Resistance of high fitness hepatitis C virus to lethal mutagenesis

Virology 523 (2018) 100–109

Contents lists available at ScienceDirect

Virology journal homepage: www.elsevier.com/locate/virology

Resistance of high fitness hepatitis C virus to lethal mutagenesis a,b

b,c,d

c

T

a

Isabel Gallego , Josep Gregori , María Eugenia Soria , Carlos García-Crespo , Mónica García-Álvareza, Alfonso Gómez-Gonzáleza,1, Rosalie Valierguea, Jordi Gómezb,e, ⁎ ⁎ Juan Ignacio Estebanb,c,f, Josep Querb,c,f, Esteban Domingoa,b, , Celia Peralesa,b,c, Centro de Biología Molecular “Severo Ochoa” (CSIC-UAM), Consejo Superior de Investigaciones Científicas (CSIC), Campus de Cantoblanco, Madrid, Spain Centro de Investigación Biomédica en Red de Enfermedades Hepáticas y Digestivas (CIBERehd) del Instituto de Salud Carlos III, Madrid, Spain c Liver Unit, Internal Medicine Hospital Universitari Vall d′Hebron, Vall d′Hebron Institut de Recerca (VHIR), Barcelona, Spain d Roche Diagnostics, S.L., Sant Cugat del Vallés, Barcelona, Spain e Instituto de Parasitología y Biomedicina 'López-Neyra' (CSIC), Parque Tecnológico Ciencias de la Salud, Armilla, Granada, Spain f Universitat Autónoma de Barcelona, Barcelona, Spain a

b

A B S T R A C T

Viral fitness quantifies the degree of virus adaptation to a given environment. How viral fitness can influence the mutant spectrum complexity of a viral quasispecies subjected to lethal mutagenesis has not been investigated. Here we document that two high fitness hepatitis C virus populations display higher resistance to the mutagenic nucleoside analogues favipiravir and ribavirin than their parental, low fitness HCV. All populations, however, exhibited a mutation transition bias indicative of active mutagenesis. Resistance to the analogues was associated with a limited expansion of mutant spectrum complexity, as evidenced by several diversity indices used to characterize mutant spectra. The results are consistent with a replicative site-drug competition mechanism that was previously proposed for HCV fitness-associated resistance to non-mutagenic inhibitors. Other alternative, non-mutually exclusive mechanisms are considered. The results introduce viral fitness as a relevant parameter to evaluate the response of viruses to lethal mutagenesis, with implications for antiviral designs.

1. Introduction Since the introduction of the concept of fitness in virology (Holland et al., 1991; Martínez et al., 1991), this parameter has proven important to interpret the connection between population dynamics, viral pathogenesis and treatment responses (Agol and Gmyl, 2018; Andino and Domingo, 2015; Dolan et al., 2018; Domingo and Holland, 1997; Domingo et al., 2012; Farci, 2011; Martinez-Picado and Martinez, 2008; Quinones-Mateu and Arts, 2006; Wargo and Kurath, 2012). Fitness is a multifactorial parameter that recapitulates the overall capacity of viruses to produce infectious progeny. It is generally measured in growth-competition experiments between the virus population to be tested and a reference preparation of the same virus (Domingo and Holland, 1997; Wargo and Kurath, 2012). In connection with antiviral treatments, high fitness and high viral load were associated with a lower frequency of extinction events of foot-and-mouth disease virus (FMDV) and human immunodeficiency virus type 1 (HIV-1) by mutagenic base and nucleoside analogues (Pariente et al., 2001; Sierra et al., 2000; Tapia et al., 2005). Resistance to extinction was interpreted according to quasispecies theory as the requirements that genomes

replicating under high mutation rates must fulfill to elude crossing an error threshold for loss of information (Eigen and Schuster, 1979; Schuster, 2016; Sierra et al., 2000). One of the parameters that render a replicative system resistant to extinction is the fitness superiority of the master (dominant) genomes in a population, relative to the surrounding mutant spectrum (Schuster and Swetina, 1988). Recent work with hepatitis C virus (HCV) has documented that high fitness is responsible of increased resistance to antiviral agents that target viral or cellular functions (Gallego et al., 2016; Moreno et al., 2017; Sheldon et al., 2014). Here we address the effect of HCV fitness in the virus response to lethal mutagenesis, and if fitness may alter the modifications undergone by the mutant spectrum. These questions are justified by the occurrence of viral fitness variations during disease processes (Domingo et al., 2012), and the increasing impact of lethal mutagenesis in antiviral therapy (Arias et al., 2014; Guedj et al., 2018; Mullins et al., 2011; Qiu et al., 2018). Here we compare the response to lethal mutagenesis to favipiravir (T-705; 6-fluoro-3-hydroxy-2-pirazinecarboxamide) and ribavirin (1-β-D-ribofuranosyl-1-H-1,2,4-triazole-3-carboxamide) of three HCV populations that differ in relative fitness: a parental HCV p0

⁎

Corresponding authors at: Centro de Biología Molecular “Severo Ochoa” (CSIC-UAM), Consejo Superior de Investigaciones Científicas (CSIC), Campus de Cantoblanco, Madrid, Spain. E-mail addresses: [email protected] (E. Domingo), [email protected] (C. Perales). 1 Present address: Institute of Molecular Life Sciences, University of Zurich, Zurich, Switzerland. https://doi.org/10.1016/j.virol.2018.07.030 Received 29 June 2018; Received in revised form 30 July 2018; Accepted 30 July 2018 0042-6822/ © 2018 Elsevier Inc. All rights reserved.

Virology 523 (2018) 100–109

I. Gallego et al.

MiSeq sequencing of the NS5B-coding region of the different HCV populations (Tables S1–S5 [http://babia.cbm.uam.es/~lab121/ SupplMatGallego.pdf]). There is a dominance of synonymous over non-synonymous and of transition over transversion mutations in all non-mutagenized and mutagenized populations (summarized in Table S6 [http://babia.cbm.uam.es/~lab121/SupplMatGallego.pdf]). Examination of amino acid substitutions based on their probability of occurrence according to the PAM 250 substitution matrix (Feng and Doolittle, 1996) indicates preponderance of substitutions with PAM 250 ≥ 0 for both, non-mutagenized (80% of all substitutions) and mutagenized (92% of all substitutions) populations. Mutagenesis did not lead to an increase of amino acid substitutions with low acceptability score. Mutagenesis of HCV by favipiravir and ribavirin increases the proportion of G→A and C→U transitions in the mutant spectra (de Avila et al., 2016; Ortega-Prieto et al., 2013). In all populations passaged in the presence of favipiravir or ribavirin the proportion of G→A and C→U transitions increased significantly relative to the populations passaged in absence of drug, as evidenced by the deep sequencing data of three amplicons (Fig. 3), and by molecular cloning-Sanger mutational analysis of the entire NS5B-coding region (data not shown). The bias was reflected in an increase of the ratio [(G→A) + (C→U)] / [(A→G) + (U→C)] which was always below 1 (range 0.24–0.56) in the populations passaged in absence of drug, and above 1 (range 1.09–8.89) in the populations passaged in presence of drug; these values represent a 2.5- to 24.7-fold increases in the populations passaged in the presence of drug relative to the corresponding populations passaged in absence of drug. Thus, irrespective of fitness levels, all HCV mutagenized populations exhibited similar shifts in their mutant spectra towards higher frequencies of A and U residues (Fig. 3). Absence of extinction was not due to attenuation of mutational bias in the mutant spectra of resistant HCV.

