GFP-tagged Apolipoprotein E: a useful marker for the study of hepatic lipoprotein egress

Abstract

Apolipoprotein E (ApoE), a component of very-low-density and high-density lipoproteins, participates in many aspects of lipid transport in the bloodstream. Underscoring its important functions, ApoE isoforms have been associated with metabolic and circulatory disease. ApoE is also incorporated into hepatitis C virus (HCV) particles, and promotes their production and infectivity. Live cell imaging analysis of ApoE behavior during secretion from producing cells thus has the potential to reveal important details regarding lipoprotein and HCV particle biogenesis and secretion from cells. However, this approach requires expression of fluorescently tagged ApoE constructs that need to faithfully reproduce known ApoE behaviors. Herein, we evaluate the usefulness of using an ApoE-GFP fusion protein in studying hepatocyte-derived, ApoE-containing lipoproteins and HCV particles. We show that while ApoE-GFP alone is not sufficient to support infectious HCV production, it nonetheless colocalizes intracellularly and associates with secreted untagged lipoprotein components. Furthermore, its rate of secretion from hepatic cells is indistinguishable from that of untagged ApoE. ApoE-GFP thus represents a useful marker for ApoE-containing hepatic lipoproteins.

Graphical Abstract

Introduction

Expression of fluorescent protein (FP) tagged constructs is a powerful method used to study spatial and temporal dynamics of proteins, membranes and organelles in live cells. The method relies on fusing, in frame, a DNA fragment encoding one of an ever-growing list of fluorescent proteins¹ to the 5'-, 3'-, or middle of a DNA fragment encoding a protein whose dynamics are to be studied, followed by expression of the resulting chimeric gene. Powerful spatio-temporal studies may be carried out and their results properly interpreted if the FP-tagging does not detectably interfere with the process studied². Unfortunately, FP-tagging may also result in dominant-negative, non-functional, or mislocalized fusion proteins, as outlined in greater detail in a previous study³. Thus, to avoid collecting and interpreting artifactual data caused by expression of aberrantly-behaving FP-fusion proteins, an initial battery of functional tests should be performed³. In this toolbox report, we present such a characterization of an ApoE-GFP fusion protein for use in the investigation of lipoprotein and HCV secretion from hepatic cells.

Plasma lipoproteins transport lipids (triglycerides, phospholipids, cholesterol, and cholesteryl esters) in the bloodstream and are classified in part by their buoyant density. They include very-low-density lipoproteins (VLDL), low-density lipoproteins (LDL) and high-density lipoproteins (HDL)⁴. VLDL, formed by the large apolipoprotein B100 (ApoB100), shuttle lipids (mainly triglycerides) from the liver to the periphery. HDL, in turn, shuttle cholesterol from the periphery back to the liver in a process known as reverse cholesterol transport and may be formed by several smaller apolipoproteins, including ApoE⁵. ApoE is expressed by and secreted from several cell types, including hepatocytes⁶, macrophages⁷ and astrocytes⁸. A 299-amino-acid, O-glycosylated apolipoprotein^9,10, ApoE is composed of a C-terminal domain rich in amphipathic helices and possessing lipid binding ability, and a N-terminal globular domain which mediates binding to members of the LDL receptor (LDLR) family⁹. Through its LDLR binding activity, ApoE promotes clearance of cholesterol-rich lipoprotein particles from circulation^11–13. Human ApoE occurs as several major isoforms, termed ApoE2, ApoE3 and ApoE4, respectively^14,15. While ApoE3 is the most common isoform and is considered “neutral” with respect to disease association, ApoE2 is associated with type III hyperlipoproteinemia and ApoE4 is associated with type V hyperlipoproteinemia^14,16,17. Furthermore, the minor isoform ApoE3-Leiden is a dominant predictor of type III hyperlipoproteinemia¹⁸. Type III disease association is correlated with low clearance rate of ApoE-containing lipoproteins due to defects in LDLR binding¹⁹.

Some of the lipid metabolic functions of ApoE are mediated through its interaction with ApoB100-containing VLDL^20,21. ApoB100 is essential for VLDL production as it is VLDL’s structural apolipoprotein component. ApoE may interact however with VLDL particles as an exchangeable apolipoprotein component. Hepatocyte-made ApoE, in particular, associates intracellularly with VLDL particles^22–24, and stimulates the secretion of VLDL-associated triglycerides^25,26. Furthermore, ApoE is incorporated into HCV particles that assemble at the endoplasmic reticulum (ER) of infected hepatocytes^27–33. This ApoE-HCV association is important for efficient production of HCV particles and for the infectivity of released HCV particles^32,34–39. Given these hepatocyte-specific functions of ApoE in lipoprotein and HCV particle formation and release, we set out to determine whether ApoE could be functionally FP-tagged and thus subsequently used in analyzing the spatio-temporal dynamics of ApoE-containing lipoprotein and HCV particles. Such imaging studies may help identify the subcellular compartments that mediate lipoprotein and HCV secretion, and may also yield quantitative, single particle focused kinetic measurements describing these processes.

Previous studies have described ApoE-GFP constructs that were used to image microtubule-dependent ApoE secretion from macrophages⁴⁰, or to colocalize ApoE with fluorescently-labeled HCV entities^41,42. In the macrophage study, ApoE-GFP secretion from cells was comparable to that of untagged ApoE, and its localization at the ER, Golgi, and in secretory vesicles, and its movement along microtubules were consistent with behaviors expected of this secreted protein⁴⁰, suggesting that this ApoE-GFP is a good marker for the secretion of ApoE from macrophages. Whether ApoE-GFP is also a useful marker for monitoring ApoE secretion from hepatic cells, in the presence or absence of HCV infection, has not been formally addressed to date.

Here, we inquired whether ApoE-GFP reproduced the behavior of untagged ApoE with respect to lipoprotein and infectious HCV particle release from hepatic cells. We tested whether: i) ApoE-GFP was properly expressed in cells; ii) ApoE-GFP colocalized with untagged ApoE and ApoB100; iii) ApoE-GFP was secreted from cells with similar efficiency as untagged ApoE; iv) ApoE-GFP associated with secreted lipoprotein or HCV particles; and v) ApoE-GFP supported infectious HCV production. Together, our data indicate that ApoE-GFP faithfully reproduces the known involvement of ApoE in hepatic lipoprotein secretion, and supports its use in future imaging studies aimed at elucidating dynamic spatio-temporal aspects of lipoprotein secretion.

Results

Expression and detection of ApoE-GFP

In this study, we fused monomeric enhanced GFP (mEGFP)⁴³ to the carboxyl terminus of full length human ApoE3 (hApoE3). We introduced silent mutations in the ApoE-coding sequence to confer resistance to shRNA-mediated knockdown. The linker between ApoE and GFP is predicted to be identical with that found in an ApoE-GFP construct that was characterized in macrophages⁴⁰. From here on, we refer to the shRNA-resistant hApoE3-mEGFP construct that we made as ApoE-GFP.

