ABSTRACT
Viral quasispecies evolution upon long-term virus replication in a noncoevolving cellular environment raises relevant general issues, such as the attainment of population equilibrium, compliance with the molecular-clock hypothesis, or stability of the phenotypic profile. Here, we evaluate the adaptation, mutant spectrum dynamics, and phenotypic diversification of hepatitis C virus (HCV) in the course of 200 passages in human hepatoma cells in an experimental design that precluded coevolution of the cells with the virus. Adaptation to the cells was evidenced by increase in progeny production. The rate of accumulation of mutations in the genomic consensus sequence deviated slightly from linearity, and mutant spectrum analyses revealed a complex dynamic of mutational waves, which was sustained beyond passage 100. The virus underwent several phenotypic changes, some of which impacted the virus-host relationship, such as enhanced cell killing, a shift toward higher virion density, and increased shutoff of host cell protein synthesis. Fluctuations in progeny production and failure to reach population equilibrium at the genomic level suggest internal instabilities that anticipate an unpredictable HCV evolution in the complex liver environment.
IMPORTANCE Long-term virus evolution in an unperturbed cellular environment can reveal features of virus evolution that cannot be explained by comparing natural viral isolates. In the present study, we investigate genetic and phenotypic changes that occur upon prolonged passage of hepatitis C virus (HCV) in human hepatoma cells in an experimental design in which host cell evolutionary change is prevented. Despite replication in a noncoevolving cellular environment, the virus exhibited internal population disequilibria that did not decline with increased adaptation to the host cells. The diversification of phenotypic traits suggests that disequilibria inherent to viral populations may provide a selective advantage to viruses that can be fully exploited in changing environments.
INTRODUCTION
RNA viruses replicate as complex and dynamic mutant distributions that are termed viral quasispecies (1–4). Quasispecies theory was initially formulated for mutant distributions in equilibrium as a theory of the origin of primitive replicative forms (2). Currently, quasispecies embraces a broad range of theoretical and experimental investigations aimed at explaining the origin, self-organization, survival strategies, and evolutionary dynamics of replicating elements, as well as the genotype-to-phenotype relationship (reviewed in reference 4). Regarding pathogenic entities, quasispecies have provided a theoretical framework to understand the molecular basis of adaptability of RNA viruses, some DNA viruses, and non-nucleic-acid-containing pathogens such as prions (1, 5–10).
The introduction of quasispecies to virology initially described the mutant spectrum nature of RNA virus populations and subsequently emphasized internal interactions among components of the mutant distributions (reviewed in references 4, 11, and 12). The connection of this view of RNA viruses with theoretical quasispecies has been strengthened by several recent developments: (i) extensions of quasispecies theory to finite populations of replicating entities under nonequilibrium conditions (13, 14), (ii) quantifications of mutant spectrum complexity by deep-sequencing methodologies (15), and (iii) experimental evidence of the existence of an error threshold for virus survival in realistic fitness landscapes (4, 16). Several questions that bear on the general understanding of viral quasispecies and disease features of viral pathogens remain. Here, we examine some of these issues using long-term replication of hepatitis C virus (HCV) in human hepatoma cells in culture. Specifically, we address the capacity of the virus population to attain a population equilibrium (steady, constant progeny production and distribution of mutant genomes), the rate of accumulation of mutations, and phenotypic diversification during extensive replication in the heterogeneous but noncoevolving environment provided by human hepatoma cells in culture.
The choice of HCV is justified because several problems for the control of HCV infections have their origins in the capacity of mutant spectra to overcome barriers to their replication (3, 17, 18). Chronic HCV infections have a global impact (19, 20). Despite average sustained response rates approaching 98% with new treatments based on direct-acting antiviral (DAA) combinations, several limitations for the prevention and control of HCV-associated diseases remain, even for human populations with access to treatment. There is an increasing frequency of circulating inhibitor-resistant mutants, a fraction of patients do not respond to the new treatments, DAAs may evoke hepatocarcinoma recurrence, and no preventive or therapeutic vaccines are available (21–28). Studies on HCV dynamics in infected patients are relevant to establish relationships between population complexity and clinical parameters (reviewed in references 3 and 18). However, the evolution of mutant spectra in the liver (or extrahepatic tissues) is perturbed by physiological and immunological changes that occur during virus multiplication (intrahost constraints) (3, 11). It is not known how HCV would evolve if left unperturbed while replicating for a long time in the same cellular environment. Such exploration can be carried out due to the availability of effective cell culture systems to replicate the complete HCV genome (29–32). Adaptive mutations have been identified in infectious HCV clones or subgenomic replicons upon multiplication in cell culture (33–37). For the present study, the starting virus was HCVcc, originated by transcription of plasmid Jc1FLAG2(p7-nsGluc2A) and cell transfection (38) and amplified in human hepatoma cells to obtain HCV p0 (36). The latter has been previously used in our laboratory to derive high-fitness populations that display resistance to several anti-HCV inhibitors (36, 39, 40). Here, we show that upon subjecting HCV p0 to 200 serial passages in Huh-7.5 reporter cells in an experimental design that avoids host cell evolution, the viral population increased its replicative capacity, with continuous mutational waves. The system generated internal instabilities that precluded attaining population equilibrium. Mutations affected all genomic regions and not particularly those proven more variable among clinical isolates and that served historically to define hypervariable regions in the HCV genome. The virus underwent a number of phenotypic changes, some of which affected the virus-host relationship. Possible mechanisms for population disequilibrium (variations in progeny production and persistence of molecular waves) are analyzed in the light of quasispecies theory. We assess the rate of accumulation of mutations in the cell culture scenario in connection with previously reported problems derived from the application of the molecular-clock hypothesis to rapidly evolving RNA viruses in nature (41, 42). Implications for HCV dynamics in vivo are discussed.
