Distinct functions of NS5A in hepatitis C virus RNA replication uncovered by studies with the NS5A inhibitor BMS-790052

Abstract

BMS-790052, targeting nonstructural protein 5A (NS5A), is the most potent hepatitis C virus (HCV) inhibitor described to date. It is highly effective against genotype 1 replicons and also displays robust genotype 1 anti-HCV activity in the clinic (M. Gao et al., Nature 465:96-100, 2010). BMS-790052 inhibits genotype 2a JFH1 replicon cells and cell culture infectious virus with 50% effective concentrations (EC(50)s) of 46.8 and 16.1 pM, respectively. Resistance selection studies with the JFH1 replicon and virus systems identified drug-induced mutations within the N-terminal region of NS5A. F28S, L31M, C92R, and Y93H were the major resistance mutations identified; the impact of these mutations on inhibitor sensitivity between the replicon and virus was very similar. The C92R and Y93H mutations negatively impacted fitness of the JFH1 virus. Second-site replacements at NS5A residue 30 (K30E/Q) restored efficient replication of the C92R viral variant, thus demonstrating a genetic interaction between NS5A residues 30 and 92. By using a trans-complementation assay with JFH1 replicons encoding inhibitor-sensitive and inhibitor-resistant NS5A proteins, we provide genetic evidence that NS5A performs the following two distinct functions in HCV RNA replication: a cis-acting function that likely occurs as part of the HCV replication complex and a trans-acting function that may occur outside the replication complex. The cis-acting function is likely performed by basally phosphorylated NS5A, while the trans-acting function likely requires hyperphosphorylation. Our data indicate that BMS-790052 blocks the cis-acting function of NS5A. Since BMS-790052 also impairs JFH1 NS5A hyperphosphorylation, it likely also blocks the trans-acting function.

Fig. 1.

Suppression of NS5A hyperphosphorylation by BMS-790052. Con1 and JFH1 subgenomic replicon clones without cell culture adaptive mutations were transiently expressed in a vaccinia virus-T7 expression system. Cells were treated with 500 nM BMS-790052 (+) or DMSO (−). After 16 h, cell lysates were fractionated by SDS-PAGE, followed by immunoblot analysis with anti-NS5A antibody. hyper-PO4, hyperphosphorylated NS5A; basal-PO4, basally phosphorylated NS5A.

Fig. 2.

Replication of JFH1 viruses with C92R and Y93H mutations. Huh-7.5 cells were treated with tissue culture media harvested from cells transfected with in vitro-synthesized RNA prepared from the indicated JFH1 viral clones. Cultures were split, and aliquots of cells were collected for luciferase assays (relative light units [RLU]) at the times indicated in the plots. (A) Cultures were infected with media harvested 72 h after RNA transfection, and subconfluent cultures were maintained for 23 days. Cultures were also split on day 16, but cells were not collected for luciferase readings at this time. (B) The media used to infect the cells were recovered 6 days after the RNA transfection. The C92R virus-infected culture was discontinued on day 15, at which time substantial cytopathogenicity was observed. (C) Cultures were infected with media from cells transfected with RNA from WT or Y93H viral clones. Media from the WT-transfected cells were diluted with regular growth media so that initial luciferase readings (4 h postinfection) of the infected cultures were very similar between the wild-type virus and the Y93H variant (see Materials and Methods). Cultures infected with WT virus were discontinued at day 15. Substantial cytopathogenicity was observed at this time in the WT virus-infected cultures but was not noted in the Y93H variant-infected cultures. The data shown are from two independent experiments.

Fig. 3.

Transient expression of luciferase replicons in neomycin-resistant replicon cells. JFH1-WT or JFH1-L31M Renilla luciferase reporter (luc) replicons were transiently expressed in cells harboring BMS-790052-resistant (JFH1-052R-neo) or -sensitive (JFH1-WT-neo) replicons. Replicons in the host cells expressed a neo gene, and cells were maintained with media containing 0.5 mg/ml G418. The JFH1-052R-neo cells were derived from selection with 20 nM BMS-790052 (Table 1). Dose response curves for transiently expressed replicons were generated with data obtained from luciferase assays performed ∼72 h posttransfection. For comparison, curves obtained from nontransfected host replicon cells, generated with data from FRET assays measuring NS3 protease activity, are also shown. The dose response is plotted relative to luciferase or FRET activity in the absence of inhibitor (100% control). The data shown are the means ± standard deviations from two independent experiments performed in duplicate. The x axis values at 50% inhibition (EC₅₀s) are indicated. Replication windows (uninhibited RLU/background RLU) for the JFH1-WT-luc replicon expressed in BMS-790052-resistant replicon cells (top) and the JFH1-L31M-luc replicon expressed in wild-type replicon cells (bottom) were 62 ± 4 and 32 ± 4, respectively. Data obtained from expression of the JFH1-L31M-luc replicon in JFH1-WT-neo cells could not be fitted to a curve due to the increase in the luciferase signal observed with increasing BMS-790052.

Fig. 4.

Effect of amino acid substitutions at residue 2204 on hyperphosphorylation and replication. JFH1 replicon clones with the indicated amino acids at residue 2204 (NS5A residue 232) were transiently expressed in a vaccinia virus-T7 expression system. Cell lysates were fractionated by SDS-PAGE, followed by immunoblot analysis with anti-NS5A antibody. The positions of basally phosphorylated and hyperphosphorylated NS5A proteins are indicated. Replication levels of the clones measured from transient replicon assays are shown below the blot. Values are relative to that of the parental control (serine at 2204) and are the means ± standard deviations from at least three experiments.