Endosomal egress and intercellular transmission of hepatic ApoE-containing lipoproteins and its exploitation by the hepatitis C virus

Liver-generated plasma Apolipoprotein E (ApoE)-containing lipoproteins (LPs) (ApoE-LPs) play central roles in lipid transport and metabolism. Perturbations of ApoE can result in several metabolic disorders and ApoE genotypes have been associated with multiple diseases. ApoE is synthesized at the endoplasmic reticulum and transported to the Golgi apparatus for LP assembly; however, ApoE-LPs transport from there to the plasma membrane is largely unknown. Here, we established an integrative imaging approach based on a fully functional fluorescently tagged ApoE. We found that ApoE-LPs accumulate in CD63-positive endosomes of hepatocytes. In addition, we observed the co-egress of ApoE-LPs and extracellular vesicles (EVs) along the late endosomal trafficking route. Moreover, complexes of ApoE-LPs and CD63-positive EVs were found to be transmitted from cell to cell. Given the important role of ApoE in viral infections, we studied the hepatitis C virus (HCV) and found that the viral replicase protein NS5A is enriched in ApoE-containing intraluminal vesicles. Interaction between NS5A and ApoE is required for the efficient release of EVs containing viral RNA. These vesicles are transported along the endosomal ApoE egress pathway. Taken together, our data argue for endosomal egress and transmission of hepatic ApoE-LPs, a pathway that is hijacked by HCV. Given the more general role of EV-mediated cell-to-cell communication, these insights provide new starting points for research into the pathophysiology of ApoE-related metabolic and infection-related disorders. Author Summary The post-Golgi egress pathway of hepatocyte-derived ApoE-containing lipoproteins (ApoE-LPs) is largely unknown. By using integrative imaging analyses, we show that ApoE-LPs are enriched in CD63-positive endosomes suggesting that these endosomes might be a central hub for the storage of ApoE-LPs from which they are released into the circulation. In addition, we provide evidence for the co-egress of ApoE-LPs with extracellular vesicles (EVs) along the late endosomal route and their transfer from cell to cell. This pathway is hijacked by the hepatitis C virus that induces the production of ApoE-associated EVs containing viral RNA. Given the important role of ApoE in multiple metabolic, degenerative and infectious diseases, and the role of EVs in cell-to-cell communication, these results provide important information how perturbations of ApoE might contribute to various pathophysiologies.


Introduction
7 these FPs, especially under the acidic conditions in late endosomes where ApoE is expected to 137 reside. Therefore, we tagged ApoE with mTurquoise2 (mT2) and eYFP. Consistent with our 138 assumption, ApoE mT2 and ApoE eYFP were not fragmented (Fig S1A, upper right). Because mT2 is a 139 rapidly-maturing cyan monomer with very low acid sensitivity (pKa = 3.1) (39), we selected this 140 tagged ApoE for functional validation.

142
ApoE mT2 was efficiently secreted into the cell culture supernatant (Fig 1A). Moreover, the 143 association of secreted LPs with ApoE mT2 was well comparable to the one with wildtype (wt) ApoE 144 as determined by separation of LPs using sucrose density gradient centrifugation (peak density of 145 ApoE mT2 and ApoE wt = 1.05 vs. 1.04 g/ml, respectively) ( Fig 1B). Moreover, in addition to a weak 146 and diffuse ER-like pattern, ApoE mT2 formed strong and dotted puncta characteristic for LPs and 147 colocalized with ApoB, a well-established LP marker ( Fig 1C). Of note, ApoE puncta detected by 148 immunofluorescence in fixed cells were much dimmer than those containing mT2, thus increasing 149 sensitivity of our analyses, especially in live-cell imaging (Fig 1C). We further investigated ApoE mT2 150 subcellular distribution in nonhepatic cells having undetectable levels of ApoE such as HEK293T 151 and Hela cells (30,40). Upon ectopic expression of ApoE mT2 , we observed a dot-like pattern in both 152 cell lines, which was well comparable to the one detected in Huh7-Lunet cells (Fig S1B).