RESULTS
Progeny production during long-term serial passages of hepatitis C virus.To study long-term HCV evolution in a noncoevolving cellular environment, HCV p0 was subjected to 200 serial passages in Huh-7.5 reporter cells with naive cells infected at each passage. Viral populations are identified with a p followed by the passage number (e.g., HCV p100 is HCV p0 passaged 100 times in Huh-7.5 reporter cells). The average values of infectious progeny and viral RNA shed into the culture medium indicated a sustained infection over 200 passages (equivalent to about 700 days of intracellular replication) (Fig. 1A, B, and C). At passage 60, there were cytopathic signs (detachment of about 10% of the cells from the monolayer) that were transiently diminished by reducing the infectious dose at passages 60 to 100. The amount of infecting virus ranged from 4 × 104 to 3 × 106 50% tissue culture infective doses (TCID50) for passages 1 to 60 and passages 100 to 200 and from 1 × 104 to 5 × 105 TCID50 for passages 60 to 100 (Fig. 1D). Virus-induced cytopathology increased at late passages in an MOI-dependent manner (see Fig. S1 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]).
Evolution of HCV infectivity during 200 passages in cell culture. The initial clonal population (HCVcc) was obtained by electroporation of Lunet cells by a transcript from plasmid Jc1FLAG2(p7-nsGluc2A). (A) Extracellular infectious-virus titer. (B) Extracellular viral RNA, quantified by real-time RT-PCR. (C) Specific infectivity (values from panels A and B). (D) Infectious dose at each passage; the decrease between p60 and p100 is due to dilution of the inoculum to limit cell lysis. The data on virus titers and viral RNA up to passage 100 were previously reported (39) and are included here for completeness. (E) Coefficient of variation of the amplitude of fluctuations of viral titer (values in panel A) and of viral RNA (values in panel B). The slope of the coefficient of variation as a function of the passage number was statistically significant for the infectivity level (P < 0.0001) but not for the RNA level (P = 0.7323). The origin of HCVcc, conditions for infections, titration of HCV infectivity, and quantification of HCV RNA, as well as the positive and negative controls included in the assays, are described in Materials and Methods.
A fluctuating pattern mathematically characterized by a bimodal distribution was observed for the progeny infectivity but not for the viral RNA in the culture medium, resulting in fluctuations of specific infectivity (Fig. 1A, B, and C; see Fig. S2 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). The amplitude of the fluctuations increased with the passage number, as shown by a 3-fold increase in the coefficient of variation of the progeny infectivity, which was not traceable for viral RNA production (Fig. 1E). The results suggest that the amount of HCV progeny was dependent on the infection context in the sense that passages with low progeny production predicted that the next passage would have high progeny production, and vice versa.
Replicative parameters of passaged hepatitis C virus populations.One of the questions raised by the fluctuating pattern is whether the virus increased its net replicative capacity with extended passaging. To address this point, the amounts of infectious progeny produced at different times postinfection were compared in single infections with HCV p0, HCV p100, and HCV p200 (Fig. 2). Both HCV p100 and HCV p200 belong to the upper range of progeny values in the serial infections (compare Fig. 1A). Following a lag period of 12 h, the progeny production measured extracellularly or intracellularly increased for HCV p100 and HCV p200 relative to HCV p0 (Fig. 2 and Table S1 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). The increased rate of progeny production of HCV p100 and HCV p200 was also observed in the course of 5 serial passages over a 1,000-fold range of the initial multiplicity of infection (MOI); the maximum titers attained beyond passage 3 for HCV p100 and HCV p200 were indistinguishable (Fig. 3A and B; see Table S1 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). HCV p0 and HCV p200 displayed the same thermal stability (Fig. 3C), excluding possible confounding effects derived from differential stability at the extracellular stage of the HCV populations under comparison. The results are in agreement with the previously reported fitness increase of HCV p45 and HCV p100 relative to HCV p0, measured independently with growth competition experiments (39). Thus, HCV p0 increased its rate of progeny production upon passaging in Huh-7.5 reporter cells, but the maximum progeny levels attained at passages 3 to 5 were similar for HCV p100 and HCV p200.
Kinetics of extracellular and intracellular progeny production by HCV p0, HCV p100, and HCV p200. (A and B) Production at early times postinfection and standard deviations from triplicate titrations. (C and D) Progeny production and standard deviations at 24 h to 72 h postinfection. The values are averages of 3 independent experiments, each with titrations in triplicate. Cells were mock infected or infected at an MOI of 0.03 TCID50/cell (4 × 105 Huh-7.5 cells infected with 1.2 × 104 TCID50). The mean values and standard deviations were calculated with the transformed data to the logarithm of the values in ordinate. Statistical significance: *, P < 0.05; **, P < 0.005; ***, P < 0.0005 (ANCOVA).
