Cooperation between the Hepatitis C Virus p7 and NS5B Proteins Enhances Virion Infectivity

ABSTRACT

The molecular mechanisms that govern hepatitis C virus (HCV) assembly, release, and infectivity are still not yet fully understood. In the present study, we sequenced a genotype 2A strain of HCV (JFH-1) that had been cell culture adapted in Huh-7.5 cells to produce nearly 100-fold-higher viral titers than the parental strain. Sequence analysis identified nine mutations in the genome, present within both the structural and nonstructural genes. The infectious clone of this virus containing all nine culture-adapted mutations had 10-fold-higher levels of RNA replication and RNA release into the supernatant but had nearly 1,000-fold-higher viral titers, resulting in an increased specific infectivity compared to wild-type JFH-1. Two mutations, identified in the p7 polypeptide and NS5B RNA-dependent RNA polymerase, were sufficient to increase the specific infectivity of JFH-1. We found that the culture-adapted mutation in p7 promoted an increase in the size of cellular lipid droplets following transfection of viral RNA. In addition, we found that the culture-adaptive mutations in p7 and NS5B acted synergistically to enhance the specific viral infectivity of JFH-1 by decreasing the level of sphingomyelin in the virion. Overall, these results reveal a genetic interaction between p7 and NS5B that contributes to virion specific infectivity. Furthermore, our results demonstrate a novel role for the RNA-dependent RNA polymerase NS5B in HCV assembly.

IMPORTANCE Hepatitis C virus assembly and release depend on viral interactions with host lipid metabolic pathways. Here, we demonstrate that the viral p7 and NS5B proteins cooperate to promote virion infectivity by decreasing sphingomyelin content in the virion. Our data uncover a new role for the viral RNA-dependent RNA polymerase NS5B and p7 proteins in contributing to virion morphogenesis. Overall, these findings are significant because they reveal a genetic interaction between p7 and NS5B, as well as an interaction with sphingomyelin that regulates virion infectivity. Our data provide new strategies for targeting host lipid-virus interactions as potential targets for therapies against HCV infection.

INTRODUCTION

Hepatitis C virus (HCV), a positive-sense single-stranded RNA virus and a member of the Flaviviridae family, infects nearly 170 million people worldwide (1, 2). HCV infection is the leading cause of chronic hepatitis, liver cirrhosis, and hepatocellular carcinoma (3). While the previous interferon-based therapies for treatment of hepatitis C were effective in only 40% of those infected with genotype 1 virus, the most prevalent HCV genotype in the United States (4), newly developed therapies for hepatitis C now use interferon-free direct-acting antiviral (DAA) regimens, with over 90% success rates (5). However, antiviral resistance is still a problem, and a vaccine is not yet available for prevention of HCV transmission.

The HCV RNA genome is translated into a single polyprotein from an internal ribosome entry site located in the 5′ untranslated region of the genome. The resulting HCV polyprotein is co- and posttranslationally processed by host and viral proteases into 10 viral proteins, including three structural proteins (core, E1, and E2), the p7 protein, and six nonstructural proteins (NS2, NS3, NS4A, NS4B, NS5A, and NS5B, which is the RNA-dependent RNA polymerase [RdRp]) (6). The minimal set of viral proteins necessary for RNA replication is the nonstructural proteins NS3, NS4A, NS4B, NS5A, and NS5B (7). However, viral assembly is mediated by both the structural and nonstructural proteins, which have been described to have both physical and functional interactions with each other that contribute to viral replication and assembly (8–10). In particular, the HCV p7, NS2, and NS5A proteins have all been found to play major roles in viral assembly, although recently other viral proteins have also been found to contribute to viral assembly (8). The formation and release of fully infectious virions require HCV interactions with host lipid pathways, especially the very-low-density lipoprotein (VLDL) pathway (9, 11, 12). In fact, virions incorporate host apolipoproteins, cholesteryl esters, and phospholipids, including phosphatidylcholine and sphingomyelin, to become fully mature and infectious (10, 13, 14).

Studies of the mechanisms underlying HCV assembly and the production of infectious virions have been facilitated by the establishment of the HCV cell culture system, which was initially developed by using JFH-1, a genotype 2A isolate of HCV (15, 16). Since the development of this system, fully infectious clones of HCV have been generated from other HCV strains and genotypes, including intra- and intergenotypic chimeras (17). While JFH-1 does make infectious virus, overall its viral titers are quite low. Therefore, several groups have generated cell culture-adapted variants of JFH-1 by passaging the virus many times, yielding viruses with increased viral titers and/or replication (17). Studies on the mechanisms underlying the increased replication of these cell culture-adapted viruses have provided new insights into the HCV life cycle, including identifying mutations that enhance viral stability, delay viral polyprotein processing to enhance replication, increase virion assembly/release and infectivity, and characterize genetic interactions between viral proteins to promote these processes (18–28). Interestingly, many of the culture-adapted amino acids identified in these studies were unique to each study, suggesting that the full spectrum of mutations that promote HCV culture adaptation has not yet been described.