[derived from plasmid Jc1FLAG2(p7-nsGluc2A) (Marukian et al., 2008; Perales et al., 2013)] and two populations, HCV p100 and HCV p200 that were obtained by passaging HCV p0 one hundred and two hundred times, respectively, in Huh-7.5 reporter cells; HCV p100 and HCV p200 increased their fitness 2.2-fold relative to HCV p0 (Moreno et al., 2017; Sheldon et al., 2014). Fitness increase was not associated with the time required for virus adsorption to host cells or with an increase of thermal stability of viral particles (Moreno et al., 2017; Sheldon et al., 2014). The main difference noted between HCV p100 and HCV p200 relative to HCV p0 was a significant increase of the rate and maximum level of infectious virus and viral RNA production determined in single and serial infections over a 1000-fold range of multiplicity of infection (Moreno et al., 2017). Favipiravir is used as antiviral agent (Furuta et al., 2017; Jordan et al., 2018) and has been shown to be mutagenic for several RNA viruses (Arias et al., 2014; Baranovich et al., 2013; de Avila et al., 2017; de la Torre, 2018; Escribano-Romero et al., 2017; Guedj et al., 2018; Qiu et al., 2018), including HCV (de Avila et al., 2016). Ribavirin is extensively used as antiviral agent, and it displays mutagenic activity with several RNA viruses (Beaucourt and Vignuzzi, 2014; Crotty et al., 2000). It is mutagenic for HCV in cell culture (Galli et al., 2018; Ortega-Prieto et al., 2013), and there is also evidence of its mutagenic activity for HCV in vivo (Asahina et al., 2005; Cuevas et al., 2009; Dietz et al., 2013). Both purine analogues exhibit a preference for G→A and C→U transitions in the HCV genome (de Avila et al., 2016; Ortega-Prieto et al., 2013). Here we show with four mutagenized HCV populations that HCV fitness did not modify the mutational bias evoked by favipiravir and ribavirin on HCV RNA. Resistance to extinction of high fitness HCV was consistently associated with a limitation in the expansion of mutant spectra evoked by favipiravir and ribavirin, as quantified by several diversity indices. The results support a replicative site-mutagen competition mechanism previously suggested for nonmutagenic inhibitors (Sheldon et al., 2014), but bring in an alternative, non-mutually exclusive, mechanism consisting of delocalization of the viral population in sequence space associated with fitness gain. The results introduce viral fitness as a relevant parameter in lethal mutagenesis-based antiviral designs.

2.3. Mutant spectrum complexity of mutagenized, high fitness HCV populations The vulnerability of HCV p0, and resistance to extinction of HCV p100 and HCV p200 to favipiravir and ribavirin despite a similar mutational bias evoked by mutagens, raised the question of possible mutant spectrum modifications that might distinguish low and high fitness HCV. The mutant spectrum of the NS5B-coding region of populations HCV p0, HCV p100 and HCV p200 and their derivatives passaged in absence or presence of favipiravir or ribavirin was analyzed by molecular cloning-Sanger sequencing. While the mutagenesis-driven increase of maximum and minimum mutation frequency (Mfmax and Mfmin) was statistically significant for HCV p0, it was not for some HCV p100 and HCV p200 populations (Table 1). In particular, the limited increase of Mfmax and Mfmin after 10 passages of HCV p100 in the presence of ribavirin was unexpected from previous quantifications of ribavirin mutagenesis of HCV p0 (Ortega-Prieto et al., 2013). To confirm this result and to examine if HCV p100 and HCV p200 had a limitation in the increase of other diversity indices that characterize mutant spectra (Gregori et al., 2016), three NS5B amplicons of four mutagenized and the corresponding non-mutagenized viral populations were analyzed by Illumina MiSeq deep sequencing [amplicon A1: residues 7626–7962; amplicon A2: 7941–8257; amplicon A3: 8229–8653; numbering according to the JFH-1 genome (GenBank accession number #AB047639)]. Two groups of diversity indices were calculated: abundance indices (that consider the counts of entities and their frequency), and incidence indices, that consider only counts of entities; both categories are also divided into non-functional and functional indices (the latter being those that consider differences among haplotypes) (Gregori et al., 2016). The indices are the following: Shannon entropy, Gini-Simpson, maximum mutation frequency (Mfmax), nucleotide diversity (π), number of polymorfic sites (N. poly. sites), number of haplotypes (N. hpl.), minimum mutation frequency

2. Results 2.1. Resistance of high fitness hepatitis C virus to favipiravir and ribavirin HCV p0, HCV p100 and HCV p200 were passaged in the absence or presence of 400 µM favipiravir or 100 µM ribavirin (Fig. 1). The drug concentrations were chosen for their capacity to extinguish HCV p0 in four to five passages, under the infection conditions used in the experiments described here (de Avila et al., 2016; Ortega-Prieto et al., 2013) (Fig. 2). Measurement of viral titers and intracellular viral RNA levels indicated that while the infectivity of HCV p0 was lost at passage 4 and 5 in presence of ribavirin and favipiravir, respectively, no extinction of HCV p100 and HCV p200 occurred with the same drug doses and multiplicity of infection (Fig. 2). Loss of infectivity preceded loss of viral RNA, as expected from the lethal defection model of virus extinction by enhanced mutagenesis (Grande-Pérez et al., 2005). The lower specific infectivity of HCV p200 than HCV p100 passaged in absence of drugs agrees with the gradual decrease of specific infectivity previously quantified from passage 100 to 200 (Moreno et al., 2017). Unexpected differences in the relative sensitivity of HCV p100 and HCV p200 to favipiravir and ribavirin were noted (see Discussion). Thus, fitness favors resistance of HCV to extinction by nucleotide analogues. 2.2. Mutation repertoire of high fitness HCV as a result of ribavirin and favipiravir treatment To approach the molecular basis of high fitness HCV resistance to extinction, mutations and deduced amino acid substitutions were identified by molecular cloning-Sanger sequencing and by Illumina 101

Virology 523 (2018) 100–109

I. Gallego et al.

Fig. 1. Scheme of HCV passages in absence or presence of favipiravir or ribavirin. The origin of HCV p0, HCV p100 and HCV p200 and serial passage conditions in Huh-7.5 reporter cells have been previously described (Moreno et al., 2017; Perales et al., 2013). Arrows indicate passages in absence of drugs (no drug) or presence of 400 µM favipiravir or 100 µM ribavirin; p followed by a number indicates passage number. Boxed populations are those whose mutant spectrum has been analyzed; since HCV p0 was extinguished by passage 5 in the presence of favipiravir or ribavirin, p10 for this virus could nor be analyzed. Infection conditions are described in Section 4.

SupplMatGallego.pdf]). Thus, high replicative fitness limits the mutant spectrum expansion of HCV by mutagenic agents.

