A functional genomic screen identifies cellular cofactors of hepatitis C virus replication

Abstract

Hepatitis C virus (HCV) chronically infects 3% of the world's population, and complications from HCV are the leading indication for liver transplantation. Given the need for better anti-HCV therapies, one strategy is to identify and target cellular cofactors of the virus lifecycle. Using a genome-wide siRNA library, we identified 96 human genes that support HCV replication, with a significant number of them being involved in vesicle organization and biogenesis. Phosphatidylinositol 4-kinase PI4KA and multiple subunits of the COPI vesicle coat complex were among the genes identified. Consistent with this, pharmacologic inhibitors of COPI and PI4KA blocked HCV replication. Targeting hepcidin, a peptide critical for iron homeostasis, also affected HCV replication, which may explain the known dysregulation of iron homeostasis in HCV infection. The host cofactors for HCV replication identified in this study should serve as a useful resource in delineating new targets for anti-HCV therapies.

Figure 1. Bioinformatics of the RNAi Screen

(A) Replicate Z scores from the primary screen. Each point represents the Z score of an individual siRNA pool in one experimental pool plotted against the Z score of the same pool in the experimental replicate well. A Z score is the number of standard deviations of the experimental luciferase activity above the median plate value. Negative Z scores indicate inhibition of HCV replication. (B) Gene ontology biological process terms significantly (p < 0.05) overrepresented by the genes that scored in the screen. The terms are ordered clockwise by ascending p values. (C) Functional mapping of HCV dependency factors to lipid- and membrane-related processes that support the formation of a membranous web structure needed for viral replication. Dotted gray lines indicate protein-protein interactions; dotted blue lines indicate processes that impinge on membranous web structures; green boxes indicate cellular processes; and black boxes refer to HCV proteins. Supporting references are listed in Table S4.

Figure 2. PI4KA Is Essential for Hepatitis C Virus Replication

(A) PI4KA silencing with four individual siRNA duplexes blocks replication of the full-length HCV replicon OR6 (black bars) without significant cytotoxicity as measured by cellular ATP content (gray bars). Values were obtained from quadruplicate wells in two independent experiments and are mean ± SD. (B) siRNAs against PI4KA block replication of the infectious JFH1 HCV strain. Cells were infected with JFH1 48 hr posttransfection, and HCV RNA was quantified by quantitative PCR (qPCR) at 24 hr postinfection. Values were obtained from triplicate qPCR replicates from duplicate wells and are mean ± SD. (C) The magnitude of HCV inhibition parallels the degree of PI4KA silencing. PI4KA transcripts were quantified by qPCR 48 hr posttransfection. Values are mean± SD. (D) PI4KA silencing depletes HCV nonstructural proteins in OR6 replicon cells as determined by immunoblotting with the indicated antibodies. Cell lysates were prepared from OR6 cells transfected with the indicated siRNAs for 72 hr. (E) Wortmannin inhibits HCV replication with an IC₅₀ consistent with a type III PI 4-kinase. OR6 replicon cells were treated with the indicated concentrations of wortmannin for 24 hr. Replicate wells were assayed for luciferase activity (diamonds) or cellular ATP content (triangles). Values were obtained from quadruplicate wells in two independent experiments and are mean ± SD. (F) PI4KA shRNAs block HCV replication and can be rescued by an RNAi-resistant PI4KA construct but not by PI4KB. OR6 replicon cells were transduced with MMLV retroviral vectors encoding GFP, PI4KA, or PI4KB cDNAs lacking the 3′UTR. The cells were then transduced 24 hr later with lentiviral vectors encoding a nontargeting shRNA (black bars) or two different shRNAs targeting the PI4KA 3′UTR (gray and white bars). Cells were assayed 96 hr after lentiviral transduction. Values were obtained from quadruplicate wells in two independent experiments and are mean ± SD. (G) PI4KA silencing induces the formation of abnormally large NS5A-positive membrane structures. UHCVcon57.3 cells were transduced with the indicated shRNA constructs 4 days prior to induction of HCV polyprotein expression by withdrawal of tetracycline from the growth medium. Twenty-four hours after induction, cells were fixed and stained with a monoclonal antibody to HCV NS5A (green) and counterstained with DAPI to highlight cell nuclei (blue).

Figure 3. The COPI Complex Supports HCV Replication Early in the Viral Lifecycle

(A) Individual siRNA duplexes against COPZ1 block replication of the full-length HCV replicon OR6 in a concentration-dependent manner. OR6 cells were transfected with the indicated siRNA duplexes for 64 hr and then assayed for luciferase activity. Values were obtained from quadruplicate wells in two independent experiments and are mean ± SD. (B) Individual siRNA duplexes against COPZ1 deplete COPZ1 transcripts in a dose-dependent manner without detectable cytotoxicity. OR6 cells transfected with the indicated siRNA duplexes as in (A) were assayed for cellular ATP content (open symbols) or for COPZ1 transcript levels by quantitative real-time PCR (closed symbols). Values were obtained from triplicate qPCR measurements from duplicate wells in two independent experiments and are mean ± SD. (C) Brefeldin A (BFA), a pharmacologic inhibitor of COPI function, blocks replication of the OR6 full-length replicon. OR6 cells were treated with the indicated concentrations of BFA for 24 hr and then assayed for luciferase activity (triangles) or cellular ATP content (diamonds). Values were obtained in quadruplicate and are mean ± SD. (D) BFA blocks replication of the infectious JFH1 isolate early in viral infection and blocks HCV secretion late in infection. BFA was added at 100 ng/mL to Huh7.5.1 cells at the indicated times relative to the time of JFH1 infection. Intracellular HCV RNA was quantified by qPCR at 20 hr of BFA treatment and normalized to HCV RNA from mock treated cells. Values were obtained from triplicate qPCR replicates from duplicate wells and are mean ± SD. (E) Preformed NS5A-positive structures are resistant to BFA treatment. OR6 cells were treated with 100 μg/mL BFA (lower panels) or 1% ethanol (upper panels) for 4 hr and then processed for immunofluorescence for HCV NS5A (left panels) and HCV NS5A (middle panels). Nuclei were counterstained with DAPI (blue). BFA causes dissociation of β-COP from the Golgi apparatus within minutes of treatment.

Figure 4. The Iron Regulatory Hormone Hepcidin Supports HCV Replication

(A) HAMP silencing blocks replication of the full-length HCV replicon OR6. OR6 cells were transfected with the indicated siRNAs for 72 hr and then assayed for luciferase activity (black bars), HAMP transcript levels (white bars), and cellular ATP content (gray bars). Values were obtained from quadruplicate wells in three independent experiments and are mean ± SD. (B) HAMP silencing blocks replication of the infectious JFH1 HCV strain in Huh7 cells. Two-tailed p value was calculated by indepdendent two-sample t test. Values were obtained from triplicate qPCR replicates from duplicate wells in two independent experiments and are mean ± SD.