To characterize ApoE-GFP, we stably expressed it in human hepatoma Huh-7.5 cells. We used these cells because they secrete both ApoB100- and ApoE-containing lipoproteins, support the complete HCV live cycle^44,45, and are easy to grow and passage, thus making them amenable to large scale biochemical experiments or to genetic manipulations. Huh-7.5/ApoE-GFP cells expressed both ApoE-GFP (62 kDa predicted unglycosylated molecular weight) and untagged ApoE (predicted 34 kDa), as detected using a polyclonal goat α-ApoE antibody (Fig. 1A, left panel). The ApoE-GFP fusion was also detected using an α-GFP antibody (Fig. 1A, middle panel). ApoE-GFP was not detected in the parental Huh-7.5 cell line, nor in the empty vector (EV) transduced control cell line Huh-7.5/EV Hygro (Fig. 1A). Similar amounts of cell lysate from each of the three cell lines were loaded, as detected by an α-actin antibody (Fig. 1A, bottom panel).

Figure 1.

Expression and detection of ApoE-GFP. (A.) Lysates from Huh-7.5 cells (left lane), from Huh-7.5/EV Hygro cells (middle lane), and from Huh-7.5/ApoE-GFP cells (left lane) were processed for Western blotting with the following primary antibodies: α-ApoE (AB947, goat polyclonal, left panel), α-GFP (JL-8, middle panel), α-ApoE (ab52607, rabbit monoclonal clone EP1374Y, right panel) and α-actin (clone AC-74, bottom panel). Molecular weights (kDa) are listed at the left. The detected untagged and GFP-tagged ApoE species are labeled at the right. A likely degradation product of ApoE-GFP is indicated by a star (*) sign on the goat polyclonal α-ApoE blot (B.) Immunofluorescence on HeLa cells (top row), HeLa cells expressing ApoE-GFP (middle row), and HeLa cells expressing GFP-ApoE (bottom row), using the rabbit α-ApoE monoclonal antibody clone EP1374Y (right column) and the α-GFP antibody clone JL-8 (middle column). Brightfield images are shown in the left column. The epifluorescence images were not deconvolved. Scale bars, 50 µm.

We also detected a ∼45 kDa band in the lysates obtained only from Huh-7.5/ApoE-GFP cells and stained with the polyclonal α-ApoE antibody (marked by a star in Fig. 1A, left panel). This band was not detected using the α-GFP antibody (Fig. 1A, middle panel), was not secreted (Fig. 3C, see below), and was rapidly degraded as observed in a radioactivity pulse chase experiment (data not shown). As such, this band likely represents a non-fluorescent, unstable polypeptide produced in the ApoE-GFP cells that is unlikely to interfere with the imaging of ApoE-GFP.

Figure 3.

ApoE-GFP is secreted from Huh-7.5 cells and associates with lipoprotein particles. (A.) The rates of secretion of ApoE and of ApoE-GFP were measured using a radioactive pulse-chase experiment. At each time point, the amount of secreted radiolabeled cargo is shown as percent of the amount of total (cell-associated + secreted) radiolabeled cargo recovered at that time point. ApoE amounts were measured during secretion from Huh-7.5/EV Hygro (black trace) and Huh-7.5/ApoE-GFP cells (red trace), and the ApoE-GFP amounts were measured in Huh-7.5/ApoE-GFP cells (blue trace). (B.) ApoE-GFP associates with secreted untagged ApoE and ApoB100. Media was conditioned by Huh-7.5/ApoE-GFP cells or by Huh-7.5/EV Hygro cells, then IP with antibodies listed above the figure was performed. The pulled down material was then blotted against ApoB100 (AB742, top panel) and ApoE (AB947, bottom panel). Letters between the two panels correspond to the lane labeling described in the text. (C.) Media used in the IP experiment in panel B. was conditioned by similar amounts of cells and contained comparable amounts of secreted cargo. Cell lysates (left panels) and total media (right panels) were blotted against ApoB100 (AB742, top panel), ApoE (AB947, middle panel), and β-actin (AC-74, lower panel).

Interestingly, when we probed the same lysates discussed above with a monoclonal rabbit α-ApoE antibody (clone EP1374Y) that was raised against the carboxyl terminus of ApoE (Abcam), we only detected the endogenously expressed untagged ApoE (Fig. 1A, right panel). We suspected that the GFP tag, when fused to or near this antibody’s epitope, either obscures or modifies the epitope, rendering ApoE-GFP unrecognizable by this α-ApoE antibody. Epitope masking by the GFP tag has been previously documented in other contexts⁴⁶. We confirmed this was the case by performing immunofluorescence experiments in HeLa cells, which do not express endogenous ApoE⁴⁷. We transduced these cells with lentiviruses expressing ApoE-GFP, which has the carboxyl terminal epitope fused to the GFP tag, or GFP-ApoE, which has a free carboxyl terminus. As predicted, ApoE-GFP was not stained with the α-ApoE monoclonal antibody, while GFP-ApoE was stained brightly (Fig. 1B, right column). The GFP signal was detected in both GFP-ApoE and ApoE-GFP expressing cells (Fig. 1B, middle column), while no GFP or ApoE staining was detected in untransduced HeLa cells (Fig. 1B, top row). We conclude therefore that this monoclonal α-ApoE antibody does not recognize ApoE-GFP and can therefore be used to distinguish between the untagged ApoE and ApoE-GFP.