(Mfmin), mutation frequency at entity level (Mfe), nucleotide diversity at entity level (πe) (see Materials and Methods for definitions). Indices were calculated for each population individually; values obtained for each amplicon and population are compiled in Tables S7–S9 [http:// babia.cbm.uam.es/~lab121/SupplMatGallego.pdf]. Pairwise differences between populations are expressed as ratios of index values. Two groups of comparisons were considered: those involving HCV p0, HCV p100 and HCV p200, without participation of favipiravir or ribavirin mutagenesis (Fig. 4), and those with each of the three populations passaged in presence or absence of favipiravir or ribavirin (Fig. 5). The comparison among HCV p0, HCV p100 and HCV p200 shows a large increase of Mfmax and π in the evolution from HCV p0 to HCV p100, that was highly reduced in the evolution from HCV p100 to HCV p200 (Fig. 4A). Increases were much more modest for the other indices, with the exception of Mfmin for HCV p100 and more prominently for HCV p200. Haplotypes with more than one mutation were not detected in HCV p0 while haplotypes with up to 20 mutations were scored for HCV p100 and HCV p200 (Fig. 4B). A different picture was obtained when comparing mutagenized versus the corresponding non-mutagenized populations (Fig. 5). Except Mfe and πe, that remained invariant in all cases, for other indices the increase in diversity as a consequence of favipiravir or ribavirin treatment was significantly larger for HCV p0 than for HCV p100 or HCV p200, with some differences in the effect of the two mutagenic agents. Particularly relevant is that Mfmax, a commonly used parameter to quantify differences in mutant spectrum complexity, showed limited sensitivity to the consequences of mutagenesis of HCV p100 and HCV p200, in agreement with the molecular cloning-Sanger sequencing analyses (Table 1). Ratios of some incidence indices that do not consider abundances (N. poly. sites, N. hpl., Mfmin) increased slightly (especially in HCV p100) which suggests a reorganization of haplotypes despite Mfmax not being modified. The ratio increase from passage 3 to passage 10 was significant in 50% of the cases (values included in Fig. 5). Comparison of the number of haplotypes and number of different mutations per haplotype revealed that these two parameters ─that were expanded in the evolution from HCV p0 to HCV p100 and HCV p200 (Fig. 4B)─ did not increase (or increased only modestly) as a result of the mutagenic activity of favipiravir and ribavirin acting on the three populations (Figs. S1–S3 [http://babia.cbm.uam.es/~lab121/

3. Discussion The present comparative study with two high fitness HCV populations subjected to favipiravir and ribavirin mutagenesis has established viral fitness as a relevant parameter in the response of HCV to lethal mutagenesis by favipiravir and ribavirin, under the cell culture conditions of our study. HCV fitness did not modify the mutation type bias typical of the mutagens, but limited the mutant spectrum expansion as judged by several diversity indices. This result has been established with a total of four high fitness viral populations from two independent evolutionary lineages that share a common clonal ancestor (Moreno et al., 2017; Perales et al., 2013; Sheldon et al., 2014). In contrast to the high fitness populations, the parental HCV p0 virus exhibited an increase of several diversity indices as a result of mutagenesis with the same mutagenic purine analogue doses. Such an increase in diversity is expected from previous measurements of Mf, π and Shannon entropy with HCV p0 and other RNA viruses passaged in the presence of base and nucleotide analogues under comparable conditions (Agudo et al., 2010; Airaksinen et al., 2003; Grande-Pérez et al., 2002; Moreno et al., 2011; Pariente et al., 2001; Sierra et al., 2007, 2000). A higher rate of genome replication and maximum level of infectious progeny production distinguished HCV p100 and HCV p200 from HCV p0 (Moreno et al., 2017). This difference suggests a straightforward interpretations of the decrease of mutant spectrum expansion upon mutagenesis of HCV p100 and HCV p200, as an extension of the replication site-drug competition model previously proposed for non-mutagenic inhibitors (Moreno et al., 2017). The reduction in amount of mutagenic nucleotide per RNA replication event is expected to maintain the mutation bias typical of the mutagenic agent but to evoke a lower number of mutational events thus allowing virus survival. We are currently analyzing the effect of higher favipiravir and ribavirin doses on the survival of high fitness HCV. Despite this seemingly likely course of events, there is an alternative, non-mutually exclusive mechanism, that may also contribute to the resistance of HCV p100 and HCV p200 to mutagenesis-driven extinction. This second mechanism is suggested by the continued presence 102

Virology 523 (2018) 100–109

I. Gallego et al.

Fig. 2. Progeny production of HCV p0, HCV p100 and HCV p200 in absence or presence of favipiravir and ribavirin. The top three panels give viral titers in the cell culture supernatant. The statistical significance (proportion test) between infectivity values was: HCV p0: no drug versus favipiravir p < 0.0001; no drug versus ribavirin p < 0.0001; favipiravir versus ribavirin p > 0.99. HCV p100: no drug versus favipiravir p = 0.048; no drug versus ribavirin p = 0.23; favipiravir versus ribavirin p = 0.99. HCV p200: no drug versus favipiravir p = 0.72; no drug versus ribavirin p = 0.002; favipiravir versus ribavirin p = 0.30. Values of p < 0.05 are considered significant. The middle panels include the quantification of viral RNA in the corresponding cell culture supernatants. The statistical significance (proportion test) between viral RNA values was: HCV p0: no drug versus favipiravir p < 0.0001; no drug versus ribavirin p < 0.0001; favipiravir versus ribavirin p > 0.81. HCV p100: no drug versus favipiravir p = 0.0001; no drug versus ribavirin p = 0.0002; favipiravir versus ribavirin p > 0.99. HCV p200: no drug versus favipiravir p = 0.94; no drug versus ribavirin p < 0.0001; favipiravir versus ribavirin p = 0 < 0.0001. Bottom panels give the calculated value of specific infectivity (ratio of values of the two upper panels). The horizontal discontinuous lines indicate the limit of detection. The experimental design and infection conditions are described in Fig. 1 and Section 4.

possibility is that the high fitness populations occupy multiple points of sequence space (Fig. 6). Delocalization of sequence space over a plateau is suggested by the number of heterogeneity points (counted as positions in the genome where a double nucleotide peak is present, with the minor peak amounting to 10% to < 50% of the total) in the consensus sequences, as determined by molecular cloning-Sanger sequencing. The number of heterogeneity points is 1, 19 and 70 per genome for HCV p0, HCV p100 and HCV p200, respectively [comparison included in Table S6 (http://babia.cbm.uam.es/~lab121/SupplMatGallego.pdf), and based on the data reported in Moreno et al. (2017)]. Multiple initial points on a fitness plateau to confront a mutational increase should

mutational waves (multiple mutations that increase or decrease in frequency) in the HCV mutant spectra, even when approaching passage 200 (Moreno et al., 2017). Such waves denote absence of population equilibrium, and suggest that the common fitness landscape view that HCV p200 settled at a fitness peak might be a simplification. Contrary to what is found experimentally for HCV p100 and HCV p200, a viral population located at a high fitness peak would be vulnerable to mutations because many of them would be expected to cause the genomes to descend from the peak. This prediction is based on studies with digital organisms, and it is known as “advantage of the flattest” (Schuster and Swetina, 1988; Wilke et al., 2001). In the case of HCV, an appealing 103

