with membranes in an EIPR1-independent manner. In both WT and Eipr1 KO 832/13 cell fractions, VPS54::Myc was equally distributed between the postnuclear P100 membrane fraction and the S100 cytosolic fraction. (D) VPS50 and VPS51 levels are reduced in Eipr1 KO cells. Protein extracts from 832/13 (WT), Eipr1 KO 832/13 (Eipr1 KO), and Eipr1 KO 832/13 cells expressing WT Eipr1 (EIPR1[+]) were blotted with antibodies against VPS50 and VPS51. ß-Tubulin served as a loading control. (E) VPS50 interacts with VPS51 in an EIPR1-independent way. EGFP-tagged VPS50 was coexpressed with 13Myc-tagged VPS51 in 832/13 (WT) and Eipr1 KO cells. Immunoprecipitation of VPS50::GFP pulled down VPS51::Myc independently of EIPR1, but did not pull down ß-tubulin. IN, input; IP, immunoprecipitated. (F) VPS50 interacts with VPS53 in an EIPR1-independent way. EGFP-tagged VPS50 was coexpressed with 13Myc-tagged VPS53 in 832/13 (WT) and Eipr1 KO cells. Immunoprecipitation of VPS50::GFP pulled down VPS53::Myc independently of EIPR1, but did not pull down ß-tubulin. IN, input; IP, immunoprecipitated. (G) VPS51 interacts with VPS54 in an EIPR1-independent way. EGFP-tagged VPS54 was coexpressed with 13Myc-tagged VPS51 in 832/13 (WT) and Eipr1 KO cells. Immunoprecipitation of VPS54::GFP pulled down VPS51::Myc independently of EIPR1, but did not pull down ß-tubulin. IN, input; IP, immunoprecipitated. (H) VPS50 and EIPR1 cofractionate on a linear 8–30% sucrose velocity gradient. Fractions from 832/13 (WT) cell lysate were blotted with antibodies against VPS50 and EIPR1. ß-Tubulin served as a control soluble protein. (I) VPS50 fractionates similarly from cell lysates of 832/13 (WT) and Eipr1 KO cells on a linear 8–25% sucrose velocity gradient. Fractions from WT and Eipr1 KO cell lysates were blotted with antibodies against VPS50.