Importantly, in cells treated with 50× EC50 AL-9 (corresponding to 10× EC90 of BMS-553 as determined by inhibition of HCV replication), DMVs were still clearly detectable ( Figure 5A and Supplementary Figure 12), with no difference between post- or co-treatment conditions (compare Figure 4 and Supplementary Figure 12). These results suggest that abrogation of web formation is a distinct phenotype that is not mediated by PI4KIIIα inhibition.

We also found no further impact of BMS-553 co-treatment on intracellular PI4P pools ( Supplementary Figure 9B) compared with posttreatment ( Figure 3D). To exclude the possibility that we VX 809 had missed virus-induced membrane rearrangements in inhibitor-treated cells due to low/no expression of viral proteins in analyzed cells, correlative light-electron microscopy was applied. We used NS3-5B wt or Y93H polyproteins containing a green Belnacasan in vivo fluorescent protein insertion in DIII of NS5A that does not interfere with replication competence.33 Upon expression of these constructs in Huh7-Lunet/T7 cells, no distinct difference in NS5A fluorescence patterns was detected between wt and resistance mutant in cells co-treated with BMS-553 (Figure 5B and C, respectively).

However, in case of wt, no DMVs or other virus-induced membrane rearrangements were detected in subcellular regions containing low or high NS5A amounts ( Figure 5B), corroborating profound inhibition of MW formation by the NS5A inhibitor. In contrast, in case of Y93H polyprotein, DMVs were readily detected ( Figure 5C). Importantly, the sites of strong NS5A accumulation corresponded to areas containing lipid droplets surrounded by DMV clusters, consistent with the observed accumulation of NS5A in close proximity of lipid droplets in inhibitor-treated cells ( Supplementary Figure 13). 18 To confirm inhibition of MW formation in a replication-based system, we analyzed cells transfected with the Jc1 genome and treated with BMS-553. A reduction of DMV number and size was observed already

at 4 hours, and more dramatic at 12 hours after BMS-553 treatment, which was not found in Jc1 Y93H-transfected cells (Figure 6A and B). Importantly, during these time periods, abundance and subcellular distribution of NS5A were not affected ( Figure 6C and D). In addition, cells Mirabegron co-treated directly after Jc1 transfection completely inhibited wt, but not the Y93H-resistant mutant. In summary, these results demonstrate that a potent daclatasvir-like NS5A inhibitor blocks biogenesis of the MW, the presumed sites of HCV replication. Disruption of the biogenesis of a virus-induced replication factory, which is a central element for the replication of all plus-strand RNA viruses,6 is a novel paradigm in antiviral therapy. We show here that a daclatasvir-like inhibitor abrogates formation of the MW, the presumed replication site of HCV.