Sec16A is critical for both conventional and unconventional secretion of CFTR

Abstract

CFTR is a transmembrane protein that reaches the cell surface via the conventional Golgi mediated secretion pathway. Interestingly, ER-to-Golgi blockade or ER stress induces alternative GRASP-mediated, Golgi-bypassing unconventional trafficking of wild-type CFTR and the disease-causing ΔF508-CFTR, which has folding and trafficking defects. Here, we show that Sec16A, the key regulator of conventional ER-to-Golgi transport, plays a critical role in the ER exit of protein cargos during unconventional secretion. In an initial gene silencing screen, Sec16A knockdown abolished the unconventional secretion of wild-type and ΔF508-CFTR induced by ER-to-Golgi blockade, whereas the knockdown of other COPII-related components did not. Notably, during unconventional secretion, Sec16A was redistributed to cell periphery and associated with GRASP55 in mammalian cells. Molecular and morphological analyses revealed that IRE1α-mediated signaling is an upstream regulator of Sec16A during ER-to-Golgi blockade or ER stress associated unconventional secretion. These findings highlight a novel function of Sec16A as an essential mediator of ER stress-associated unconventional secretion.

Figure 1. Sec16A, but not core COPII components, is required for unconventional secretion of CFTR.

(a–d) Surface biotinylation of CFTR. HEK293 cells were transfected with plasmids expressing wild-type (WT) (a,b) or ΔF508 CFTR (c,d), and a surface biotinylation assay was performed 24 h after transfection. Some cells were cotransfected with the Arf1-Q71L plasmid to induce ER-to-Golgi blockade. The cells were pretreated with scrambled or Sec16A-specific siRNAs (100 nM) 24 h before plasmid transfection. Cell surface-specific labeling of proteins was confirmed by the absence of the cytosolic protein aldolase A in the biotinylated fraction. Representative surface biotinylation assays are shown (a,c), and the results of multiple experiments (n = 4) are summarized (b,d). Sec16A knockdown abolished both the conventional cell-surface trafficking of the Golgi complex-glycosylated WT-CFTR (band C) and the unconventional cell-surface trafficking of the ER core-glycosylated WT, as well as ΔF508 CFTRs (band B) induced by Arf1-Q71L. (e,f) Effects of COPII depletion on the Arf1-Q71L-induced cell-surface expression of ΔF508-CFTR. Effects of individual knockdown of each isoform are shown in (e) and those of combinatorial knockdown of the same gene family are shown in (f). The results of multiple experiments (n = 3–7) are summarized, and images of representative surface biotinylation assays are shown in Figure S2. Knockdown of Sec16B and core COPII components (Sec23, Sec24, Sec13, and Sec31) did not affect the unconventional cell-surface transport of ΔF508-CFTR. The depletion of Sec16A and COPII components by siRNA was confirmed by immunoblotting or mRNA quantitation (Figure S3). Data are shown as mean ± SEM. ns: not significant, **P < 0.01.

Figure 2. Sec16A silencing inhibits unconventional ΔF508-CFTR trafficking to the cell surface.

(a–d) Representative immunofluorescence images of wild-type (WT) or ΔF508-CFTR. Extracellular loop HA-tagged WT (a) and ΔF508 (b–d) CFTRs were expressed in HeLa cells. CFTR at the cell surface was immunostained with anti-HA antibodies before membrane permeabilization (green), and then the total CFTR was stained with anti-R4 CFTR antibodies after permeabilization (red). Some cells coexpressed Myc-Sar1-T39N to induce unconventional surface expression of ΔF508-CFTR (c,d). The expression of the Sar1-T39N mutant was confirmed by anti-Myc staining. The cells were pretreated with scrambled (b,c) or Sec16A-specific (d) siRNAs (100 nM) 24 h before plasmid transfection. Cells expressing Sar1-T39N are marked with white dotted lines. Arrows indicate surface expression of WT or ΔF508 CFTRs and arrowheads indicate cells that do not express surface CFTR. (e) Quantification of surface CFTR intensity. Data are shown as mean ± SEM from three independent experiments (each comprising analyses of 20–30 cells). Sec16A depletion abolished the cell-surface expression of ΔF508-CFTR induced by Sar1-T39N. Scale bar: 10 μm, **P < 0. 01.

