Trafficking of hepatitis C virus core protein during virus particle assembly

Abstract

Hepatitis C virus (HCV) core protein is directed to the surface of lipid droplets (LD), a step that is essential for infectious virus production. However, the process by which core is recruited from LD into nascent virus particles is not well understood. To investigate the kinetics of core trafficking, we developed methods to image functional core protein in live, virus-producing cells. During the peak of virus assembly, core formed polarized caps on large, immotile LDs, adjacent to putative sites of assembly. In addition, LD-independent, motile puncta of core were found to traffic along microtubules. Importantly, core was recruited from LDs into these puncta, and interaction between the viral NS2 and NS3-4A proteins was essential for this recruitment process. These data reveal new aspects of core trafficking and identify a novel role for viral nonstructural proteins in virus particle assembly.

Figure 1. Fluorescent labeling of functional HCV core protein.

(A) Schematic of HCV genome and polyprotein. The open bullet represents a signal peptide peptidase cleavage site; closed bullets represent signal peptidase cleavage sites; the open arrowhead represents the NS2–3 cysteine autoprotease cleavage site; closed arrowheads represent NS3-4A serine protease cleavage sites. (B) Schematic of TC tag inserted into the C terminus of core to create Jc1/core(TC). (C) Time course of Jc1 (closed circle) and Jc1/core(TC) (open circle) virus production after RNA electroporation. Values are averages of results from three independent transfections ± SEM. The grey line shows the limit of detection of this assay; the asterisk denotes the timepoint at which infectivity titers were significantly different (Student's unpaired t-test, p <0.05); NS denotes timepoints where the differences in infectivity titers were not statistically significant (p >0.05). (D) Western blot analysis of core protein expression from cells lysed at 48 h post-electroporation with the indicated viral transcript. Actin is shown as a loading control. (E) Live cell imaging of core protein. Huh-7.5 cells were infected with Jc1 or Jc1/core(TC) and labeled with FlAsH at 48 h post-infection. TC-tagged core was frequently observed in crescent structures (top panel, arrowhead) and in small puncta. Background signals were minimal with untagged Jc1 (bottom panel). (F) Effect of FlAsH labeling on virus assembly. Huh-7.5 cells were infected with Jc1/core(TC) and labeled with FlAsH or a mock label at 48 h post-infection. Supernatants were collected at 2, 4, 8, and 24 h post-labeling and titrated for infectivity. Values are averages from three replicates ± SEM. The grey line shows the limit of detection of this assay. No statistically significant difference (unpaired Student's t-test, p >0.05) was observed between FlAsH and mock-labeled cells at any time point. (G) Specificity of biarsenical labeling for core protein. Huh-7.5 cells were infected with Jc1 or Jc1/core(TC), labeled with FlAsH at 48 h post-infection, then fixed and stained for IF with anti-core antibody. Co-localization was observed in the merged image. For all micrographs, scale bars represent 10 µm.

Figure 2. Localization and kinetics of core protein in live cells.

(A) Localization of core in Huh-7.5 cells. Cells expressing GFP-ADRP were labeled with ReAsH at 48 h post-infection. Three forms of core were observed; i) ER-associated core; ii) core associated with ADRP-positive LDs; (iii) ADRP-independent core puncta. Panels show a magnification of ADRP-associated core as a cap on LDs (ii), and ADRP-independent core puncta (iii). Contrast was enhanced to show ER-associated core. Nuclei (N) are marked for reference. (B) Kinetics of intracellular core protein movement. The velocities of three representative LD-associated core particles (from Figure 2Aii) and core puncta (from Figure 2Aiii) were determined over the course of 45 minutes (0.75 frame/min) and the velocity at each timepoint shown. Individual particles were tracked by using Volocity Quantitation software. (C) Core puncta are sensitive to treatment with Nocodazole. Cells expressing GFP-ADRP were labeled with ReAsH at 48 h post-infection, then treated with 50 µM Nocodazole for 1 h. The velocity of individual ADRP-independent core particles in treated and untreated samples (n = 26) was determined over 10 frames (1 frame/min) by using Volocity Quantitation software. Values show average velocity of each particle over 10 frames, horizontal bars represent mean values of each group. The mean velocity of Nocodazole-treated cells was significantly different to non-treated cells (unpaired Student's t-test). (D–E) Co-localization of core with E2 and NS3. Cells expressing ADRP-CFP were labeled with FlAsH at 48 h post-infection, then fixed and stained for IF with anti-E2 (D) or anti-NS3 (E) antibodies. Nuclei (N) are shown for reference in the merged images. (D) Puncta of core co-localized with E2 puncta in areas adjacent to LDs (arrowheads). (E) NS3 localized to LDs containing caps of core (arrowheads). (F) Localization of core in ADRP-CFP-expressing cells electroporated with the indicated viral transcript (WT, wild type). Panels show ADRP, core, and merged images. The D88 mutant showed enhanced accumulation of core around LDs. For all micrographs, scale bars represent 10 µm.