ApoE-GFP colocalizes with endogenous ApoE, ApoB100, and the Golgi

We proceeded to characterize the localization of ApoE-GFP in the Huh-7.5/ApoE-GFP cells. The ApoE-GFP signal displayed reticular and punctate distributions (Fig. 2), consistent with expected ER and secretory vesicle localization⁴⁰. ApoE-GFP also concentrated in perinuclear ribbons of signal, reminiscent of a Golgi localization and expected of a secreted protein. Indeed, these ribbons of ApoE-GFP signal co-stained with an antibody raised against GM130, a structural component of the Golgi (Fig, 2A and 2B). We next processed the Huh-7.5/ApoE-GFP cells for immunofluorescence using an α-GFP antibody to boost the signal from the GFP, and with the monoclonal α-ApoE antibody (clone EP1374Y) that does not recognize ApoE-GFP (Fig. 1). The resulting ApoE-GFP and ApoE signals overlapped, particularly within puncta that likely represent secretory vesicles (Fig. 2C). Of 442 ApoE-GFP puncta from 12 cells, 71% colocalized with ApoE puncta. Furthermore, cell-wide measured Pearson’s colocalization coefficients yielded a high value (0.53 ± 0.09, n=7 fields), consistent with colocalization between the ApoE and ApoE-GFP signals. Since ApoE may also associate intracellularly with nascent VLDL particles^22–24, we inquired whether ApoE-GFP colocalized with ApoB100. In cells co-stained with an α-ApoB100 antibody (Fig. 2D), 73% of the ApoE-GFP puncta overlapped with ApoB100 puncta (n=426 ApoE-GFP puncta from 9 cells), consistent with ApoE-GFP and ApoB100 being found in close proximity. As a control, we also colocalized ApoE-GFP with mCherry⁴⁸ , which we targeted to the lumen of the secretory pathway by fusing it to the signal peptide of human serum albumin (Fig 2E). This mCherry construct is thus expected to model a generic secreted cargo found in the lumen of the secretory pathway, including its vesicles. Only 41% of the ApoE-GFP puncta (n=573 puncta from 10 cells) colocalized with mCherry puncta, indicating that colocalization between ApoE-GFP and ApoE or ApoB100 is specific, rather than a consequence of all three molecules being secretory cargoes loaded into the same vesicular carrier. ApoE-GFP thus likely decorates the hepatic lipoproteins found within the lumen of the secretory pathway.

Figure 2.

Intracellular localization of ApoE-GFP. (A. and B.) ApoE-GFP colocalizes with the Golgi. Huh-7.5/ApoE-GFP cells were stained with an α-GM130 antibody and imaged using a dipping lens epifluorescence microscope. A single deconvolved image of a wide field of cells (A., scale bar, 50 µm) or a higher magnification image of a single cell (B., scale bar, 10 µm) are shown. (C.) ApoE-GFP colocalizes with ApoE. Huh-7.5/ApoE-GFP cells were stained with α-ApoE (clone EP1374Y) and with α-GFP (clone JL-8), and imaged on a Deltavision microscope. A deconvolved slice (scale bar, 10 µm), and a magnified view of the marked region (scale bar, 1 µm) are shown. (D.) ApoE-GFP colocalizes with ApoB100. Huh-7.5/ApoE-GFP cells were stained with α-ApoB100 and imaged using the dipping lens epifluorescence microscope. A single deconvolved image of a cell, and a magnified view of the marked region are shown (scale bars, 10 µm). Yellow stars are placed immediately below ApoE-GFP puncta that colocalized with ApoB100 puncta, while cyan stars are placed below ApoE-GFP puncta that did not colocalize with ApoB100. (E.) ApoE-GFP shows sparse colocalization with lumenal mCherry. Huh-7.5/SP-mCherry cells, expressing mCherry fused to the signal peptide of human serum albumin, and thus directed to the lumen of the secretory pathway, were imaged using the dipping lens epifluorescence microscope. A single deconvolved image of a cell, and a magnified view of the marked region are shown (scale bars, 10 µm). Yellow stars are placed immediately below ApoE-GFP puncta that colocalized with mCherry puncta, while cyan stars are placed below ApoE-GFP puncta that did not colocalize with mCherry.

ApoE-GFP and endogenous ApoE secretion rates are indistinguishable

ApoE is a secreted protein. To characterize ApoE-GFP’s kinetics of secretion from cells, we performed radioactive pulse-chase experiments in the Huh-7.5/ApoE-GFP cells and in the control Huh-7.5/EV Hygro cells. After a short radioactive pulse, we chased the cells in the absence of label while measuring ApoE- and ApoE-GFP- associated radioactivity in both media and cell lysates at regular intervals. The percent of total ApoE-GFP (Fig. 3A, blue trace) that was recovered from the media was at all times undistinguishable from the percent of endogenously-expressed ApoE recovered from the same Huh-7.5/ApoE-GFP conditioned media (Fig. 3A, red trace) or from media conditioned by control Huh-7.5/EV Hygro cells (Fig. 3A, black trace). The percent not secreted at each time point represents protein that remained cell-associated. ApoE-GFP thus possesses the same capacity to be secreted from the Huh-7.5 cell line as endogenous ApoE.

ApoE-GFP associates with secreted ApoE and ApoB100

ApoE is secreted from cells as lipoprotein particles of various sizes and lipid compositions. These include ApoB100-containing VLDL/LDL and ApoB100-free HDL particles. To be a good marker of lipoprotein egress, ApoE-GFP should retain ApoE’s ability to associate with itself and with secreted ApoB100. To test if this was the case, we performed immuno-precipitation (IP) assays on media conditioned by either Huh-7.5/ApoE-GFP cells or control Huh-7.5/EV Hygro cells. These experiments were done in the absence of detergent, to preserve the integrity of the lipoprotein particles. IP of Huh-7.5/ApoE-GFP conditioned media with an α-GFP antibody pulled down ApoE-GFP, as expected, but also untagged ApoE and ApoB100 (Fig. 3B, lane a). Reciprocal IP of the same media with α-ApoB100 pulled down ApoB100, ApoE-GFP, and untagged ApoE (Fig. 3B, lane b). These results indicate that secreted ApoE-GFP associates with both ApoB100 and untagged ApoE, likely as part of lipoprotein particles. To establish the specificity of the IP assay, we performed control IP with normal species-matched IgG, and IP of media conditioned by Huh-7.5/EV Hygro cells with the same sets of antibodies. IP of ApoE-GFP-free media with the α-ApoB100 antibody resulted in recovery of only untagged ApoE, as expected (Fig. 3B, lane e), while the other conditions resulted in minimal or no recovery of ApoE, ApoE-GFP or ApoB100 (Fig. 3B, lanes c, d, and f). The media samples used in these assays were conditioned by similar numbers of cells (Fig. 3C, left panels), and contained comparable total amounts of ApoB100 and ApoE (Fig. 3C, right panels). All in all, our localization, kinetic and biochemical assays establish ApoE-GFP as a useful marker for the analysis of the secretion of ApoE-containing lipoproteins.

Live cell imaging of ApoE-GFP transport

We imaged live Huh-7.5/ApoE-GFP cells with a microscope setup that allows us to excite the field of view using either epifluorescence or total internal reflection fluorescence microscopy (TIR-FM) illumination⁴⁹. Using this system, we identified numerous ApoE-GFP puncta that displayed rapid, saltatory movements, generally towards the periphery of the cell (Movies S1–S3). The paths that the ApoE-GFP puncta followed in their movements were generally straight (Fig. 4A–B). This was even more apparent when we performed a maximum intensity projection of all the images from a time series onto a single plane, thus highlighting the trajectories of particle movement (Fig. 4A and 4B, last panel). This type of dynamics are expected of secretory vesicles containing ApoE-GFP and moving along cytoskeleton tracks, likely microtubules, as previously documented for ApoE-GFP in macrophages⁴⁰, or for HCV⁴¹. Our results thus show that ApoE-GFP can readily be used in live cell imaging experiments to analyze lipoprotein secretion by hepatocytes.

