Fig 1.
Establishment and characterization of fully functional fluorescently tagged ApoE.
(A) Secretion of ApoEmT2. Lysates and supernatants of unmodified Huh7-Lunet cells (naïve) and Huh7-Lunet cells in which endogenous ApoE was depleted by stable knock-down and substituted by stable expression of mTurquoise2-tagged ApoE (mT2) were harvested one day after seeding. Samples were analyzed by Western blot using ApoE-specific antibody; α-tubulin served as a loading control for cell lysates. The ratios of secreted to total ApoE are given below the lanes. The value of ApoEwt was set to 1. (B) Density of secreted ApoEmT2. Upper panel: conditioned media of naïve Huh7-Lunet cells and Huh7-Lunet cells expressing ApoEmT2, both from (A), were subjected to 10–50% iodixanol isopycnic centrifugation and fractions were analyzed by Western blot using ApoE-specific antibody. I: input; M: protein marker lane. Lower panel: signal intensities of the Western blot image were quantified and values were normalized to total ApoE amount in all fractions. Densities of fractions are specified on the right Y-axis (g/ml). Densities of peak fractions are given. (C) Normal lipid-binding property of ApoEmT2. Immunofluorescent staining of ApoEmT2 in ApoEmT2 reconstituted Huh7-Lunet/ApoE-KD cells using ApoE- and ApoB-specific antibodies. Two-row images on the right show magnified views of boxed areas in the left overview image. Arrowheads in cropped images point to signal overlaps of ApoEmT2 and ApoB; plot profiles in the right panels are along the lines indicated with white arrows in the merge images. (D) Functionality of ApoEmT2 as determined by the rescue of infectious HCV particle production. Left panel: Huh7-Lunet/ApoE-KD cells were transduced with either an empty vector (Empty V.), or ApoEwt, or ApoEmT2 and stably selected. Cells were then electroporated with in vitro transcripts of the Renilla luciferase (RLU) HCV reporter genome JcR2a. At 24, 48 and 72 h post-electroporation, amounts of extracellular core protein present in culture supernatants were determined by chemiluminescence assay. Right panel: amount of infectious HCV particles released into the culture supernatant of electroporated cells. At the indicated time points supernatants were harvested, naïve Huh7.5 cells were inoculated and 72 h later, luciferase activity was determined. Values were normalized to HCV RNA replication in each cell line to exclude replication effects. Data are medians (range) from three independent experiments. P-value was determined using unpaired Student’s t-test. N.s: not statistically significant (P>0.05).
Fig 2.
Detection of ApoE in CD63-positive late endosomes, intracellular endosomal trafficking of ApoE and egress from hepatocytes.
(A) Colocalization of ApoEmT2 with markers of the ER (PDI), Golgi (GM130), and intraluminal vesicles/endosomes (CD63). Proteins specified on the top of each panel were detected in Huh7-Lunet/ApoEmT2 cells by immunostaining and cells were analyzed by confocal microscopy. Boxed areas in the left panels are shown as enlarged views in the panels on the right of each row. Profiles on the right of each panel were taken along the lines indicated with white arrows in cropped images. (B-C) ApoE foci predominantly form independent of autophagy and re-internalization of secreted ApoE via endocytosis. (B) Majority of ApoEmT2 foci is devoid of LC3 puncta. Huh7-Lunet/ApoEmT2 cells were lentivirally transduced with the LC3mCherry expression construct. At 24 h post-transduction, cells were fixed and analyzed by confocal microscopy. Boxed area in the left panel is shown as enlarged view in the panels on the right. Dashed circles mark LC3-free ApoEmT2 foci. Arrowheads point to LC3-positive ApoEmT2 signals. The numbers of total ApoE, LC3, and ApoE-LC3 double-positive foci in single cells were determined (right). P-values were determined using Mann-Whitney test. (C) Majority of ApoE-CD63 double-positive signals corresponds to newly synthesized structures rather than secreted and re-internalized structures. Schematic representation of used approach to visualize newly synthesized and re-internalized ApoE. (i) Huh7-Lunet/ApoE-KD cells were transduced with lentiviruses encoding ApoEmT2 for 4 h (donor/D). After being washed twice with PBS, the cells were added to Huh7-Lunet/ApoE-KD cells expressing the membrane marker eYFPCaaX (recipient/R) at a 1:4 (D:R) ratio. After 12 h, cells were fixed, subjected to immunostaining for CD63, and analyzed by confocal microscopy. (ii) Huh7-Lunet/ApoE-KD cells, lentivirally transduced with the ApoESNAPf expression construct for 4 h, were cultured in medium containing 5 μM cell-nonpermeable SNAP-surface substrate for 12 h. Thereafter, cells were fixed, subjected to ApoE and CD63 immunostaining, and analyzed by confocal microscopy. Right panel: quantitative analysis of the two assays. (i) Number of total ApoE, CD63, and ApoE-CD63 double-positive foci in single donor cells (D) and donor-transmitted ApoE-positive foci in single recipient cells (R). Values of total CD63+ signals were divided by 4 to account for their high abundance; (ii) number of re-internalized ApoE foci in single cells. P-values in both approaches were determined using Mann-Whitney test. (D) Endosomal localization of ApoE-CD63 double-positive structures. Huh7-Lunet/ApoEmT2 cells expressing CD63mCherry were analyzed by CLEM using lipid droplets (LDs) stained with lipidTox as fiducial markers. The overlay image is shown on the left. Middle and right panels: magnified EM micrographs from an area with ApoE-positive, CD63-negative signals showing Golgi stacks and vesicles (crop 1) and from an area with ApoE-CD63 double-positive endosomes (crop 2), respectively.