Virology 523 (2018) 100–109

I. Gallego et al.

Fig. 3. Mutation repertoire in HCV populations. The filled top box in each group of panels indicates the virus population. Mutation data were obtained by Illumina MiSeq deep sequencing (Tables S3–S5 [http://babia.cbm.uam.es/~lab121/SupplMatGallego.pdf]. Mutation types or sum of mutation types are indicated in abbreviated form in each abscissa (i.e. AG is A→G, and GA + CU is the sum of G→A and C→U transitions). The number of mutations is indicated in ordinate. The box on the right of each group of panels serves as color code for absence of drug (no drug) or favipiravir (FPV) or ribavirin (RIB) mutagenesis; it includes also the indicated transition mutation ratio. Procedures for nucleotide sequencing and data processing are described in Section 4.

ribavirin mutagenesis, but of the prior history of adaptation of HCV to the cell culture environment. The unexpected difference in the response of HCV p100 and HCV p200 to favipiravir and ribavirin is now under investigation; it may relate to the mechanisms of activity of the two drugs, such as the RNAchain terminator activity of favipiravir that has not been documented for ribavirin, or the inhibition of inosine monophosphate dehydrogenase by ribavirin-monophosphate that results in depletion of GTP (Streeter et al., 1973). Relevant unknowns are the effective concentrations of nucleoside-triphosphate that reach the replication complexes, and their affinity for incorporation relative to the standard nucleotide substrates; this is particularly relevant for ATP and GTP, because they directly compete with the mutagenic purine nucleotides. We cannot exclude that drug-induced changes in host cell gene expression could also modify the drug susceptibility of the different HCV populations, a point that requires further investigations. Treatments based on lethal mutagenesis with nucleotide analogues licensed for administration to humans could be considered as an alternative for patients who fail to clear the virus with direct acting antiviral

facilitate escape to the mutagenic activity through a choice of those genome subpopulations capable of continuing replication with acquisition of a limited number of mutations, despite extinction of many other genomes. The proposal is also consistent with previous studies that indicated that RNA viruses can find multiple mutational pathways for fitness gain (Escarmís et al., 1999), and that extensive passage of FMDV in a cell type led to an expansion rather than to a specialization of host cell tropism (Ruiz-Jarabo et al., 2004), suggesting delocalization both in sequence and phenotypic space. Although fitness gain of HCV p0 involved a broadering of the mutant spectrum, individual biological clones of HCV p100 displayed similar mutant spectrum complexity than HCV p0 but a high fitness typical of the entire HCV p100 population (Sheldon et al., 2014). Therefore, a broad mutant spectrum is not necessary for HCV p100 and probably also HCV p200 to display high replicative fitness. Irrespective of the mechanisms underlying resistance to extinction, what the HCV results have disclosed is that mutational robustness (understood both as a reduction in mutant spectrum diversification and evasion of extinction) was not a consequence of the favipiravir or 104

Virology 523 (2018) 100–109

I. Gallego et al.

Table 1 Mutant spectrum analysis of the NS5B-coding region of hepatitis C virus populations. HCV populationa

HCV HCV HCV HCV HCV HCV HCV HCV HCV HCV

p0, [no drug] p3 p0, [favipiravir] p3 p0, [no drug] p3 p0, [ribavirin] p3 p100, [no drug] p10 p100, [favipiravir] p10 p100, [ribavirin] p10 p200, [no drug] p10 p200, [favipiravir] p10 p200, [ribavirin] p10

N. of nucleotides analyzed (clones/haplotypes)b

31,968 35,520 22,691 16,676 24,864 24,864 21,312 31,968 28,416 23,088

(18/9) (20/20) (18, 10/5, 3) (12, 11/7, 8) (14/14) (14/14) (12/12) (18/16) (16/16) (13/13)

N. of different (total) mutationsc

17 (17) 69 (71) 6 (16) 21 (22) 26 (55) 72 (100) 35 (49) 37 (53) 76 (94) 60 (108)

Mutation frequency Minimumd

Maximume

5.3 × 10−4 1.9 × 10−3 2.6 × 10−4 1.3 × 10−3 1.0 × 10−3 2.9 × 10−3 1.6 × 10−3 1.2 × 10−3 2.7 × 10−3 2.6 × 10−3

5.3 × 10−4 2.0 × 10−3 7.1 × 10−4 1.3 × 10−3 2.2 × 10−3 4.0 × 10−3 2.3 × 10−3 1.7 × 10−3 3.3 × 10−3 4.7 × 10−3

a The populations analyzed are those schematically represented in Fig. 1, and their origin is detailed in Materials and Methods; [no drug] means passages in absence of drug; [favipiravir] means passages in the presence of 400 µM favipiravir; [ribavirin] means passages in the presence of 100 µM ribavirin; p indicates passage number. The values for HCV p0, [no drug] p3 and HCV p0, [favipiravir] p3, HCV p0, [no drug] p3 and HCV p0, [ribavirin] p3 given in the first four rows were previously published (De Ávila et al., 2016; Ortega-Prieto et al., 2013), and are included here for comparative purposes. b The genomic region analyzed by molecular cloning-Sanger sequencing spans residues 7667–9442 (entire NS5B-coding region); the residue numbering corresponds to the JFH-1 genome (GenBank accession number #AB047639). The values in parenthesis indicate the number of clones analyzed followed by the number of haplotypes (number of different RNA sequences) found in the mutant spectrum. c Number of different and total mutations identified by comparing the sequence of each individual clone with the consensus sequence of the corresponding population. d Data represent the average number of different mutations (counted relative to the consensus sequence of the corresponding population) per nucleotide in the components of the mutant spectrum. The statistical significance of the differences between two populations (proportion test) is the following: HCV p0, [no drug] p3 versus HCV p0, [favipiravir] p3: p = 2.5 × 10−7; HCV p0, [no drug] p3 versus HCV p0, [ribavirin] p3: p = 2.0 × 10−4; HCV p100, [no drug] p10 versus HCV p100, [favipiravir] p10: p = 2.7 × 10−6; HCV p100, [no drug] p10 versus HCV p100, [ribavirin] p10: p = 0.051; HCV p200, [no drug] p10 versus HCV p200, [favipiravir] p10: p = 1.3 × 10−5; HCV p200, [no drug] p10 versus HCV p200, [ribavirin] p10: p = 5.3 × 10−5. e Data represent the average number of total mutations (counted relative to the consensus sequence of the corresponding population) per nucleotide in the components of the mutant spectrum relative to the consensus sequence of the corresponding population. The statistical significance of the differences between two populations (proportion test) is the following: HCV p0, [no drug] p3 versus HCV p0, [favipiravir] p3: p = 1.2 × 10−7; HCV p0, [no drug] p3 versus HCV p0, [ribavirin] p3: p = 0.03; HCV p100, [no drug] p10 versus HCV p100, [favipiravir] p10: p = 2.0 × 10−4; HCV p100, [no drug] p10 versus HCV p100, [ribavirin] p10: p = 0.46; HCV p200, [no drug] p10 versus HCV p200, [favipiravir] p10: p = 2.8 × 10−5; HCV p200, [no drug] p10 versus HCV p200, [ribavirin] p10: p = 8.0 × 10–11.