Growth rates and maximum viral titers of HCV p0, HCV p100, and HCV p200 in the course of five serial passages in human hepatoma cells. (A) Huh-7.5 reporter cells were either mock infected or infected with HCV p0, HCV p100, or HCV p200 (filled boxes above each panel) at initial MOI of 0.03, 0.003, 0.0003, and 0.00003 TCID50/cell (upper box) (4 × 105 Huh-7.5 cells infected with 1.2 × 104, 1.2 × 103, 1.2 × 102, or 1.2 × 101 TCID50, respectively); in successive passages, the MOI ranged from 0.01 to 1 TCID50/cell for HCV p0, 0.01 to 10 TCID50/cell for HCV p100, and 0.1 to 100 TCID50/cell for HCV p200. The values are the results of triplicate experiments. Infections were allowed to proceed for 72 to 96 h. Viral titers were determined in the cell culture supernatants. The dashed horizontal lines indicate the limits of infectivity detection. The dashed curve of progeny production by HCV p200 at an MOI of 0.00003 TCID50/cell (right) represents the average of the results of two experiments (two values are given for each passage), since the third experiment did not yield progeny virus. (B) Average growth rates and maximum titer values (filled boxes above each panel) for HCV p0, HCV p100, and HCV p200 calculated with all the MOIs shown in panel A. The mean values and standard deviations were calculated with the transformed data to the logarithm of the values in ordinate. Statistical significance: ns, not significant; ***, P < 0.0001; ANCOVA and t test (for growth rate and maximum titer, respectively). (C) Thermal inactivation of HCV p0 and HCV p200 (inset). Virus samples in DMEM were incubated at 45°C for the indicated amounts of time and titrated in triplicate. The thermal inactivation constants (k) at 45°C were 0.0125 ± 0.0015 (corresponding to a half-life of 55 min) for HCV p0 and 0.0122 ± 0.0014 (corresponding to a half-life of 57 min) for HCV p200 (P = 0.37; ANOVA).
Genetic diversification and mutation waves upon long-term passage of hepatitis C virus.The genetic diversification of HCV p0 upon passage in Huh-7.5 reporter cells was studied through determination of the consensus sequence and characterization of the mutant spectrum of sequential populations. Comparison of the consensus genomic nucleotide sequences of HCV p0, HCV p45, HCV p100, and HCV p200 indicated accumulation of mutations, a limited number of reversions, and multiple points of heterogeneity (more than 1 nucleotide and amino acid at a given position) (Fig. 4A and B; see Table S2 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). The average ratio of synonymous to nonsynonymous mutations was 0.9 to 1.1. The number of accumulated mutations as a function of the passage number deviated slightly from linearity (Fig. 4C), as implied by a similar fit of the experimental points to an exponential and a linear function (see Fig. S3A [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). A similar deviation was found by counting only synonymous mutations. From passages 1 to 100, the average rate of accumulation of total mutations was 2.7 × 10−5 mutations per nucleotide and passage (m/nt/p) (which is equivalent to 7.9 × 10−6 mutations per nucleotide and day [m/nt/day]); the corresponding value obtained from passages 100 to 200 was 4.9 × 10−5 m/nt/p (equivalent to 1.4 × 10−5 m/nt/day), a 1.8-fold difference (P < 0.05; chi-square test). A similar difference was calculated using the number of new mutations between successive analyses (Fig. 4A, red lines).
Accumulation of mutations in the consensus sequence of HCV p0 passaged in Huh 7.5 reporter cells. (A) Scheme of the HCV genome with the RNA residue numbers that delimit the encoded proteins. A Gaussia luciferase gene inserted between coding region p7 and NS2 (38) has been omitted; the marker was lost prior to passage 45 (36). The viral populations are indicated on the right. Mutations relative to the sequence of Jc1FLAG2(p7-nsGluc2A) are indicated by vertical lines below the genome; the red lines indicate new mutations, and the black lines indicate mutations present in the preceding population analyzed. A caret at the tip of a line represents heterogeneity at that nucleotide position (two peaks in the same sequence position). The asterisks indicate mutations that reverted relative to the previous passage analyzed. (B) Amino acid substitutions deduced from the nucleotide sequences given in panel A. The viral populations are indicated on the left. Viral proteins and amino acid (single-letter code) residue numbers are given at the top. Amino acids separated by a slash indicate a mixture of amino acids at the indicated position; an asterisk next to an amino acid indicates that the amino acid is present at a proportion lower than 25% of the total. (C) Number of accumulated mutations as a function of the passage number. All mutations (irrespective of their proportions at a given position) and dominant mutations (present at a frequency of 50% or higher at a given position) are indicated. (Left) Mutations in the entire genome. (Right) Mutations in the NS5A-coding region. A complete description of all mutations is given in Table S2 (http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ).