To generate an infectious clone of JFH-1 with enhanced replication, we sequenced a cell culture-adapted strain of JFH-1 that had increased viral titers compared to wild-type JFH-1. We identified nine amino acid changes in this virus and reconstructed the infectious clone containing these nine culture-adapted mutations to make JFH-1-M9. These adaptive mutations led to a 100-fold improvement in viral titer. We found that a synergistic interaction between p7 and NS5B promoted the production of HCV particles with increased infectivity. Our studies indicated that JFH-1 virions with only these identified culture-adapted mutations in p7 and NS5B had reduced levels of sphingomyelin, leading to the enhanced virion infectivity compared to wild-type JFH-1. Overall, these findings provide novel insights into the molecular mechanisms of HCV assembly by revealing that cooperation between the p7 and NS5B proteins can alter how the virus interacts with host lipids to promote viral infectivity.

MATERIALS AND METHODS

Cell lines.

Human hepatoma Huh-7.5 cells (29) were grown in Dulbecco's modification of Eagle's medium (DMEM; Mediatech) supplemented with 10% fetal bovine serum (FBS; Thermo Scientific) and 1% penicillin, streptomycin, and l-glutamine (Life Technologies). The identity of the Huh-7.5 cells used in this study was verified by using the Promega GenePrint STR kit (DNA Analysis Facility, Duke University).

Cell culture-adapted HCV.

The culture-adapted JFH-1 strain of HCV genotype 2A, which was generated by >30 passages in Huh-7.5 cells, was kindly provided by Michael Gale, Jr. (University of Washington). To identify the sequence of this culture-adapted strain of JFH-1, RNA was isolated from HCV-infected Huh-7.5 cells by using an RNeasy kit (Qiagen) and reverse transcribed by using the iScript cDNA synthesis kit (Bio-Rad). Subsequently, sequence analysis was performed with a set of primers covering the complete HCV genome (Tacgen). In most cases, three clones were sequenced for each amplicon, and only mutations identified in more than one clone were considered for further study.

Plasmids and site-directed mutagenesis.

All nucleotide and amino acid positions refer to the JFH-1 genome (GenBank accession number AB047639). Mutations in the HCV coding region were introduced into pENTR-SJ*, a molecular clone of JFH-1 that has the sequence of JFH-1 and a T7 promoter (pENTR-SJ*) (30), by using site-directed mutagenesis (QuikChange Lightning kit; Stratagene). Following sequence verification, correct sequences were subcloned back into pENTR-SJ* as described below and verified again by DNA sequencing. The following constructs were made: psJFH-1-M9 with nine identified culture-adapted amino acid changes was generated by site-directed mutagenesis, and its sequence was verified (GenScript); psJFH-1-M9-GND was generated by site-directed mutagenesis of psJFH-1-M9 to insert the NS5B D318N mutation, followed by subcloning of the HindIII-XbaI fragment; psJFH-1-S (core G32S, E1 A378T, E2 H614D, and p7 C766Y) was generated by subcloning the AgeI-KpnI fragment of psJFH-1-M9 into psJFH-1; psJFH-1-NS (NS5A D227V, S2338G, L2410P, and V2440L and NS5B R2676K) was generated by subcloning the AgeI-KpnI fragment of psJFH-1 into psJFH-1-M9; psJFH-1-S+NS5A-#4 and psJFH-1-NS5A-#4 were generated by site-directed mutagenesis of psJFH-1-NS or psJFH-1, respectively, to insert the NS5A V2440L mutation, followed by subcloning of the SpeI-HindIII fragment; psJFH-1-p7 and psJFH-1-p7+NS were generated by site-directed mutagenesis of psJFH-1 or psJFH-1-NS, respectively, to insert the p7 C766Y mutation, followed by subcloning of the BsiW1-NotI fragment; psJFH-1-5B and psJFH-1-p7+NS5B were generated by subcloning of the HindIII-XbaI fragment of psJFH-1-M9 into psJFH-1 and psJFH-1-p7, respectively, to insert the NS5B R2676K mutation. The following oligonucleotides were used for mutagenesis: for p7 C766Y, 5′-GCTGCGAGTGCGGCTAACTATCATGGCCTCCTATATTTTG-3′ and 5′-CAAAATATAGGAGGCCATGATAGTTAGCCGCAC-TCGCAGC-3′; for NS5A V2440L, 5′-GAGGACGATACCACCTTGTGCTGCTCCATGT-3′ and 5′-ACATGGAGCAGCACAAGGTGGTATCGTCCTC-3′; for core K78E, 5′-GGGGCGACCTGGCTCTCCCCAGGCCTTGC-3′ and 5′-GCAAGGCCTGGGGAGAGCCAGGTCGCCCC-3′; and for E2 T563I, 5′-CGCGCCACAAGTCTTGATGA-AACCAGTGGAGTT-3′ and 5′-AACTCCACTGGTTTCATCAAGACTTGTGGCGCG-3′.

In vitro transcription and electroporation of HCV RNA.

Plasmid DNA encoding HCV (pENTR-SJ* [30]) was linearized by using XbaI. Purified linearized DNA was used as a template for in vitro transcription with a MEGAscript T7 transcription kit (Promega). RNA was purified to be free of DNA and transfected into Huh-7.5 cells via electroporation, as follows: 5 μg of RNA was mixed with 4 × 10⁶ Huh-7.5 cells in phosphate-buffered saline (PBS) and electroporated at 250 V and 950 μF with a Gene Pulser Xcell system (Bio-Rad). At 4 or 24 h postelectroporation, cells were washed extensively with PBS and complete DMEM (cDMEM).