The consensus genomic nucleotide sequences of HCV p0, HCV p100 and HCV p200 have been deposited in GENBank with accession numbers KC595606, KC595609 and KY123743.

treatments; in fact, ribavirin is still a component of some of the current treatments (AASLD-ISDA, 2018). In summary, the present study introduces viral fitness as a relevant parameter to interpret the response of viruses to lethal mutagenesis.

4.2. Virus titration

4. Materials and methods

For titration of infectious HCV, cell culture supernatants were serially diluted and applied to Huh-7.5 cells that had been seeded in 96well plates at 6400 cells/well 16 h earlier. Three days postinfection, cells were washed with PBS, fixed with ice-cold methanol, and stained to detect NS5A using an anti-NS5A monoclonal antibody 9E10, as described previously (Lindenbach et al., 2005; Perales et al., 2013). Virus titers are expressed as TCID50/ml (Reed and Muench, 1938).

4.1. Cells and viruses The origin of Huh-7.5, Huh-7-Lunet, Huh-7.5 reporter cell lines, and procedures for cell growth in Dulbecco's modification of Eagle's medium (DMEM), have been previously described (Blight et al., 2002; Jones et al., 2010; Perales et al., 2013); cells were cultured at 37 °C and 5% CO2. Huh-7.5 cells were used for titration of virus infectivity, while Huh-7.5 reporter cells were used for standard infections and serial passages of HCV. Cells were periodically thawed from a large frozen stock and passaged a maximum of 30 times at a split ratio of 1:3 before use in the experiments. The viruses used in the experiments were rescued from plasmid Jc1FLAG2(p7-nsGluc2A) (a chimera of J6 and JFH-1 from genotype 2a) and amplified to yield HCV p0, and from plasmid GNNFLAG2(p7nsGluc2A), termed GNN (which carries a mutation in the NS5B RNAdependent RNA polymerase that renders the virus replication defective) (Marukian et al., 2008); GNN was used as a negative infection control. The preparation of the initial virus, HCV p0, has been previously described (Perales et al., 2013). HCV p100 and HCV p200 resulted from population HCV p0 passaged 100 and 200 times, respectively, in Huh7.5 reporter cells, as described (Moreno et al., 2017). To control for the absence of contamination, the supernatants of mock-infected cells, which were maintained in parallel with the infected cultures, were titrated; no infectivity in the mock-infected or GNN-infected cultures was detected in any of the experiments.

4.3. Treatments with ribavirin and favipiravir Solutions of favipiravir (Atomax Chemicals Co. Ltd) and ribavirin (Sigma) were prepared at concentrations of 20 mM, and 100 mM in H2O, and PBS, respectively; they were sterilized by filtration, and stored at − 70 °C. Prior to use, the stock solutions were diluted in Dulbecco's modification of Eagle's medium (DMEM) to reach the desired concentration. Huh-7.5 reporter cells were pretreated with the appropriate concentrations (or with DMEM without drug) during 16 h prior to infection. Then, 4 × 105 Huh-7.5 reporter cells were infected (or mock infected) with 1.2 × 104 TCID50 of HCV p0, HCV p100 and HCV p200 (MOI of 0.03 TCID50/cell); the adsorption time was 5 h, and the infection continued for 72–96 h in the absence or presence of the drugs. For successive viral passages, 4 × 105 Huh-7.5 reporter cells were infected with 0.5 ml of the supernatant from the previous infection; the MOI ranged between 4.6 × 10−5 and 5.8 TCID50/cell; each MOI can be calculated from the infectivity values given for each experiment. 105

Virology 523 (2018) 100–109

I. Gallego et al.

Fig. 4. Diversity of non-mutagenized HCV populations. (A) Ratio of diversity indices for the pairwise comparison of the mutant spectrum of populations HCV p0, HCV p100 and HCV p200 (identification code in top box) as determined from MiSeq Illumina sequencing. Values are the average obtained for the three NS5B amplicons; individual index values for each amplicon are given in Tables S1–S3 [http://babia.cbm.uam.es/~lab121/ SupplMatGallego.pdf]. Indices are defined and explained in Gregori et al. (2016). (B) Comparison of the percentage of haplotypes with number of mutations (color coded; black means no mutations relative to the corresponding consensus sequence) for the three amplicons analyzed in the mutant spectra of HCV p0, HCV p100 and HCV p200 (indicated on the left). Procedures for derivation of sequencing reads and their bioinformatics processing are described in Section 4.

controls without template RNA were included to ascertain absence of contaminating templates. Control to ensure an excess of template for molecular cloning and Sanger sequencing were performed as previously described (Airaksinen et al., 2003; Moreno et al., 2017). Amplified DNA was ligated to pGEM-T Vector System I (Promega), and the products were used to transform E.coli DH5α, and DNA from individual clones was amplified and sequenced (Agudo et al., 2010) using the 23 ABI 3730 XLS sequencer (Macrogen, Inc.). Only mutations detected in the two strands of amplified DNA were considered valid for diversity calculations. For Illumina deep sequencing, PCR products were purified (QIAquick Gel Extraction Kit, QIAgen), quantified (Pico Green assay),

4.4. RNA extraction, cDNA synthesis, and PCR amplification and nucleotide sequencing Intracellular viral RNA was extracted from infected cells using the Qiagen RNeasy kit according to the manufacturer's instructions (Qiagen, Valencia, CA, USA). RT-PCR amplification was carried out using AccuScript (Agilent), as specified by the manufacturers. NS5B genomic region was amplified using the specific oligonucleotides Jc1NS5B F1 and Jc1-NS5B R4 (Table S10 [http://babia.cbm.uam.es/ ~lab121/SupplMatGallego.pdf]). Amplification products were analyzed by agarose gel electrophoresis, using Gene Ruler 1 Kb Plus DNAladder (Thermo Scientific) as molar mass standard. Negative

106

Virology 523 (2018) 100–109

I. Gallego et al.

Fig. 5. Effect of mutagenesis on the diversity of mutant spectra of HCV populations. Abundance and incidence indices (indicated at the top and on the abscissa of the bottom panel) are defined in the text and in Gregori et al. (2016). Values are the average obtained for the three NS5B amplicons. Ratios correspond to the value obtained in the presence of favipiravir or ribavirin divided by the corresponding value (same virus population, same number of passages) in absence of drug. The HCV population is indicated on the left, and the passage number (p) is written in each panel. The statistical significance of the increase of the ratio (drug/control) for HCV p0, p3 versus HCV p100 (p3 and p10) or HCV p200 (p3 and p10) was evaluated by applying the Wilcoxon test considering the combined ratio values obtained with favipiravir and ribavirin. All comparisons gave a significant larger ratio increase for HCV p0 than HCV p100 or HCV p200 (p < 0.05 or p < 0.01) except in the comparison of Gini Simpson (HCV p0, p3 versus HCV p100, p10; HCV p0, p3 versus HCV p200, p3; HCV p0, p3 versus HCV p200, p10), Mfmax (HCV p0, p3 versus HCV p100, p10; HCV p0, p3 versus HCV p200, p3), Mfe (HCV p0, p3 versus HCV p100, p3; HCV p0, p3 versus HCV p100, p10; HCV p0, p3 versus HCV p200, p3), and πe (HCV p0, p3 versus HCV p100, p10; HCV p0, p3 versus HCV p200, p3). The Wilcoxon test was also used to compare the significance of the effect of presence of mutagenesis for HCV p100 and HCV p200 considering the combined values for p3 and p10; indices for which differences were significant (p < 0.05) for HCV p100 were: Shannon entropy, Mfmin, N. poly. sites, and N. hpl.; for HCV p200 all indices gave significant differences (p < 0.05) except for Mfe and πe. Individual index values are given in Tables S7–S9, [http://babia.cbm.uam. es/~lab121/SupplMatGallego.pdf], and bioinformatic procedures are described in Section 4.

