Molecular determinants of ER-Golgi contacts identified through a new FRET-FLIM system

Abstract

ER-TGN contact sites (ERTGoCS) have been visualized by electron microscopy, but their location in the crowded perinuclear area has hampered their analysis via optical microscopy as well as their mechanistic study. To overcome these limits we developed a FRET-based approach and screened several candidates to search for molecular determinants of the ERTGoCS. These included the ER membrane proteins VAPA and VAPB and lipid transfer proteins possessing dual (ER and TGN) targeting motifs that have been hypothesized to contribute to the maintenance of ERTGoCS, such as the ceramide transfer protein CERT and several members of the oxysterol binding proteins. We found that VAP proteins, OSBP1, ORP9, and ORP10 are required, with OSBP1 playing a redundant role with ORP9, which does not involve its lipid transfer activity, and ORP10 being required due to its ability to transfer phosphatidylserine to the TGN. Our results indicate that both structural tethers and a proper lipid composition are needed for ERTGoCS integrity.

Figure 1.

Morphometric analysis of ERTGoCS. (A) Electron micrograph showing ER (white arrowhead) closely apposed (red dashed line) to the TGN (black arrowheads) in a HeLa cell expressing TGN46-HRP (visualized by the DAB reaction that produces a dark product in the lumen of the TGN). Golgi cis elements are indicated. (B) Measurement of ERTGoCS distance (left; y axis) and length (right; y axis). Mean ± SD of 40 and 22 Golgi stacks, respectively, representative of three independent experiments. (C) Percentage of Golgi stacks negative (white bar) or positive (black bar) for ERTGoCS (left part of graph). In the ERTGoCS-positive Golgi stacks, the frequency distribution of the ERTGoCS surface area was calculated and expressed as a percentage of the total TGN surface (Materials and methods). n = 130 Golgi stacks. (D) FIB-SEM images of a HepG2 cell. Four images from Video 1 taken every 60 nm representing a total depth of 240 nm. The ER is colored in red and the TGN in green. The small arrow indicates a clathrin-coated profile that identifies the trans pole of the Golgi.

Figure 2.

FRET–FLIM analysis of ERTGoCS. (A) Schematic representation of the TGN (TGN46-GFP; top construct) and ER (C tail of CytB5, Cb5; bottom construct) membrane reporter proteins used in this study (see Materials and methods for a detailed description of individual domains). (B) Graphic representation of the topology of the donor (TGN46-GFP) and the acceptor (mCherry-Cb5) fluorophores at the ERTGoCS under different conditions: left, normal-length linkers (16 amino acids used in the majority of the experiments; Várnai et al., 2007); middle, longer helical linkers (61 amino acids; Várnai et al., 2007); and right, 2 min treatment with 200 nM rapamycin. The mean FRET efficiency measured in the ROI is indicated for each condition. (C) The FLIM showing donor (TGN46-GFP) mean lifetime in the presence of the acceptor (mCherry-Cb5) in a defined ROI (top) compared with the whole TGN area (bottom) in cells expressing reporter proteins with normal-length linkers. The τ values are represented by a pseudocolor scale ranging from 1.8 to 2.5 ns. (D) FLIM (in color) and immunofluorescence (IF; gray) images of HeLa cells expressing the TGN46-GFP (donor) alone (top), TGN46-GFP and mCherry-Cb5 (donor+acceptor; middle), or TGN46-GFP with free cytosolic mCherry (bottom). The FLIM images show the spatial variation of the mean fluorescence lifetime (τ) of TGN46-GFP: τ values are represented by a pseudocolor scale ranging from 1.8 to 2.5 ns. (E) Mean τ value distribution curves of TGN46-GFP (donor) in cells expressing TGN46-GFP alone (top; red line), TGN46-GFP and mCherry-Cb5 (middle; green line), or TGN46-GFP with free mCherry (bottom; orange line). (F) The spatial distribution of the FRET efficiency (FRET [%] map) is depicted (color scale ranges from 0 to 30%).

Figure 3.