HCV infection.

JFH-1 was propagated, and infectivity was determined by using anti-NS5A (9E10) antibody (gift of Charles Rice, Rockefeller University), by focus formation assay as described previously (15, 31). Intracellular infectivity was determined following four freeze-thaw cycles of infected cell pellets followed by focus formation assay, as described previously (32). All HCV infections were done at a multiplicity of infection (MOI) of 0.2. Specific infectivity was calculated as a ratio of extracellular titer (focus-forming units [FFU] per milliliter) to extracellular HCV RNA copies per milliliter at 72 h postinfection.

Quantification of HCV RNA.

Viral RNAs were isolated from cellular supernatants by using a QIAamp viral RNA kit (Qiagen) or from cells by using an RNeasy kit (Qiagen), as recommended by the manufacturer. HCV RNA copy number was measured in triplicate by quantitative real-time PCR (qRT-PCR) by using a TaqMan one-step assay (TaqMan Fast Virus 1-Step Mix; Qiagen) with an HCV-specific probe targeting the 5′ untranslated region of HCV (assay identifier [ID] Pa03453408_s1), as previously described (33). Analysis was done with an Applied Biosystems Step One Plus RT-PCR system. The copy number of HCV was calculated by comparison to a standard curve of a full-length in vitro-transcribed HCV RNA.

Determination of sphingomyelin content.

Sphingomyelin contents of culture supernatants harvested at 72 h post-transfection of in vitro-transcribed HCV RNA into Huh-7.5 cells were measured using a fluorometric sphingomyelin assay kit (Abcam). Supernatants were spun down to pellet virions, filtered, or directly measured, all resulting in similar sphingomyelin content readings. Subsequent supernatants were therefore measured directly. The sphingomyelin content of untransfected Huh-7.5 cell supernatant was measured and subtracted as a control. The relative sphingomyelin content was determined by dividing the sphingomyelin content (micromolar) by the viral RNA (copies per milliliter) in the supernatant.

Immunofluorescence analysis and confocal microscopy.

Cells were fixed in 4% paraformaldehyde in PBS, permeabilized with 0.2% Triton X-100 in PBS, and blocked with 10% FBS in PBS. Slides were stained with HCV core antibody (Abcam; 1:500 dilution), washed three times with PBS, and stained with conjugated Alexa Fluor 633 secondary antibody (Life Technologies) along with 4′,6-diamidino-2-phenylindole (DAPI; Life Technologies). Samples were incubated with boron-dipyrromethene (BODIPY) (493/503; Invitrogen; 1:1,000 dilution) for the last 10 min of secondary staining, washed, and mounted with ProLong Gold (Invitrogen). Imaging was performed by using a Zeiss 710 laser scanning confocal microscope (Light Microscopy Core Facility, Duke University). All images were processed with NIH Fiji/ImageJ (34).

Quantitative imaging analysis.

Lipid droplet size was calculated by using a macro written in Fiji. Lipid droplets were counted as particles in each of 8 to 15 cells, and the area of each was measured in square micrometers. Lipid droplets were grouped into those with areas smaller than 0.5 μm², between 0.5 μm² and 1.5 μm², and greater than 1.5 μm². The percentage of total lipid droplets with an area greater than 1.5 μm² for each cell is represented with a box-and-whisker plot. The circularity of lipid droplets was measured in Fiji by using the circularity function, where a value of 1 would indicate perfect circularity.

Statistical analysis.

Student's unpaired t test and one-way analyses of variance (ANOVAs) were used for statistical analysis of the data. Graphed values are presented as means ± standard deviations (SD) (n = 3 or as indicated) (*, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.001).

RESULTS

Cell culture adaptation of JFH-1 leads to enhanced infectious virion production.

To compare the replication of a cell culture-adapted JFH-1 virus that had been generated by >30 passages in Huh-7.5 cells, a human hepatoma cell line that supports high levels of HCV replication (35), to that of the parental JFH-1 virus, we infected Huh-7.5 cells with JFH-1 and the culture-adapted JFH-1 strain of HCV at an MOI of 0.2. Three days after infection, we harvested and determined the titer of the cellular supernatants on naive Huh-7.5 cells by using a focus formation assay (15, 31). We found that culture-adapted JFH-1 had nearly a 100-fold increase in extracellular titer compared to the parental JFH-1 strain (Fig. 1A). In addition, at 3 days after infection at an MOI of 0.2, nearly three times as many Huh-7.5 cells were HCV positive, as measured by analysis of fixed cells immunostained for the HCV NS5A protein (Fig. 1B). To determine if culture adaptation increased the intracellular viral RNA levels of JFH-1, we measured intracellular HCV RNA copy numbers by qRT-PCR at various time points after infection with JFH-1 and the culture-adapted JFH-1. The results indicate that the culture-adapted JFH-1 had higher levels of intracellular HCV RNA than the parental JFH-1 virus at all time points tested (Fig. 1C). However, the magnitude of increase in viral RNA levels of the culture-adapted JFH-1 compared to the parental JFH-1 was not as great as the difference found between the extracellular titers of the two viruses. This result revealed that the culture-adapted JFH-1 produced more virus than the parental JFH-1. This result also suggested that the culture-adapted virus had acquired mutations that permitted this increased growth, as had been seen previously (18–28, 36).