4.5. Quantification of HCV RNA using real-time RT-PCR

and tested for quality (Bioanalyzer). Three amplicons of NS5B spanning genomic residues 7626–7962, 7941–8257, and 8229–8653 were analyzed from intracellular viral RNA using the Illumina MiSeq platform, with the 2 × 300 mode with v3 chemistry; the fastq files were analyzed as previously described (Gregori et al., 2016, 2014) to obtain the forward and reverse consensus haplotypes with abundances at or above 0.1% median coverage 147,000 reads, and IQR 75570–226100. Files were further down-sampled to the minimum coverage (40,000 reads), and haplotypes below 0.2% were excluded. The resulting median coverage was of 139200 with IQR 71,480-210,600 reads. Procedures for read cleaning, and controls for reliable mutant detection have been described (Gregori et al., 2016, 2014; Quer et al., 2017). The cut-off value mutant detection was set at 0.2%, and only mutations found in two DNA strands were considered. Diversity indices were calculated as described previously (Gregori et al., 2016).

Real-time quantitative RT-PCR (qRT-PCR) of HCV RNA was carried out using the Light Cycler RNA Master SYBR green I kit (Roche) (Perales et al., 2013). The 5′ untranslated region (UTR) of the HCV genome was amplified using as primers oligonucleotides HCV-5UTR-F2 and HCV-5UTR-R2 (Table S10 [http://babia.cbm.uam.es/~lab121/ SupplMatGallego.pdf]). Quantification was relative to a standard curve obtained with known amounts of HCV RNA synthesized by in vitro transcription of plasmid GNNFLAG2(p7-nsGluc2A). The specificity of the reaction was monitored by the denaturation curve of the amplified DNAs. Negative controls (without template RNA and RNA from mock-infected cells) were run in parallel with each amplification reaction to ascertain the absence of contamination with undesired templates. 107

Virology 523 (2018) 100–109

I. Gallego et al.

Acknowledgments We thank Dr. Charles M. Rice for the supply of plasmid Jc1FLAG2(p7-nsGluc2A) and helpful advice for the implement of HCV replicon in cell culture, and to Mercedes Guerrero for help with the statistics. Work in Barcelona was funded by Instituto de Salud Carlos III, by grants PI13/00456, PI15/00829, and PI16/00337 cofinanced by the European Regional Development Fund (ERDF), and by CDTI (Centro para el Desarrollo Tecnológico Industrial), Spanish Ministry of Economics and Competitiveness (MINECO), IDI-20151125. C.P. is supported by the Miguel Servet program of the Instituto de Salud Carlos III (CP14/00121) cofinanced by the European Regional Development Fund (ERDF). CIBERehd (Centro de Investigación en Red de Enfermedades Hepáticas y Digestivas) is funded by Instituto de Salud Carlos III. The work in Madrid was supported by grants BFU-201123604, SAF2014-52400-R, SAF2017-87846-R by Spanish MINECO and S2013/ABI-2906 (PLATESA from Comunidad de Madrid/FEDER). Institutional grants from the Fundación Ramón Areces and Banco Santander to the CBMSO are also acknowledged.

Fig. 6. Schematic representation of delocalization of sequence space as HCV p0 gains fitness in the evolution towards HCV p100 and HCV p200. The broader the fitness plateaus the larger the number of points in sequence space from which the virus can undertake effective replication with limited mutational increase (see text).

4.6. Diversity indices

Appendix A. Supporting information

Diversity indices are used to summarize the composition of a viral population by numeric values. They provide simple ways to approach the understanding of very complex populations. The information required to fully characterize a viral quasispecies is included in the aligned sequences (reads), and their abundances in the population. Considering the number of entities in the population (independently of their abundances), incidence-based indices (number of polymorphic sites, number of haplotypes) can be obtained. Considering haplotype abundances, irrespective of the differences among them, abundancebased indices (Shannon entropy, Gini-Simpson index) can be calculated. When the differences among genomes are taken into account, we define functional indices that in turn can be divided in incidence-based indices (differences among genomes, irrespective of their abundances) (Mfmin, Mfe, πe,) or abundance-based indices (differences are weighted by the corresponding genome abundances) (Mfmax, π). Definitions of the indices used are the following: Shannon entropy, measure of uncertainty in assigning a randomly sampled sequence to an haplotype; Gini-Simpson, probability that two genomes taken at random belong to a different haplotype; Maximum mutation frequency or Mfmax, average number of mutations per nucleotide in the quasispecies with respect to the dominant haplotype, nucleotide diversity or π, average number of nucleotide differences between any two genomes of the quasispecies, number of polymorfic sites (N. poly. sites), number of positions with mutations; number of haplotypes (N. hpl.), number of different genomes; minimum mutation frequency (Mfmin), average number of different mutations per nucleotide; mutation frequency at entity level (Mfe), fraction of mutated residues with respect to the dominant haplotype in the multiple alignment of genomes (irrespective of haplotype abundances); nucleotide diversity at entity level (πe), average number of substitutions among pairs of haplotypes in the multiple alignment of genomes (irrespective of haplotype abundances).