Figure 3. Relocalization of Sec16A in Arf1-Q71L-induced unconventional secretion of ΔF508-CFTR.

(a,b) Intracellular localization of Sec16A was analyzed using immunocytochemistry. HeLa cells were transfected with plasmids expressing extracellular loop HA-tagged ΔF508-CFTR with or without Arf1-Q71L. CFTR at the cell surface was immunostained with anti-HA antibodies before membrane permeabilization (green). Then, the total CFTR was stained with anti-M3A7 CFTR antibodies (grey), and Sec16A was stained with anti-KIAA0310 Sec16A antibody (red). In control cells (a), Sec16A puncta were mostly localized in the juxtanuclear area (arrowhead). (b) ER-to-Golgi blockade by Arf1-Q71L evoked the cell-surface expression of ΔF508-CFTR (green) and the redistribution of Sec16A to entire cellular area (arrows, red). (c,d) Sec16A redistribution in cells coexpressing HA-Arf1-Q71L (white dotted line) was compared with that in cells not expressing HA-Arf1-Q71L. Representative immunofluorescence images are shown (c), and the results of multiple experiments (n = 5, each comprising analyses of 5–10 cells, **P < 0.01) are summarized (d). Scale bar: 5 μm, **P < 0.01.

Figure 4. Relocalization of Sec16A in GRASP-mediated unconventional secretion of ΔF508-CFTR.

(a–c) Cellular distribution of Sec16A was analyzed using immunocytochemistry. HeLa cells were transfected with mock or GRASP55-Myc plasmids. Representative immunofluorescence images are shown. Endogenous GRASP55 (a) and a low level of GRASP55-Myc (b), which does not induce unconventional secretion (Figure S5a), were expressed at the perinuclear Golgi apparatus, and Sec16A was enriched around that region. In contrast, Sec16A puncta and GRASP55 were redistributed to the entire cellular area in cells with a high level of GRASP55-Myc expression (c), which does induce unconventional secretion of ΔF508-CFTR (Figure S5b). (d) Quantification of the ratio of the Sec16A (+) area versus the total cell area in multiple experiments (mean ± SEM, n = 5, each comprising analyses of 5–10 cells) are summarized. Scale bar: 10 μm, **P < 0.01: difference from control.

Figure 5. Relocalized Sec16A colocalizes with the ER marker proteins.

(a–c) Cellular distribution of Sec16A was analyzed using immunocytochemistry in HeLa cells. Cells were transfected with plasmids encoding the ER marker protein ER-yellow fluorescent protein (ER-YFP) with or without Arf1-Q71L. Representative immunofluorescence images are shown (a,b), and the results of multiple experiments are summarized (c). Analyses using Manders’ colocalization coefficient (MCC) show that ER-to-Golgi blockade by Arf1-Q71L increased the extent of correlation between ER-YFP and Sec16A (c; mean ± SEM, n = 6, each comprising analyses of 5–10 cells). (d–g) Cells were transfected with mock (d), GRASP55-Myc (e), or Myc-Sar1-T39N plasmids (f), and the ER marker protein calnexin was co-immunostained with Sec16A. GRASP55-Myc and Myc-Sar1-T39N were immunostained with anti-Myc antibody. Representative immunofluorescence images are shown (d–f), and the results of multiple experiments are summarized (g). White dotted line shows cell periphery of GRASP55-Myc or Sar1-T39N expressing cells. Arrows indicated the redistribution of Sec16A to entire cellular area by ER-to-Golgi blockade or GRASP55 overexpression. The MCC analyses show that GRASP55 overexpression and ER-to-Golgi blockade by Sar1-T39N increased the extent of correlation between calnexin and Sec16A (g; mean ± SEM, n = 6, each comprising analyses of 5–10 cells). Scale bar: 10 μm. **P < 0.01: difference from each control.

Figure 6. Sec16A colocalizes and interacts with GRASP55.

