Extended-synaptotagmin 1 engages in unconventional protein secretion mediated via SEC22B+ vesicle pathway in liver cancer

Abstract

Protein secretion in cancer cells defines tumor survival and progression by orchestrating the microenvironment. Studies suggest the occurrence of active secretion of cytosolic proteins in liver cancer and their involvement in tumorigenesis. Here, we investigated the identification of extended-synaptotagmin 1 (E-Syt1), an endoplasmic reticulum (ER)-bound protein, as a key mediator for cytosolic protein secretion at the ER-plasma membrane (PM) contact sites. Cytosolic proteins interacted with E-Syt1 on the ER, and then localized spatially inside SEC22B⁺ vesicles of liver cancer cells. Consequently, SEC22B on the vesicle tethered to the PM via Q-SNAREs (SNAP23, SNX3, and SNX4) for their secretion. Furthermore, inhibiting the interaction of protein kinase Cδ (PKCδ), a liver cancer-specific secretory cytosolic protein, with E-Syt1 by a PKCδ antibody, decreased in both PKCδ secretion and tumorigenicity. Results reveal the role of ER-PM contact sites in cytosolic protein secretion and provide a basis for ER-targeting therapy for liver cancer.

Keywords: PKCδ; SEC22B; cytosolic protein secretion; endoplasmic reticulum (ER); liver cancer.

Fig. 1.

BioID screen to identify putative membrane interactors with cytosolic PKCδ reveals E-Syt1. (A) The 2DICAL analysis of biotinylated proteins purified from the membrane fraction of two HepG2 cells stably expressing PKCδ–BioID2 (clone E5 or A8). The 360 counts of 2DICAL were calculated by comparing doxycycline-treated cells with untreated cells that do not express PKCδ–BioID2. Specificity is defined as having threefold more spectral counts in both E5 and A8 clones. The ER-bound proteins (e.g., E-Syt1 and CLCC1) and the mitochondrial intermembrane protein (mitochondrial antiviral-signaling protein) are identified. Other proteins found at comparable levels were abundant cytoplasmic or nuclear proteins (i, PRKAR2A; ii, TPD52L2; and iii, LBR). (B) Immunoblot analysis of biotinylated proteins purified with streptavidin from lysates in clone E5 or A8 cells; n = 3; independent experiments. Cells were treated with or without 1 μg/mL doxycycline to induce PKCδ–BioID2 expression. Actin was used as the loading control. (C) Confocal micrographs to depict colocalization of endogenous PKCδ and E-Syt1 in HepG2 cells. Images are representative of three independent experiments. (Scale bars, 10 μm.) (Inset) Magnified view of the region in the white box (x 5.5). (D) Doxycycline-inducible PKCδ-GFP-expressing stable HepG2 cells were incubated with 0.5 μg/mL doxycycline for 24 h, fixed, and stained with an antibody to Sec61β as an ER marker. The stained cells were imaged with superresolution microscopy (3D SIM). Images from a single plane (x–y) and 3D reconstructed images (Inset: magnified view of the region in the white box, y–z) are shown. Arrowheads indicate the colocalization between PKCδ-GFP and the ER. Images are representative of three independent experiments. (Scale bars, 5 μm.) (E) Confocal micrographs to detect the interaction with PKCδ and E-Syt1 in HepG2 and AGS cells. Each cell is fixed, reacted with combinations of mouse anti-PKCδ and rabbit anti–E-Syt1 antibodies (PKCδ × E-Syt1), or mouse IgGs and rabbit IgGs (control × control), or mouse anti-PKCδ antibody and rabbit anti-PKCδ antibody (PKCδ × PKCδ), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (n = 20 for HepG2 cells and n = 20 for AGS cells), *P < 0.0001 (two-tailed Mann–Whitney test). Images are representative of three independent experiments. (Scale bars, 10 μm.) (F) Confocal micrographs to detect the interaction with PKCδ and Sec61β (an ER marker) in parental HepG2 (control), E-Syt1 KO HepG2 (E-Syt1 KO), or AGS cells. Each cell is fixed, reacted with a combination of mouse anti-PKCδ and rabbit anti-Sec61β antibodies (PKCδ × Sec61β), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (n = 20 for control HepG2 cells and n = 29 for E-Syt1 KO HepG2 cells and n = 20 for AGS WT cells), *P < 0.0001 (two-tailed Mann–Whitney test). Images are representative of three independent experiments. (Scale bars, 10 μm.)