154
Next, we validated the functionality of ApoE mT2 by probing its capacity to rescue the production of 155 infectious HCV, which was used as readout because this virus incorporates ApoE into virions 156 intracellularly to increase viral infectivity (32, 41).To facilitate the analysis, we employed the HCV 8 nonhepatic cell line HEK293T-miR122, which does not express endogenous apolipoproteins but 165 supports HCV RNA replication (40), arguing that the expression of non-ApoE LPs in Huh7-derived 166 cells compensates, at least in part, for ApoE deficiency (43, 44) ( Fig S2). Also in these cells, ApoE wt 167 and ApoE mT2 rescued infectious HCV particle production ( Fig S2). Taken together, our data show 168 that mT2 is a novel and well-applicable tag for labeling and functional analyses of ApoE.

Endosomal trafficking and egress of ApoE in hepatocytes 171
Having established a suitable FP-tagged ApoE, we employed light and electron microscopy 172 methods to study the trafficking and egress route of ApoE in hepatocytes. First, we confirmed the 173 conventional trafficking route of ApoE, which starts at the ER where it is co-translationally delivered 174 into the lumen to enter the secretory pathway (23, 24). Consistently, in Huh7-Lunet/ApoE mT2 cells 175 we detected reticular ApoE mT2 signals overlapping with the ER marker PDI (Fig 2A, top row). In 176 addition, we observed condensed ApoE mT2 puncta in the Golgi area containing GM130, a marker 177 of the Golgi apparatus, consistent with the assembly of ApoE-LPs at this site (Fig 2A, middle row).

179
To determine whether hepatocyte-derived ApoE-LPs are released via an endosomal egress 180 pathway, we initially determined its colocalization with CD63, the commonly used marker of ILVs 181 that are sorted into late endosomes (45, 46). We detected numerous ApoE mT2 -containing structures 182 in Golgi-devoid areas and these signals predominantly overlapped with CD63, indicating 183 accumulation of ApoE in late endosomes (Fig 2A, bottom row). Consistently, a fraction of ApoE 184 signals overlapped strongly with Rab7 (a marker of late endosomes), but rarely with ADRP (a 185 marker of lipid droplets) (Fig S3). We further identified the ultrastructure of ApoE-CD63 positive 186 signals by correlative light and electron microscopy (CLEM) using lipid droplets as fiducial markers, 187 because they are easy to detect in both light and electron microcopy and have a unique distribution 188 and size in each Huh7 cell (Fig 2B). We found that ApoE-CD63 double-positive signals 189 predominantly corresponded to regions containing electron-dense vesicles of ~500 nm in diameter, 190 which is a typical feature of endosomal compartments (47) (Fig 2B, right panel).

192
With the aim to track and record the dynamics of ApoE association with CD63, we took advantage 193 of ApoE mT2 and conducted time-lapse confocal microscopy (Movie S1). We observed co-trafficking    Co-secretion and cell-to-cell co-transmission of ApoE 217 and endosome-derived extracellular vesicles 218 A recent study by Busatto and colleagues demonstrated that EVs in crude plasma frequently bind 219 to and fuse with LPs arguing for a physiological interaction between these two nano-particle species 220 (20). Given the cotrafficking of intracellular hepatic ApoE with CD63 and the secretion of ApoE-221 associated CD63 (Fig 2), we speculated that extracellular hepatic ApoE might associate with CD63-222 positive EVs via LPs. Given the difficulties to separate EVs from LPs (21, 52-54), we employed 223 ApoE-specific pull-down to isolate ApoE from the supernatant of Huh7-Lunet cells that had been 224 cultured in EV-depleted medium. Captured complexes were eluted under native conditions and 225 analyzed by EM revealing predominantly small vesicles, which had the size of regular LDL or large 226 HDL particles (mean diameter ~25 nm) ( Fig 3A). Of note, we detected in much lower quantity co-

