A Quantitative High-Resolution Genetic Profile Rapidly Identifies Sequence Determinants of Hepatitis C Viral Fitness and Drug Sensitivity

Abstract

Widely used chemical genetic screens have greatly facilitated the identification of many antiviral agents. However, the regions of interaction and inhibitory mechanisms of many therapeutic candidates have yet to be elucidated. Previous chemical screens identified Daclatasvir (BMS-790052) as a potent nonstructural protein 5A (NS5A) inhibitor for Hepatitis C virus (HCV) infection with an unclear inhibitory mechanism. Here we have developed a quantitative high-resolution genetic (qHRG) approach to systematically map the drug-protein interactions between Daclatasvir and NS5A and profile genetic barriers to Daclatasvir resistance. We implemented saturation mutagenesis in combination with next-generation sequencing technology to systematically quantify the effect of every possible amino acid substitution in the drug-targeted region (domain IA of NS5A) on replication fitness and sensitivity to Daclatasvir. This enabled determination of the residues governing drug-protein interactions. The relative fitness and drug sensitivity profiles also provide a comprehensive reference of the genetic barriers for all possible single amino acid changes during viral evolution, which we utilized to predict clinical outcomes using mathematical models. We envision that this high-resolution profiling methodology will be useful for next-generation drug development to select drugs with higher fitness costs to resistance, and also for informing the rational use of drugs based on viral variant spectra from patients.

Author Summary

The emergence of drug resistance during antiviral treatment limits treatment options and poses challenges to pharmaceutical development. Meanwhile, the search for novel antiviral compounds with chemical genetic screens has led to the identification of antiviral agents with undefined drug mechanisms. Daclatasvir, an effective NS5A inhibitor, is one such example. In traditional methods to identify critical residues governing drug-protein interactions, wild type virus is passaged under drug treatment pressure, enabling the identification of resistant mutations evolved after multiple viral passages. However, this method only characterizes a fraction of the positively selected variants. Here we have simultaneously quantified the relative change in replication fitness as well as the relative sensitivity to Daclatasvir for all possible single amino acid mutations in the NS5A domain IA, thereby identifying the entire panel of positions that interact with the drug. Using mathematical models, we predicted which mutations pose the greatest risk of causing emergence of resistance under different scenarios of treatment compliance. The mutant fitness and drug-sensitivity profiles obtained can also inform the patient-specific use of Daclatasvir and may facilitate the development of second-generation drugs with a higher genetic barrier to resistance.