Fig. 2.

E-Syt1 is necessary for PKCδ secretion. (A) Immunoblot analysis of lysates and media from control or E-Syt1 KO HepG2 cells showed that depletion of E-Syt1 reduces PKCδ level in media; n = 3; independent experiments. The relative signal density was quantified, and data are shown as mean ± SD, *P < 0.0001 (two-tailed Student’s t test); n.s., not significant. GAPDH and Ponceau-S stain were used as the loading control for lysates or media, respectively. (B) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr) or E-Syt1 siRNA for 24 h; n = 4 independent experiments. Luminescence was measured after cells were recultured with a medium containing 0.5 μg/mL doxycycline for 24 h. Data are shown as mean ± SD, *P = 0.0053 (Welch’s t test). (C) Schematics of human PKCδ WT and deletion mutants (Δ451–676 and Δ601–676). These PKCδ constructs are fused with BioID2 and HA-epitope tag. (D) Immunoblot analysis of biotinylated proteins purified with streptavidin and lysates in doxycycline-inducible HepG2 cells transfected with empty or WT, Δ451–676 or Δ601–676 of PKCδ–BioID2–hemagglutinin (HA) vector to show a weak interaction with E-Syt1 when the C-terminal region of PKCδ is deleted; n = 3; independent experiments. Cells were treated with 1 μg/mL doxycycline for 24 h. Representative blot is shown. Actin was used as the loading control. (E) Confocal micrographs of doxycycline-inducible HepG2 cells transfected with WT, or Δ601–676 of PKCδ–BioID2–HA vector showing lack of colocalization between PKCδ and Sec61β (an ER marker). Images are representative of two independent experiments. (Scale bar, 10 μm.) (F) Immunoblot analysis of lysates and media in doxycycline-inducible HepG2 cells transfected with WT or Δ601–676 of PKCδ–BioID2–HA vector to show less detection of Δ601–676 of PKCδ in media; n = 3; independent experiments. Cells were treated with 1 μg/mL doxycycline for 24 h. Actin was used as the loading control. (G) PKCδ KO HepG2 cells expressing PKCδ WT or Δ601–676 were inoculated subcutaneously into nude mice (n = 6 per group). Tumor size was monitored. (H) Microscopic images and tumor weight of PKCδ KO HepG2 expressing PKCδ WT or Δ601–676 (n = 6 per group) tumors. Error bars, mean ± SD, *P = 0.022 (two-tailed Mann–Whitney test).

Fig. 3.