Figure 4.

Live cell imaging of ApoE-GFP. (A.) A maximum intensity projection of all frames recorded during the time lapse imaging of a Huh-7.5/ApoE-GFP cell using the TIR-FM setup. A total of 141 frames acquired every 0.5 s were projected. Trajectories of single ApoE-GFP puncta are clearly visible. The cell is the same one depicted in Movie S1. Scale bar, 15 µm. (B.) Selected consecutive frames depicting the movement of an ApoE-GFP punctum, corresponding to the inset from panel A. Frame numbers are listed at the bottom of the frames (1 frame is 0.5 s). A maximum intensity projection of all 33 frames is shown as the last frame. Scale bar, 1 µm.

ApoE-GFP and infectious HCV egress

Since ApoE is also a functionally important component of infectious HCV particles^32,34–39, we next investigated whether ApoE-GFP expression supported infectious HCV production. To test this, we performed rescue experiments in the context of ApoE knockdown. Two Huh-7.5 derived ApoE knockdown cell clones (ApoE KD1 and KD2, respectively) expressed barely detectable levels of ApoE, compared to parental Huh-7.5 cells, or to a control knockdown cell line, EV KD (data not shown), consistent with previous reports^32,35,37,50. We transduced these cell lines with shRNA-resistant untagged ApoE, ApoE-GFP, or control EV lentiviruses. As expected, ApoE expression was not rescued by transduction of the cells with the empty lentiviral expression vector (Fig. 5A). In contrast, transduction with the lentivirus expressing untagged ApoE resulted in rescue of ApoE expression, and transduction with the lentivirus expressing ApoE-GFP resulted in comparable levels of expression of the fusion protein (Fig. 5A). We next launched HCV infection in these cells by HCV RNA electroporation and measured supernatant HCV infectivity titers as well as intracellular HCV RNA levels. The expression of ApoE-GFP in the EV KD background, where endogenous ApoE remains expressed (Fig. 5A), did not significantly change HCV infectivity release, compared to control, EV-transduced cells (Fig. 5B, left pair of bars). This indicates that ApoE-GFP does not act as a dominant negative factor with respect to the release of infectious HCV particles. Unfortunately, ApoE-GFP expression in the ApoE knockdown cell lines did not rescue infectious HCV particle release (Fig. 5B, compare second and third black bars to the second and third white bars, respectively). In these cells, the release of infectious HCV particles was indistinguishable from that observed when the ApoE KD clones were mock rescued by transduction with an EV (Fig. 5B, white bars). In contrast, exogenous expression of untagged ApoE was able to partially rescue HCV infectious particle release (Fig. 5B, compare second and third gray bars to the second and third white bars, respectively). In all but one cell population, intracellular HCV RNA accumulated to similar levels (Fig. 5C), and comparable HCV RNA amounts were delivered into these cells, as quantified at 6 h post electroporation (Fig. 5D). Overall, these findings ruled out a major inhibitory effect of ApoE-GFP expression on HCV genome replication, while establishing that ApoE-GFP cannot support the production of infectious HCV particles in the absence of untagged ApoE.

Figure 5.

ApoE-GFP does not support the production of infectious HCV particles in the absence of untagged ApoE. (A.) Huh-7.5 cells, Huh-7.5 engineered to express shRNA targeting endogenous ApoE protein expression (clones ApoE KD1 and ApoE KD2) and control knockdown cells (clone EV KD) were transduced with the following rescue vectors: EV, ApoE, or ApoE-GFP, as indicated, and were then selected. Lysates from the cells were processed for Western blotting using α-ApoE (AB947, top panel) and α-actin (bottom panel) antibodies. (B.) Infectivity of HCV particles released by the indicated ApoE knockdown cell lines at 72h post electroporation. The cells expressed the following rescue vectors: EV (white bars), ApoE (gray bars) or ApoE-GFP (black bars). Shown are means ± s.e.m obtained from 2 or 3 independent electroporations, with 3 virus samples analyzed for each electroporation. (C. and D.) Cell-associated HCV RNA copies quantified by qRT-PCR at 72 h (C.) or 6 h (D.) post electroporation in samples from the experiment presented in panel B. The lower RNA levels in the ApoE KD2/ApoE-GFP cells both at 6 h (panel D) and at 72 h (panel C) likely reflected a lower electroporation efficiency in that cell background. (E-G.) Pull down of HCV RNA. HCV RNA was electroporated into Huh-7.5/EV Hygro, Huh-7.5/SP-GFP, and Huh-7.5/ApoE-GFP cells. After 72 h, the media was harvested and subjected to IP or processed by Western blotting. (E.) Western blot of input media used for IP of HCV RNA. (F.) IP was performed on media conditioned by HCV-producing Huh-7.5/EV Hygro, Huh-7.5/SP-GFP, or Huh-7.5/ApoE-GFP cells using the indicated antibodies: normal rabbit IgG (white bars), rabbit α-GFP (gray bars), or rabbit α-ApoE (black bars, clone EP1374Y). The recovered HCV amounts were quantified and are expressed relative to the amounts pulled down by the normal IgG control. Two sets of electroporations with two IP per sample were performed. Shown are means ± s.e.m. (G.) HCV RNA amounts in the samples used for the pull downs from panel F. Left graph, total HCV RNA levels (Log₁₀ of RNA copy number) in 1/10 of the input sample. Right graph, relative HCV amounts in the same input samples. Shown are means ± s.e.m. Statistical differences (Student’s t-test: ns, p>0.05; *, p<0.05; **, p<0.01; ***, p<0.001). (H.) Co-IP of media conditioned by Huh-7.5/EV Hygro, Huh-7.5/SP-GFP, and Huh-7.5/ApoE-GFP cells in the absence of HCV infection. The cell types are indicated at the top of the blots, the IP antibodies at the bottom, the proteins detected at the right of the blots, and the antibodies used for Western blotting at the left of the blots.

