Direct acting antivirals (DAAs) provide efficient hepatitis C virus (HCV) therapy and clearance for a majority of patients, but are not available or effective for all patients. They risk developing HCV-induced hepatocellular carcinoma (HCC), for which the mechanism remains obscure and therapy is missing. Annexin A2 (AnxA2) has been reported to co-precipitate with the non-structural (NS) HCV proteins NS5B and NS3/NS4A, indicating a role in HCC tumorigenesis and effect on DAA therapy. Surface plasmon resonance biosensor technology was used to characterize direct interactions between AnxA2 and HCV NS5B, NS3/NS4 and RNA, and the subsequent effects on catalysis and inhibition. No direct interaction between AnxA2 and NS3/NS4A was detected, while AnxA2 formed a slowly dissociating, high affinity (K D = 30 nM), complex with NS5B, decreasing its catalytic activity and affinity for the allosteric inhibitor filibuvir. The RNA binding of the two proteins was independent and AnxA2 and NS5B interacted with different RNAs in ternary complexes of AnxA2:NS5B:RNA, indicating specific preferences. The complex interplay revealed between NS5B, AnxA2, RNA and filibuvir, suggests that AnxA2 may have an important role for the progression and treatment of HCV infections and the development of HCC, which should be considered also when designing new allosteric inhibitors.

Allosteric inhibitors of hepatitis C virus (HCV) non-structural protein 5B (NS5B) polymerase are effective for treatment of genotype 1, although their mode of action and potential to inhibit other isolates and genotypes are not well established. We have used biophysical techniques and a novel biosensor-based real-time polymerase assay to investigate the mode-of-action and selectivity of four inhibitors against enzyme from genotypes 1b (BK and Con1) and 3a. Two thumb inhibitors (lomibuvir and filibuvir) interacted with all three NS5B variants, although the affinities for the 3a enzyme were low. Of the two tested palm inhibitors (dasabuvir and nesbuvir), only dasabuvir interacted with the 1b variant, and nesbuvir interacted with NS5B 3a. Lomibuvir, filibuvir and dasabuvir stabilized the structure of the two 1b variants, but not the 3a enzyme. The thumb compounds interfered with the interaction between the enzyme and RNA and blocked the transition from initiation to elongation. The two allosteric inhibitor types have different inhibition mechanisms. Sequence and structure analysis revealed differences in the binding sites for 1b and 3a variants, explaining the poor effect against genotype 3a NS5B. The indirect mode-of-action needs to be considered when designing allosteric compounds. The current approach provides an efficient strategy for identifying and optimizing allosteric inhibitors targeting HCV genotype 3a.