The VAPs are required to establish/maintain the ERTGoCS. (A) Electron micrograph of a VAP-KD cell and quantification of ERTGoCS in WT and VAP-KD cells. n = 40 Golgi stacks. (B) Quantification of average donor lifetime in donor alone or donor+acceptor VAP-KD and VAP-KO HeLa cells, in cells after treatment with rapamycin (2 min, 200 nM), or after transfection with FLAG-tagged PH-FFAT of OSBP1. The corresponding FRET efficiency is shown on the right. n = 20; data are means ± SD. **, P < 0.01; ***, P < 0.001; Student’s t test. (C–E) Representative immunofluorescence (IF; gray) and FLIM (in color) images showing donor lifetime in VAP-KD and VAP-KO cells (C), in rapamycin-treated cells (D), and in PH-FFAT–expressing cells (E). The pseudocolor scale representing the τ values of TGN46-GFP is the same as in Fig. 2 D. The graphs show the mean τ value distribution curves of TGN46-GFP under the indicated conditions.

Figure 4.

ORP10 has an essential role, while OSBP and ORP9 play a redundant role in establishing/maintaining the ERTGoCS. (A) Graphical representation of the candidate proteins analyzed in the siRNA-based screen. The candidates possess a Golgi targeting motif represented by a PI4P-binding PH domain and an ER targeting motif represented by a FFAT motif (OSBP, ORP9, and CERT; Murphy and Levine, 2016) or a putative FFAT-like motif (FAPP2 [Mikitova and Levine, 2012] and ORP10 and ORP11 [Murphy and Levine, 2016]) that mediates the interaction with the VAP ER membrane proteins. FAPP1 possesses two potential FFAT motifs: 165 TFITTLE 171 (if phosphorylated on T165 and T168) and 242 TYSDTDS 248 (if phosphorylated on T242 and S248). (B) FLIM images of HeLa cells expressing TGN46-GFP and mCherry-Cb5 (donor+acceptor) transfected with the indicated siRNAs. The τ values are represented by a pseudocolor scale ranging from 1.8 to 2.5. Scale bar, 10 µm. (C and D) Quantification of average TGN-GFP lifetime in HeLa cells expressing TGN46-GFP and mCherry-Cb5 (donor+acceptor) and transfected with the indicated single (C) or combined (D) siRNAs. Donor indicates cells transfected with TGN46-GFP alone; CTRL indicates cells transfected with nontargeting siRNAs. The corresponding FRET efficiency is shown on the right. n = 20; data are means ± SD. *, P < 0.05; **, P < 0.01; Student’s t test. (E) Electron micrographs of ORP10-KD, OSBP1-KD, ORP9-KD, and OSBP1+ORP9-KD cells. (F) Corresponding quantification of the percentage of Golgi stacks engaged in ERTGoCS (ERTGoCS occurrence). n = 20–30 Golgi stacks. Scale bar, 200 nm.

Figure 5.

The integrity of the ERTGoCS requires the PS transfer activity of ORP10. (A) Subcellular localization of mCherry-tagged WT-ORP10 and of the Golgi marker Golgin-97. Scale bar, 10 µm. (B) Residues in ORP5/ORP8 involved in PS/PI4P exchange are conserved in the ORP10 ORD domain. (C) Subcellular localization of ORD-domain mutant forms (L418D and H535/536A) of mCherry-tagged ORP10. Scale bar, 10 µm. (D) Quantification of ERTGoCS in WT and ORP10-KD cells and in ORP10-KD cells expressing siRNA-resistant GFP-tagged WT-ORP10 or ORD-domain mutant forms of ORP10. The experiment was performed in HeLa cells stably expressing the TGN46-FRB-HA and mCherry-T2A-FKBP-Cb5 reporter proteins as shown in Fig. S3 A and specified in the corresponding legend (see also Materials and methods). Means ± SD of three independent experiments; n > 150; **, P < 0.01; ***, P < 0.001; Student’s t test. (E) The PS distribution in control and ORP10-KD cells using the C2 domain of lactadherin (C2-Lact-GFP, a marker for PS). Insets: Golgi marker Golgin-97. Scale bar, 10 µm. Right: Fluorescence intensity ratio of C2-Lact-GFP measured in an equivalent area of the Golgi and the PM in mock and ORP10-KD cells. Data are means ± SD. n = 50. ***, P < 0.001; Student’s t test. (F and G) Quantification of ERTGoCS in control cells and in ORP10-KD cells untreated or treated with PS (DOPS; F), or in cells expressing PS synthase 1 (PTDSS1) in its WT and constitutively active (P269S) forms (G). Means ± SD of three independent experiments; n > 150. *, P < 0.05; ***, P < 0.001; Student’s t test. (H) Quantification of ERTGoCS in HeLa cells and in ORP9-KD HeLa cells treated with itraconazole. n > 150; three independent experiments. Data are means ± SD. *, P < 0.05; ns, not significant with respect to untreated control cells; Student’s t test.