249
While the ApoE-E1/E2 interaction appears to be critical for HCV particle production, NS5A has 250 been detected in purified EV preparations (59, 60), raising the question of whether NS5A follows 251 the ApoE endosomal egress pathway. To address this question, we monitored ApoE, NS5A and 252 E2 trafficking in HCV-replicating Huh7-Lunet cells stably expressing ApoE mT2 . FPs for NS5A and 253 E2 were chosen to allow clear spectral separation from each other and from ApoE mT2 . In each case, 254 fusion with the FP did not affect the functionality of the protein as shown here for ApoE mT2 , and 255 earlier for tagged NS5A and E2 (61, 62). To allow live-cell imaging under low biosafety conditions, 256 we took advantage of the HCV trans-complementation system (63) in which the HCV genome is 257 genetically split into a stably expressed unit encoding the viral assembly factors (core-E1-E2 eYFP -258 p7-NS2) and a self-replicating subgenomic replicon encoding the viral replicase proteins (NS3-4A-259 4B-5A mCherry -5B) (Fig 4A). To determine the overall subcellular distribution of FP-tagged ApoE mT2 , 260 NS5A mCherry , and E2 eYFP during the course of HCV infection, we acquired time-lapse images by 261 confocal spinning disc microscopy in 30 min intervals between 5 and 54 h post-electroporation 262 using minimum laser exposure to avoid phototoxicity. Prior to electroporation of the subgenomic 263 replicon, E2 eYFP showed a reticular ER-like pattern consistent with its ER retention (64). Around 25 264 h post-electroporation, E2 eYFP subcellular distribution began to change and NS5A mCherry -E2 eYFP abundance increased significantly over time (Fig 4C), much higher as compared to NS5A mCherry -270 E2 eYFP positive foci. We confirmed the high number of NS5A mCherry -ApoE mT2 double-positive foci at 271 a late stage of infection by live-cell imaging using a shorter time interval (10 sec/frame). Under this 272 imaging condition, NS5A mCherry -ApoE mT2 foci were readily detectable (Movie S5).    (Fig 4). Importantly, we 304 found that about half of ApoE -NS5A double-positive foci also contained CD63 ( Fig 5C, Fig S4E).

305
When we visualized NS5A and ApoE by super-resolution STED microscopy, in addition to the 306 reticular ER and the ring-like lipid droplet staining patterns of NS5A CLIPf , we detected ~100-200 nm 307 diameter NS5A CLIPf -positive structures that were decorated with ApoE SNAPf at CD63-positive sites 308 ( Fig 5D, arrows).

310
To determine the ultrastructure of ApoE-NS5A double-positive sites, we employed CLEM using

338
To verify the presence of NS5A in the ApoE-captured complexes, we transfected Huh7-Lunet cells 339 with a subgenomic replicon RNA encoding Nanoluciferase (Nluc)-tagged NS5A to allow its 340 sensitive detection in cell culture supernatants (Fig 7B, upper). In agreement with a previous report 341 (60), we observed time-dependent secretion of NS5A Nluc into the cell culture supernatant (Fig 4F).

342
Importantly, Nluc activity was clearly detected upon ApoE-specific immunocapture indicating a 343 direct or indirect association between NS5A and ApoE ( Fig 7B, lane 2). The specificity of the pull-structures correspond, at least in part, to EVs that were frequently associated with LP-like structures 349 ( Fig 7C, arrows).

418
to the best of our knowledge, the association between liver-generated LPs and endosome-derived

419
EVs is not well documented and their possible intercellular co-transmission has been unknown.

421
Our data suggest that in naïve and HCV-infected hepatocytes, ApoE-LPs and endosome-derived 422 CD63-positive ILVs/EVs not only share a common intracellular late endosomal trafficking route, but 423 also are partially co-secreted. These particle complexes forming intracellularly co-enter target cells,

424
arguing for a stable interaction between ApoE-LPs and CD63-positive ILVs/EVs. This would explain   Reagents and resources used in this study are provided in Table S1. DMEMcplt. FCS devoid of extracellular vesicles (EVs) was prepared as previously described (59).

487
The full names of constructs used in this study are given in the Supporting Table 1.

Antibodies and immunofluorescence reagents 490
All antibodies and immunofluorescence reagents used in this study are listed in S1    to be colocalized with a CD63 particle, if the ApoE particle has a nearest CD63 particle within a 765 maximum distance for at least a minimum number of consecutive frames. Otherwise, the ApoE 766 particle is considered as non-colocalized with a CD63 particle. We used a maximum distance of 5

1129
(D) Functionality of ApoE mT2 as determined by the rescue of infectious HCV particle production.