SFT-4/Surf4 control ER export of soluble cargo proteins and participate in ER exit site organization

Abstract

Lipoproteins regulate the overall lipid homeostasis in animals. However, the molecular mechanisms underlying lipoprotein trafficking remain poorly understood. Here, we show that SFT-4, a Caenorhabditis elegans homologue of the yeast Erv29p, is essential for the endoplasmic reticulum (ER) export of the yolk protein VIT-2, which is synthesized as a lipoprotein complex. SFT-4 loss strongly inhibits the ER exit of yolk proteins and certain soluble cargo proteins in intestinal cells. SFT-4 predominantly localizes at ER exit sites (ERES) and physically interacts with VIT-2 in vivo, which suggests that SFT-4 promotes the ER export of soluble proteins as a cargo receptor. Notably, Surf4, a mammalian SFT-4 homologue, physically interacts with apolipoprotein B, a very-low-density lipoprotein core protein, and its loss causes ER accumulation of apolipoprotein B in human hepatic HepG2 cells. Interestingly, loss of SFT-4 and Surf4 reduced the number of COPII-positive ERES. Thus, SFT-4 and Surf4 regulate the export of soluble proteins, including lipoproteins, from the ER and participate in ERES organization in animals.

Figure 1.

SFT-4 is required for efficient ER export of VIT-2 from intestinal cells. (A–C) Loss of sft-4 caused high accumulation of VIT-2 in intestinal cells. (A–A″) In mock-treated animals, VIT-2–GFP is detected on certain small punctate structures in intestinal cells and is prominently accumulated in oocytes (arrows). (B–B″) VIT-2–GFP is highly accumulated in sft-4(RNAi) intestinal cells (arrowheads) but is almost undetectable in oocytes (arrows). (C) The amount of VIT-2–GFP in intestinal cells was significantly increased after knockdown of sft-4. Fluorescence signal intensities per unit area were measured and statistically analyzed using Student’s t test; ***, P < 0.001; error bars: SEM (n = 34 and 31 intestines from mock and sft-4(RNAi) animals, respectively). Dotted lines indicate the outlines of worm bodies (A and B) or intestines and oocytes (other panels). Regions surrounded by squares are enlarged (16×) in insets. Bars: (A, A′, B, and B′) 50 µm; (A″ and B″) 10 µm; (insets) 5 µm. (D and E) Subcellular localization of VIT-2 in mock and sft-4(RNAi) intestinal cells. Intestines of transgenic animals coexpressing VIT-2–tdimer2 and GFP-SP12 (D) or ssGFP-HDEL (E) are shown. L3 larvae were treated with RNAi for 2 d in the case of sar-1(RNAi). VIT-2–tdimer2 localizes to granular structures in the ER lumen labeled with ssGFP-HDEL in sft-4(RNAi) or sar-1(RNAi) intestinal cells. Dotted lines indicate the outlines of intestines. Regions surrounded by squares are enlarged (16×) in insets. Bars: 10 µm; (insets) 5 µm.

Figure 2.

Loss of sft-4 results in VIT-2 accumulation in the ER lumen. (A and B) The intestines of transgenic worms expressing VIT-2–GFP or ssGFP-HDEL were dissected and stained with LipidTOX Red, which labels neutral lipids. In mock animals, LipidTOX Red mainly stained cytosolic LDs (A and B) and weakly stained the ER lumen (arrowheads). In sft-4(RNAi) animals, the accumulated VIT-2–GFP and ssGFP-HDEL overlapped with LipidTOX Red–positive regions (A and B, arrows). Cytosolic LDs, which appeared as multiple punctate structures, were superficially normal. Dotted lines indicate the outlines of intestines. Regions surrounded by squares are enlarged (16×) in insets. Bars: 10 µm; (insets) 5 µm. (C) Electron microscopy analysis of intestinal cross sections from WT and sft-4(RNAi) animals. Mock-treated animals showed normal intestinal organelle distribution, such as the tubular structures of the ER, orderly alignment of microvilli, and several vesicles and granules. The large electron-dense granules (asterisks) likely contain yolk components. In contrast, sft-4(RNAi) animals showed increased numbers of electron-dense granules juxtaposed to electron-lucent granules (arrowheads) in the cytoplasm. Notably, these structures appear to be surrounded by a membrane. Bars, 1 µm.

Figure 3.

SFT-4 is required for transport from the ER of CPL-1, but not SYN-1 or PGP-1. (A) ER export of CPL-1–YFP is impaired by RNAi of sft-4. Subcellular localization of CPL-1–YFP and DsRed-KDEL was examined in the intestines of mock- and sft-4(RNAi)–treated animals. (B) Intestines of transgenic animals expressing VIT-2–GFP were stained with an antibody against SYN-1. When sft-4 was depleted, SYN-1 was not markedly affected: although VIT-2–GFP was strongly accumulated in the ER (arrowheads), SYN-1 mainly localized to the basolateral membrane (arrows). (C) Subcellular localization of GFP–PGP-1 in the intestinal cells of mock- or sft-4(RNAi)–treated animals. GFP–PGP-1 was mostly transported to the apical plasma membrane even in sft-4(RNAi) animals. Dotted lines indicate the outlines of intestines. Regions surrounded by squares are enlarged (4× in A and 16× in B) in insets. Bars: 10 µm; (insets) 5 µm.

Figure 4.