(a–d) The cellular localization of Sec16A and GRASP55 was analyzed using immunocytochemistry in HeLa cells. Representative immunofluorescence images are shown. Endogenous GRASP55 was labeled with fluorophore-tagged antibodies (green), and Sec16A was labeled with fluorophore-tagged antibodies (red). Some cells were cotransfected with Arf1-Q71L (c) or Sar1-T39N (d) to induce ER-to-Golgi blockade. The induction of ER stress (thapsigargin) (b) or ER-to-Golgi blockade (Arf1-Q71L or Sar1-T39N) increased the colocalization of Sec16A and GRASP55 in both the perinuclear (green box) and the peripheral (red box) regions. (e) Analyses using Manders’ colocalization coefficient (MCC) show that ER stress (thapsigargin) and ER-to-Golgi blockade (Arf1-Q71L or Sar1-T39N) significantly increased the extent of correlation between Sec16A and GRASP55 in both the perinuclear and peripheral regions (mean ± SEM, n = 5, each comprising analyses of 5–10 cells). (f,g) Coimmunoprecipitation experiments with Sec16A and GRASP55 were performed in HEK293 cells. A representative coimmunoprecipitation assay is shown (f), and the results of multiple experiments (n = 3) are summarized (g). Protein samples were precipitated with anti-Sec16A (KIAA0310) and blotted with anti-GRASP55. ER-to-Golgi blockade by Arf1-Q71L increased the association between Sec16A and GRASP55. Scale bar: 10 μm, **P < 0.01: difference from each control.

Figure 7. ER-to-Golgi blockade and GRASP55 overexpression do not relocalize Sec31A.

(a–d) The cellular localization of Sec31A, a core COPII component, was compared with that of Sec16A using immunocytochemistry in HeLa cells. Representative immunofluorescence images are shown. Some cells were cotransfected with GRASP55-Myc (b), Arf1-Q71L (c), or Sar1-T39N (d) to induce Sec16A redistribution. (e) Quantification of the ratio of the Sec16A (+) or Sec31A (+) area versus the total cell area in multiple experiments (mean ± SEM, n ≥ 5, each comprising analyses of 5–10 cells) are summarized. GRASP55 overexpression, Arf1-Q71L, or Sar1-T39N-induced ER-to-Golgi blockade caused no significant alterations of Sec31A localization. Scale bar: 10 μm, **P < 0.01: difference from each control.

Figure 8. IRE1α depletion inhibits unconventional ΔF508-CFTR secretion by down-regulating Sec16A.

(a,c) Surface biotinylation of CFTR was performed in HEK293 cells after the induction of unconventional ΔF508-CFTR secretion by Arf1-Q71L (a) or by GRASP55 overexpression. (c) The cells were pretreated with scrambled or IRE1α-specific siRNAs (100 nM) 24 h before plasmid transfection. (b,d) The results of multiple experiments (n = 3) are summarized. IRE1α depletion by siRNA inhibited the Arf1-Q71L-induced or GRASP55-induced surface expression of ΔF508-CFTR. Supplementation with exogenous Sec16A rescued the effects of Arf1-Q71L and GRASP55, which induced the cell-surface expression of ΔF508-CFTR. **P < 0.01.

Figure 9. IRE1α depletion inhibits Sec16A redistribution.

(a–d) The cellular localization of Sec16A was analyzed using immunocytochemistry in control HeLa cells (a,b) and in cells after the induction of unconventional ΔF508-CFTR secretion by Arf1-Q71L (c,d). The cells were pretreated with scrambled or IRE1α-specific siRNAs (100 nM) 24 h before plasmid transfection. The intracellular distribution of Sec16A in IRE1α-depleted cells was analyzed using confocal images taken with five times higher acquisition gain (Gain X5) because IRE1α depletion reduced the overall Sec16A expression levels (b,d). (e) The results of multiple experiments (n ≥ 5, each comprising analyses of 5–10 cells) are summarized. IRE1α depletion inhibited the Arf1-Q71L-induced redistribution of Sec16A. Scale bar: 10 μm, **P < 0.01: difference from control.