Supplementary data associated with this article can be found in the online version at doi:10.1016/j.virol.2018.07.030. References AASLD-ISDA, 2018. Hepatitis C virus guidance: recommendations for testing, managing, and treating hepatitis C. 〈http://www.hcvguidelines.org〉 (Accessed 30 July 2018). Agol, V.I., Gmyl, A.P., 2018. Emergency services of viral RNAs: repair and remodeling. Microbiol. Mol. Biol. Rev. 82. Agudo, R., Ferrer-Orta, C., Arias, A., de la Higuera, I., Perales, C., Perez-Luque, R., Verdaguer, N., Domingo, E., 2010. A multi-step process of viral adaptation to a mutagenic nucleoside analogue by modulation of transition types leads to extinctionescape. PLoS Pathog. 6, e1001072. Airaksinen, A., Pariente, N., Menendez-Arias, L., Domingo, E., 2003. Curing of foot-andmouth disease virus from persistently infected cells by ribavirin involves enhanced mutagenesis. Virology 311, 339–349. Andino, R., Domingo, E., 2015. Viral quasispecies. Virology 479–480, 46–51. Arias, A., Thorne, L., Goodfellow, I., 2014. Favipiravir elicits antiviral mutagenesis during virus replication in vivo. eLife 3, e03679. Asahina, Y., Izumi, N., Enomoto, N., Uchihara, M., Kurosaki, M., Onuki, Y., Nishimura, Y., Ueda, K., Tsuchiya, K., Nakanishi, H., Kitamura, T., Miyake, S., 2005. Mutagenic effects of ribavirin and response to interferon/ribavirin combination therapy in chronic hepatitis C. J. Hepatol. 43, 623–629. Baranovich, T., Wong, S.S., Armstrong, J., Marjuki, H., Webby, R.J., Webster, R.G., Govorkova, E.A., 2013. T-705 (Favipiravir) induces lethal mutagenesis in influenza A H1N1 viruses in vitro. J. Virol. 87, 3741–3751. Beaucourt, S., Vignuzzi, M., 2014. Ribavirin: a drug active against many viruses with multiple effects on virus replication and propagation. Molecular basis of ribavirin resistance. Curr. Opin. Virol. 8, 10–15. Blight, K.J., McKeating, J.A., Rice, C.M., 2002. Highly permissive cell lines for subgenomic and genomic hepatitis C virus RNA replication. J. Virol. 76, 13001–13014. Crotty, S., Maag, D., Arnold, J.J., Zhong, W., Lau, J.Y., Hong, Z., Andino, R., Cameron, C.E., 2000. The broad-spectrum antiviral ribonucleoside ribavirin is an RNA virus mutagen. Nat. Med. 6, 1375–1379. Cuevas, J.M., González-Candelas, F., Moya, A., Sanjuán, R., 2009. Effect of ribavirin on the mutation rate and spectrum of hepatitis C virus in vivo. J. Virol. 83, 5760–5764. de Avila, A.I., Gallego, I., Soria, M.E., Gregori, J., Quer, J., Esteban, J.I., Rice, C.M., Domingo, E., Perales, C., 2016. Lethal mutagenesis of hepatitis C virus induced by favipiravir. PLoS One 11, e0164691. de Avila, A.I., Moreno, E., Perales, C., Domingo, E., 2017. Favipiravir can evoke lethal mutagenesis and extinction of foot-and-mouth disease virus. Virus Res. 233, 105–112. de la Torre, J.C., 2018. Extending the antiviral value of favipiravir. J. Infect. Dis. 218, 509–511. Dietz, J., Schelhorn, S.E., Fitting, D., Mihm, U., Susser, S., Welker, M.W., Fuller, C., Daumer, M., Teuber, G., Wedemeyer, H., Berg, T., Lengauer, T., Zeuzem, S., Herrmann, E., Sarrazin, C., 2013. Deep sequencing reveals mutagenic effects of ribavirin during monotherapy of hepatitis C virus genotype 1-infected patients. J. Virol. 87, 6172–6181. Dolan, P.T., Whitfield, Z.J., Andino, R., 2018. Mapping the evolutionary potential of RNA viruses. Cell Host Microbe 23, 435–446. Domingo, E., Holland, J.J., 1997. RNA virus mutations and fitness for survival. Annu. Rev. Microbiol. 51, 151–178. Domingo, E., Sheldon, J., Perales, C., 2012. Viral quasispecies evolution. Microbiol. Mol. Biol. Rev. 76, 159–216.

4.7. Statistical analyses To determine the statistical significance of differences in infectivity and RNA, two-way analysis of variance (ANOVA) was carried out using Prism 6 software (GraphPad). For multiple comparisons, Bonferroni's correction was applied. The statistical significance of differences between mutation frequencies was evaluated by the proportion test. The Wilcoxon test was applied to determine the statistical significance of differences in index ratios. To quantify differences in the type of substitutions, the proportion test was carried out.

108

Virology 523 (2018) 100–109

I. Gallego et al.

Manrubia, S., Quer, J., Esteban, J.I., Rice, C.M., Gomez, J., Gastaminza, P., Domingo, E., Perales, C., 2017. Internal disequilibria and phenotypic diversification during replication of hepatitis C virus in a noncoevolving cellular environment. J. Virol. 91, e02505–e02516. Moreno, H., Gallego, I., Sevilla, N., de la Torre, J.C., Domingo, E., Martin, V., 2011. Ribavirin can be mutagenic for arenaviruses. J. Virol. 85, 7246–7255. Mullins, J.I., Heath, L., Hughes, J.P., Kicha, J., Styrchak, S., Wong, K.G., Rao, U., Hansen, A., Harris, K.S., Laurent, J.P., Li, D., Simpson, J.H., Essigmann, J.M., Loeb, L.A., Parkins, J., 2011. Mutation of HIV-1 genomes in a clinical population treated with the mutagenic nucleoside KP1461. PLoS One 6, e15135. Ortega-Prieto, A.M., Sheldon, J., Grande-Perez, A., Tejero, H., Gregori, J., Quer, J., Esteban, J.I., Domingo, E., Perales, C., 2013. Extinction of hepatitis C virus by ribavirin in hepatoma cells involves lethal mutagenesis. PLoS One 8, e71039. Pariente, N., Sierra, S., Lowenstein, P.R., Domingo, E., 2001. Efficient virus extinction by combinations of a mutagen and antiviral inhibitors. J. Virol. 75, 9723–9730. Perales, C., Beach, N.M., Gallego, I., Soria, M.E., Quer, J., Esteban, J.I., Rice, C., Domingo, E., Sheldon, J., 2013. Response of hepatitis C virus to long-term passage in the presence of alpha interferon: multiple mutations and a common phenotype. J. Virol. 87, 7593–7607. Qiu, L., Patterson, S.E., Bonnac, L.F., Geraghty, R.J., 2018. Nucleobases and corresponding nucleosides display potent antiviral activities against dengue virus possibly through viral lethal mutagenesis. PLoS Negl. Trop. Dis. 12 (e0006421). Quer, J., Rodriguez-Frias, F., Gregori, J., Tabernero, D., Soria, M.E., Garcia-Cehic, D., Homs, M., Bosch, A., Pinto, R.M., Esteban, J.I., Domingo, E., Perales, C., 2017. Deep sequencing in the management of hepatitis virus infections. Virus Res. 239, 115–125. Quinones-Mateu, M.E., Arts, E.J., 2006. Virus fitness: concept, quantification, and application to HIV population dynamics. Curr. Top. Microbiol. Immunol. 299, 83–140. Reed, L.J., Muench, H., 1938. A simple method for estimating fifty per cent endpoint. Am. J. Hyg. 27, 493–497. Ruiz-Jarabo, C.M., Pariente, N., Baranowski, E., Dávila, M., Gómez-Mariano, G., Domingo, E., 2004. Expansion of host-cell tropism of foot-and-mouth disease virus despite replication in a constant environment. J. Gen. Virol. 85, 2289–2297. Schuster, P., 2016. Quasispecies on fitness landscapes. Curr. Top. Microbiol. Immunol. 392, 61–120. Schuster, P., Swetina, J., 1988. Stationary mutant distributions and evolutionary optimization. Bull. Math. Biol. 50, 635–660. Sheldon, J., Beach, N.M., Moreno, E., Gallego, I., Pineiro, D., Martinez-Salas, E., Gregori, J., Quer, J., Esteban, J.I., Rice, C.M., Domingo, E., Perales, C., 2014. Increased replicative fitness can lead to decreased drug sensitivity of hepatitis C virus. J. Virol. 88, 12098–12111. Sierra, M., Airaksinen, A., González-López, C., Agudo, R., Arias, A., Domingo, E., 2007. Foot-and-mouth disease virus mutant with decreased sensitivity to ribavirin: implications for error catastrophe. J. Virol. 81, 2012–2024. Sierra, S., Dávila, M., Lowenstein, P.R., Domingo, E., 2000. Response of foot-and-mouth disease virus to increased mutagenesis. Influence of viral load and fitness in loss of infectivity. J. Virol. 74, 8316–8323. Streeter, D.G., Witkowski, J.T., Khare, G.P., Sidwell, R.W., Bauer, R.J., Robins, R.K., Simon, L.N., 1973. Mechanism of action of 1- -D-ribofuranosyl-1,2,4-triazole-3-carboxamide (Virazole), a new broad-spectrum antiviral agent. Proc. Natl. Acad. Sci. USA 70, 1174–1178. Tapia, N., Fernandez, G., Parera, M., Gomez-Mariano, G., Clotet, B., Quinones-Mateu, M., Domingo, E., Martinez, M.A., 2005. Combination of a mutagenic agent with a reverse transcriptase inhibitor results in systematic inhibition of HIV-1 infection. Virology 338, 1–8. Wargo, A.R., Kurath, G., 2012. Viral fitness: definitions, measurement, and current insights. Curr. Opin. Virol. 2, 538–545. Wilke, C.O., Wang, J.L., Ofria, C., Lenski, R.E., Adami, C., 2001. Evolution of digital organisms at high mutation rates leads to survival of the flattest. Nature 412, 331–333.