Since NS5A accumulated 30% of the total mutations found in the consensus sequence of HCV p200 (Fig. 4; see Table S2 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]), the mutant spectrum of the NS5A-coding regions of HCV p0, HCV p45, HCV p100, HCV p150, and HCV p200 was examined using molecular cloning-Sanger sequencing (MCS) and ultradeep pyrosequencing (UDPS) (Tables 1 and 2; see Tables S3 and S4 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). A comparison of the maximum and minimum mutation frequencies indicated a significant increase from HCV p0 to HCV p100 but not among later populations (Table 1; see Fig. S3B [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). A similar conclusion was reached from a comparison of diversity indices derived from UDPS data (Table 2). Visualization of the frequency variation of the mutations found by both methods relative to the reference (HCVcc) sequence (22% of the total number of mutations scored by either of the two methods) as a function of the passage number revealed multiple mutational waves (Fig. 5A). The fact that similar patterns were obtained by MCS and UDPS renders it extremely unlikely that the mutational waves were the result of a bias at the level of the virus sample taken for RNA extraction, amplification, and sequencing. MCS and UDPS yielded very similar distributions of mutations occurring temporarily (transient mutations that were found only in one population, mutations that were transient and reemergent, mutations maintained since their first appearance, and mutations unique to HCV p200 [Fig. 5B]) and comparable distributions of mutation types in the passaged populations (see Fig. S3C [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). The number of clean reads obtained by UDPS permitted the calculation of six diversity indices (D0, D1, D2, and Dinf, which are Hill numbers; Mf, maximum mutation frequency; π, sample nucleotide diversity) for three NS5A amplicons, following the procedures described previously (43) (see Fig. S4 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). The results of the individual indices and a principal-component analysis indicated an increase of diversity relative to each corresponding consensus sequence up to passage 100 and then a leveling off, and even some decrease, after passage 100. A limitation of intrapopulation diversity after passage 100 did not entail a decrease of the mutational waves (relative to the parental HCVcc sequence) that continued beyond passage 100 (Fig. 5A). Thus, increase of HCV progeny production occurred through complex changes in mutant frequencies, despite replication in a noncoevolving cellular environment and reaching plateaus for maximum progeny production and overall intrapopulation complexity.
Mutant spectrum analysis by molecular cloning and Sanger sequencing of the NS5A-coding region of hepatitis C virus passaged in Huh-7.5 reporter cells
Ultradeep pyrosequencing analysis of the NS5A-coding region of hepatitis C virus passaged in Huh 7.5 reporter cells
Mutant spectrum dynamics of hepatitis C virus passaged in Huh-7.5 reporter cells. (A) Pictorial representation of the variation in mutation frequency of the mutations (color codes are on the right) found both by MCS and UDPS. The colored lines join with an arbitrary outline of the mutation frequency values determined at passages 0, 45, 100, 150, and 200. Note that mutations that contribute to waves (mutations that increase or decrease in frequency relative to the previous or subsequent passage analyzed) are present in each of the four passaged populations analyzed. (B) Distributions of mutations according to the HCV populations where they were found: transient, observed only in some intermediate population; transient and reemergent, detected in a population and then not observed in a subsequent population and detected again in a later population; maintained, detected since their first appearance; and unique to HCV p200.
Variation and heterogeneity of phenotypic traits.The complex dynamics at the genome level raised the question whether, in addition to altered cell killing and replication rate (Fig. 2 and 3; see Fig. S1 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]), other phenotypic traits also varied upon long-term passage of HCV p0. To probe virion properties, we chose to examine the virus density profile by sucrose gradient sedimentation, because changes in density sedimentation properties were described for HCV after long-term persistence in Huh-7.5.1 cells (44). HCV p100 and HCV p200 modified their profiles toward higher density relative to HCV p0 (Fig. 6). Evidence of intrapopulation heterogeneity was suggested by differences in specific infectivity along the gradient; the high-density fractions of HCV p100 and HCV p200 displayed specific infectivities that were significantly higher than those of the corresponding fractions of HCV p0 (Fig. 6D; see Fig. S5 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). Thus, the genetic diversification and adaptation of HCV p0 to Huh-7.5 reporter cells was accompanied by modifications of the density properties of virus particles.
Sedimentation properties of HCV populations. (A to C) The viral populations indicated at the top were subjected to sucrose gradient sedimentation under the conditions detailed in Materials and Methods. The virus titer (bars; the dashed horizontal lines mark the limits of detection) and viral RNA (lines above the bars) were determined for each fraction in triplicate. The results are the averages of the determinations with two gradients per sample run in parallel. (D) Specific infectivity (expressed as the logarithm of the ratio of infectivity to viral RNA) of the virus present in each fraction. Viruses are indicated at the top. The dashed line indicates the sucrose density values (an average of six determinations per fraction; three gradients with each sample were run in duplicate). The asterisks denote a statistically significant difference between the specific infectivity of HCV p0 and either HCV p100 or HCV p200 (*, P < 0.05; two-tailed ANOVA). The error bars indicate standard deviations.