Fig 3.
Endosome-dependent trafficking and secretion of ApoE.
(A-C) Microtubule-dependent motility of intracellular ApoE-CD63 double-positive structures. (A) Mean squared displacement (MSD) of general ApoE and CD63 trafficking in mock and colchicine-treated cells. Huh7-Lunet/ApoEmT2 cells expressing CD63mcherry were either mock-treated or treated with 80 μM colchicine for 1 h to depolymerize microtubules and analyzed by live-cell confocal microscopy. (B) Sizes of ApoE-CD63 double-positive structures from (A) were measured. P-value was determined using Mann-Whitney test. Ns, non-significant. (C) Trafficking velocities of ApoE-CD63 double-positive structures were computed (see methods for detail). P-value was determined using Mann-Whitney test. (D-E) Peripheral distribution of ApoE-CD63 double-positive structures depends on intact microtubules. (D) Representative images of cells analyzed by live-cell confocal microscopy from (A). Boxed areas in the left panels are shown as enlarged views in the panels on the right of each row. Arrowheads point to peripheral ApoE-CD63 double-positive signals in mock-treated cells. (E) The mean radial signal intensities of CD63 in the peripheral regions of cells from (D) were determined (top) and quantified (bottom) by using the CellProfiler module “MeasureObjectRadialDistribution”. P-value was determined using Mann-Whitney test. (F-H) Secretion of ApoE via the endosome-dependent route. (F) Huh7-Lunet cells expressing CD63nLuc were either mock-treated or treated with 5 μM colchicine for 1 h to depolymerize microtubules. Extracellular ApoE in the supernatant was quantified by ApoE Elisa. Data are means (SD) from 5 independent experiments. P-value was determined using unpaired Student’s t-test. (G) The extracellular CD63nLuc activity in the supernatants from (F) was quantified by nLuc assay. Data are means (SD) from 5 independent experiments. P-value was determined using unpaired Student’s t-test. (H) Secretion of ApoE-positive ILVs visualized by pHluorin-tagged CD63. Huh7-Lunet cells expressing ApoEmT2 and CD63pHluorin were cultured in imaging medium (pH 7.4) and analyzed by time-lapse confocal microscopy with a focus on plasma membrane resident CD63-fluorescent signals. (Right panel) Maximum fluorescence intensity of CD63pHluorin and associated ApoE in the selected dashed area indicated in S3 Movie. Images taken at indicated time points are displayed on the bottom and they correspond to initiation (t1), peak (t2), and late-secretion (t3), respectively.
Fig 4.
Association of ApoE-containing lipoproteins with CD63-positive extracellular vesicles and their endocytosis.
(A) Analysis of fractions enriched for ApoE-containing LPs. Huh7-Lunet cells were cultured in EV-depleted medium and ApoE-associated vesicles released into the culture medium were captured using ApoE-specific antibody. Immunocomplexes were analyzed by TEM after negative staining. Arrowheads: ~5–10 nm vesicles; arrows: ~20–30 nm vesicles; stars: ~50–60 nm vesicles. Vesicles in the electron micrographs were segmented by using Ilastik to allow quantification of vesicle diameters shown in the histogram below the micrograph. (B) Detection of CD63-positive EVs in isolated ApoE-containing LP fractions. Purified ApoE-associated vesicles from (A) were immunogold-labeled with ApoE- (upper) and CD63-specific antibodies (lower). Arrowheads point towards gold particles. A zoom image of a CD63-positive gold particle is shown on the top. Note that immuno-gold labeling of ApoE alters the vesicular shape of LPs, most likely because of distortion of ApoE by used antibodies and protein A-gold incubation during the labeling procedure known to destabilize the LP structure. (C-E) Visualization of the uptake of ApoE and CD63 by recipient cells. (C) Schematic representation of used approach. Huh7-Lunet/ApoEmT2/CD63mCherry served as donor cells; Huh7-Lunet cells expressing eYFP-tagged CaaX (the farnesylation signal from human HRAS) as recipients. (D) Donor and recipient cells from (C) were co-cultured for 16 h and analyzed by live-cell confocal imaging (refers to S4 Movie). D: donor; R: recipient. Arrows in cropped sections on the bottom indicate transferred ApoE-CD63 signals; stars: transferred CD63-only signals. (E) Donor and recipient cells from (C) were co-cultured and fixed at 24 h and 48 h post-seeding. The numbers of ApoE+ CD63- and ApoE- CD63+ single positive structures as well as ApoE+ CD63+ double-positive structures in single recipient cells were quantified. Each dot represents a single cell. P-value was determined using Mann-Whitney test.