FIG 1.

Increase in HCV-JFH-1 titers with cell culture adaptation. (A) Viral titers were determined by focus formation assay (focus-forming units [FFU]/ml) at 72 h after infection with JFH-1 or cell culture-adapted JFH-1 (C. Adap). (B) The percentage of HCV NS5A-positive cells was determined at 72 h postinfection by counting over 100 cells from each of three different fields of view. (C) The intracellular HCV RNA levels in infected cells were measured by qRT-PCR at the indicated time points. Values are presented as means ± SD (n = 3). ***, P ≤ 0.001 by unpaired Student t test (A and B) or by one-way ANOVA (C). The data presented in panels A and C are representative of three different experiments. (D) Schematic diagram of the JFH-1 open reading frame. The positions of the 9 identified cell culture-adapted mutations are indicated with triangles. Table 1 gives the exact positions in the HCV genome.

To identify the amino acid changes that confer this increased virus production in the culture-adapted JFH-1, we isolated RNA from infected cells, reverse transcribed the RNA, amplified the cDNA with HCV gene-specific primers, and then sequenced the amplicons. Analysis of the resulting sequence revealed that the culture-adapted JFH-1 contained 8 synonymous changes and 9 nonsynonymous changes in the sequence of the viral RNA (Fig. 1D). In this study, we report the characterization of the function of the 9 nonsynonymous changes. Of the identified culture-adapted mutations, only NS5A V2440L has been previously identified in JFH-1 viruses that have been culture adapted (19, 20, 24).

Culture-adapted mutations in JFH-1 increase viral titer and specific infectivity.

To confirm that this set of nine amino acid changes was responsible for the enhanced virus production of the culture-adapted JFH-1, we introduced the whole set of mutations into the JFH-1 parental genome (JFH-1-M9). We then transfected the in vitro-transcribed HCV RNA containing the JFH-1-M9 mutations or the parental JFH-1 (JFH-1-Wt) into Huh-7.5 cells by electroporation and measured HCV production by using several different assays. First, we measured intracellular HCV RNA levels over time. We found that the JFH-1-M9 virus had a slight increase (less than 10-fold) in intracellular viral RNA levels over JFH-1-Wt (Fig. 2A), suggesting that the nine cell culture-adapted mutations had only modest effects on HCV intracellular RNA levels. Importantly, the viral RNA replication-deficient virus JFH-1-M9 GND, which has an inactivating mutation in the NS5B polymerase active site, had decreased viral RNA levels over time, as expected (Fig. 2A).

FIG 2.

The JFH-1-M9 infectious clone with culture-adapted mutations has increased viral titer and specific infectivity compared to JFH-1. (A) Intracellular HCV RNA levels in Huh-7.5 cells transfected with in vitro-transcribed JFH-1 RNA constructs (wild-type [Wt], 9 culture-adapted amino acid changes [M9], or with an inactivating mutation in the NS5B active site [GND]) were measured by qRT-PCR at the indicated time points. (B) Time course of HCV RNA release into the supernatant was determined by qRT-PCR following transfection of indicated constructs. (C) Release of HCV virions into the supernatant from transfected cells was determined at 72 h postinfection by focus formation assay. (D) Intracellular titer of HCV at 72 h after transfection of the indicated RNA constructs. (E) Specific infectivity of the indicated viruses was calculated as a ratio of the titer to the HCV RNA in the supernatant. Values are presented as means ± SD (n = 3). *, P ≤ 0.05; **, P ≤ 0.01; and ***, P ≤ 0.001, by unpaired Student t test (C to E) or by one-way ANOVA comparing Wt to M9 (A and B). Panels shown are representative of three independent experiments.

To test if these engineered mutations affected HCV production, we measured both the extracellular HCV RNA levels and the extracellular viral titer of both JFH-1-Wt and JFH-1-M9. Compared to JFH-1-Wt, the transfected JFH-1-M9 RNA produced a >2-log-higher viral titer at 72 h (Fig. 2C), while it had less than 1-log-higher levels of extracellular viral RNA over the time course (Fig. 2B). The intracellular titer of JFH-1-M9 was also higher than that of JFH-1-Wt, suggesting that there were no major differences in viral release between the two viruses (Fig. 2D). Importantly, we found that the extracellular specific infectivity of the JFH-1-M9 virus was nearly 100-fold higher than that of JFH-1-Wt (Fig. 2E). In summary, while the JFH-1-M9 virus showed only modest increases in intracellular viral RNA and extracellular viral RNA compared to JFH-1-Wt, it displayed a significantly higher increase in virus titer and specific infectivity.

Culture-adapted mutations in either the structural or nonstructural genes of JFH-1 only slightly increase infectious viral titer and specific infectivity.