Our previous studies showed that HCV p100 exhibited an increase of shutoff host cell protein synthesis relative to the parental HCV p0 and that the increase was associated with enhanced phosphorylation of protein kinase R (PKR) and eIF2α and paralleled HCV protein expression (36, 39), in agreement with previous investigations (45). Since the evolution from HCV p100 to HCV p200 resulted in an increase of replication kinetics, but not of maximum viral production (Fig. 2 and 3), we compared the shutoff evoked by HCV p100 and HCV p200. The results revealed a significant increase of shutoff of host cell protein synthesis by HCV p200 relative to HCV p100 and HCV p0 that was more accentuated at late times postinfection, concomitant with an increase of viral progeny production (Fig. 7A and B). At 72 h postinfection, the ratio of the viral protein level for HCV p200 relative to HCV p100 was 1.42 for NS5A and 0.87 for core (Fig. 7C). This limitation of the core level suggests differential regulation of viral protein synthesis or degradation to achieve unequal relative protein levels (see Discussion). The genetic wandering parallels a progressive directional divergence of phenotypic traits, particularly increasing cytopathology, a shift toward higher virion density, and enhanced shutoff of host cell protein synthesis.
Effects of HCV p0, HCV p100, and HCV p200 infection on host cell protein synthesis and accumulation of viral proteins. (A) (Top) Huh-7.5 cells were either mock infected or infected with HCV p0, HCV p100, or HCV p200 at an MOI of 0.03 TCID50/cell (4 × 105 Huh-7.5 cells infected with 1.2 × 104 TCID50), and the virus was allowed to adsorb to the cells for 5 h. At 24, 48, or 72 h postinfection (p.i.), the cells were labeled with [35S]Met/Cys for 1 h, and cell extracts were subjected to SDS-PAGE and visualized by autoradiography. (Bottom) The total amount of labeled protein was calculated by densitometry of the corresponding autoradiograms and expressed as a percentage of the amount of actin, taking the amount of actin in uninfected cells as 100%. (B) Viral titers in the cell culture supernatants determined at 24, 48, and 72 h p.i. (C) HCV protein expression at 24, 48, and 72 h p.i. The protein extracts from infected Huh-7.5 cells were those from the experiment described in panel A. HCV NS5A and core were stained by Western blotting (WB) using monoclonal antibodies specific for the indicated proteins. The amount of cellular proteins was normalized to the amount of actin, visualized by Western blotting. The asterisks denote statistically significant differences for the comparisons (*, P < 0.05; **, P < 0.005; ***, P < 0.0005; ns, not significant; unpaired t test).
DISCUSSION
The objective of the present study was a quantitative evaluation of the capacity of HCV to evolve genetically and phenotypically when allowed to replicate extensively in a noncoevolving cellular environment. This question cannot be answered by comparing sequential HCV populations from infected patients due to immunological and physiological host alterations. The Huh-7.5 cells used for this study are not homogeneous because they display genetic instability typical of transformed cells (46–48). They are, however, noncoevolving in the sense that fresh cells are used for each passage, in contrast with long-term persistence of HCV or its subgenomic replicons, in which the cells may coevolve with the virus (44, 49). The study has documented adaptation of HCV p0 to replicate with increasing efficiency in the Huh-7.5 reporter cells, as expected of a virus whose origin is a molecular clone derived from an isolate from a patient (31). The degree of adaptation did not preclude internal instabilities, as indicated by a passage-to-passage fluctuation in infectious-progeny production whose amplitude increased at late passages, as well as sustained mutational waves throughout the experiment. The mutational flow dynamics suggest that multiple HCV sublineages (clouds of related mutants) that show minor incremental replicative advantages relative to parental lineages frequently arise. This gradual accommodation to a new environment suggests that HCV has available multiple alternative evolutionary pathways to gain fitness, as previously documented with foot-and-mouth disease virus (FMDV) (50).
Several not mutually exclusive mechanisms may account for the origin of the fluctuations in progeny production observed with HCV. Early work documented a similar pattern during the cyclic dominance of vesicular stomatitis virus (VSV) and its defective interfering (DI) particles during serial infections at high MOI (5, 51, 52). Also, fluctuations in infectious-progeny production were reported upon large population passages of VSV that were attributed to a population size limitation for further fitness gain (53, 54). Fluctuations in infectivity were exhibited by FMDV clones that attained very low fitness values as a consequence of many serial plaque-to-plaque passages (55, 56). In this case, fluctuations were attributed to the stochasticity caused by extreme bottlenecks, which resulted in high unpredictability of the fitness of the founder genome after each passage (14, 57, 58). In the current case, an alternation between dominance of interference (due to a transient increase of defective mutants) and complementation (known to occur within replicative units) may play a role (3, 4, 8, 11, 12, 59). Although cellular functions may also be involved, fluctuations cannot be attributed to Huh-7.5 cell heterogeneity, which was a constant parameter throughout the experiment, while the amplitude of the fluctuations increased at late passages. A key role of the accumulation of classic DI particles seems unlikely for two reasons: (i) there is only a slight trend toward a decrease of specific infectivity from passages 100 to 200, which may even be influenced by a larger amount of viral RNA in cellular debris due to increased cytopathology, and (ii) the MOI fell periodically below 1 TCID50/cell. A role of other types of defective genomes, even if they do not accumulate, generated during intracellular rounds of replication is unlikely because the high-density fractions of HCV p100 and HCV p200 displayed significantly higher specific infectivity than the equivalent fractions of HCV p0. Only two point deletions in heteropolymeric regions could be reliably identified in the RNA of HCV p150 (see Tables S3 and S4 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]), precluding statistical evaluation of deletions as possible markers of defectiveness. We are further investigating the observed HCV fluctuations to attempt to produce a realistic model for their occurrence.