We also inquired whether ApoE-GFP associated with the HCV particles. We launched HCV infection by RNA electroporation into Huh-7.5/EV Hygro cells, Huh-7.5/ApoE-GFP cells, and Huh-7.5/SP-GFP cells. The Huh-7.5/SP-GFP cells express a fusion between the signal peptide of calnexin and GFP, which results in translocation of GFP into the lumen of the ER and its subsequent secretion (Fig. 5E). We allowed these cells to secrete HCV for 72 h post electroporation, and then performed IP assays with the rabbit α-ApoE monoclonal antibody EP1374Y, which does not recognize ApoE-GFP, with rabbit α-GFP, or with control normal rabbit IgG. We then quantified the HCV RNA levels pulled down under these conditions. The α-ApoE pull down resulted in between 2.5 and 4-fold more HCV RNA being IP-ed compared to the normal IgG pull down (Fig. 5F), as expected given that ApoE is a component of the HCV particle²⁷. The α-GFP also pulled down 3-fold more HCV RNA than the normal IgG from the media conditioned by ApoE-GFP-expressing cells, but only pulled down background levels of HCV RNA from the media conditioned by the EV Hygro or SP-GFP-expressing cells (Fig. 5F). This indicates that ApoE-GFP physically associates with the HCV particle, and that this association is not mediated by the GFP tag. Comparable levels of HCV RNA were present in the media subjected to the IP reaction, although the amount of HCV RNA in the ApoE-GFP samples was slightly (1.5 times) higher than in the EV control sample (Fig. 5G).

Lastly, we expanded this analysis by inquiring whether GFP alone could associate with the lipoprotein particles. We subjected media conditioned by Huh-7.5/EV Hygro, Huh-7.5/SP-GFP, and Huh-7.5/ApoE-GFP cells to IP with α-ApoE, α-GFP, or normal rabbit IgG, as above. These samples were obtained in the absence of HCV infection. We found that α-ApoE pulled down ApoE from all the samples, as well as ApoE-GFP from the corresponding media sample, but not free GFP from the SP-GFP sample (Fig. 5H). In turn, α-GFP pulled down free GFP from the SP-GFP sample, and ApoE-GFP from the ApoE-GFP sample (Fig. 5H). ApoE was pulled down by the α-GFP only from the sample containing ApoE-GFP (Fig. 5H), further confirming the specificity of the association between ApoE-GFP and ApoE-containing lipoproteins or HCV particles.

Discussion

Previous studies have shown that knockdown of ApoE expression results in decrease of infectious HCV release from Huh-7.5 cells^{32,35,37,50,51}, and that infectivity release may be at least partly rescued by re-expression of knockdown-resistant ApoE^35,37,50. Our findings are consistent with these studies. In our hands, ApoE-GFP was unable to rescue HCV infectivity release when expressed in ApoE-depleted cells. ApoE-GFP nonetheless was incorporated into HCV particles. The infectivity defects of ApoE-GFP-containing HCV particles could thus be explained if ApoE-GFP prevents binding of HCV to its entry co-receptors, or if the particles containing ApoE-GFP are otherwise defective. We attempted to test whether our ApoE-GFP construct could bind LDLR, an ApoE and HCV receptor, but our investigations yielded inconclusive results. It thus remains possible that a putative defect of LDLR binding by ApoE-GFP may render ApoE-GFP-containing HCV particles non-infectious. Alternatively, the GFP tag might sterically clash with the E1E2 glycoprotein ectodomains on the surface of the HCV particle, or might mask glycoprotein domains involved in entry receptor interaction. Interestingly, HCV produced in the presence of ApoE-GFP and untagged ApoE is fully infectious. Presumably, whatever amounts of untagged ApoE become loaded onto an ApoE-GFP containing HCV particle, they are sufficient to preserve enough receptor binding such that the overall process of entry is not detectably perturbed. Alternatively, ApoE and ApoE-GFP could be loaded onto different HCV and lipoprotein particles, although this scenario is not favored in light of the results of our IP experiments.

Given these findings, could ApoE-GFP still be used to label and image HCV particles? If the structure and rate of production of the ApoE-GFP-containing HCV non-infectious particles were indistinguishable from the structure and rate of production of infectious ApoE-containing HCV particles, then ApoE-GFP might still be used in the analysis of HCV particle secretion. Unfortunately, only a small portion of released HCV particles are infectious, and they appear to be difficult to purify and characterize structurally^27,52. As such, showing equivalence between infectious ApoE-containing HCV particles and the non-infectious ApoE-GFP-containing HCV particles is challenging. We thus conclude that ApoE-GFP cannot be unambiguously used to mark and image infectious HCV particles during secretion from hepatic cells.

Nonetheless, our results document a behavior of ApoE-GFP that closely mirrors that of untagged ApoE with respect to lipoprotein release. We showed that ApoE-GFP colocalized intracellularly with untagged ApoE as well as with ApoB100, and was secreted from cells at a rate undistinguishable from that of untagged ApoE. We further showed that secreted ApoE-GFP interacted with both ApoE and ApoB100, as expected of a proper lipoprotein particle-associated marker. This association could of course be in part due to exchange of ApoE-GFP between lipoprotein particles, however, our intracellular colocalization experiments suggest that ApoE-GFP associates with the lipoprotein particles in the lumen of the secretory pathway, and not solely after secretion into the extracellular space.

Overall, this is, to our knowledge, the first in depth analysis of the usefulness of ApoE-GFP as a marker of hepatocyte lipoprotein secretion. Previous studies have characterized ApoE-GFP in macrophages⁴⁰, or have utilized ApoE-GFP in hepatocytes without performing a functional characterization of the constructs^41,42. We have performed thus a detailed analysis of ApoE-GFP’s behavior. Based on our findings, we propose that ApoE-GFP (or similarly made constructs) may be used in studies aiming, for example, to identify the route(s) of vesicular transport which shuttle ApoE-containing lipoproteins out of producing cells. Furthermore, quantitative kinetic imaging studies, including the study of whether and how the various ApoE isoforms affect the rates of lipoprotein secretion, may be performed using such FP-tagged ApoE constructs. Besides characterizing ApoE-GFP’s behavior, our study provides a framework for testing other fluorescently tagged markers of lipoprotein particles. Lastly, since ApoE has also been involved in neurodegenerative diseases^53,54 and cancer⁵⁵, ApoE-GFP may be used in disease-specific cellular contexts to potentially answer cell biology questions relevant for the understanding of those pathologies.