Fig 5.
Enrichment of NS5A in ApoE-positive structures and co-trafficking of ApoEmT2 with NS5A and E2 independent of HCV assembly.
(A) Experimental approach. Fluorescently tagged ApoEmT2, HCV proteins supporting assembly (C to NS2 with eYFP-tagged E2), and a subgenomic replicon (dotted lines indicate 5’ and 3’ NTRs) are shown from top to bottom; the experimental approach is depicted below. Cells stably expressing ApoEmT2 and C-NS2/E2eYFP were electroporated with the replicon RNA encoding mCherry-tagged NS5A. Cells were subjected to confocal time-lapse live-cell imaging to monitor signal overlaps of the various fluorescent proteins: NS5A + E2; ApoE only; ApoE + NS5A; ApoE + NS5A + E2. (B) Time-dependent enrichment of NS5A-ApoE double-positive structures in HCV-replicating cells. Huh7-Lunet/ApoEmT2 cells expressing HCV Core-NS2/E2eYFP and containing the subgenomic replicon were subjected to live-cell confocal imaging from 5 to 54 h post-electroporation (30 min/frame) to observe ApoE, NS5A, and E2 signals. A series of still images taken at time points after electroporation specified on the top are shown. White arrowheads: NS5A-E2 foci; yellow arrows: ApoE-NS5A foci; stars: ApoE-NS5A-E2 triple-positive foci. (C) Quantification of NS5A-ApoE and E2-NS5A double-positive foci detected in single cells in (B). Ten single cells were analyzed. P-value was determined using Mann-Whitney test. (D) Assembly-independent enrichment of NS5A in ApoE-positive foci. Huh7-Lunet/ApoEmT2 cells expressing the C-p7 construct (NS2-deletion; upper panel) were electroporated with in vitro-transcripts of the subgenomic replicon sgJFH1/NS5AmCherry and analyzed by confocal microscopy to observe ApoE, NS5A, and E2 signals. A representative image showing ApoE-NS5A double-positive foci (arrowheads) and diffuse E2 signal at 72 h post-electroporation is shown. Images on the right show magnified views of the boxed area in the left overview image. (E) Quantification of NS5A-ApoE double-positive foci detected in 100 single cells in (D) at 24, 48, and 72 h post-electroporation. Data are medians (range) of the number of detected foci. P-value was determined using Mann-Whitney test.
Fig 6.
Colocalization of NS5A and ApoE with the intraluminal vesicle marker CD63 as revealed by super resolution microscopy.
(A) Experimental approach. Schematic representations of SNAPf-tagged ApoE and the subgenomic replicon encoding CLIPf-tagged NS5A are shown on the top. Huh7-Lunet cells were lentivirally transduced with the ApoE expression vector and transfected with the subgenomic replicon RNA. ApoE and NS5A were detected by STED microscopy and CD63 by immunofluorescence confocal microscopy. (B) Colocalization of ApoESNAPf and NS5ACLIPf. Huh7-Lunet/ApoESNAPf cells were electroporated with subgenomic replicon RNA encoding NS5ACLIPf and after 48 h, cells were labeled with SNAPSiR647 and CLIPATTO590 for 1 h, fixed, and subjected to confocal microscopy. Arrowheads: colocalized ApoE-NS5A signals. (C) Quantification of CD63-positive ApoE-NS5A double-positive foci. Cells from (B) harvested at 72 h post-electroporation were fixed, permeabilized, and incubated with anti-CD63AF488 antibody. To determine the correlation between ApoE-NS5A double-positive foci and how many of them colocalized with CD63, 100 cells were analyzed. Each dot represents one cell and displays the number of ApoE-NS5A double-positive foci (x-axis) and the number of CD63-ApoE-NS5A triple-positive foci (y-axis). The R-squared value is given on the plot. (D) STED-resolved ApoE-NS5A double-positive structures colocalizing with the intraluminal vesicle marker CD63. Huh7-Lunet/ApoESNAPf cells were electroporated with the subgenomic replicon RNA encoding NS5ACLIPf. After 48 h, cells were labeled with SNAPSiR647 and CLIPATTO590 for 1 h, fixed, and incubated with anti-CD63AF488 antibody. ApoE, NS5A, and CD63 fluorescent signals were sequentially imaged using confocal and STED microscopy, the latter to achieve a higher resolution of ApoE and NS5A signals that were deconvoluted using Huygens. Arrows: ~100–200 nm-sized ApoE-NS5A-CD63 positive structures; star: ~500 nm-sized ring-like NS5A positive structure.