Autophagy-related proteins are utilized for PKCδ secretion. (A) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells preincubated with 0.5 μg/mL doxycycline in 10% FBS-containing medium for 24 h, and stimulated with 10% FBS-containing medium or EBSS medium for 6 h; n > 3 independent experiments. Data are shown as mean ± SD, *P < 0.0001 (two-tailed Student’s t test). (B) Flow cytometric analysis of doxycycline-treated HepG2 cells incubated under 10% FBS-containing (gray area), 0.1% FBS-containing (black dashes), or EBSS (pink dashes) medium conditions for 6 h to show no leakage of PKCδ by disruption of the plasma membrane under each medium condition. Representative data are shown from three individual experiments. (C) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr), ATG5, ATG7, ATG16L1, p62, LC3, or PKCδ (as a positive control) siRNAs (2 nM) for 48 h; n = 3 independent experiments. Luminescence was measured after the cells were recultured with a medium containing 0.5 μg/mL doxycycline for 24 h. Data are shown as mean ± SD, *P < 0.001; **P = 0.0292; ***P = 0.0022; ****P < 0.0001 (ANOVA). (D) Immunoblot analysis of lysates and media from doxycycline-inducible HepG2 cells treated with scrambled (Scr), ATG5, or LC3 siRNAs (2 nM) for 48 h to show comparable inhibition of secretion on PKCδ-HiBiT fusion and endogenous PKCδ. Representative images are shown from three individual experiments. GAPDH and Ponceau-S stain were used as the loading control for lysates and media, respectively. (E) Confocal micrographs showing localization of endogenous PKCδ and LC3B in HepG2 cells cultured with 10% FBS-containing medium. Images are representative of three independent experiments. (Inset) Magnified view of the region in the white box (x 10). Black arrows indicate colocalization between PKCδ and LC3B. (Scale bars, 10 μm.) (F) Confocal micrographs to show the interaction with PKCδ, E-Syt1, and LC3B in doxycycline-inducible PKCδ–GFP-expressing HepG2 cells. Cells were treated with 1 μg/mL doxycycline for 24 h, fixed, reacted with a combination of mouse anti-LC3B and rabbit anti–E-Syt1 antibodies (LC3B × E-Syt1), and subjected to Duolink in situ PLA. White arrows indicate colocalization of PKCδ and PLA signals. Images are representative of two independent experiments. (Scale bars, 10 μm.) (G) Immunoblot analysis of biotinylated proteins purified with streptavidin in clone E5 cells; n = 3; independent experiments. Cells were treated with or without 1 μg/mL doxycycline. (H) Confocal micrographs to detect the interaction with PKCδ and LC3B (an autophagosome marker) in HepG2 or AGS cells. Each cell is fixed, reacted with a combination of mouse anti-PKCδ and rabbit anti-LC3B antibodies (PKCδ × LC3B), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (n = 20 for HepG2 cells and n = 20 for AGS cells), *P < 0.0001 (two-tailed Mann–Whitney test). Images are representative of three independent experiments. (Scale bars, 10 μm.) (I) Confocal micrographs to detect the interaction with LC3B and SEC22B or Sec61β (an ER marker) in parental HepG2 (control), E-Syt1 KO HepG2 (E-Syt1 KO) cells. Each Cell is fixed, reacted with a combination of mouse anti-LC3B and rabbit anti-SEC22B or Sec61β antibodies (LC3B × SEC22B or LC3B × Sec61β), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (LC3B × SEC22B; n = 26 for control cells and n = 26 for E-Syt1 KO cells, LC3B × Sec61β; n = 25 for control cells and n = 24 for E-Syt1 KO cells), *P < 0.0001 (two-tailed Mann–Whitney test). Images are representative of two independent experiments. (Scale bars, 10 μm.)

Fig. 4.