Materials and Methods

Expression vectors

The lentiviral vectors pLVX Phi, pLVX Phi3 and pLVX Hhi3 were derived from pLVX Puro (Clontech) and allow for cytomegalovirus promoter driven expression of genes of interest and puromycin (Phi and Phi3) or hygromycin (Hhi3) selection of transduced cells, respectively. pCMV6-XL5 hCalnexin was obtained from Origene, and pDNR-LIB hAlbumin was obtained from Open Biosystems.

pLVX Hhi3 hApoE3

A DNA sequence encoding an shRNA-resistant version of hApoE3 was synthesized, PCR-amplified using NT628 (5'-TATAACGCGTGCCACCATGGTCAAGGTTCTGTG-3') and NT636 (5'-GAGAGGATCCTATTAGTGATTGTCGC TGGGCACAGGG-3'), digested with MluI and BamHI and inserted into the MluI/BamHI sites of pLVX Hhi3.

pLVX Hhi3 hApoE3-mEGFP

hApoE3 (shRNArez) was amplified with NT117 (5'-TATATAGGTACCGC CACCATGGTCAAGGTTCTGTGGGCTGCGTTG-3') and NT637 (5'-ATAGGATCCGCGTGATTGTCG CTGGGCACAGG-3'), digested with KpnI and BamHI and inserted into the KpnI/BamHI sites of pmEGFP-N1 (Clontech). The resulting hApoE3-mEGFP was amplified using NT628 and NT644 (5'-GTGGTATGGCTGATTATGATCTAGAGTCGCGGC-3'), digested with MluI and XbaI and cloned into the MluI/XbaI sites of pLVX Hhi3.

pLVX Phi3 mEGFP-hApoE3

The following fragments were assembled into the cloning vector pSL1180 (Amersham): a DNA sequence encoding amino acids 1–22 of calnexin (its signal peptide) were amplified using NT356 (5'-TATATACGCGTG CCACCATGGAAGGGAAGTGGTTGC-3') and NT357 (5'-CAGACGGCCGATCATGAGCCTCAAC AATAGC-3') and digested with MluI and EagI; mEGFP was amplified from pmEGFP-N1 using primers NT344 (5'-TATACGGCCGATGGTGAGCA AGGGCGAGG-3') and NT358 (5'-CGACGCCTAG GCTTGTACAGCTCGTCCATGC-3') and digested with EagI and AvrII; mature hApoE3 coding sequence (lacking the signal peptide) was amplified with NT635 (5'-TATACCTAGGGGAGGAGGTAAG GTGGAGCAAGCGGTGG-3') and NT636 and digested with AvrII and BamHI. The resulting SP-mEGFP-hApoE3 was excised using MluI and BamHI and inserted into the MluI/BamHI sites of pLVX Phi3.

pLVX Phi SP-mEGFP

The signal peptide of calnexin described above was fused to the amino terminus of mEGFP via a EagI restriction enzyme site. The fusion, followed by a STOP codon, was inserted into the MluI/BamHI sites of pLVX Phi.

pLVX Phi3 SP-mCherry

The coding sequence of human albumin was amplified from pDNR-LIB hAlbumin using primers NT647 (5'-GAGCTCGAGC CACCATGAAGTGGGTAACCTTTATTTCC-3') and NT648 (5'-TGAGAATTCCTAA GCCTAAGGCAGC TTGACTTGCAGC-3'), and was digested with XhoI and EcoRI. The mCherry coding sequence was excised from pmCherry-N1 as an EcoRI-NotI fragment. The albumin and mCherry fragments were assembled between the XhoI and NotI sites of pLVX Phi3. The albumin ectodomain was then removed by site directed mutagenesis using primers NT1158 (5'-CGGCTTATTCCAGGGGTGTGGTGAGCAAGGGCGAGGAGG-3') and NT1159 (5'-CCTCCTCGCCCTTGCTCACCACACCCCTGGAATAAGCCG-3'). The expressed polypeptide contains the first 21 amino acids of human pre-proalbumin (its signal peptide) fused to the amino terminus of mCherry.

All cloning steps were performed using standard molecular biology techniques. GC-rich ApoE cDNA was amplified using Takara PrimeSTAR HS polymerase.

Cell lines and growth conditions

The cell lines used in this study and the lentiviral vectors used for their construction are listed in Table 1. All cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco), supplemented with L-glutamine, sodium pyruvate, 1% (v/v) non-essential amino-acids (NEAA, Gibco) (from here-on referred to as DMEM+) and 10% (v/v) fetal bovine serum (FBS, Sigma or HyClone) in humidified incubators at 37°C and in a 5% CO₂ atmosphere. Antibiotic selection was performed using puromycin (Sigma; 2–5 µg/mL for Huh-7.5 and 1 µg/mL for HeLa), and/or hygromycin (Invivogen, 150 µg/mL for Huh-7.5 or 250 µg/mL for HeLa). Selection was started two days after transduction and was maintained throughout subsequent growth.

Table 1.

Cell lines used in this study

Cell line	Parental Cell Line	Viral vector used for construction	Source / Reference
HEK293T	HEK293	N/A	Paul Bieniasz
Huh-7.5	Huh-7	N/A	59
Huh-7.5/EV Hygro	Huh-7.5	pLVX Hhi3	This Study
Huh-7.5/ApoE-GFP	Huh-7.5	pLVX Hhi3 hApoE3-mEGFP	This Study
Huh-7.5/ApoE-GFP/ SP-mCherry	Huh-7.5/ApoE-GFP	pLVX Phi3 SP-mCherry	This Study
Huh-7.5/SP-mEGFP	Huh-7.5	pLVX Phi SP-mEGFP	This Study
EVKD	Huh-7.5	pLKO.1 (EV)	This Study
EVKD/EV Hygro	EVKD	pLVX Hhi3	This Study
EVKD/ApoE-GFP	EVKD	pLVX Hhi3 hApoE3-mEGFP	This Study
EKD1	Huh-7.5	pLKO.1 shApoE	This Study
EKD1/EV Hygro	EKD1	pLVX Hhi3	This Study
EKD1/ApoE	EKD1	pLVX Hhi3 hApoE3	This Study
EKD1/ApoE-GFP	EKD1	pLVX Hhi3 hApoE3-mEGFP	This Study
EKD2	Huh-7.5	pLKO.1 shApoE	This Study
EKD2/EV Hygro	EKD2	pLVX Hhi3	This Study
EKD2/ApoE	EKD2	pLVX Hhi3 hApoE3	This Study
EKD2/ApoE-GFP	EKD2	pLVX Hhi3 hApoE3-mEGFP	This Study
HeLa	N/A	N/A	Paul Bieniasz
HeLa/ApoE-GFP	HeLa	pLVX Hhi3 hApoE3-mEGFP	This Study
HeLa/GFP-ApoE	HeLa	pLVX Phi3 mEGFP-hApoE3	This Study

Generation of ApoE knockdown cell lines

Huh-7.5 cells were transduced with lentiviral particles obtained from pLKO.1 vectors expressing either an ApoE-targeting shRNA (Broad Institute’s Genetic Perturbation Platform; target sequence NM_000041; clone ID TRCN0000010913) or EV negative control (Dharmacon). The cells were selected with puromycin before undergoing single cell sorting into 96-well plates on a BD FACSAriaII sorter. Individual cell clones were expanded in DMEM+ supplemented with 10% FBS and 1% Pen/Strep and screened for HCV pseudoparticle entry and HCV replication competence as well as ApoE knockdown by PCR and Western blotting.

