Human genome-wide RNAi screen reveals host factors required for enterovirus 71 replication

Abstract

Enterovirus 71 (EV71) is a neurotropic enterovirus without antivirals or vaccine, and its host-pathogen interactions remain poorly understood. Here we use a human genome-wide RNAi screen to identify 256 host factors involved in EV71 replication in human rhabdomyosarcoma cells. Enrichment analyses reveal overrepresentation in processes like mitotic cell cycle and transcriptional regulation. We have carried out orthogonal experiments to characterize the roles of selected factors involved in cell cycle regulation and endoplasmatic reticulum-associated degradation. We demonstrate nuclear egress of CDK6 in EV71 infected cells, and identify CDK6 and AURKB as resistance factors. NGLY1, which co-localizes with EV71 replication complexes at the endoplasmatic reticulum, supports EV71 replication. We confirm importance of these factors for EV71 replication in a human neuronal cell line and for coxsackievirus A16 infection. A small molecule inhibitor of NGLY1 reduces EV71 replication. This study provides a comprehensive map of EV71 host factors and reveals potential antiviral targets.

Figure 1. Genome-wide siRNA screen setup and functional classification of screening hits.

(a) Graphical summary of the study workflow from genome-wide screen to eventual functional validation of select hits. (b) Classification of screening hits using Gene Ontology (GO) database revealed overrepresented pathways, shown here as the log of the given false discovery rate (FDR) with all shown pathways having P<0.005 (Hypergeometric test).

Figure 2. Evaluation of select hits against EV71 or CA16 in RD cells and EV71 in the neuronal cell line SK-N-SH.

(a) RD cells were infected with EV71 or CA16 at MOI 1 before relative infection rates were determined at 12 h.p.i. based on immunofluorescence quantification and normalized against NT controls. (b) SK-N-SH cells were infected with EV71 at MOI 1 and evaluated at 24 h.p.i. using immunofluorescence-based quantification of infection rates normalized against NT controls. (c) qRT-PCR detection of viral RNA at 12 h.p.i. showing the effects of siRNA knockdown of CDK6, AURKB, NGLY1 or RDX in RD cells on EV71 replication. Fold changes relative to NT-treated cells were calculated and results show box representing the upper quartile, median and lower quartile from three independent experiments while bars indicate highest and lowest reading for each sample. P-values are shown for two-tailed, Student's T-test performed against NT controls. (d) Representative images of the effect of gene knockdown on EV71 replication in SK-N-SH are shown, with the green signal staining for VP0/VP2 and the blue for nucleus. Representative scale bar for all micrographs in this panel can be found in the NT micrograph with an indicated length of 0.1 cm. (a) and (b) show means of results from three independent experiments and error bars represent standard deviation. P-values shown are for one-tailed, Student's T-test performed for each sample against NT controls.

Figure 3. CDK6 and AURKB restrict EV71 infection.

(a) Dose-dependent inhibition of CDK6 using PD00332291 or (d) AURKB using AZD1152-HQPA, for 16 h before infection and throughout infection resulted in increased rate of EV71 infection. Bars represent mean infection rate relative to 0.1% DMSO and error bars represent standard deviation. (b) EV71 infection in cells stably overexpressing CDK6 or its kinase inactive mutant CDK6 DN or (e) AURKB and its kinase inactive mutant AURKB DN. Bars represent mean infection rate relative to V5 (empty vector) and error bars represent standard deviation. (c) Immunofluorescence tracking of CDK6 in mock-infected and EV71-infected RD cells at 12 h.p.i.. Red signal indicates CDK6, green for dsRNA and blue for nuclei. (f) Diagram showing the relative functional roles of AURKB, CDK6 and inhibitors used in the mitotic cell cycle. (g) Cell cycle inhibition of RD cells at specific stages enhances EV71 replication. All inhibitor studies here involved a 16 h pre-infection treatment followed by continued treatment after infection-adsorption until 12 h.p.i.. Bars represent mean relative infection rates and error bars show standard deviations. All results shown are of two-independent experiments. Two-tailed, Student's T-test was performed against controls, *P<0.05 and **P<0.001. Representative scale bars for all micrographs in (a,d,g) can be found in the control 0.1% DMSO micrograph with an indicated length of 0.5 cm. The representative scale bars for (b,e) can be found in the control V5 micrograph with an indicated length of 0.1 cm.

Figure 4. NGLY1 and VCP functions are required for EV71 infection.

(a) Functional domains of NGLY1 and VCP and their relative interactions at the ER membrane. (b,c) Specific inhibition of NGLY1 and VCP inhibits EV71 replication. Bars represent mean relative infection rates from three-independent experiments, normalized against 0.1% DMSO controls. Error bars show standard deviations. (d) EV71 infection in cells stably overexpressing NGLY1, ΔN NGLY1, NGLY1 R401X, VCP or VCPK524M. Bars show average relative infection rates measured against V5 (empty vector) expressing cell line and error bars represent standard deviation. Data are from two-independent experiments. (e) Immunofluorescence tracking of NGLY1 or (f) VCP in mock-infected RD cells and EV71-infected RD cells at 12 h.p.i.. Red signal indicates (e) NGLY1 or (f) VCP, green for dsRNA and blue for nuclei. Two-tailed, Student's T-test was performed against controls, *P<0.05 and **P<0.001. Representative scale bars for all micrographs in (b,d) can be found in the controls 0.1% DMSO and V5 micrographs respectively, with an indicated length of 0.1 cm. The representative scale bars for (c) can be found in the control 0.1% DMSO micrograph with an indicated length of 0.5 cm.

Figure 5. RDX is needed for EV71 infection.

(a) Stable cell lines expressing RDX, RDX T564A or RDX ΔP were infected with EV71 and relative infection rates at 12 h.p.i. were compared. Bars show average infection rate normalized against V5 (empty vector) control and error bars represent standard deviation. Data are obtained from two-independent experiments and two-tailed, Student's T-test was performed against controls, *P<0.05 and **P<0.001. Representative scale bar for all micrographs in this panel can be found in the V5 micrograph with an indicated length of 0.1 cm. (b) RDX was tracked in immunofluorescence microscopy in mock-infected or EV71-infected RD cells at 12 h.p.i. Red signal indicates RDX, green for dsRNA and blue for nuclei.