To identify which of the culture-adapted mutations contributed to the enhanced virus production of JFH-1-M9, we inserted various combinations of the culture-adapted mutations into JFH-1 (Fig. 3A). In our first analysis, we made constructs that contained the previously identified NS5A V2440L culture-adapted mutation (JFH-1-NS5A-#4) (19, 20, 24), all of the mutations of JFH-1 found in the structural genes and p7 (JFH-1-S; amino acid changes 1 to 4), or the nonstructural genes (NS; amino acid changes 5 to 9) (Table 1). We then determined the impact of these mutations on HCV intracellular RNA levels. HCV RNA was in vitro transcribed and transfected into Huh-7.5 cells, and 72 h later, intracellular HCV RNA levels were measured by qRT-PCR (Fig. 3B). We found that JFH-1-NS5A-#4, JFH-1-S, and JFH-1-NS all had increased levels of intracellular HCV RNA compared to JFH-1 (Fig. 3B). Interestingly, addition of the previously identified replication-enhancing NS5A V2440L mutation (19) to the JFH-1-S construct (JFH-1-S+NS5A-#4) resulted in levels of viral RNA even lower than those of the parental JFH-1 (Fig. 3B).

FIG 3.

Culture-adapted mutations in both structural and nonstructural genes of JFH-1 slightly enhance specific infectivity. (A) Schematic diagram of the JFH-1 open reading frame with mutations indicated by asterisks. Table 1 has the exact positions in the HCV genome. (B to E) At 72 h after transfection of the indicated RNA constructs into Huh-7.5 cells, HCV intracellular RNA levels were determined by qRT-PCR (B), extracellular titer was determined by focus formation assay (C), HCV RNA release into the supernatant was determined by qRT-PCR (D), and specific infectivity was determined (E). Values are presented as means ± SD (n = 3). Panels shown are representative of three independent experiments.

TABLE 1.

Amino acid substitutions identified in cell culture-adapted JFH-1^c

Mutation	Amino acid substitutiona	Protein	Amino acid substitutionb	Domain location
1	G32S	Core	G32S	Domain 1
2	A378T	E1	A187T	Transmembrane domain
3	H614D	E2	H231D	Core ectodomain
4	C766Y	p7	C16Y	N-terminal helix 1
5	D2227V	NS5A-1	D251V	Domain II
6	S2338G	NS5A-2	S362G	Domain III
7	L2410P	NS5A-3	L435P	Domain III
8	V2440L	NS5A-4	V464L	Domain III
9	R2676K	NS5B	R234K	Interconnecting region of palm domain

^aPosition of the substitution in polyprotein.

^bPosition in the individual protein.

^cSynonymous substitutions were as follows: R502 (AGG to GGG), T791 (ACC to ACT), P1118 (CCC to CCT), L1232 (TTG to CTG), G1450 (GGA to GGG), V1460 (GTA to GTG), L1679 (CTG to TTG), and F1709 (TTT to TTC).

To determine how these culture-adapted mutations affected the production of HCV, we measured the HCV titer and RNA levels present in the cellular supernatant 72 h after HCV RNA transfection. The results indicate that addition of the structural amino acid changes (JFH-1-S) led to the greatest increase in HCV titer compared to JFH-1 (Fig. 3C). However, none of these constructs resulted in the same extracellular viral titer as did JFH-1-M9, suggesting that an interaction between the structural and nonstructural genes is required for the increased titer of JFH-1-M9. Addition of NS5A V2440L to JFH-1-S (JFH-1-S+NS5A-#4) also did not result in a further increase in titer for the JFH-1-S construct, which was likely a result of the decreased intracellular viral RNA levels observed with this construct (Fig. 3C and B). For all of these HCV constructs tested, the amount of viral RNA released into the supernatant mirrored the amount of viral titer found in the supernatant (compare Fig. 3D to C), and thus, only minor differences in the specific infectivities of these viruses were observed, with none reaching the level of specific infectivity displayed by JFH-1-M9 (Fig. 3E).

Culture-adapted mutations in p7 and NS5B provide the minimal set of mutations for enhanced viral titers and specific infectivity of JFH-1.

Because the JFH-1-M9 virus had increased titer and specific infectivity compared to JFH-1-Wt, we hypothesized that the culture-adapted mutations in the structural genes would be responsible for this phenotype. While, following transfection into Huh-7.5 cells, the JFH-1-S RNA did produce the highest viral titer, it was still not as high as JFH-1-M9, however. Therefore, neither mutations in the structural HCV genes nor mutations in the nonstructural HCV genes alone could recapitulate the increase in titer and specific infectivity seen with JFH-1-M9. This suggested that cooperation between the structural and nonstructural genes was responsible for the increased virus production of JFH-1-M9 compared to JFH-1. Indeed, interactions between the HCV p7 protein and the nonstructural proteins have been shown to be important for assembly of infectious HCV particles (36–40). Further, p7 has been shown to be important for HCV capsid assembly and envelopment (41), and recently, the HCV RNA-dependent RNA polymerase (RdRp) protein NS5B has been found to have a role in viral assembly, although the mechanism was not described (40). Therefore, to test if interactions between the identified culture-adapted mutations in p7 and NS5B contributed to the increased specific infectivity of JFH-1-M9, we generated HCV constructs that contained the culture-adapted p7 (C766Y) or NS5B (R2676K) mutations alone or in combination or p7 along with the complement of nonstructural gene mutations (JFH-1-p7+NS) (Fig. 4A). All of these RNAs displayed less than 1-log difference of intracellular viral RNA and produced similar levels of extracellular viral RNA (Fig. 4B and D). The JFH-1-p7+NS virus was the most similar to JFH-1-M9 in all the parameters of HCV infection tested (Fig. 4B to D). While the JFH-1-p7 and JFH-1-NS5B viruses had viral titers and extracellular RNA levels similar to those of JFH-1, addition of both the p7 and NS5B mutations to JFH-1 increased the titer and specific infectivity of JFH-1 to nearly the same level as that of JFH-1-M9 (Fig. 4C to E). This demonstrated that these 2 amino acid changes together were responsible for the majority of the increased specific infectivity of JFH-1-M9 compared to JFH-1.