Lentivirus transduction

Low-passage HEK293T cells (60–80% confluent) were co-transfected using polyethylenimine (Polysciences Inc.) or Fugene (Promega) and a plasmid cocktail composed of a lentiviral plasmid, an HIV-1 GagPol plasmid, and a vesicular stomatitis virus glycoprotein plasmid. Virus-producing cells were maintained in DMEM+ supplemented with 3% FBS. Virus-containing media was collected, cleared through a 0.45 µm filter, and used to transduce target cells in the presence of 4 µg/mL Polybrene (Millipore).

HCV

RNA transcripts of the infectious HCV clone J6/JFH1 were generated from plasmids as previously described⁴⁴. Briefly, plasmid DNA was linearized by digestion with XbaI, templates were purified by Minelute column (Qiagen), and 1 µg DNA was transcribed using the T7 RiboMAX™ Express Large Scale RNA Production System (Promega). Template DNA was removed by digestion with 1 U DNase I and RNA was cleaned up by RNeasy kit (Qiagen) with an additional on-column DNase I digestion step (Qiagen). RNA was quantified by absorbance at 260 nm, its integrity was verified by agarose gel electrophoresis, and 5 µg aliquots were stored at −80°C.

HCV RNA was transfected into Huh-7.5 cells by electroporation as described previously⁴⁴. Briefly, Huh-7.5 cells were trypsinized, washed twice with ice-cold Hanks’ Balanced Salt Solution (HBSS, Ca²⁺/Mg²⁺-free, Gibco), and resuspended to 1.5 × 10⁷ cells/mL in cold phosphate buffered saline (PBS). For each electroporation, 5 µg of HCV RNA were mixed with 6 × 10⁶ cells and immediately pulsed using an ElectroSquare Porator ECM 830 (BTX, Holliston, MA; 860 V, 99 µsec, five pulses). Electroporated cells were diluted in 30 mL of DMEM+ supplemented with 10% FBS, and plated in 6-well plates. After 6–8 h the cells were washed twice with HBSS and media was replenished. Supernatants were collected at 72 h, cleared through 0.45 µm filter, stored at −80°C before infectivity was determined by limiting dilution assay and TCID50 calculation by the method of Reed and Müench, as previously described⁴⁴. Cells were washed with cold PBS, and harvested in RLT buffer (Qiagen) containing 0.14 M β-mercaptoethanol, processed through Qiashredder columns and stored at −80°C before RNA was isolated using the RNeasy kit (Qiagen) with the additional on column DNase I digestion, and stored again at −80°C. HCV RNA levels were quantified against a standard curve using a one-step quantitative RT-PCR assay (Multicode-RTx HCV RNA kit, Luminex Corp.) targeting the 3' UTR of the viral genome and a Roche Light Cycler 480 and were expressed as copies of RNA/ng total cellular RNA.

Immunofluorescence

The cells were grown on glass-bottom MatTek dishes and processed as follows at room temperature (RT), unless otherwise indicated. The cells were fixed with paraformaldehyde (4% w/v in PBS, 10–15 min), washed once with PBS, treated with 250 mM Tris-HCl pH 8.0 for 10 min, washed twice with PBS, permeabilized with 0.2% (v/v) Triton X-100 in PBS for 20 min, blocked with 5% (w/v) bovine serum albumin (BSA) and 0.1% (w/v) Tween-20 in PBS for 20 min, stained with primary antibody in blocking buffer for 1 h at RT or overnight at 4°C, washed thrice with PBS, incubated with secondary antibody in blocking buffer for 1–1.5 h, washed thrice with PBS, stained with Hoechst 1:1000 in PBS for 1 or 20 min, washed once in PBS and imaged. Alternatively, permeabilization and blocking were done for 1 h at RT with 1% BSA, 2.5% donkey serum, 0.1% Triton X-100. Images were acquired on the following systems: an Olympus IX70 inverted microscope equipped with a 10X UplanFL N.A. 0.3 air objective and a Hamamatsu ORCA-ER camera and Metamorph software (Molecular Devices), an Olympus BX61WI upright microscope equipped with a UMPlanFL 60X N.A. 1.00 water dipping objective and a Hamamatsu Orca Flash 4.0 digital CMOS camera and Metamorph software, or a Deltavision system equipped with a 60X N.A. 1.42 oil objective and SoftWoRx software (Applied Precision). Deconvolution of fluorescence images using measured point-spread function was done in the SoftWoRx software. Alternatively, deconvolution was done with Autoquant (Media Cybernetics) with standard settings for each filter set using blind adaptive point-spread function. Pearson’s colocalization coefficient measurements were done in Imaris (Bitplane) and final images were processed using FiJi or Metamorph software. To determine colocalization of ApoE-GFP puncta with ApoE, ApoB100, or mCherry signals, the ApoE-GFP puncta were identified without regard to the distribution of signal in the other channel, then the channels were overlayed and the total number of ApoE-GFP puncta as well as the number of puncta that showed colocalization with a punctate signal in the opposite channel were counted.

Live cell imaging

Cells were imaged on a custom-built microscope⁴⁹, based on an Olympus IX-81 frame and equipped with a UApoN 100X, 1.49 N.A. objective. Cells were imaged with a 488 nm diode laser (Omicron, 100 mW) and an ET525/50M emission filter (Chroma). The temperature was maintained at 37°C throughout imaging using custom-built housing. During imaging, the media was replaced with cell imaging media (Hanks BBS [Sigma H-1387], 10 mM Hepes, pH 7.4), supplemented with 1% FBS. Images were collected every 0.5 seconds with a 200 ms exposure time. Cells were imaged under both azimuthal TIR-FM⁵⁶ and epifluorescence microscopy.

SDS-PAGE and Western blotting

Samples were lysed and denatured in reducing SDS sample buffer (SDS-SB) supplemented with Complete protease inhibitors (Roche). The proteins were separated on 4–12% Bis-Tris SDS-PAGE gels (Invitrogen), then transferred onto nitrocellulose membranes which were blocked in 5% (w/v) non-fat milk in TBS-T (Tris-buffered saline + 0.1 % w/v Tween-20), immunoblotted with primary antibodies and matched horseradish peroxidase (HRP)-conjugated secondary antibodies, treated with chemiluminescence substrates (ECL Prime, Amersham or West Pico or West Femto, Pierce), before being exposed to Amersham Hyperfilm MP autoradiography film and developed.