Eigen, M., Schuster, P., 1979. The Hypercycle. A Principle of Natural Self-organization. Springer, Berlin. Escarmís, C., Dávila, M., Domingo, E., 1999. Multiple molecular pathways for fitness recovery of an RNA virus debilitated by operation of Muller's ratchet. J. Mol. Biol. 285, 495–505. Escribano-Romero, E., Jimenez de Oya, N., Domingo, E., Saiz, J.C., 2017. Extinction of West Nile virus by favipiravir through lethal mutagenesis. Antimicrob. Agents Chemother. 61 (pii: e01400-17). Farci, P., 2011. New insights into the HCV quasispecies and compartmentalization. Sem. Liver Dis. 31, 356–374. Feng, D.F., Doolittle, R.F., 1996. Progressive alignment of amino acid sequences and construction of phylogenetic trees from them. Methods Enzymol. 266, 368–382. Furuta, Y., Komeno, T., Nakamura, T., 2017. Favipiravir (T-705), a broad spectrum inhibitor of viral RNA polymerase. Proc. Jpn. Acad. Ser. B Phys. Biol. Sci. 93, 449–463. Gallego, I., Sheldon, J., Moreno, E., Gregori, J., Quer, J., Esteban, J.I., Rice, C.M., Domingo, E., Perales, C., 2016. Barrier-independent, fitness-associated differences in sofosbuvir efficacy against hepatitis C virus. Antimicrob. Agents Chemother. 60, 3786–3793. Galli, A., Mens, H., Gottwein, J.M., Gerstoft, J., Bukh, J., 2018. Antiviral effect of ribavirin against HCV associated with increased frequency of G-to-A and C-to-U transitions in infectious cell culture model. Sci. Rep. 8, 4619. Grande-Pérez, A., Lázaro, E., Lowenstein, P., Domingo, E., Manrubia, S.C., 2005. Suppression of viral infectivity through lethal defection. Proc. Natl. Acad. Sci. USA 102, 4448–4452. Grande-Pérez, A., Sierra, S., Castro, M.G., Domingo, E., Lowenstein, P.R., 2002. Molecular indetermination in the transition to error catastrophe: systematic elimination of lymphocytic choriomeningitis virus through mutagenesis does not correlate linearly with large increases in mutant spectrum complexity. Proc. Natl. Acad. Sci. USA 99, 12938–12943. Gregori, J., Perales, C., Rodriguez-Frias, F., Esteban, J.I., Quer, J., Domingo, E., 2016. Viral quasispecies complexity measures. Virology 493, 227–237. Gregori, J., Salicru, M., Domingo, E., Sanchez, A., Esteban, J.I., Rodriguez-Frias, F., Quer, J., 2014. Inference with viral quasispecies diversity indices: clonal and NGS approaches. Bioinformatics 30, 1104–1111. Guedj, J., Piorkowski, G., Jacquot, F., Madelain, V., Nguyen, T.H.T., Rodallec, A., Gunther, S., Carbonnelle, C., Mentre, F., Raoul, H., de Lamballerie, X., 2018. Antiviral efficacy of favipiravir against Ebola virus: a translational study in cynomolgus macaques. PLoS Med. 15, e1002535. Holland, J.J., de la Torre, J.C., Clarke, D.K., Duarte, E., 1991. Quantitation of relative fitness and great adaptability of clonal populations of RNA viruses. J. Virol. 65, 2960–2967. Jones, C.T., Catanese, M.T., Law, L.M., Khetani, S.R., Syder, A.J., Ploss, A., Oh, T.S., Schoggins, J.W., MacDonald, M.R., Bhatia, S.N., Rice, C.M., 2010. Real-time imaging of hepatitis C virus infection using a fluorescent cell-based reporter system. Nat. Biotechnol. 28, 167–171. Jordan, P.C., Stevens, S.K., Deval, J., 2018. Nucleosides for the treatment of respiratory RNA virus infections. Antivir. Chem. Chemother. 26 (2040206618764483). Lindenbach, B.D., Evans, M.J., Syder, A.J., Wolk, B., Tellinghuisen, T.L., Liu, C.C., Maruyama, T., Hynes, R.O., Burton, D.R., McKeating, J.A., Rice, C.M., 2005. Complete replication of hepatitis C virus in cell culture. Science 309, 623–626. Martinez-Picado, J., Martinez, M.A., 2008. HIV-1 reverse transcriptase inhibitor resistance mutations and fitness: a view from the clinic and ex vivo. Virus Res. 134, 104–123. Martínez, M.A., Carrillo, C., Gonzalez-Candelas, F., Moya, A., Domingo, E., Sobrino, F., 1991. Fitness alteration of foot-and-mouth disease virus mutants: measurement of adaptability of viral quasispecies. J. Virol. 65, 3954–3957. Marukian, S., Jones, C.T., Andrus, L., Evans, M.J., Ritola, K.D., Charles, E.D., Rice, C.M., Dustin, L.B., 2008. Cell culture-produced hepatitis C virus does not infect peripheral blood mononuclear cells. Hepatology 48, 1843–1850. Moreno, E., Gallego, I., Gregori, J., Lucia-Sanz, A., Soria, M.E., Castro, V., Beach, N.M.,

109