Serial passages are expected to favor virus adaptation to a cell culture environment, and adaptation has been reported for many virus-host systems, including HCV replicons (60–62; reviewed in reference 11). Coexistence of viral evolutionary sublineages has been previously documented with several viruses, both in cell culture and in vivo, particularly in response to external selective constraints (63–68). What may be significant for the evolutionary potential of HCV is that the frequency of sublineages did not tend to decrease with adaptation to cells. We considered that the changes of mutant frequency might not occur in the viral population but might be an artifact due to biases derived from the limited number of genomic sequences sampled in each analysis. This possibility was rendered unlikely because of the agreement of results using MCS and UDPS that involved two independent sampling events, viral genomes copied with different sets of primer pairs, and different nucleic acid amplification conditions. We cannot evaluate whether mutations that vary upon virus passage belong to the same or different genomic RNA molecules or the possible contribution of hitchhiking to the increase in frequency of some mutants (11). Establishing mutation linkage will require full-genome deep sequencing.
The distribution of mutations along the HCV genome differed from the distribution observed when comparing clinical isolates (69–72). In particular, it is noteworthy that hypervariable regions 1 and 2 (HVR1 and HVR2) in HCV p200 include only 1 amino acid replacement relative to HCV p0 (1.78% of the total versus 66% of the total within NS5A and NS5B). These differences probably reflect the direct and prolonged selection for improved replication of HCV p0 in the absence of an immune response and emphasize the extent to which attribution of constant and variable regions in viral genomes is dependent on the environments in which replication took place.
The acquisition of mutations upon prolonged replication in an externally unperturbed environment constitutes an ideal scenario to test the molecular-clock hypothesis with an RNA virus. Despite the limited number of data points obtained for the consensus sequences of the entire genome or the NS5A-coding region, the rate of accumulation of mutations could be fitted equally to a linear or to an exponential function, with a significant trend of the rate to increase at late passages. The clock acceleration may reflect the increased replication rate of HCV p100 and HCV p200 relative to HCV p0 (Fig. 2 and 3). Molecular-clock irregularities are frequent in general evolution (73, 74) and can be very dramatic in the case of viruses (41, 42). Irregularities have been explained by mechanisms as diverse as differences in replication rate, time-dependent differences in intensity of selection, variations of virus population size, or different mutational inputs due to alterations of the replication machinery, among others (73–75). In the HCV cell culture system, we do not have any evidence of changes in copying fidelity of the HCV replicative machinery with passage number. The two main arguments are (i) that the HCV polymerase from HCV p200 does not include any amino acid substitution that we could relate to catalytic properties of the enzyme—no mutations have been found around the active site (which spans residues G188 to D225 and T287 to V370) of NS5B in HCV p200—and (ii) more directly, biological clones isolated from HCV p0 and HCV p200 display indistinguishable mutant spectrum complexities (I. Gallego and C. Perales, a unpublished results). We cannot exclude the possibility that clusters of mutations that affect other HCV proteins could alter replicative properties other than copying fidelity. An alternative possibility is that the clock deviation was triggered by the transient 2-log-unit decrease of the infectious dose (and consequently of the MOI) from passages 60 to 100, introduced to attenuate the increase of cytopathology. If this alteration were responsible for the deviation, its mechanism would be unclear, because the modest bottleneck should still allow selection with minimal drift (76), and bottlenecks are expected to result in loss of diversity, which was not observed at passage 100. The fact that a strict molecular clock was not followed in the experimental design of our study suggests that the alterations in population size and selective constraints that viruses undergo in nature must impose severe limitations on a molecular-clock-like behavior (41, 42).
The exploration of sequence space by HCV was obviously not a walk totally in neutral regions. It appeared as a wavy path toward adaptation, as suggested by at least six relevant phenotypic modifications. The increase in the replication rate may be associated with cytopathology and increased shutoff of host cell protein synthesis (77). The still limited HCV cytopathology at late passages, and the requirement for multiple host cell factors for HCV replication (78–80), may explain the limitation of the core protein level that accompanied the enhanced shutoff of host cell protein synthesis at passage 200. One of the functions of HCV core is to downregulate host protein expression (81), but too high a level of core as a consequence of increased replication may limit the input of host proteins needed for sustained replication. How modulation of core levels is achieved can only be speculated at present, since HCV core interacts with about 80 host proteins (82, 83).
The evolution toward increased particle density and specific infectivity of HCV in cell culture has been previously associated with adaptive single amino acid replacements in E2 that decrease the dependency of virus entry on the levels of some HCV receptors and diminish the association of virions with low-density lipoprotein or very low-density lipoproteins (44, 84–88). Although the amino acid substitutions found in E2 of HCV p100 or HCV p200 are not the same as in previous reports, it is likely that particle affinity for lipids may be influenced by several E1 and E2 residues, thus altering the observed density profile (Fig. 7). Evidence of an additional potential phenotypic impact of the mutations scored is that two of them (found at a frequency of >10% in the mutant spectrum) correspond to inhibitor resistance mutations (T24A and F28I in NS5A for HCV p150 and HCV p200, both detected by MCS and UDPS, and T179A in NS5B for HCV p100, detected in the consensus sequence, a substitution that is often associated with S282T in NS5B to confer sofosbuvir resistance [89]) (90) (see Tables S4 and S5 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]).