Figure 3. Effects of Brefeldin A on core localization.

(A) Schematic of experimental design. Huh-7.5 cells expressing fluorescent protein-tagged ADRP were FlAsH-labeled at 48 h post-infection. Parallel sets of cells were treated with 1 µg/ml BFA for 4 h before drug wash out. Cells were fixed before (0 h) and after (4 h) BFA treatment and also during washout (6 h and 8 h), as indicated by open circles. (B) BFA treatment caused an increase in ADRP-associated core. Images of cells fixed after BFA treatment (4 h) and after washout (6 h). (C–E) Quantitation of images from each time point (n≥19). Image quantification was performed as described in Materials and Methods. All values show mean ± SEM, p values were calculated by using an unpaired Student's t-test. (C) BFA treatment causes accumulation of core-containing LDs. The mean number of core-containing LDs per cell was calculated for each time point. (D) BFA treatment increases the amount of core on each LD. The mean proportion of LD surface area occupied by core (expressed as a percentage) was calculated for each time point. (E) BFA washout increases the number of non-LD core puncta. The mean number of core puncta per cell was calculated for each time point. (F–G) Effects of BFA treatment on infectious HCV production. Triplicate samples of cells (F) or supernatants (G) that had been treated with BFA (1 µg/ml) were collected at the times shown in Figure 3A. Mean virus titers for untreated (black) or BFA-treated (blue) samples are shown for each time point (± SEM). The grey line shows the limit of detection of each assay. (H) Formation of core puncta after BFA washout. Cells expressing tagged-ADRP were treated with BFA (1 µg/ml) for 3 h and labeled with FlAsH. BFA was removed by multiple washes with HBSS immediately prior to imaging. Images were collected every hour from the time BFA was washed out (0 h). Arrowheads show newly formed puncta at each time point. For all micrographs, scale bars represent 10 µm. (I) Core motility during BFA treatment and shortly after drug washout. Three core puncta were randomly chosen from representative time courses (see Videos S5 and S6); their velocities were calculated by using Volocity Quantitation software and plotted over time.

Figure 4. Long term dynamics of core trafficking.

(A) Schematic of experimental design. Parallel cultures of Huh-7.5 cells expressing ADRP-CFP were labeled with FlAsH at 48 h post-infection (green bar) then incubated for 2, 8, or 24 h. At the indicated times, cells were labeled with ReAsH (red bar) and imaged (open circle). (B) Localization of newly synthesized core protein over time. Old core labeled with FlAsH (left column) and new core labeled with ReAsH (middle column) are shown together with ADRP in merged images (right column) at each timepoint. For the 8 h timepoint, cells were incubated±30 µM cycloheximide (CHX) to halt protein synthesis. Newly synthesized core was not detected in CHX-treated cells. For all images, scale bar is 10 µm and nuclei (N) are shown for reference. (C) Quantification of old and new core localization over time. The mean number of LD-associated core (top panel) and core puncta (bottom panel) per cell at each timepoint were calculated as described in Materials and Methods. For both graphs, values show mean number of particles at each time point ± SEM (n = 74 to 98). Statistical analysis was used to compare the change in each particle type over time. Asterisks indicate that a statistically significant change was observed in comparison to the previous time point (Student's t-test, p<0.05).

Figure 5. The interactions between NS2 and NS3 are essential for recruiting core from LDs.

(A) Localization of core in cells expressing ADRP-CFP electroporated with the indicated viral transcript. Panels show ADRP, core, and merged images; scale bars are 10 µm. (B) The NS2 K27A mutant shows enhanced accumulation of core around LDs. Quantification was performed as described in Materials and Methods; values show mean percent LD surface area occupied by core ± SEM (n≥10). (C) Enhanced accumulation of intracellular core in the presence of NS2 mutations (left panel) ± NS3 Q221L suppressor (right panel). Quantification was performed as described in Materials and Methods (n>38). Horizontal bars represent mean values, asterisks indicate statistically significant differences from WT (Student's t-test, p<0.05). (D–E) Dual labeling of LD-associated core and core puncta in the presence of NS2 Y39A mutation ± NS3 Q221L. Cells were electroporated with the indicated viral transcript then labeled with FlAsH followed by ReAsH after an 8 h chase. The mean number of LD-associated core (left panel) and core puncta (right panel) per cell at each timepoint were calculated as described in Materials and Methods. For all graphs, values show mean number of particles at each time point ± SEM (n = 23 to 71). For comparison with WT, please refer to Figure 4C.

Figure 6. Models of core trafficking and virus assembly.

(A) Our model of core protein trafficking is illustrated here. The question mark indicates that the process of virus assembly is still poorly understood. (B) Our model of virus particle assembly. Virus particle assembly likely requires the simultaneous recruitment of core protein from the surface of LDs and viral RNA from replication complexes, while nascent virus particles bud into the ER lumen. NS2 helps to coordinate this process by bringing together viral structural and NS proteins.