FIG 4.

Culture-adapted mutations in p7 and NS5B enhance the titer and specific infectivity of JFH-1. (A) Schematic diagram of the JFH-1 open reading frame with mutations indicated by asterisks. Table 1 gives exact positions in the HCV genome. (B to E) At 72 h after transfection of the indicated RNA constructs into Huh-7.5 cells, HCV intracellular RNA levels were determined by qRT-PCR (B), extracellular titer was determined by focus formation assay (C), HCV RNA release into the supernatant was determined by qRT-PCR (D), and specific infectivity was determined (E). Values are presented as means ± SD (n = 3). Panels shown are representative of three independent experiments.

Specific HCV culture-adapted mutations alter lipid droplet size and circularity.

JFH-1 viral assembly is known to take place in association with endoplasmic reticulum (ER) membranes around cytosolic lipid droplets (8, 42), and the localization of the HCV core protein in relationship to lipid droplets has been shown to correlate with viral infectivity (36). Therefore, we analyzed the localization of the HCV core protein in relationship to lipid droplets in Huh-7.5 cells by immunostaining and confocal microscopy following transfection of the HCV RNA constructs containing various culture-adapted mutations. However, we did not observe any remarkable changes to the localization of the HCV core protein in relationship to lipid droplets under any condition, even in the JFH-1-M9 virus that had the highest viral titers and virion specific infectivity (Fig. 5A). Interestingly, we did find that the lipid droplets were larger in JFH-1-M9-transfected cells than in JFH1-Wt-transfected cells (Fig. 5A; quantified in Fig. 5B). In addition, the lipid droplets were also larger in cells that had been transfected with any construct containing the culture-adapted mutation in p7, C766Y. While strict calculation of lipid droplet size suggested that the NS5B R2676K change also increased lipid droplet size, we noticed that this was a result of many smaller lipid droplets in close proximity, giving the appearance of larger, nonspherical lipid droplets (Fig. 5A and B). This lipid droplet aggregation gives the illusion of large lipid droplets with reduced average circularity in our calculations (Fig. 5C). Taken together, these results suggest that the C776Y culture-adapted mutation in p7 promotes an increase in lipid droplet size but that the NS5B R2676K mutation alone does not promote a similar increase in lipid droplet size and instead induces aggregation of small lipid droplets.

FIG 5.

Specific culture-adapted mutations in HCV affect lipid droplet size and circularity. (A) Confocal micrographs of Huh-7.5 cells transfected with indicated in vitro-transcribed HCV RNAs that were stained for lipid droplets (BODIPY, green), HCV core protein (red), and nuclei (DAPI, blue). Crop images are taken from the region with the white box in the merged image. Bar, 10 μm. (B) Percentage of lipid droplets larger than 1.5 μm² is graphed in box-and-whisker plots, representing the minimum, first quartile, median, third quartile, and maximum. (C) The average circularity of lipid droplets is graphed in box-and-whisker plots. Values are presented as in panel B. For panels B and C, all quantifications were conducted on at least 8 cells per HCV clone. **, P ≤ 0.01, and ***, P ≤ 0.001, by unpaired Student t test.

JFH-1-p7+NS5B virions have decreased sphingomyelin content.

While p7 C766Y did promote a change in lipid droplet size during HCV replication, this amino acid change is not sufficient to explain the increase in virion specific infectivity that we saw with both culture-adapted amino acid changes in p7 and NS5B. Interestingly, the NS5B R2676 amino acid residue is in the sphingomyelin-binding pocket of NS5B, in close proximity to the exit tunnel of the RdRp (43, 44). As sphingomyelin is known to be present in the HCV virion (14), we hypothesized that the increased infectivity of the p7+NS5B virus could be due to alterations in sphingomyelin virion content compared to that of JFH-1. To test this, we used a sphingomyelin assay to directly determine the sphingomyelin levels in supernatants harvested 72 h following HCV RNA electroporation into Huh-7.5 cells (Fig. 6A). We found that the relative amount of sphingomyelin (normalized to viral RNA in the supernatant) was decreased in the supernatants of JFH-1-p7+NS5B-transfected cells, compared to JFH-1-transfected cells (Fig. 6A). Our results suggest that the culture-adapted mutations in p7 and NS5B are required to enhance the infectivity of JFH-1 by decreasing the sphingomyelin content of the virion.

FIG 6.

Role of sphingomyelin in the infectivity of HCV. At 72 h after transfection of the indicated constructs into Huh-7.5 cells, the sphingomyelin content of virions was determined by a fluorometric sphingomyelin assay and the sphingomyelin content relative to the extracellular RNA was calculated (A), the extracellular titer was determined by focus formation assay (B), HCV RNA release into the supernatant was measured by qRT-PCR (C), and the specific infectivity was determined (D). Values are presented as means ± SD (n = 3). ***, P ≤ 0.001, by one-way ANOVA comparing p7+NS5B to the other viruses.