Antibodies

The following antibodies were used for Western blotting: mouse α-GFP (clone JL-8, Clontech); mouse α-(β)actin (clone AC-74, Sigma A5316); goat α-ApoE (Millipore, AB947); rabbit α-ApoE (Abcam ab52607, clone EP1374Y); goat α-ApoB100 (Millipore AB742); goat α-mouse IgG (HRP-conjugated, Sigma A9917); goat α-rabbit IgG (HRP-conjugated, Sigma A0545); and rabbit α-goat (HRP-conjugated, Sigma A5420). For TCID50, mouse α-HCV NS5A (clone 9E10) and goat α-mouse (HRP-conjugated, Jackson Immuno-Research, 115-035–146) were used. For IP, the following antibodies were used: rabbit α-ApoB100 (Abcam ab50069), an in-house made rabbit α-GFP⁵⁷, rabbit α-ApoE (clone EP1374Y, Abcam) purified normal rabbit IgG (Jackson ImmunoResearch or Thermo Scientific) or goat α-ApoE (Millipore AB947). For immunofluorescence, primary antibodies were rabbit α-ApoE (Abcam ab52607, 1:250), goat α-ApoB100 (Millipore AB742, 1:250), rabbit α-GM130 (clone EP892Y, Abcam ab52649, 1:250), and mouse α-GFP (Clontech clone JL8, 1:250). Secondary antibodies were goat α-mouse (AF488), goat α-rabbit (AF594), donkey α-rabbit (DyLight 594) or donkey α-goat (DyLight 680) from Molecular Probes (1:1000).

Secretion assay and immunoprecipitation

Cells were washed once with DMEM+ containing 1% FBS, allowed to secrete cargo in the same media for 5h, then media and cells were harvested. Media was cleared by a 3 min, 500 x g spin, then an aliquot of the media was mixed with SDS-SB, denatured, and stored at −20 °C. IP of 1.8 mL of cleared media was done as follows. All tubes were pre-coated with 1% BSA, 0.1% Tween-20, PBS for 1 hour, then washed 3 times with PBS. Media was incubated with 29 µg of rabbit α-GFP, rabbit α-ApoB100, or normal rabbit IgG, overnight at 4 °C. Protein G Dynabeads were from Thermo Scientific. A volume of 50 µL beads, pre-washed with PBS were added to the media and incubated 1h at RT, then washed 4 times with PBS, then denatured in SDS-SB. Samples were then processed by SDS-PAGE and Western blotting. For HCV IP assays, cells were electroporated with HCV RNA, the media was changed at 6 h, and then harvested at 72 h. The media samples were first incubated with Protein G Dynabeads, then the beads were discarded. Fresh beads (25 µL) were conjugated to 5 µg antibody for 30 min at 4°C. The beads were then washed 3 times with DMEM+ containing 10% FBS, then incubated with 400 µL HCV supernatant sample, overnight at 4°C. The beads were washed 3 times with PBS, then HCV RNA was extracted from half of the sample using QIAamp viral RNA mini kit (Qiagen). This co-IP was also repeated in the absence of HCV infection, where the eluted samples were analyzed by Western blotting.

Radioactive pulse-chase

This procedure was adapted from a previous study⁵⁸ as follows. Huh-7.5/EV Hygro and Huh-7.5/ApoE-GFP cells were plated at 3 × 10⁵ cells/well in 6-well plates. The next day, the cells were washed in PBS and then pulse labeled with ³⁵S-Cys/Met (120 µCi/well) in DMEM containing no cold Met and Cys, and supplemented with 1% FBS and 1% NEAA for 20 min. The cells were then washed twice in PBS and incubated for 0, 20, 40, 60, or 120 min of chase in DMEM containing cycloheximide (Millipore, 50 µM) and excess Met (1.5 mg/mL) and Cys (0.5 mg/mL). At each time point, the media was removed and the cells were washed once with PBS. The cells were lysed by rocking at 4°C for 30 min in 1 mL of Lysis buffer: 6.1 mM Na₂HPO₄, 4.5 mM NaH₂PO₄, 88.4 mM NaCl, 36.58 mM LiCl, 24.1 mM sodium deoxycholate, 1% (v/v) Triton X-100, 1% (w/v) SDS, pH 7.4, supplemented with 4 µL/mL phenylmethanesulfonyl-fluoride (Sigma, 0.1 M in ethanol) and Complete protease inhibitors. Total protein-incorporated radioactivity in the samples collected at each time point was determined after trichloroacetic acid precipitation, using scintillation fluid. IP of ApoE from media and cell lysates was done by incubation with 5 µL α-ApoE antibody (Millipore AB947) and 50 µL protein A Sepharose 4B beads (Invitrogen) in 1X NET buffer (0.15 M NaCl, 5 mM EDTA, 50 mM Tris, pH 7.5) supplemented with 1% (v/v) Triton X100, overnight at 4°C. The beads were then washed 3 times in 1X NET buffer supplemented with 1% (v/v) Triton X100 and 1% (w/v) SDS, then the bound protein was eluted in 1X NET buffer containing 1% (w/v) SDS. Samples were resuspended in denaturing buffer: 125 mM Tris-HCl, pH 6.8, 6M urea, 1 mM EDTA, 10 mM dithiothreitol, 4% (w/v) SDS, 25 mM β-mercaptoethanol, and separated by SDS-PAGE on 4–12% Bis-Tris gels. The gels were fixed, incubated with Autofluor (National Diagnostic) and exposed in a phosphorimager and the intensity of each band was quantified. Relative amounts of radiolabeled amino acid incorporation were calculated by dividing the intensity of the protein-specific band at the end of the pulse period by the measured total protein-contained radioactivity in the corresponding sample. Secretion of newly synthesized ApoE or ApoE-GFP was assessed by expressing the amount of radioactively labeled protein present in the media at a given time point as a percent of the total (secreted + cell-associated) radioactively labeled protein measured at the same time point.

Supplementary Material

Supp MovieS1

Movie S1 Timelapse of ApoE-GFP vesicle movement in a Huh-7.5/ApoE-GFP cell imaged using TIR-FM illumination settings. Frames are 0.5 seconds apart. Time elapsed is listed as min:sec. Scale bar is 15 μm. The same cell is shown in Fig. 4A.

Supp MovieS2

Movie S2 Timelapse of ApoE-GFP vesicle movement in a Huh-7.5/ApoE-GFP cell imaged using TIR-FM illumination settings. Frames are 0.5 seconds apart. Time elapsed is listed as min:sec. Scale bar is 15 μm. The same cell was also imaged using epifluorescence illumination settings and is shown in Movie S3.

Supp MovieS3

Movie S3 Timelapse of ApoE-GFP vesicle movement in a Huh-7.5/ApoE-GFP cell imaged using epifluorescence illumination settings. Frames are 0.5 seconds apart. Time elapsed is listed as min:sec. Scale bar is 15 μm. The same cell was also imaged using TIR-FM illumination settings and is shown in Movie S2.

Synopsis.

ApoE-GFP reproduces the behavior of untagged ApoE with respect to lipoprotein secretion from hepatic cells but does not support infectious hepatitis C virus production and is a useful marker for ApoE-containing hepatic lipoproteins.