SFT-4 localizes at ERES and affects the ERES organization. (A and A′) GFP–SFT-4 localizes to reticular network structures in the middle of intestinal cells (A) and punctate structures beneath the cell surface (A′; arrows). GFP–SFT-4 expression was driven by sft-4 promoter. (B) Immunostaining of endogenous SFT-4 shows reticular and punctate (arrows) structures as observed with GFP–SFT-4. (C–C″) GFP–SFT-4 localizes to the ER membrane, labeled here with mC-SP12. GFP–SFT-4 and mCherry-SP12 (mC-SP12) were coexpressed in the intestine and examined. (D–D″) SFT-4 localizes at ERES. The intestines of transgenic animals expressing SEC-23–GFP were stained with anti–SFT-4 antibody. SFT-4 colocalizes with SEC-23–GFP (D″; arrows). (E–G) Loss of SFT-4 reduces the number and size of SEC-23–GFP–positive ERES. SEC-23–GFP was localized to punctate structures, which are presumably ERES in mock animals (E). The number and size of SEC-23–GFP–positive ERES was reduced in sft-4(RNAi) intestinal cells (E, arrowheads). The number (F) and size (G) of SEC-23–GFP–positive ERES in mock and sft-4(RNAi) animals were measured (n = 7 and 10 intestines from mock and sft-4(RNAi) animals, respectively). **, P < 0.05 (Student’s t test); error bars indicate SEM (F). Dotted lines indicate the outlines of intestines. Regions surrounded by squares are enlarged (16× in A–D″ and 9× in E) in insets. Bars: 10 µm; (insets) 5 µm.

Figure 5.

SFT-4 interacts with VIT-2 in vivo. (A) Subcellular localization of VIT-2 and SFT-4 in intestinal cells. VIT-2–GFP partially colocalized with mC–HA–SFT-4 on punctate structures (arrowheads). (B and C) SFT-4 interacts with VIT-2. Lysates of whole animals coexpressing VIT-2–GFP and mC–HA–SFT-4 were immunoprecipitated with anti-HA (B; upper panel) and anti-GFP (C) antibodies. The precipitates and 0.3% of the total lysate were immunoblotted with anti-HA and anti-GFP antibodies. SFT-4 does not interact with ssGFP-HDEL when lysates of whole animals coexpressing ssGFP-HDEL and mC–HA–SFT-4 were immunoprecipitated with anti-HA antibody (B; lower panel). CAV-1 (C; arrowheads) was not coimmunoprecipitated with VIT-2–GFP. (D) Secretion of ssGFP from the intestine into the body cavity was impaired in sft-4(RNAi) animals. Dotted lines indicate the outlines of intestines, oocytes, and embryos. Regions surrounded by squares are enlarged (4×) in insets. Bars, 10 µm.

Figure 6.

Surf4, a mammalian homologue of SFT-4, is required for ER export of ApoB in HepG2 cells. (A) Whole-cell lysates were immunoblotted with antibodies against Surf4 and β-actin. Surf4 expression was efficiently silenced. (B) Surf4 loss causes ApoB accumulation in the ER. HepG2 cells were transiently transfected for 3 d with control (mock) or surf4 siRNA (Surf4 KD) and then fixed and stained with anti-ApoB and anti-PDI antibodies. In Surf4 KD cells, ApoB was accumulated in the ER and was colocalized with PDI. Bars, 10 µm. (C) Colocalization quantifications were calculated by using Pearson’s correlation coefficient and statistically analyzed using Student’s t test; ***, P < 0.001; error bars: SEM (n = 96 and 87 of mock and Surf4 KD cells, respectively). (D) Loss of Surf4 reduced the number and size of Sec31-positive ERES. HepG2 cells were transiently transfected for 3 d with control (mock) or surf4 siRNA (Surf4 KD) and then fixed and stained with anti-ApoB and anti-Sec31 antibodies. Regions surrounded by squares are enlarged (9×) in insets. Bars: 10 µm; (insets) 5 µm. (E) The quantifications of the number and signal intensity of Sec31-positive punctate structures spread in the cytoplasm were measured and statistically analyzed using Student’s t test; **, P < 0.05; error bars: SEM (left panel, n = 70 and 50 of mock and Surf4 KD cells for the number of puncta/unit area; right panel, n = 62 and 66 of mock and Surf4 KD cells for the signal intensity/puncta, respectively). (F) The pattern of total secreted proteins. Whole-cell lysates and culture medium were immunoblotted with anti-ApoB. Coomassie Brilliant Blue staining of culture medium is also shown in middle panel. The secretion was not generally inhibited by the loss of Surf4. The arrowhead presumably indicates BSA conjugated with oleic acids. (G and H) ApoB amount was quantified through densitometric scanning of band intensities, and the relative amount was determined. The amount of ApoB secreted from Surf4-depleted cells was significantly decreased as compared with that secreted from mock-treated cells (G), whereas the amount of ApoB in whole cell lysates was significantly increased (H). Results were analyzed using Student’s t test; **, P < 0.05; ***, P < 0.001; error bars: SEM, n = 6 (G) or 4 (H) are shown. (I) Lysates from vector control or GFP-Surf4 transfected HepG2 cells were immunoprecipitated with anti-GFP antibody. The precipitates and 1% of the total lysate were immunoblotted with anti-ApoB antibody. (J) HepG2 cell lysates were immunoprecipitated with anti-Surf4 antibody. The precipitates and 1% of the total lysate were immunoblotted with anti-ApoB antibody.