PKCδ is secreted via SEC22B⁺ vesicle. (A) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr) or SEC22B siRNAs (2 nM) for 48 h; n = 3; independent experiments. Luminescence was measured after cells were recultured with a medium containing 0.5 μg/mL doxycycline for 24 h. Data are shown as mean ± SD, *P = 0.0223 (two-tailed Student’s t test). (B) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible SEC22B KO HepG2 cells; n = 3; independent experiments. Luminescence was measured after cells were recultured with a medium containing 0.5 μg/mL doxycycline for 24 h. Data are shown as mean ± SD, *P = 0.0002 (two-tailed Student’s t test). (C) Confocal micrographs to indicate the proximity with PKCδ and SEC22B in parental HepG2 (control), E-Syt1 KO HepG2 (E-Syt1 KO), or AGS cells. Each cell is fixed, reacted with a combination of mouse anti-PKCδ and rabbit anti-SEC22B antibodies (PKCδ × SEC22B), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (n = 20 for control HepG2 cells and n = 31 for E-Syt1 KO HepG2 cells, and n = 20 for AGS WT cells), *P < 0.0001 (two-tailed Mann–Whitney test). Images are representative of three independent experiments. (Scale bars, 10 μm.) (D) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr) or GRASP55 siRNAs (2 nM) for 48 h; n = 4 independent experiments. Luminescence was measured after cells were recultured with normal 10% FBS-containing medium or EBSS with 0.5 μg/mL doxycycline for 24 h. Data are shown as mean ± SD, *P = 0.0027, **P = 0.0112, ***P = 0.0061 (ANOVA), n.s., not significant. (E) Electron micrographs of HepG2 cells showing the existence of PKCδ (arrows) in SEC22B⁺ (arrowheads) vesicles at the vicinity of the PM (Left) and moment of the vesicle-PM fusion and secretion of PKCδ (Right). Images are representative and two similar independent experiments were performed. (Scale bars, 100 nm.) (F) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr), SNAP23, STX3, STX4, or a mixture with STX3 and STX4 siRNAs (2 nM) for 48 h; n > 3; independent experiments. Luminescence was measured after the cells were recultured with a medium containing 0.5 μg/mL doxycycline for 24 h. Data are shown as mean ± SD, *P < 0.0001 (ANOVA). (G) Images of tumor tissues of patients with hepatocellular carcinoma to indicate tumor cell-specific colocalization of PKCδ with SEC22B using Duolink in situ PLA (Center and Right). Tumor (T) and nontumor (NT) lesion were defined by evaluating H&E staining of each tumor section (Left). Two lines of images are representative in sections of five patients with liver cancer. (Scale bars, 50 μm.) (H) Confocal micrographs to detect the interaction with PKCδ and SEC22B or Stx17 in HepG2 cells treated with DMSO or BafA1 (100 nM) for 6 h. Each Cell is fixed, reacted with a combination of mouse anti-PKCδ and rabbit anti-SEC22B or Stx17 antibodies (PKCδ x SEC22B or PKCδ × Stx17), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (PKCδ × SEC22B; n = 30 for DMSO-treated and n = 30 for BafA1-treated cells, PKCδ × Stx17; n = 43 for DMSO-treated and n = 44 for BafA1-treated cells,), *P < 0.0001 (two-tailed Student’s test), n.s., not significant. Images are representative of two independent experiments. (Scale bars, 10 μm.)

Fig. 5.

Other cytoplasmic proteins are secreted via the same autophagy-related system. (A) Confocal micrographs showing colocalization of endogenous importin α1 with PDI or Calnexin in HepG2 cells. Images are representative of three independent experiments. (Scale bars, 10 μm.) (Insets) Magnified views of the regions in the white boxes, Upper: (x 10), Lower: (x 5). (B) Confocal micrographs showing colocalization of endogenous NCL with ER using ERseeing in HepG2 cells. Images are representative of three independent experiments. White arrows indicate colocalization of NCL with the ER. (Scale bar, 10 μm.) (C) Importin α1 secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr), E-Syt1, ATG5, LC3, or SEC22B siRNAs (2 nM) for 48 h; n = 4 independent experiments. Luminescence was measured after cells recultured with medium containing 1 μg/mL doxycycline for 24 h. Data are shown as means ± SD, *P < 0.0002; **P = 0.0018; ***P = 0.0104; ****P = 0.0001 (ANOVA). (D) NCL secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with scrambled (Scr), E-Syt1, ATG5, LC3, or SEC22B siRNAs (2 nM) for 48 h; n = 4 independent experiments. Luminescence was measured after cells recultured with medium containing 1 μg/mL doxycycline for 24 h. Data are shown as means ± SD, *P < 0.001; **P = 0.0022; ***P = 0.0007; ****P = 0.0003 (ANOVA). (E) Confocal micrographs to detect the interaction with importin α1 and indicated proteins in HepG2 cells. Cells are fixed, reacted with combination of mouse anti-importin α1 or LC3B and rabbit antibodies, and subjected to Duolink in situ PLA. The images are representative of two independent experiments. (Scale bars, 10 μm.) (F) Confocal micrographs to detect the interaction with NCL and LC3B or SEC22B in HepG2 cells. Cells are fixed, reacted with combination of mouse anti-LC3B or SEC22B and a rabbit anti-NCL antibody, and subjected to Duolink in situ PLA. The images are representative of two independent experiments. (Scale bars, 10 μm.) (G) A model for the role of E-Syt1 as a link between the ER and vesicle-mediated secretion of cytosolic proteins. Packaging of cytosolic proteins into SEC22B⁺ vesicles is apparently dependent on the expression of E-Syt1 and autophagy-related proteins. E-Syt1 may form a specific complex for cytosolic protein secretion at the ER in liver cancer cells.