To determine if other published cell culture-adapted mutations in JFH-1 had mechanisms of increased infectivity similar to those of the JFH-1-p7+NS5B virus, we constructed viruses with a mutation in either core (K78E) or E2 (T563I). JFH-1 containing either of these mutations has previously been shown to have an increased specific infectivity compared to the wild-type virus (24, 25). Similar to these prior studies, we found that each of these amino acid changes resulted in an increased specific infectivity of the virions compared to JFH-1 (Fig. 6B to D). However, we found that the sphingomyelin content of the supernatants from Huh-7.5 cells transfected with these viruses was similar to that of JFH-1 and was increased relative to that of JFH-1-p7-NS5B (Fig. 6A). These results suggest that the decrease in sphingomyelin content of virions from JFH-1 is not a general mechanism to increase virion specific infectivity but a unique result of the culture-adapted mutations in p7 and NS5B.

DISCUSSION

Our results demonstrate that the mechanism of culture adaptation of a JFH-1 strain of HCV is primarily due to the concerted actions of the p7 and NS5B proteins to promote enhanced virion infectivity. These two culture-adapted amino acid changes in p7 and NS5B contributed to 90% of the increase of infectious titers of the culture-adapted variant compared to the parental JFH-1, while neither amino acid on its own contributed very much to the increased infectivity of the virions. In addition, we found that culture-adapted amino acid changes in p7 and NS5B reduce sphingomyelin incorporation into the virion, making the viral particles more infectious. Therefore, p7 and NS5B together contribute to the specific infectivity of the JFH-1 virus, providing evidence for a cooperative role of these viral proteins in HCV particle maturation.

Previous studies on cell culture adaptation of JFH-1 have identified many different culture-adapted mutations and various mechanisms responsible for the culture adaptation (18–28). While our culture-adapted virus had nine amino acid substitutions within the genome, only one of these amino acid changes (NS5A V2440L) had been identified previously. This amino acid change in NS5A had been found to delay the processing of the NS5A/NS5B junction in the viral polyprotein, resulting in increased viral replication (19, 20, 24). Our work confirms these findings, in that JFH-1 having only the NS5A V2440L change does indeed have higher levels of intracellular viral RNA and a higher viral titer than JFH1-Wt. However, we found that our culture-adapted virus with all 9 amino acid changes has a 2-log-higher viral titer than JFH-1 with only NS5A V2440L, demonstrating that this single-amino-acid change at NS5A does not play a large role in the increased infectivity of this virus compared to that of the parental JFH-1. Unexpectedly, we also found that addition of the NS5A V2440L amino acid to JFH-1-S resulted in decreased levels of intracellular HCV RNA compared to JFH-1 or JFH-1-S. This result supports the finding that the structural and nonstructural genes can act together to regulate HCV RNA levels (36, 38).

The complete mechanistic details underlying HCV assembly are not yet understood. However, the viral p7 transmembrane protein plays an important role in this process, as it is required for efficient virion assembly, release, and envelopment (27, 41, 45, 46). The 63-amino-acid p7 protein oligomerizes to form a proton channel, and mutational analysis of p7 has revealed that this proton channel contributes to virion infectivity at a postassembly step (47). Interestingly, amino acid changes within p7 have been identified in many of the previous studies on culture adaptation of JFH-1 as contributing to increased viral titers (19, 22, 26, 27, 36). Similarly, our culture-adapted JFH-1 also identified an amino acid change in p7, C766Y, that contributed to the higher viral titers of a culture-adapted JFH-1. This C766Y mutation is directly adjacent to the N765D mutation identified previously as contributing to JFH-1 cell culture adaptation (19, 22, 26). Both of these residues are located in the N-terminal transmembrane helix of p7 directly adjacent to a flexible hinge that connects this helix to transmembrane helix 2 in the protein (48). Further, these residues line the inner pore of the proton channel formed by these helices, and they are required for proton channel activity (47, 48). We found that the C766Y change in p7 increased the overall size of lipid droplets in the infected cell. Of note is the fact that the NS5B culture-adapted mutation did not induce a similar increase in lipid droplet size; instead, this mutation caused lipid droplet aggregation. Therefore, this result indicates that the p7 mutation plays the primary role in increasing lipid droplet size during infection with this culture-adapted JFH-1 virus.

Lipid droplets, in association with ER membranes, are the sites of HCV particle assembly (42). While HCV infection does induce transcriptional activation of lipogenic genes that enhance viral assembly (49), we did not find that the p7 C766Y virus induced the transcriptional activation of these same genes (data not shown), suggesting another mechanism for regulating lipid droplet size in our system. Lipid droplet size is regulated by cellular enzymes involved in the biosynthesis of lipids and also by the number of ER connections with lipid droplets, including during HCV infection (50–52). We hypothesize that the identified p7 C766Y protein alters the targeting of lipid synthesis enzymes to lipid droplets to regulate the size, as well as the phospho- and sphingolipid contents, of the lipid droplet (53). We further hypothesize that the C776Y amino acid change in p7 modulates its ion channel activity, resulting in an increased size of lipid droplets. Under this hypothesis, the altered p7 ion channel activity in the JFH-1-p7 virus would change the intracellular pH to target lipid synthesis enzymes to lipid droplets that modulate their size. This hypothesis is supported by the fact that intracellular pH has been described to regulate protein binding and targeting to certain lipids (54). Alternatively, the C766Y amino acid change could directly increase ER-lipid droplet contacts through yet-unknown membrane-protein interactions via the p7 transmembrane domain (55).

