Fig 2: Erythroid defects observed in SEC23B-deficient HUDEP-2 cells are rescued in a dose-dependent manner by increased expression of SEC23A using CRISPR-activation (CRISPRa).(A) Two independent SEC23A-targeting CRISPRa gRNAs were individually transduced into a HUDEP-2 reporter cell line expressing GFP-tagged SEC23A from the endogenous SEC23A locus. GFP fluorescence was quantified by flow cytometry at days 0 and 11 of differentiation and compared to cells transduced with a NT control gRNA. (B) Viability and (C) fold expansion of SEC23B-deficient HUDEP-2 cells (n = 4 independent clonal lines) are rescued by increased expression of SEC23A, in a dose-dependent manner. (D and E) SEC23B-deficient HUDEP-2 cells were evaluated at days 7 and 11 of differentiation and found to exhibit increased annexin V staining, impaired up-regulation of surface CD233, and increased % of cells at the G2-M phase of the cell cycle (consistent with increased binucleated cells). All these defects were ameliorated by activation of SEC23A, using either sgRNA1 or sgRNA2, mostly in a dose-dependent manner. Statistical tests: Two-sided, unpaired t tests were used to compare means. P values: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001. P values not significant (n.s.) unless otherwise marked. Data represent means ± SD.

Fig 3: Mice with combined erythroid-specific SEC23A and SEC23B deficiency.(A) Mice with combined erythroid-specific (EpoR-Cre) Sec23a/Sec23b deletion die at mid-embryogenesis (E12.5) exhibiting a smaller size and pale appearance compared to their WT littermate controls and (B) dyserythropoiesis (black arrows) on histological evaluation (E10.5).

Fig 4: Misfolded protein engages DR5 at the ER-Golgi intermediate compartment, inducing active DR5 signaling clusters.(A) Top: Immunofluorescence of HCT116 cells transfected with MPZ-GFP (green) for 24 hr and stained with anti-DR5 (red, scale bar 5 µm). Bottom: Enlargements of the inset stepping through the z-plane in 0.5 µm increments (scale bar 2 µm). (B) Subcellular fractionation of lysate expressing MPZ-GFP, where IRE1 marks the ER, Giantin marks the Golgi, Sec31A and Sec23A mark COPII vesicles, and ERGIC53 and RCAS1 correspond to ER-Golgi intermediate compartment. Bands of the expected size are indicated by “– “and bands that may represent a modified or degraded protein are indicated by *. (C) Average caspase activity of each fraction from subcellular gradient centrifugation in (B) normalized to total lysate (input) measured by caspase 8 substrate luminescence (n = 3 biological replicates, error bars = SEM; ns1 indicates p=0.079, * denotes p=0.015, and ns indicates p=0.31 from unpaired t-tests with equal SD). (D) Top: Immunostaining of DR5 and ERGIC53 in fixed HCT116 cells transfected with MPZ-GFP for 24 hr as in (A). Bottom: Merged images with ERGIC53 in magenta or cyan to depict overlapping signal as white (scale bar = 5 µm, insets scale bar = 2 µm). (E) Immunostaining of DR5 and giantin in fixed HCT116 cells expressing MPZ-GFP. Giantin is magenta in the overlay with MPZ (green) or cyan in the overlay with DR5 (red). Bottom row enlarges the inset marked in the merges images to show little overlapping signal with giantin (scale bar = 5 µm, inset scale bar = 1 µm). (F) Box-whisker plots quantifying the Pearson’s correlation per cell between DR5 and ERGIC53 (mean = 0.61 ± 0.03) or giantin (mean = 0.14 ± 0.02) within MPZ-positive cells (N > 55), where whiskers correspond to minimum and maximum values of the data (**** indicates p<0.001).Figure 2—source data 1.Caspase activity for fractions of iodixanol gradient.This excel file contains the caspase glo 8 luminescent units of the fractionation samples (n = 3 biological replicates) shown in Figure 2C.

Fig 5: Erythroid phenotypes of mice with deletion of three of the four Sec23 alleles using the EpoR-cre.(A) Mice with combined erythroid-specific Sec23a haploinsufficiency and Sec23b deletion (Sec23ahet + Sec23bko) die in late embryogenesis, with E14.5 embryos exhibiting pale appearance, (B) reduced fetal liver erythropoiesis by both H&E staining and (C) flow cytometry analysis, as well as (D) impaired erythroid maturation in the fetal liver as determined by flow cytometry (see fig. S2B for gating strategy); n = 6 for WT controls, n = 4 for erythroid Sec23ahet + Sec23bko. (E) Mice with combined erythroid-specific Sec23a deletion and Sec23b haploinsufficiency (Sec23ako + Sec23bhet) are viable and exhibit a ~1.0 g/dl reduction in hemoglobin level compared to WT littermate controls and (F) impaired terminal erythroid maturation in the BM as determined by flow cytometry (n = 5 for each group). (G) SEC23B is more abundant than SEC23A in murine erythroid cells. Data were obtained from Gautier et al. (28). The 0 hour of differentiation consisted primarily of erythroid colony-forming units and proerythroblasts, while hour 60 of differentiation consisted primarily of orthochromatic erythroblasts and reticulocytes. (H) The severity of the erythroid phenotype in the mouse is inversely correlated with the total SEC23 level in erythroid cells. Statistical tests: Two-sided, unpaired t tests were used in (C) and (E); two-way ANOVA with Sidak’s multiple comparisons were used in (D) and (F). P values: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. P values not significant (n.s.) unless otherwise marked. Data represent means ± SD.