Fig 7.
Detection of HCV-produced intraluminal vesicles in NS5A-ApoE double-positive endosomes.
(A) Huh7-Lunet/ApoEmT2 cells expressing HCV Core-NS2/E2eYFP and containing the subgenomic replicon sgJFH1/NS5AmCherry (Fig 5A) were investigated with the CLEM method at 48 h post-electroporation. Lipid droplets stained with lipidTox were used as fiducial markers to correlate light and electron micrographs. Dashed squares in the overlay image (right panel) refer to NS5A-ApoE double-positive structures. The left panels show single-channel light or EM micrographs of the enlarged overlay image on the right. For ease of visualization, endosome peripheries are marked with dashed lines. (B) Magnified views of regions indicated in the dashed squared areas in the overlay image in (A). Cropped areas 1, 2, 3: ApoE-NS5A double-positive ILVs in endosomes; crop 4: an ApoE-positive, NS5A-negative endosome; cropped areas 5 and 6: NS5A-E2 double-positive areas containing numerous DMVs. Orange arrowheads point to ILVs in crops 1, 2 and 3; and DMVs in crops 5 and 6. LD, lipid droplet.
Fig 8.
Association of ApoE-lipoproteins with endosome-derived extracellular vesicles containing HCV NS5A and RNA and evidence for intercellular co-transmission.
(A) Virion-free release of HCV RNA in association with ApoE. Huh7 cells harboring a subgenomic HCV replicon and control cells were cultured in a medium containing 1% FCS for 6 h. Culture supernatants were subjected to immunoprecipitation using ApoE-specific or IgG control antibodies. Immuno-complexes were analyzed by HCV-specific RT-qPCR. Data are means (SD) from 4 independent experiments. P-value was determined using one-way ANOVA and unpaired Student’s t-tests. (B-C) Association of secreted ApoE with NS5A-containing EVs. (B, top panel) Schematic of the Nanoluciferase (Nluc)-tagged NS5A subgenomic replicon construct. (B, middle and bottom panel) Huh7-Lunet cells were electroporated with subgenomic replicon RNA encoding the NLuc-tagged NS5A and 72 h post-electroporation, culture supernatant was subjected to immunoprecipitation using ApoE-, or NS5A-, or control TIA1-specific antibodies. NS5A contained in captured immuno-complexes was quantified by measuring NLuc activity (middle panel). ApoE contained in captured complexes was analyzed by Western blot (bottom panel). Data are means (SD) from two biologically independent experiments. P-value was determined using unpaired Student’s t-test. (C) Captured complexes from (B) were visualized by negative staining and analyzed by TEM. Turquoise arrows point to LP-like particles (~20 nm) attached to EVs that were captured with antibodies specified on the top of each panel. (D) Reduced virion-free secretion of HCV RNA with the ApoE-binding deficient NS5AAPK99AAA mutant. Total RNA contained in Huh7-Lunet cells with stable wildtype or mutant subgenomic replicon was extracted and HCV RNA was quantified by RT-qPCR. In addition, total RNA in culture supernatants was isolated and HCV RNA contained therein was quantified by RT-qPCR. Ratios of secreted to total HCV RNA are shown. Data are means (SD) from three independent experiments. P-value was determined using unpaired Student’s t-test. (E-F) Uptake of ApoE-associated, virion-free released HCV RNA by HCV-negative bystander cells. (E) Experimental approach. Huh7-Lunet cells expressing tagged ApoE and CD63 and containing a subgenomic replicon (constructs on the top) served as donor cells. Huh7-Lunet-derived recipient cells expressed eYFP, fused to the farnesylation signal from human HRAS protein (CaaX) to visualize cellular membranes. Donor and recipient cells were co-cultured for 24 h, fixed, and HCV RNA in recipient cells was detected by using smFISH with Hulu probes. (F) The number of total HCV RNA positive structures and ApoE-CD63 double-positive HCV RNA containing structures detected in single recipient cells (n = 16) from (E) was quantified using ColocQuant and ColocJ software suit and data are shown as medians (range). Note that non-specific signals in the cell nuclei were excluded from the analysis.