The C766Y culture-adapted mutation in p7 on its own does not substantially increase viral titer. Indeed, we found that a genetic interaction between p7 and NS5B contributes to virion infectivity. It is known that p7 interacts with other viral proteins to facilitate its role in viral assembly (36, 38, 40, 46, 56). Further, a previous study has suggested that p7 and NS5B might act in a coordinated function to promote viral assembly (40). Our work confirms this finding and importantly shows that only a single amino acid change in each protein is required for the two proteins to enhance virion infectivity in a coordinated fashion. We found that the mechanism for these proteins participating in virion infectivity is through decreased sphingomyelin content of the virus containing both the culture-adapted p7 and NS5B. The R2676K culture-adapted mutation in NS5B is directly in the sphingomyelin-binding pocket of the enzyme (43). In this sphingomyelin-binding pocket within NS5B, the R2676 amino acid residue is located in an amphipathic α-helix antiparallel to another amphipathic α-helix. It forms a bidentate salt bridge with a glutamine located in the antiparallel α-helix (57). While mutation of this arginine to a lysine could alter the affinity of NS5B for sphingomyelin, how this would alter the sphingomyelin content of the virion is unclear, as NS5B is not known to be part of the virion. More likely, the changes in sphingomyelin in the virion are the result of interactions between p7 and NS5B. Our data did not reveal any increase in intracellular viral RNA levels by JFH-1 containing the NS5B R2676K mutation, supporting previous data that have shown that the RdRp activity of JFH-1 NS5B is not affected by sphingomyelin (44). While sphingomyelin has been shown to be required for the infectivity of JFH-1 virions as a result of increased internalization of HCV particles containing sphingomyelin during infection (13), the mechanism of how a decrease in sphingomyelin would increase the virion specific infectivity requires further investigation. Interestingly, a study of the lipid composition of HCV and VLDL particles found that VLDL has a reduced sphingomyelin content compared to HCV particles (14), and so perhaps the lipid content of JFH-1-p7+NS5B particles is altered to more closely resemble that of VLDL. Our results support the hypothesis that virions produced from JFH-1 with p7 and NS5B culture-adapted mutations have decreased sphingomyelin content that likely enhances their infectivity, but we note that this could be a result of broad changes in the lipid composition of the viral particles.

While other studies of culture-adapted viruses have shown that relocalization of core protein from lipid droplets to the ER facilitates an increase in viral titer (36), we did not see any changes in the localization of core in relationship to lipid droplets in cells following transfection with any of our HCV RNA constructs. Besides p7 and NS5B, the other culture-adapted amino acid changes that we identified in JFH-1 may very well contribute to the full culture adaptation of the virus. However, after the culture adaptation has occurred, these other amino acid changes do not seem to play a major role in the increased infectivity of the culture-adapted JFH-1. As the JFH-1-p7+NS virus does have higher replication levels, titers, and virion specific infectivity than JFH-1-p7+NS5B, the culture-adapted amino acid changes in NS5A clearly are playing some role in these increases. In addition to the V2440L mutation, we identified two amino acid changes in domain 3 of NS5A, a functional domain known to be important for HCV assembly (58), which could be contributing to the increased replication of the JFH-1-p7+NS virus.

We propose the following mechanism for how amino acid changes in both p7 and NS5B contribute to HCV particle infectivity. First, the C766Y change in p7 results in an increased size of lipid droplets as a result of phospho- and sphingolipid trafficking to the lipid droplets, likely as a result of increased contacts between the ER and lipid droplets or by specific p7 protein-protein interactions at these sites (51). These interactions alter the local concentration of lipids in or near the lipid droplet. We hypothesize that the NS5B R2676K mutation alters the affinity of NS5B for lipids, perhaps near the lipid droplet, and thereby contributes to the relative distribution of lipids within the ER near sites of HCV assembly. The assembling and/or maturing virion would then incorporate an altered lipid profile as it matures through the VLDL secretion pathway in the Golgi compartment for release from the infected cell. Importantly, these virions with decreased sphingomyelin would have an altered lipid profile that would ultimately increase their infectivity. Taken together, our data support the hypothesis that interactions between the structural and nonstructural HCV proteins with host lipids and lipid-associated machinery work collectively to increase virus specific infectivity. Indeed, this idea is supported by a recent study that found that a lipid regulatory enzyme, phosphatidylserine-specific phospholipase A1, enhances interactions between the viral structural and nonstructural proteins to promote viral assembly (59). Therefore, the combination of culture-adaptive mutations in both p7 and NS5B facilitates an increase in lipid droplet size and reduced sphingomyelin incorporation into the virion that results in an increase in virion infectivity. Importantly, modulation of these interactions of NS5B with sphingomyelin or p7 could provide novel therapeutic targets for hepatitis C treatment.