The results reported here and in previous studies (65, 67, 68) provide additional support of quasispecies dynamics during RNA virus replication with populations consisting of mutant clouds rather than defined lineages with molecular continuity (4). Successive replacements of mutant clouds are possible by the multiple evolutionary alternatives that viruses have to approach adaptation to the environment (11). Exploration of alternative pathways may be related to the internal disequilibrium, which in turn may originate within intracellular replicative units. In fact, according to quasispecies theory, the mutant spectrum per se is part of the environment (91). The replicative environment per se may have been constantly modified by a different infecting mutant swarm and by subsequent mutations during replication. Even if HCV could reach a perfect adaptation to homogeneous Huh-7.5 reporter cells, the present study suggests that the system would internally develop perturbations due to the unpredictable mutational input, and this unavoidable event may also provide an advantage for long-term virus survival.
MATERIALS AND METHODS
Cells and viruses.The origin of Huh-7.5, Huh-7 Lunet, and Huh-7.5 reporter cell lines and procedures for cell growth in Dulbecco's modified Eagle's medium (DMEM) have been previously described (36, 49, 92); 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.
Cytopathology was measured in Huh-7.5 reporter cells. Briefly, cells were seeded in 96-well plates at 70% confluence and infected with HCV populations from passages 8, 65, 95, 151, and 200 at three different MOI (0.03, 0.3, and 3 TCID50/cell). After 72 h, the cells were fixed, stained with crystal violet, and resuspended in 1% SDS. The optical density was measured at a wavelength of 595 nm. Each cell biomass was the average of 8 different determinations.
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(p7-nsGluc2A), termed GNN (which carries a mutation in NS5B that renders the virus replication defective) (38); GNN was used as a negative infection control. The preparation of the initial virus, HCV p0, has been previously described (36). No infectivity in the mock-infected or GNN-infected cultures was detected in any of the experiments.
The intracellular progeny was measured in a cell lysate; at the indicated times postinfection, cells were washed once with phosphate-buffered saline (PBS) (1.47 mM KH2PO4, 10 mM Na2HPO4, 2.7 mM KCl, 137 mM NaCl, pH 7.4) and incubated with trypsin-EDTA (Invitrogen, Carlsbad, CA) for 2 min at 37°C. The cells were resuspended in PBS and collected by centrifugation at 1,500 rpm for 3 min. The cell pellet was resuspended in DMEM-10% fetal calf serum (FCS), and the cells were lysed by three freeze-thaw cycles in dry ice and a 37°C water bath, respectively. The supernatant was collected and used for titration or stored at −80°C.
Virus titration.For titration of infectious HCV, samples were serially diluted and applied to Huh-7.5 cell monolayers in 96-well plates (6,400 cells/well seeded 16 h earlier). Three days postinfection, the cells were washed with PBS, fixed with ice-cold methanol, and stained to detect NS5A using anti-NS5A monoclonal antibody 9E10, as described previously (29, 36). Virus titers are expressed as TCID50 per milliliter (93). Titrations were performed in triplicate.
Long-term serial passage of HCV.Serial passages of HCV p0 were carried out as previously described (36). Briefly, 4 × 105 Huh-7.5 reporter cells were infected with HCV p0 at an MOI of 0.5 TCID50/cell; after a virus adsorption period of 5 h at 37°C, the inoculum was removed and 2 ml of medium was added to the cell monolayer. The infected cells were further incubated at 37°C for 72 h (in some experiments for 96 h); for each subsequent passage, 4 × 105 Huh-7.5 reporter cells were infected as indicated above using 0.5 ml of cell culture supernatant (or 0.5 ml of a dilution of the supernatant) from the previous passage. A total of 200 passages were performed. The MOI for passages 1 to 59 and 101 to 200 ranged from 0.1 to 8 TCID50/cell, due to fluctuations in progeny production; the MOI for passages 60 to 100 ranged from 0.02 to 1 TCID50/cell due to fluctuations and dilution of the inoculum to limit cytopathology. The experimental design was such that each passage involved infection of fresh cells; this means that the evolution of cells that occurs during persistent HCV infections in cell culture (44) is avoided in the present design.
RNA extraction, cDNA synthesis, PCR amplification, and nucleotide sequencing.Intracellular RNA was extracted from infected cells using the Qiagen RNeasy kit (Qiagen, Valencia, CA, USA), according to the manufacturer's instructions. RNA from cell culture supernatants or cell lysates was extracted using the Qiagen QIAamp viral RNA mini kit (Qiagen, Valencia, CA, USA). Reverse transcription (RT) was performed using avian myeloblastosis virus (AMV) reverse transcriptase (Promega), and PCR amplification of specific HCV genomic regions was carried out using AccuScript (Agilent Technologies), with specific oligonucleotide primers (36) (see Table S5 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). The amplification products were analyzed by agarose gel electrophoresis, with HindIII-digested Φ-29 DNA as a molecular mass standard. Negative controls without template RNA were included in parallel to ascertain the absence of cross-contamination by template nucleic acids. Nucleotide sequences of genomic HCV RNA were determined on the two strands of an amplified cDNA copy (36, 94); only mutations detected in the two strands were counted. To evaluate the complexity of mutant spectra by MCS, HCV RNA was extracted as described above and subjected to RT-PCR to amplify the NS5A-coding regions, as previously described (95). To ensure an excess of template in the RT-PCR amplifications for mutant spectrum analysis and to avoid complexity biases due to redundant amplifications of the same initial RNA templates, amplifications were carried out with template preparations diluted 1:10 and 1:100; only when the 1:100-diluted template produced a visible DNA band was molecular cloning pursued, using the DNA amplified from undiluted template (96). Controls to ascertain that mutation frequencies were not affected by the basal error rate during amplification have been previously described (97). Amplified DNA was ligated to pGEM-4Z (Amersham) and used to transform Escherichia coli DH5α, and individual colonies were picked for PCR amplification and nucleotide sequencing, as previously described (94). No difference in the consensus sequence or in the complexity of the mutant spectrum was observed with intracellular or extracellular HCV p100 RNA (unpublished results).