Fig. 6.

Targeting interaction between PKCδ and E-Syt1 inhibits liver cancer cell growth. (A) Images of intracellular delivery of PKCδ antibody (Ab) (C-20) using Ab-DeliverIN transfection regent (DeliverIN; as a drug delivery system) in HepG2 cells. The epitope of the C-20 Ab on PKCδ is a C-terminal sequence of PKCδ, which was required for binding to E-Syt1 and the secretion (Fig. 2 D and F and SI Appendix, Fig. S8A). No intracellular accumulation of the Ab in HepG2 cells treated with only C-20 Ab is confirmed in SI Appendix, Fig. S8B. (Scale bars, 10 μm.) (B) Confocal micrographs of HepG2 cells treated with DeliverIN and control IgGs or C-20 Ab showing the inhibition of the interaction with PKCδ and E-Syt1 in intracellular delivery of the C-20 Ab. Each cell is fixed, reacted with a combination of mouse anti-PKCδ and rabbit anti–E-Syt1 antibodies (PKCδ × E-Syt1), and subjected to Duolink in situ PLA. Data are shown as mean ± SD (n = 20 for the group of treatment with DeliverIN and control IgG, n = 20 for the group of treatment with DeliverIN and the C-20 Ab), *P < 0.0001 (two-tailed Student’s t test). Images are representative of three independent experiments. (Scale bars, 10 μm.) (C) PKCδ secretion measured by HiBiT extracellular assay in doxycycline-inducible HepG2 cells treated with DeliverIN and control IgGs or C-20 Ab for 24 h to show the suppression of PKCδ secretion by the intracellular delivery of the C-20 Ab; n = 4; independent experiments. 0.5 μg/mL doxycycline was simultaneously added when DeliverIN and antibodies were treated. Data are shown as mean ± SD, *P < 0.0001 (two-tailed Student’s t test). (D–G) Proliferation assay by cell count of the indicated cell lines treated with control or C-20 Ab in the presence of DeliverIN for 48 h showing that the cell proliferative reduction by treatment with DeliverIN and the C-20 Ab specifically occurs in cell lines that secrete PKCδ. Haemocytometer was used for cell counting after trypan-blue staining to exclude dead cells, and each plot was an average of four fields of view. Similar experiments were independently repeated more than three times. Data are shown as mean ± SD (D) *P = 0.0047 (Welch’s t test); (E) *P = 0.0129 (Welch’s t test); (F) *P = 0.0153 (Welch’s t test); (G) n.s., not significant. (H) Images of HepG2 3D multicellular spheroid treated with control IgGs or C-20 Ab in the presence/absence of DeliverIN to show the anti-tumorigenic effect of the C-20 Ab with DeliverIN. (Scale bars, 20 μm.) (I) The number of HepG2 3D multicellular spheroids was counted based on the criterion of size > 25 μm. Three independent experiments were performed. Data are shown as mean ± SD, *P = 0.002 (ANOVA); n.s., not significant.