For UDPS analyses (GS-FLX or GS-Junior platform; 454 Life Sciences-Roche), RT-PCR was performed using Accuscript (Agilent); six amplicons covering the NS5A-coding region (A1, spanning genomic residues 6152 to 6454; A2, residues 6446 to 6767; A3, residues 6737 to 6954; A4, residues 6910 to 7252; A5, residues 7224 to 7550; and A6, residues 7432 to 7725) were analyzed using specific primers (see Table S5 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.pdf ]). RT-PCR amplifications were performed in triplicate and mixed equimolarly prior to the analysis. Then, the PCR products were purified (QIAquick Gel Extraction kit), quantified (Pico Green assay), and analyzed for quality (Bioanalyzer) prior to the UDPS procedure (39, 95). Negative controls (without template RNA) were run in parallel to ascertain the absence of contamination with undesired templates. The reliability of mutation frequency values was established by amplification of RNAs containing mutated residues at known frequencies. Mutations were counted only when retrieved in the two strands of the amplified DNA. These controls established a 0.5% cutoff value for mutant frequency, with a sequence depth of at least 10,000 reads per strand. Controls and procedures for the data analysis have been previously described (43, 95).
Quantification of HCV RNA using real-time RT-PCR.Real-time quantitative RT-PCR (qRT-PCR) of HCV RNA was carried out using the Light Cycler RNA Master SYBR green I kit (Roche) (29, 36). The 5′ untranslated region (UTR) of the HCV genome was amplified using as primers oligonucleotides HCV-5UTR-F2 and HCV-5UTR-R2 (see Table S5 [http://babia.cbm.uam.es/~lab121/SupplMatMoreno.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. Quantifications were carried out in triplicate.
Sedimentation equilibrium gradient analysis.For sedimentation equilibrium gradient analysis, three viral populations were prepared by infecting Huh-7.5 reporter cells with HCV p0, HCV p100, and HCV p200 at an MOI of 0.03. The three viral populations are also termed HCV p0, HCV p100, and HCV p200. Gradients were formed by overlaying 2 ml of 50%, 40%, 30% (containing the virus sample with a minimum infectivity of 104 TCID50/ml), 20%, 10%, and 0% sucrose solutions in TNE buffer (10 mM Tris-HCl [pH 8], 150 mM NaCl, 2 mM EDTA) in a 12-ml ultracentrifuge tube (SW40 rotor; Beckman). Ultracentrifugation was for 16 h at 120,000 × g at 4°C. Twelve 1-ml fractions were collected from the top of the gradient and analyzed for virus infectivity and HCV RNA as described above. The density of each fraction was determined by weighing a 100-μl sample as previously described (44).
Statistical analyses.The statistical significance of differences between mutation frequencies was evaluated by the chi-square test. To determine the statistical significance of differences in cell biomass, infectivity, and protein levels, one-way analysis of variance (ANOVA) was carried out using Prism 6 software (GraphPad). For multiple comparisons, Bonferroni's correction was applied.
The fluctuating pattern of viral titers was analyzed by adjusting the distributions of passages to Gaussian distributions using SciDAVis software (http://scidavis.sourceforge.net/ ). MATLAB script (available from A.L-S. upon request) was used for the calculation of the coefficient of variation (CV) and its dependence on the passage number.
Linear regression analysis and longitudinal data analysis of covariance (ANCOVA) using the software R was used to determine whether virus progeny production or viral RNA production as a function of time or passage number was significantly different among populations of HCV p0, HCV p100, and HCV p200.
ACKNOWLEDGMENTS
We are indebted to A. I. de Ávila for expert technical assistance and to M. Dandan, M. Enciso Vargas, C. Díez, and I. Palacios for help with some experiments.
The work in Madrid was supported by grants BFU-2011-23604, SAF2014-52400-R, and S2013/ABI-2906 (PLATESA from Comunidad Autónoma de Madrid/FEDER) and by Fundación Ramón Areces. The 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.M.R is supported by NIH R01 grants CA057973 and AI099284. E.M. is supported by grant BES-2012-052749. A.L.-S. is supported by the SVP-2014-068581 (Severo Ochoa contracts for centers of excellence Severo Ochoa). 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.
FOOTNOTES
- Received 30 December 2016.
- Accepted 28 February 2017.
- Accepted manuscript posted online 8 March 2017.
- Copyright © 2017 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
